Encyclopedia of Ocean Sciences, Second Edition, Volume 3: G-M

Citation preview

ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION

Editor-in-chief

JOHN H. STEELE

Editors

STEVE A. THORPE KARL K. TUREKIAN

Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

(c) 2011 Elsevier Inc. All Rights Reserved.

ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION

(c) 2011 Elsevier Inc. All Rights Reserved.

Subject Area Volumes from the Second Edition Climate & Oceans edited by Karl K. Turekian Elements of Physical Oceanography edited by Steve A. Thorpe Marine Biology edited by John H. Steele Marine Chemistry & Geochemistry edited by Karl K. Turekian Marine Ecological Processes edited by John H. Steele Marine Geology & Geophysics edited by Karl K. Turekian Marine Policy & Economics guest edited by Porter Hoagland, Marine Policy Center, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts Measurement Techniques, Sensors & Platforms edited by Steve A. Thorpe Ocean Currents edited by Steve A. Thorpe The Coastal Ocean edited by Karl K. Turekian The Upper Ocean edited by Steve A. Thorpe

(c) 2011 Elsevier Inc. All Rights Reserved.

ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION Volume 3: G - M Editor-in-chief

JOHN H. STEELE

Editors

STEVE A. THORPE KARL K. TUREKIAN

Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

(c) 2011 Elsevier Inc. All Rights Reserved.

Academic Press is an imprint of Elsevier 32 Jamestown Road, London NW1 7BY, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA Copyright ^ 2009 Elsevier Ltd. All rights reserved

The following articles are US government works in the public domain and are not subject to copyright: Fish Predation and Mortality; International Organizations; Large Marine Ecosystems; Ocean Circulation: Meridional Overturning Circulation; Salt Marsh Vegetation; Satellite Passive-Microwave Measurements of Sea Ice; Satellite Oceanography, History and Introductory Concepts; Satellite Remote Sensing: Ocean Color; Science of Ocean Climate Models; Wind- and Buoyancy-Forced Upper Ocean. Fish Migration, Horizontal Crown Copyright 2001 Turbulence Sensors Canadian Crown Copyright 2001 No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44)(0) 1865 843830; fax (+44)(0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier website at (http://elsevier.com/locate/permissions), and selecting Obtaining permissions to use Elsevier material

Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein, Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

Library of Congress Control Number: 2009935908

ISBN: 978-0-12-375044-0

For information on all Elsevier publications visit our website at www.elsevierdirect.com

PRINTED AND BOUND IN ITALY 09 10 11 12 13 10 9 8 7 6 5 4 3 2 1

(c) 2011 Elsevier Inc. All Rights Reserved.

Editors

Editor-in-chief John H. Steele Marine Policy Center, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA

Editors Steve A. Thorpe National Oceanography Centre, University of Southampton Southampton, UK School of Ocean Sciences, University of Bangor, Menai Bridge, Anglesey, UK Karl K. Turekian Yale University, Department of Geology and Geophysics, New Haven, Connecticut, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

v

Editorial Advisory Board John H. S. Blaxter Scottish Association for Marine Science Dunstaffnage Marine Laboratory Oban Argyll, UK Quentin Bone The Marine Biological Association of the United Kingdom Plymouth, UK Kenneth H. Brink Woods Hole Oceanographic Institution Woods Hole MA, USA Harry L. Bryden School of Ocean and Earth Science James Rennell Division University of Southampton Empress Dock Southampton, UK Robert Clark University of Newcastle upon Tyne Marine Sciences and Coastal Management Newcastle upon Tyne, UK J. Kirk Cochran State University of New York at Stony Brook Marine Sciences Research Center Stony Brook NY, USA Jeremy S. Collie Coastal Institute Graduate School of Oceanography University of Rhode Island South Ferry Road Narragansett RI, USA

Paul G. Falkowski Departments of Geological Sciences & Marine & Coastal Sciences Institute of Marine & Coastal Sciences School of Environmental & Biological Sciences Rutgers University New Brunswick NJ, USA Mike Fashamw Southampton Oceanography Centre University of Southampton Southampton UK John G. Field MArine REsearch (MA-RE) Institute University of Cape Town Rondebosch South Africa Michael Fogarty NOAA, National Marine Fisheries Service Woods Hole MA, USA Wilford D. Gardner Department of Oceanography Texas A&M University College Station TX, USA Ann Gargett Old Dominion University Center for Coastal Physical Oceanography Crittenton Hall Norfolk VA, USA

Peter J. Cook Australian Petroleum Cooperative Research Centre Canberra, Australia

vi

Robert A. Duce Departments of Oceanography and Atmospheric Sciences Texas A&M University College Station TX, USA

w

deceased

(c) 2011 Elsevier Inc. All Rights Reserved.

Editorial Advisory Board

Christopher Garrett University of Victoria Department of Physics Victoria British Columbia, Canada

Lindsay Lairdw Aberdeen University Zoology Department Aberdeen UK

W. John Gould Southampton Oceanography Centre University of Southampton Southampton UK

Peter S. Liss University of East Anglia School of Environmental Sciences Norwich, UK

John S. Grayw Institute of Marine Biology and Limnology University of Oslo Blindern Oslo, Norway

Ken Macdonald University of California Department of Geological Sciences Santa Barbara CA, USA

Gwyn Griffiths Southampton Oceanography Centre University of Southampton Southampton UK

Dennis McGillicuddy Woods Hole Oceanographic Institution Woods Hole MA, USA Alasdair McIntyre University of Aberdeen Department of Zoology Aberdeen UK

Stephen J. Hall World Fish Center Penang Malaysia Roger Harris Plymouth Marine Laboratory West Hoe Plymouth, UK Porter Hoagland Woods Hole Oceanographic Institution Woods Hole MA, USA George L. Hunt Jr. University of California, Irvine Department of Ecology and Evolutionary Biology Irvine CA, USA William J. Jenkins Woods Hole Oceanographic Institution Woods Hole MA, USA

w

deceased

W. Kendall Melville Scripps Institution of Oceanography UC San Diego La Jolla CA, USA John Milliman College of William and Mary School of Marine Sciences Gloucester Point VA, USA James N. Moum College of Oceanic and Atmospheric Sciences Oregon State University Corvallis OR, USA Michael M. Mullinw Scripps Institution of Oceanography Marine Life Research Group University of California San Diego La Jolla CA, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

vii

viii

Editorial Advisory Board

Yoshiyuki Nozakiw University of Tokyo The Ocean Research Institute Nakano-ku Tokyo Japan

Ellen Thomas Yale University Department of Geology and Geophysics New Haven CT, USA

John Orcutt Scripps Institution of Oceanography Institute of Geophysics and Planetary Physics La Jolla CA, USA Richard F. Pittenger Woods Hole Oceanographic Institution Woods Hole MA, USA Gerold Siedler Universita¨t Kiel Institut fua¨r Meereskunde Kiel Germany

Peter L. Tyack Woods Hole Oceanographic Institution Woods Hole MA, USA Bruce A. Warren Woods Hole Oceanographic Institution Woods Hole MA, USA Wilford F. Weeks University of Alaska Fairbanks Department of Geology and Geophysics Fairbanks AK, USA

Robert C. Spindel University of Washington Applied Physics Laboratory Seattle WA, USA

Robert A. Weller Woods Hole Oceanographic Institution Woods Hole MA, USA

Colin P. Summerhayes Scientific Committee on Antarctic Research (SCAR) Scott Polar Institute Cambridge, UK

w

Stewart Turner Australian National University Research School of Earth Sciences Canberra Australia

James A. Yoder Woods Hole Oceanographic Institution Woods Hole MA, USA

deceased

(c) 2011 Elsevier Inc. All Rights Reserved.

Preface to Second Print Edition The first edition of the Encyclopedia of Ocean Sciences, published in print form in 2001, has proven to be a valuable asset for the marine science community – and more generally. The continuing rapid increase in electronic access to academic material led us initially to publish the second edition electronically. We have now added this print version of the second edition because of a demonstrated need for such a product. The encyclopedia can now be accessed in print or electronic format according to the preferences and needs of individuals and institutions. In this edition there are 54 new articles, 67 revisions of previous articles, and a completely revised and improved index. We are grateful to the members of the Editorial Advisory Board, nearly all of whom have stayed with us during the lengthy process of going electronic. The transition from Academic Press to Elsevier occurred between the two editions. We thank Dr. Debbie Tranter of Elsevier for her efforts to see this edition through its final stages.

Preface to First Edition In 1942, a monumental volume was published on The Oceans by H. U. Sverdrup, M. W. Johnson, and R. H. Fleming. It was comprehensive and covered the knowledge at that time of the scientific study of the oceans. This seminal book helped to initiate the tremendous burgeoning of marine research that occurred during the following decades. The Encyclopedia of Ocean Sciences aims to embody the great growth of knowledge in a major new reference work. There have been remarkable new approaches to the study of the oceans that blur the distinctions between the physical, chemical, biological, and geological disciplines. New theories and technologies have expanded our knowledge of ocean processes. For example, plate tectonics has revolutionized our view not only of the geology and geophysics of the seafloor but also of ocean chemistry and biology. Satellite remote sensing provides a global vision as well as detailed understanding of the close coupling of ocean physics and biology at local and regional scales. Exploration, fishing, warfare, and the impact of storms have driven the past study of the seas, but we now have a great public awareness of and concern with broader social and economic issues affecting the oceans. For this reason, we have invited articles explicitly on marine policy and environmental topics, as well as encouraged authors to address these aspects of their particular subjects. We believe the encyclopedia should be of use to those involved with policy and management as well as to students and researchers. Over 400 scientists have contributed to this description of what we now know about the oceans. They are distinguished researchers who have generously shared their knowledge of this ever-growing body of science. We are extremely grateful to all these authors, whose ability to write concisely on complex subjects has generated a perspective on our science that we, as editors, believe will enhance the appreciation of the oceans, their uses, and the research ahead. It has been a major challenge for the members of the Editorial Advisory Board to cover such a heterogeneous subject. Their knowledge of the diverse areas of research has guaranteed comprehensive coverage of the ocean sciences. The Board contributed significantly by suggesting topics, persuading authors to contribute, and reviewing drafts. Many of them wrote Overviews that give broad descriptions of major parts of the ocean sciences. Clearly, it was the dedicated involvement of the Editorial Advisory Board that made this venture successful. Such a massive enterprise as a multivolume encyclopedia would not be possible without the long-term commitment of the staff of the Major Reference Works team at Academic Press. In particular, we are very grateful for the consistent support of our Senior Developmental Editor, Colin McNeil, who has worked so well with us throughout the whole process. Also, we are very pleased that new technology permits enhanced search and retrieval through the Internet. We believe this will make the encyclopedia much more accessible to individual researchers and students.

(c) 2011 Elsevier Inc. All Rights Reserved.

ix

x

Preface to Second Print Edition

In Memoriam During the creation of the Encyclopedia of Ocean Sciences and also in several cases prior to the publication of the electronic Second Edition, several Associate Editors or designated Associate Editors died. We specifically acknowledge their role in making this work an effective publication. They are Mike Fasham, John S. Gray, Lindsay Laird, Michael Mullin and Yoshiyuki Nozaki. J. H. Steele, S. A. Thorpe, and K. K. Turekian Editors

(c) 2011 Elsevier Inc. All Rights Reserved.

Guide to Use of the Encyclopedia

Introductory Points In devising the vision and structure for the Encyclopedia, the Editors have striven to unite and interrelate all current knowledge that can be designated ‘‘Ocean Sciences’’. To aid users of the Encyclopedia, this new reference work offers intuitive searching and extensive cross-linking of content. These features are explained in more detail below.

Structure of the Encyclopedia The material in the Encyclopedia is arranged as a series of articles in alphabetical order. To help you realize the full potential of the material in the Encyclopedia we have provided three features to help you find the topic of your choice.

1. Contents Lists Your first point of reference will probably be the contents list. The contents list appearing in each volume will provide you with the page number of the article. Alternatively you may choose to browse through a volume using the alphabetical order of the articles as your guide. To assist you in identifying your location within the Encyclopedia a running headline indicates the current article.

2. Cross References All of the articles in the encyclopedia have heen extensively cross referenced. The cross references, which appear at the end of each article, have heen provided at three levels: i. To indicate if a topic is discussed in greater detail elsewhere.

ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.

ii. To draw the reader’s attention to parallel discussions in other articles. ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.

(c) 2011 Elsevier Inc. All Rights Reserved.

xi

xii

Guide to Use of the Encyclopedia

iii. To indicate material that broadens the discussion.

ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.

3. Index The index will provide you with the volume and page number where the material is to be located, and the index entries differentiate between material that is a whole article, is part of an article or is data presented in a table or figure. On the opening page of the index detailed notes are provided.

4. Appendices In addition to the articles that form the main body of the encyclopedia, there are a number of appendices which provide bathymetric charts and lists of data used throughout the encyclopedia. The appendices are located in volume 6, before the index.

5. Contributors A full list of contributors appears at the beginning of volume 1.

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors Volume 1 E E Adams

N Caputi

Massachusetts Institute of Technology, Cambridge, MA, USA

Fisheries WA Research Division, North Beach, WA, Australia

T Akal

C A Carlson

NATO SACLANT Undersea Research Centre, La Spezia, Italy

University of California, Santa Barbara, CA, USA H Chamley

R Arimoto New Mexico State University, Carlsbad, NM, USA

Universite´ de Lille 1, Villeneuve d’Ascq, France R Chester

J L Bannister The Western Australian Museum, Perth, Western Australia

Liverpool University, Liverpool, Merseyside, UK V Christensen University of British Columbia, Vancouver, BC, Canada

E D Barton University of Wales, Bangor, UK

J W Dacey

N R Bates Bermuda Biological Station for Research, St George’s, Bermuda, USA

Woods Hole Oceanographic Institution, Woods Hole, MA, USA R A Duce

A Beckmann

Texas A&M University, College Station, TX, USA

Alfred-Wegener-Institut fu¨r Polar und Meeresforschung, Bremerhaven, Germany

H W Ducklow

P S Bell

The College of William and Mary, Gloucester Point, VA, USA

Proudman Oceanographic Laboratory, Liverpool, UK I Dyer G Birnbaum

Marblehead, MA, USA

Alfred-Wegener-Institut fu¨r Polar und Meeresforschung, Bremerhaven, Germany

D W Dyrssen Gothenburg University, Go¨teborg, Sweden

B O Blanton The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA E A Boyle Massachusetts Institute of Technology, Cambridge, MA, USA

S M Evans Newcastle University, Newcastle, UK I Everson Anglia Ruskin University, Cambridge, UK

P Boyle

J W Farrington

University of Aberdeen, Aberdeen, UK

Woods Hole Oceanographic Institution, MA, USA

D M Bush

M Fieux

State University of West Georgia, Carrollton, GA, USA

Universite´ Pierre et Marie Curie, Paris, France

K Caldeira

R A Fine

Stanford University, Stanford, CA, USA

University of Miami, Miami, FL, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

xiii

xiv

Contributors

K G Foote Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA L Franc¸ois University of Lie`ge, Lie`ge, Belgium M A M Friedrichs Old Dominion University, Norfolk, VA, USA T Gaston National Wildlife Research Centre, Quebec, Canada J Gemmrich University of Victoria, Victoria, BC, Canada Y Godde´ris University of Lie`ge, Lie`ge, Belgium D R Godschalk University of North Carolina, Chapel Hill, NC, USA A J Gooday Southampton Oceanography Centre, Southampton, UK A L Gordon Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA D A Hansell University of Miami, Miami FL, USA L W Harding Jr University of Maryland, College Park, MD, USA R Harris Plymouth Marine Laboratory, Plymouth, UK P J Herring Southampton Oceanography Centre, Southampton, UK B M Hickey University of Washington, Seattle, WA, USA M A Hixon Oregon State University, Corvallis, OR, USA E E Hofmann Old Dominion University, Norfolk, VA, USA S Honjo Woods Hole Oceanographic Institution, Woods Hole, MA, USA D J Howell Newcastle University, Newcastle, UK J M Huthnance CCMS Proudman Oceanographic Laboratory, Wirral, UK B Ja¨hne University of Heidelberg, Heidelberg, Germany F B Jensen SACLANT Undersea Research Centre, La Spezia, Italy A John Sir Alister Hardy Foundation for Ocean Science, Plymouth, UK

C D Jones University of Washington, Seattle, WA, USA P F Kingston Heriot-Watt University, Edinburgh, UK W Krauss Institut fu¨r Meereskunde an der Universita¨t Kiel, Kiel, Germany W A Kuperman Scripps Institution of Oceanography, University of California, San Diego, CA, USA D Lal Scripps Institute of Oceanography, University of California San Diego, La Jolla, CA, USA C S Law Plymouth Marine Laboratory, The Hoe, Plymouth, UK W J Lindberg University of Florida, Gainesville, FL, USA J R E Lutjeharms University of Cape Town, Rondebosch, South Africa P Malanotte-Rizzoli Massachusetts Institute of Technology, Cambridge, MA, USA W R Martin Woods Hole Oceanographic Institution, Woods Hole, MA, USA R P Matano Oregon State University, Corvallis, OR, USA J W McManus University of Miami, Miami, FL, USA G M McMurtry University of Hawaii at Manoa, Honolulu, HI, USA R Melville-Smith Fisheries WA Research Division, North Beach, WA, Australia P N Mikhalevsky Science Applications International Corporation, McLean, VA, USA W D Miller University of Maryland, College Park, MD, USA D Monahan University of New Hampshire, Durham, NH, USA J C Moore University of California at Santa Cruz, Santa Cruz, CA, USA A Morel Universite´ Pierre et Marie Curie, Villefranche-sur-Mer, France R Narayanaswamy The University of Manchester, Manchester, UK

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors W J Neal Grand Valley State University, Allendale, MI, USA D Pauly University of British Columbia, Vancouver, BC, Canada J W Penn Fisheries WA Research Division, North Beach, WA, Australia L C Peterson University of Miami, Miami, FL, USA S G Philander Princeton University, Princeton, NJ, USA N J Pilcher Universiti Malaysia Sarawak, Sarawak, Malaysia O H Pilkey Duke University, Durham, NC, USA

xv

D H Shull Western Washington University, Bellingham, WA, USA D K Steinberg College of William and Mary, Gloucester Pt, VA, USA L Stramma University of Kiel, Kiel, Germany R N Swift NASA Goddard Space Flight Center, Wallops Island, VA, USA T Takahashi Lamont Doherty Earth Observatory, Columbia University, Palisades, NY, USA P D Thorne Proudman Oceanographic Laboratory, Liverpool, UK

A R Piola Universidad de Buenos Aires, Buenos Aires, Argentina J M Prospero University of Miami, Miami, FL, USA S Rahmstorf Potsdam Institute for Climate Impact Research, Potsdam, Germany P C Reid SAHFOS, Plymouth, UK G Reverdin LEGOS, Toulouse Cedex, France S R Rintoul CSIRO Antarctic Climate and Ecosystems Cooperative Research Centre, Hobart, TAS, Australia J M Roberts Scottish Association for Marine Science, Oban, UK P A Rona Rutgers University, New Brunswick, NJ, USA T C Royer Old Dominion University, Norfolk, VA, USA B Rudels Finnish Institute of Marine Research, Helsinki, Finland

P L Tyack Woods Hole Oceanographic Institution, Woods Hole, USA T Tyrrell National Oceanography Centre, Southampton, UK F E Werner The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA E A Widder Harbor Branch Oceanographic Institution, Fort Pierce, FL, USA D J Wildish Fisheries and Oceans Canada, St. Andrews, NB, Canada A J Williams, III Woods Hole Oceanographic Institution, Woods Hole, MA, USA D K Woolf Southampton Oceanography Centre, Southampton, UK

W Seaman University of Florida, Gainesville, FL, USA

C W Wright NASA Goddard Space Flight Center, Wallops Island, VA, USA

F Sevilla, III, University of Santo Tomas, Manila,The Philippines

J D Wright Rutgers University, Piscataway, NJ, USA

L V Shannon University of Cape Town, Cape Town, South Africa

J R Young The Natural History Museum, London, UK

G I Shapiro University of Plymouth, Plymouth, UK

H J Zemmelink University of Groningen, Haren, The Netherlands

A D Short University of Sydney, Sydney, Australia

W Zenk Universita¨t Kiel, Kiel, Germany

(c) 2011 Elsevier Inc. All Rights Reserved.

xvi

Contributors

Volume 2 G P Arnold Centre for Environment, Fisheries & Aquaculture Science, Suffolk, UK

K Dyer University of Plymouth, Plymouth, UK

K M Bailey Alaska Fisheries Science Center, Seattle, WA, USA

M Elliott Institute of Estuarine and Coastal Studies, University of Hull, Hull, UK

J G Baldauf Texas A&M University, College Station, TX, USA

D M Farmer Institute of Ocean Sciences, Sidney, BC, Canada

J Bascompte CSIC, Seville, Spain

A V Fedorov Yale University, New Haven, CT, USA

A Belgrano Institute of Marine Research, Lysekil, Sweden

M J Fogarty Northeast Fisheries Science Center, National Marine Fisheries Service, Woods Hole, MA, USA

O A Bergstad Institute of Marine Research, Flødevigen His, Norway J H S Blaxter Scottish Association for Marine Science, Argyll, UK

R Fonteyne Agricultural Research Centre, Ghent, Oostende, Belgium

Q Bone The Marine Biological Association of the United Kingdom, Plymouth, UK

D J Fornari Woods Hole Oceanographic Institution, Woods Hole, USA

I Boyd University of St. Andrews, St. Andrews, UK

A E Gargett Old Dominion University, Norfolk, VA, USA

K M Brander DTU Aqua, Charlottenlund, Denmark and International Council for the Exploration of the Sea (ICES), Copenhagen, Denmark

C H Gibson University of California, San Diego, La Jolla, CA, USA

J N Brown Yale University, New Haven, CT, USA T K Chereskin University of California San Diego, La Jolla, CA, USA J S Collie Danish Institute for Fisheries Research, Charlottenlund, Denmark and University of Rhode Island, Narragansett, RI, USA G Cresswell CSIRO Marine Research, Tasmania, Australia

J D M Gordon Scottish Association for Marine Science, Argyll, UK J F Grassle Rutgers University, New Brunswick, New Jersey, USA S J Hall Flinders University, Adelaide, SA, Australia N Hanson University of St. Andrews, St. Andrews, UK P J B Hart University of Leicester, Leicester, UK

J Davenport University College Cork, Cork, Ireland

K R Helfrich Woods Hole Oceanographic Institution, Woods Hole, MA, USA

R H Douglas City University, London, UK

D M Higgs University of Windsor, Windsor, ON, Canada

S Draxler Karl-Franzens-Universita¨t Graz, Graz, Austria

N G Hogg Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J T Duffy-Anderson Alaska Fisheries Science Center, Seattle, WA, USA J A Dunne Santa Fe Institute, Santa Fe, NM, USA and Pacific Ecoinformatics and Computational Ecology Lab, Berkely, CA, USA

E D Houde University of Maryland, Solomons, MD, USA V N de Jonge Department of Marine Biology, Groningen University, Haren, The Netherlands

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors K Katsaros Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, FL, USA J M Klymak University of Victoria, Victoria, BC, Canada M Kucera Eberhard Karls Universita¨t Tu¨bingen, Tu¨bingen, Germany R S Lampitt University of Southampton, Southampton, UK J R N Lazier Bedford Institute of Oceanography, NS, Canada J R Ledwell Woods Hole Oceanographic Institution, Woods Hole, MA, USA P F J Lermusiaux Harvard University, Cambridge, MA, USA M E Lippitsch Karl-Franzens-Universita¨t Graz, Graz, Austria

xvii

T J Pitcher University of British Columbia, Vancouver, Canada A N Popper University of Maryland, College Park, MD, USA J F Price Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA R D Prien Southampton Oceanography Centre, Southampton, UK A-L Reysenbach Portland State University, Portland, OR, USA P L Richardson Woods Hole Oceanographic Institution, Woods Hole, MA, USA A R Robinson Harvard University, Cambridge, MA, USA M D J Sayer Dunstaffnage Marine Laboratory, Oban, Argyll, UK

B J McCay Rutgers University, New Brunswick, NJ, USA

R W Schmitt Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J D McCleave University of Maine, Orono, ME, USA

J Scott DERA Winfrith, Dorchester, Dorset, UK

D Minchin Marine Organism Investigations, Killaloe, Republic of Ireland

M P Sissenwine Northeast Fisheries Science Center, Woods Hole, MA, USA

C M Moore University of Essex, Colchester, UK K Moran University of Rhode Island, Narragansett, RI, USA G R Munro University of British Columbia, Vancouver, BC, Canada J D Nash Oregon State University, Corvallis, Oregon, OR, USA A C Naveira Garabato University of Southampton, Southampton, UK

T P Smith Northeast Fisheries Science Center, Woods Hole, MA, USA P V R Snelgrove Memorial University of Newfoundland, St John’s, NL, Canada M A Spall Woods Hole Oceanographic Institution, Woods Hole, MA, USA A Stigebrandt University of Gothenburg, Gothenburg, Sweden D A V Stow University of Southampton, Southampton, UK

J D Neilson Department of Fisheries and Oceans, New Brunswick, Canada

D J Suggett University of Essex, Colchester, UK

Y Nozakiw University of Tokyo, Tokyo, Japan

U R Sumaila University of British Columbia, Vancouver, BC, Canada

R I Perry Department of Fisheries and Oceans, British Columbia, Canada S G Philander Princeton University, Princeton, NJ, USA w

Deceased.

K S Tande Norwegian College of Fishery Science, Tromsø, Norway S A Thorpe National Oceanography Centre, Southampton, UK R S J Tol Economic and Social Research Institute, Dublin, Republic of Ireland

(c) 2011 Elsevier Inc. All Rights Reserved.

xviii

Contributors

K E Trenberth National Center for Atmospheric Research, Boulder, CO, USA J J Videler Groningen University, Haren, The Netherlands

R S Wells Chicago Zoological Society, Sarasota, FL, USA D C Wilson Institute for Fisheries Management and Coastal Community Development, Hirtshals, Denmark

Volume 3 S Ali Plymouth Marine Laboratory, Plymouth, UK

K H Coale Moss Landing Marine Laboratories, CA, USA

J T Andrews University of Colorado, Boulder, CO, USA

M F Coffins University of Texas at Austin, Austin, TX, USA

M A de Angelis Humboldt State University, Arcata, CA, USA

P J Corkeron James Cook University, Townsville, Australia

A J Arp Romberg Tiburon Center for Environment Studies, Tiburon, CA, USA

B C Coull University of South Carolina, Columbia, SC, USA

T Askew Harbor Branch Oceanographic Institute, Ft Pierce, FL, USA

R Cowen University of Miami, Miami, FL, USA

R D Ballard Institute for Exploration, Mystic, CT, USA

G Cresswell CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia

G Barnabe´ Universite´ de Montpellier II, France

D S Cronan Royal School of Mines, London, UK

R S K Barnes University of Cambridge, Cambridge, UK

J Csirke Food and Agriculture Organization of the United Nations, Rome, Italy

E D Barton University of Wales, Bangor, Menai Bridge, Anglesey, UK

G A Cutter Old Dominion University, Norfolk, VA, USA

D Bhattacharya University of Iowa, Iowa City, IA, USA

D J DeMaster North Carolina State University, Raleigh, NC, USA

F von Blanckenburg Universita¨t Bern, Bern, Switzerland

T D Dickey University of California, Santa Barbara, CA, USA

D R Bohnenstiehl North Carolina State University, Raleigh, NC, USA

D Diemand Coriolis, Shoreham, VT, USA

H L Bryden University of Southampton, Southampton, UK J Burger Rutgers University, Piscataway, NJ, USA S M Carbotte Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA G T Chandler University of South Carolina, Columbia, SC, USA M A Charette Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C S M Doake British Antarctic Survey, Cambridge, UK C M Domingues CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia C J Donlon Space Applications Institute, Ispra, Italy F Doumenge Muse´e Oce´anographique de Monaco, Monaco R A Dunn University of Hawaii at Manoa, Honolulu, HI, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors R P Dziak Oregon State University/National Oceanic and Atmospheric Administration, Hatfield Marine Science Center, Newport, OR, USA O Eldholm University of Oslo, Oslo, Norway

xix

S K Hooker University of St. Andrews, St. Andrews, UK H Hotta Japan Marine Science & Technology Center, Japan G R Ierley University of California San Diego, La Jolla, CA, USA

A E Ellis Marine Laboratory, Aberdeen, Scotland, UK C R Engle University of Arkansas at Pine Bluff, Pine Bluff, AR, USA C C Eriksen University of Washington, Seattle, WA, USA V Ettwein University College London, London, UK S Farrow Carnegie Mellon University, Pittsburgh, PA, USA M Fieux Universite´ Pierre et Marie Curie, Paris Cedex, France N Forteath Inspection Head Wharf, TAS, Australia J D Gage Scottish Association for Marine Science, Oban, UK S M Garcia Food and Agriculture Organization of the United Nations, Rome, Italy

G Ito University of Hawaii at Manoa, Honolulu, HI, USA J Jacoby Woods Hole Oceanographic Institution, Woods Hole, MA, USA M J Kaiser Bangor University, Bangor, UK A E S Kemp University of Southampton, Southampton Oceanography Centre, Southampton, UK W M Kemp University of Maryland Center for Environmental Science, Cambridge, MD, USA V S Kennedy University of Maryland, Cambridge, MD, USA P F Kingston Heriot-Watt University, Edinburgh, UK G L Kooyman University of California San Diego, CA, USA

C Garrett University of Victoria, VIC, Canada

W Krijgsman University of Utrecht, Utrecht, The Netherlands

R N Gibson Scottish Association for Marine Science, Argyll, Scotland

J B Kristoffersen University of Bergen, Bergen, Norway

M Gochfeld Environmental and Community Medicine, Piscataway, NJ, USA

K Lambeck Australian National University, Canberra, ACT, Australia

H O Halvorson University of Massachusetts Boston, Boston, MA, USA

R S Lampitt University of Southampton, Southampton, UK

B U Haq Vendome Court, Bethesda, MD, USA

M Landry University of Hawaii at Manoa, Department of Oceanography, Honolulu, HI, USA

G R Harbison Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C G Langereis University of Utrecht, Utrecht, The Netherlands A Lascaratos University of Athens, Athens, Greece

R M Haymon University of California, CA, USA

S Leibovich Cornell University, Ithaca, NY, USA

D L Hebert University of Rhode Island, RI, USA J E Heyning The Natural History Museum of Los Angeles County, Los Angeles, CA, USA P Hoagland Woods Hole Oceanographic Institution, Woods Hole, MA, USA

W G Leslie Harvard University, Cambridge, MA, USA C Llewellyn Plymouth Marine Laboratory, Plymouth, UK R A Lutz Rutgers University, New Brunswick, NJ, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

xx

Contributors

K C Macdonald Department of Geological Sciences and Marine Sciences Institute, University of California, Santa Barbara, CA, USA F T Mackenzie University of Hawaii, Honolulu, HI, USA L P Madin Woods Hole Oceanographic Institution, Woods Hole, MA, USA M Maslin University College London, London, UK G A Maul Florida Institute of Technology, Melbourne, FL, USA M McNutt MBARI, Moss Landing, CA, USA M G McPhee McPhee Research Company, Naches, WA, USA A D Mclntyre University of Aberdeen, Aberdeen, UK J Mienert University of Tromsø, Tromsø, Norway G E Millward University of Plymouth, Plymouth, UK H Momma Japan Marine Science & Technology Center, Japan J H Morison University of Washington, Seattle, WA, USA A E Mulligan Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J E Petersen Oberlin College, Oberlin, OH, USA M Phillips Network of Aquaculture Centres in Asia-Pacific (NACA), Bangkok, Thailand B Qiu University of Hawaii at Manoa, Hawaii, USA F Quezada Biotechnology Center of Excellence Corporation, Waltham, MA, USA N N Rabalais Louisiana Universities Marine Consortium, Chauvin, LA, USA R D Ray NASA Goddard Space Flight Center, Greenbelt, MD, USA M R Reeve National Science Foundation, Arlington VA, USA R R Reeves Okapi Wildlife Associates, QC, Canada A Reyes-Prieto University of Iowa, Iowa City, IA, USA P B Rhines University of Washington,Seattle, WA, USA A R Robinson Harvard University, Cambridge, MA, USA H T Rossby University of Rhode Island, Kingston, RI, USA H M Rozwadowski Georgia Institute of Technology, Atlanta, Georgia, USA

W Munk University of California San Diego, La Jolla, CA, USA

A G V Salvanes University of Bergen, Bergen, Norway

E J Murphy British Antarctic Survey, Marine Life Sciences Division, Cambridge, UK

R Schlitzer Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany

P D Naidu National Institute of Oceanography, Dona Paula, India

M E Schumacher Woods Hole Oceanographic Institution, Woods Hole, MA, USA

N Niitsuma Shizuoka University, Shizuoka, Japan

M I Scranton State University of New York, Stony Brook, NY, USA

D B Olson University of Miami, Miami, FL, USA G-A Paffenho¨fer Skidaway Institute of Oceanography, Savannah, GA, USA C Paris University of Miami, Miami, FL, USA M R Perfit Department of Geological Sciences, University of Florida, Gainsville, FL, USA

K Sherman Narragansett Laboratory, Narragansett, RI, USA M D Spalding UNEP World Conservation Monitoring Centre and Cambridge Coastal Research Unit, Cambridge, UK J Sprintall University of California San Diego, La Jolla, CA, USA J H Steele Woods Hole Oceanographic Institution, MA, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors C A Stein University of Illinois at Chicago, Chicago, IL, USA

S M Van Parijs Norwegian Polar Institute, Tromsø, Norway

C Stickley University College London, London, UK

L M Ver University of Hawaii, Honolulu, HI, USA

U R Sumaila University of British Columbia, Vancouver, BC, Canada

F J Vine University of East Anglia, Norwich, UK

S Takagawa Japan Marine Science & Technology Center, Japan

K L Von Damm University of New Hampshire, Durham, NH, USA

P K Taylor Southampton Oceanography Centre, Southampton, UK

R P Von Herzen Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A Theocharis National Centre for Marine Research (NCMR), Hellinikon, Athens, Greece

xxi

D Wartzok Florida International University, Miami, FL, USA

P C Ticco Massachusetts Maritime Academy, Buzzards Bay, MA, USA R P Trask Woods Hole Oceanographic Institution, Woods Hole, MA, USA

W F Weeks Portland, OR, USA R A Weller Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A W Trites University of British Columbia, British Columbia, Canada

J A Whitehead Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A Turner University of Plymouth, Plymouth, UK

J C Wiltshire University of Hawaii, Manoa, Honolulu, HA, USA

P L Tyack Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C Woodroffe University of Wollongong, Wollongong, NSW, Australia

G J C Underwood University of Essex, Colchester, UK

C Wunsch Massachusetts Institute of Technology, Cambridge, MA, USA

C L Van Dover The College of William and Mary, Williamsburg, VA, USA

H S Yoon University of Iowa, Iowa City, IA, USA

Volume 4 A Alldredge University of California, Santa Barbara, CA, USA D M Anderson Woods Hole Oceanographic Institution, Woods Hole, MA, USA O R Anderson Columbia University, Palisades, NY, USA

J M Bewers Bedford Institute of Oceanography, Dartmouth, NS, Canada N V Blough University of Maryland, College Park, MD, USA W Bonne

P G Baines CSIRO Atmospheric Research, Aspendale, VIC, Australia

Federal Public Service Health, Food Chain Safety and Environment, Brussels, Belgium

J M Baker Clock Cottage, Shrewsbury, UK

University of Cape Town, Cape Town, Republic of South Africa

J G Bellingham Monterey Bay Aquarium Research Institute, Moss Landing, CA, USA

R D Brodeur

G M Branch

Northwest Fisheries Science Center, Newport, OR, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

xxii

Contributors

H Burchard Baltic Sea Research Institute Warnemu¨nde, Warnemu¨nde, Germany P H Burkill Plymouth Marine Laboratory, West Hoe, Plymouth, UK Francois Carlotti C.N.R.S./Universite´ Bordeaux 1, Arachon, France K L Casciotti Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J S Grayw University of Oslo, Oslo, Norway A G Grottoli University of Pennsylvania, Philadelphia, PA, USA N Gruber Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Switzerland K C Hamer University of Durham, Durham, UK D Hammond University of Southern California, Los Angeles, CA, USA

A Clarke British Antarctic Survey, Cambridge, UK

W W Hay Christian-Albrechts University, Kiel, Germany

M B Collins National Oceanography Centre, Southampton, UK

J W Heath Coastal Fisheries Institute, CCEER Louisiana State University, Baton Rouge, LA, USA

J J Cullen Department of Oceanography, Halifax, NS, Canada D H Cushing Lowestoft, Suffolk, UK

D Hedgecock University of Southern California, Los Angeles, CA, USA C Hemleben Tu¨bingen University, Tu¨bingen, Germany

K L Denman University of Victoria, Victoria, BC, Canada S C Doney Woods Hole Oceanographic Institution, Woods Hole, MA, USA

T D Herbert Brown University, Providence, RI, USA I Hewson University of California Santa Cruz, Santa Cruz, CA, USA

J F Dower University of British Columbia, Vancouver, BC, Canada

Richard Hey University of Hawaii at Manoa, Honolulu, HI, USA

K Dysthe University of Bergen, Bergen, Norway

P Hoagland Woods Hole Oceanographic Institution, Woods Hole, MA, USA

H N Edmonds University of Texas at Austin, Port Aransas, TX, USA

N Hoepffner Institute for Environment and Sustainability, Ispra, Italy

L Føyn Institute of Marine Research, Bergen, Norway

M Hood Intergovernmental Oceanographic Commission, Paris, France

J Fuhrman University of Southern California, Los Angeles, CA, USA

M J Howarth Proudman Oceanographic Laboratory, Wirral, UK

C P Gallienne Plymouth Marine Laboratory, West Hoe, Plymouth, UK

M Huber Purdue University, West Lafayette, IN, USA

E Garel CIACOMAR, Algarve University, Faro, Portugal

J W Hurrell National Center for Atmospheric Research, Boulder, CO, USA

D M Glover Woods Hole Oceanographic Institution, Woods Hole, MA, USA S L Goodbred Jr State University of New York, Stony Brook, NY, USA J D M Gordon Scottish Association for Marine Science, Oban, Argyll, UK

D R Jackett CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia R A Jahnke Skidaway Institute of Oceanography, Savannah, GA, USA w

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors

xxiii

A Jarre University of Cape Town, Cape Town, South Africa

A F Michaels University of Southern California, Los Angeles, CA, USA

J Joseph La Jolla, CA, USA

J D Milliman College of William and Mary, Gloucester, VA, USA

D M Karl University of Hawaii at Manoa, Honolulu, HI, USA

C D Mobley Sequoia Scientific, Inc., WA, USA

K L Karsh Princeton University, Princeton, NJ, USA

M M Mullinw Scripps Institution of Oceanography, La Jolla, CA, USA

J Karstensen Universita¨t Kiel (IFM-GEOMAR), Kiel, Germany

P Mu¨ller University of Hawaii, Honolulu, HI, USA

R M Key Princeton University, Princeton, NJ, USA

L A Murray The Centre for Environment, Fisheries and Aquaculture Sciences, Lowestoft, UK

P D Killworth Southampton Oceanography Centre, Southampton, UK B Klinger Center for Ocean-Land-Atmosphere Studies (COLA), Calverton, MD, USA H E Krogstad NTNU, Trondheim, Norway I Laing Centre for Environment Fisheries and Aquaculture Science, Weymouth, UK

T Nagai Tokyo University of Marine Science and Technology, Tokyo, Japan K H Nisancioglu Bjerknes Centre for Climate Research, University of Bergen, Bergen, Norway Y Nozakiw University of Tokyo, Tokyo, Japan

G F Lane-Serff University of Manchester, Manchester, UK

K J Orians The University of British Columbia, Vancouver, BC, Canada

A Longhurst Place de I’Eglise, Cajarc, France

C A Paulson Oregon State University, Corvallis, OR, USA

R Lukas University of Hawaii at Manoa, Hawaii, USA

W G Pearcy Oregon State University, Corvallis, OR, USA

M Lynch University of California Santa Barbara, Santa Barbara, CA, USA

W S Pegau Oregon State University, Corvallis, OR, USA

M Macleod World Wildlife Fund, Washington, DC, USA E Maran˜o´n University of Vigo, Vigo, Spain S Martin University of Washington, Seattle, WA, USA S M Masutani University of Hawaii at Manoa, Honolulu, HI, USA I N McCave University of Cambridge, Cambridge, UK T J McDougall CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia C L Merrin The University of British Columbia, Vancouver, BC, Canada

T Platt Dalhousie University, NS, Canada J J Polovina National Marine Fisheries Service, Honolulu, HI, USA D Quadfasel Niels Bohr Institute, Copenhagen, Denmark J A Raven Biological Sciences, University of Dundee, Dundee, UK G E Ravizza Woods Hole Oceanographic Institution, Woods Hole, MA, USA A J Richardson University of Queensland, St. Lucia, QLD, Australia M Rubega University of Connecticut, Storrs, CT, USA w

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

xxiv

Contributors

K C Ruttenberg Woods Hole Oceanographic Institution, Woods Hole, MA, USA

K K Turekian Yale University, New Haven, CT, USA T Tyrrell University of Southampton, Southampton, UK

A G V Salvanes University of Bergen, Bergen, Norway

O Ulloa Universidad de Concepcio´n, Concepcio´n, Chile

S Sathyendranath Dalhousie University, NS, Canada

C M G Vivian The Centre for Environment, Fisheries and Aquaculture Sciences, Lowestoft, UK

R Schiebel Tu¨bingen University, Tu¨bingen, Germany F B Schwing NOAA Fisheries Service, Pacific Grove, CA, USA

J J Walsh University of South Florida, St. Petersburg, FL, USA

M P Seki National Marine Fisheries Service, Honolulu, HI, USA

R M Warwick Plymouth Marine Laboratory, Plymouth, UK

L J Shannon Marine and Coastal Management, Cape Town, South Africa

N C Wells Southampton Oceanography Centre, Southampton, UK

K Shepherd Institute of Ocean Sciences, Sidney, BC, Canada

J A Whitehead Woods Hole Oceanographic Institution, Woods Hole, MA, USA

D Siegel-Causey Harvard University, Cambridge MA, USA D M Sigman Princeton University, Princeton, NJ, USA A Soloviev Nova Southeastern University, FL, USA J H Steele Woods Hole Oceanographic Institution, MA, USA P K Takahashi University of Hawaii at Manoa, Honolulu, HI, USA L D Talley Scripps Institution of Oceanography, La Jolla, CA, USA E Thomas Yale University, New Haven, CT, USA J R Toggweiler NOAA, Princeton, NJ, USA

M Wilkinson Heriot-Watt University, Edinburgh, UK R G Williams University of Liverpool, Oceanography Laboratories, Liverpool, UK C A Wilson III Department of Oceanography and Coastal Sciences, and Coastal Fisheries Institute, CCEER Louisiana State University, Baton Rouge, LA, USA H Yamazaki Tokyo University of Marine Science and Technology, Tokyo, Japan B deYoung Memorial University, St. John’s, NL, Canada G Zibordi Institute for Environment and Sustainability, Ispra, Italy

Volume 5 D G Ainley H.T. Harvey Associates, San Jose CA, USA W Alpers University of Hamburg, Hamburg, Germany J R Apelw Global Ocean Associates, Silver Spring, MD, USA w

Deceased.

A B Baggeroer Massachusetts Institute of Technology, Cambridge, MA, USA L T Balance NOAA-NMFS, La Jolla, CA, USA R Batiza Ocean Sciences, National Science Foundation, VA, USA W H Berger Scripps Institution of Oceanography, La Jolla, CA, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors J L Bodkin US Geological Survey, AK, USA

I Everson British Antarctic Survey Cambridge, UK

I L Boyd Natural Environment Research Council, Cambridge, UK

I Fer University of Bergen, Bergen, Norway

A C Brown University of Cape Town, Cape Town, Republic of South Africa

M Fieux Universite´-Pierre et Marie Curie, Paris, France

xxv

J Burger Rutgers University, Piscataway, NJ, USA

R A Flather Proudman Oceanographic Laboratory, Bidston Hill, Prenton, UK

C J Camphuysen Netherlands Institute for Sea Research, Texel, The Netherlands

G S Giese Woods Hole Oceanographic Institution, Woods Hole, MA, USA

D C Chapman Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J M Gregory Hadley Centre, Berkshire, UK

R E Cheney Laboratory for Satellite Altimetry, Silver Spring, Maryland, USA T Chopin University of New Brunswick, Saint John, NB, Canada J A Church Antarctic CRC and CSIRO Marine Research, TAS, Australia J K Cochran State University of New York, Stony Brook, NY, USA P Collar Southampton Oceanography Centre, Southampton, UK R J Cuthbert University of Otago, Dunedin, New Zealand L S Davis University of Otago, Dunedin, New Zealand K L Denman University of Victoria, Victoria BC, Canada R P Dinsmore Woods Hole Oceanographic Institution, Woods Hole, MA, USA G J Divoky University of Alaska, Fairbanks, AK, USA

S M Griffies NOAA/GFDL, Princeton, NJ, USA G Griffiths Southampton Oceanography Centre, Southampton, UK A Harding University of California, San Diego, CA, USA W S Holbrook University of Wyoming, Laramie, WY, USA G L Hunt, Jr University of Washington, Seattle, WA, USA and University of California, Irvine, CA, USA P Hutchinson North Atlantic Salmon Conservation Organization, Edinburgh, UK K B Katsaros Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, FL, USA H L Kite-Powell Woods Hole Oceanographic Institution, Woods Hole, MA, USA M A Kominz Western Michigan University, Kalamazoo, MI, USA

L M Dorman University of California, San Diego, La Jolla, CA, USA

R G Kope Northwest Fisheries Science Center, Seattle, WA, USA

J F Dower University of British Columbia, Vancouver, BC, Canada

G S E Lagerloef Earth and Space Research, Seattle, WA, USA

J B Edson Woods Hole Oceanographic Institution, Woods Hole, MA, USA

L M Lairdw Aberdeen University, Aberdeen, UK

T I Eglinton Woods Hole Oceanographic Institution, Woods Hole, MA, USA

M Leppa¨ranta University of Helsinki, Helsinki, Finland w

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

xxvi

Contributors

E J Lindstrom NASA Science Mission Directorate, Washington, DC, USA

C T Roman University of Rhode Island, Narragansett, RI, USA

A K Liu NASA Goddard Space Flight Center, Greenbelt, MD, USA

M Sawhney University of New Brunswick, Saint John, NB, Canada

C R McClain NASA Goddard Space Flight Center, Greenbelt, MD, USA

G Shanmugam The University of Texas at Arlington, Arlington, TX, USA

D J McGillicuddy Jr Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J Sharples Proudman Oceanographic Laboratory, Liverpool, UK

W K Melville Scripps Institution of Oceanography, La Jolla CA, USA

J H Simpson Bangor University, Bangor, UK R K Smedbol Dalhousie University, Halifax, NS, Canada

D Mills Atlantic Salmon Trust, UK

L B Spear H.T. Harvey Associates, San Jose, CA, USA

P J Minnett University of Miami, Miami, FL, USA W A Montevecchi Memorial University of Newfoundland, NL, Canada W S Moore University of South Carolina, Columbia, SC, USA S J Morreale Cornell University, Ithaca, NY, USA K W Nicholls British Antarctic Survey, Cambridge, UK T J O’Shea Midcontinent Ecological Science Center, Fort Collins, CO, USA T E Osterkamp University of Alaska, Alaska, AK, USA F V Paladino Indiana-Purdue University at Fort Wayne, Fort Wayne, IN, USA C L Parkinson NASA Goddard Space Flight Center, Greenbelt, MD, USA A Pearson Woods Hole Oceanographic Institution, Woods Hole, MA, USA J T Potemra SOEST/IPRC, University of Hawaii, Honolulu, HI, USA J A Powell Florida Marine Research Institute, St Petersburg, FL, USA T Qu SOEST/IPRC, University of Hawaii, Honolulu, HI, USA

R L Stephenson St. Andrews Biological Station, St. Andrews, NB, Canada J M Teal Woods Hole Oceanographic Institution, Rochester, MA, USA K K Turekian Yale University, New Haven, CT, USA P Wadhams University of Cambridge, Cambridge, UK W F Weeks Portland, OR, USA G Wefer Universita¨t Bremen, Bremen, Germany W S Wilson NOAA/NESDIS, Silver Spring, MD, USA M Windsor, North Atlantic Salmon Conservation Organization, Edinburgh, UK S Y Wu NASA Goddard Space Flight Center, Greenbelt, MD, USA L Yu Woods Hole Oceanographic Institution, Woods Hole, MA, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

Contributors

xxvii

Volume 6 A V Babanin Swinburne University of Technology, Melbourne, VIC, Australia

S E Humphris Woods Hole Oceanographic Institution, Woods Hole, MA, USA

R T Barber Duke University Marine Laboratory, Beaufort, NC, USA

W J Jenkins University of Southampton, Southampton, UK

J Bartram World Health Organization, Geneva, Switzerland

D R B Kraemer The Johns Hopkins University, Baltimore, MD, USA

A Beckmann Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung, Bremerhaven, Germany M C Benfield Louisiana State University, Baton Rouge, LA, USA P S Bogden Maine State Planning Office, Augusta, ME, USA J A T Bye The University of Melbourne, Melbourne, VIC, Australia M F Cronin NOAA Pacific Marine Environmental Laboratory, Seattle, WA, USA A R J David Bere Alston, Devon, UK W Deuser Woods Hole Oceanographic Institution, Woods Hole, MA, USA J Donat Old Dominion University, Norfolk, VA, USA C Dryden Old Dominion University, Norfolk, VA, USA A Dufour United States Environmental Protection Agency, OH, USA C A Edwards University of Connecticut, Groton, CT, USA W J Emery University of Colorado, Boulder, CO, USA E Fahrbach Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung, Bremerhaven, Germany

S Krishnaswami Physical Research Laboratory, Ahmedabad, India E L Kunze University of Washington, Seattle, WA, USA T E L Langford University of Southampton, Southampton, UK J R Ledwell Woods Hole Oceanographic Institution, Woods Hole, MA, USA P L-F Liu Cornell University, Ithaca, NY, USA M M R van der Loeff Alfred-Wegener-Institut fu¨r Polar und Meereforschung Bremerhaven, Germany R Lueck University of Victoria, Victoria, BC, Canada J E Lupton Hatfield Marine Science Center, Newport, OR, USA L P Madin Woods Hole Oceanographic Institution, Woods Hole, MA, USA M E McCormick The Johns Hopkins University, Baltimore, MD, USA M G McPhee McPhee Research Company, Naches, WA, USA J H Middleton The University of New South Wales, Sydney, NSW, Australia P J Minnett University of Miami, Miami, FL, USA E C Monahan University of Connecticut at Avery Point, Groton, CT, USA

A M Gorlov Northeastern University, Boston, Massachusetts, USA

C Moore WET Labs Inc., Philomath, OR, USA

I Helmond CSIRO Marine Research, TAS, Australia

J H Morison University of Washington, Seattle, WA, USA

R A Holman Oregon State University, Corvallis, OR, USA

J N Moum Oregon State University, Corvallis, OR, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

xxviii

Contributors

N S Oakey Bedford Institute of Oceanography, Dartmouth, NS, Canada D T Pugh University of Southampton, Southampton, UK D L Rudnick University of California, San Diego, CA, USA H Salas CEPIS/HEP/Pan American Health Organization, Lima, Peru L K Shay University of Miami, Miami, FL, USA W D Smyth Oregon State University, Corvallis, OR, USA J Sprintall University of California San Diego, La Jolla, CA, USA

L St. Laurrent University of Victoria, Victoria, BC, Canada W G Sunda National Ocean Service, NOAA, Beaufort, NC, USA M Tomczak Flinders University of South Australia, Adelaide, SA, Australia A J Watson University of East Anglia, Norwich, UK P H Wiebe Woods Hole Oceanographic Institution, Woods Hole, MA, USA P F Worcester University of California at San Diego, La Jolla, CA, USA

(c) 2011 Elsevier Inc. All Rights Reserved.

Contents Volume 1 Abrupt Climate Change

S Rahmstorf

1

Absorbance Spectroscopy for Chemical Sensors Abyssal Currents

R Narayanaswamy, F Sevilla, III

W Zenk

Accretionary Prisms

15

J C Moore

31

Acoustic Measurement of Near-Bed Sediment Transport Processes Acoustic Noise

Acoustic Scintillation Thermography Acoustics In Marine Sediments

K G Foote

62

P A Rona, C D Jones

71

T Akal

75

P N Mikhalevsky

Acoustics, Deep Ocean

92

W A Kuperman

101

F B Jensen

112

Acoustics, Shallow Water R Chester

Agulhas Current

120

J R E Lutjeharms

Aircraft Remote Sensing

128

L W Harding Jr, W D Miller, R N Swift, C W Wright

Air–Sea Gas Exchange

38 52

Acoustic Scattering by Marine Organisms

Aeolian Inputs

P D Thorne, P S Bell

I Dyer

Acoustics, Arctic

7

B Ja¨hne

138 147

Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens J W Dacey, H J Zemmelink

157

Air–Sea Transfer: N2O, NO, CH4, CO

163

Alcidae

C S Law

T Gaston

171

Antarctic Circumpolar Current Antarctic Fishes

S R Rintoul

I Everson

191

Anthropogenic Trace Elements in the Ocean Antifouling Materials

E A Boyle

211

W Seaman, W J Lindberg

234

R A Duce

Atmospheric Transport and Deposition of Particulate Material to the Oceans R Arimoto Authigenic Deposits

226

S G Philander

Atmospheric Input of Pollutants

Baleen Whales

203

B Rudels

Atlantic Ocean Equatorial Currents

Bacterioplankton

195

D J Howell, S M Evans

Arctic Ocean Circulation Artificial Reefs

178

G M McMurtry H W Ducklow

J L Bannister

238 J M Prospero, 248 258 269 276

(c) 2011 Elsevier Inc. All Rights Reserved.

xxix

xxx

Contents

Baltic Sea Circulation Bathymetry

W Krauss

288

D Monahan

297

Beaches, Physical Processes Affecting Benguela Current

Benthic Foraminifera

316 D J Wildish

328

A J Gooday

Benthic Organisms Overview

336

P F Kingston

348

P L Tyack

357

Biogeochemical Data Assimilation

E E Hofmann, M A M Friedrichs

Biological Pump and Particle Fluxes Bioluminescence

Bioturbation

S Honjo

376

A Morel

385

D H Shull

Black Sea Circulation

395

G I Shapiro

Bottom Water Formation

401

A L Gordon

415

Brazil and Falklands (Malvinas) Currents

A R Piola, R P Matano

Breaking Waves and Near-Surface Turbulence

J Gemmrich

D K Woolf

Calcium Carbonates

L C Peterson

E D Barton

Carbon Dioxide (CO2) Cycle

467

T Takahashi

Cenozoic Climate – Oxygen Isotope Evidence Cenozoic Oceans – Carbon Cycle Models

J D Wright L Franc¸ois, Y Godde´ris

R A Fine W R Martin

J W Farrington

Coastal Zone Management

514

539 551 563

F E Werner, B O Blanton

Coastal Topography, Human Impact on Coastal Trapped Waves

502

531

H Chamley

Coastal Circulation Models

495

524

Chemical Processes in Estuarine Sediments

Coccolithophores

E E Adams, K Caldeira

P Boyle

Chlorinated Hydrocarbons

477 487

Carbon Sequestration via Direct Injection into the Ocean

Clay Mineralogy

455

C A Carlson, N R Bates, D A Hansell, D K Steinberg

CFCs in the Ocean

431

445

B M Hickey, T C Royer

Canary and Portugal Currents

Cephalopods

422

439

California and Alaska Currents

Carbon Cycle

364 371

P J Herring, E A Widder

Bio-Optical Models

Bubbles

305

L V Shannon

Benthic Boundary Layer Effects

Bioacoustics

A D Short

D M Bush, O H Pilkey, W J Neal

J M Huthnance D R Godschalk

T Tyrrell, J R Young

(c) 2011 Elsevier Inc. All Rights Reserved.

572 581 591 599 606

Contents

Cold-Water Coral Reefs Conservative Elements

J M Roberts

615

D W Dyrssen

626

Continuous Plankton Recorders Copepods

A John, P C Reid

R Harris

Coral Reefs

630 640

Coral Reef and Other Tropical Fisheries Coral Reef Fishes

xxxi

V Christensen, D Pauly

M A Hixon

655

J W McManus

660

Corals and Human Disturbance Cosmogenic Isotopes

N J Pilcher

671

D Lal

678

Coupled Sea Ice–Ocean Models Crustacean Fisheries

651

A Beckmann, G Birnbaum

688

J W Penn, N Caputi, R Melville-Smith

699

CTD (Conductivity, Temperature, Depth) Profiler Current Systems in the Atlantic Ocean Current Systems in the Indian Ocean

A J Williams, III

L Stramma M Fieux, G Reverdin

Current Systems in the Southern Ocean

A L Gordon

Current Systems in the Mediterranean Sea

P Malanotte-Rizzoli

708 718 728 735 744

Volume 2 Data Assimilation in Models Deep Convection

A R Robinson, P F J Lermusiaux

J R N Lazier

Deep Submergence, Science of

13 D J Fornari

22

K Moran

37

Deep-Sea Drilling Methodology Deep-Sea Drilling Results

1

J G Baldauf

45

Deep-Sea Fauna

P V R Snelgrove, J F Grassle

55

Deep-Sea Fishes

J D M Gordon

67

Deep-Sea Ridges, Microbiology

A-L Reysenbach

73

Deep-Sea Sediment Drifts

D A V Stow

80

Demersal Species Fisheries

K Brander

90

Determination of Past Sea Surface Temperatures Differential Diffusion

A E Gargett

Dispersion from Hydrothermal Vents Diversity of Marine Species Dolphins and Porpoises

R W Schmitt, J R Ledwell

K R Helfrich

P V R Snelgrove R S Wells

Double-Diffusive Convection

98 114

Dispersion and Diffusion in the Deep Ocean

Drifters and Floats

M Kucera

R W Schmitt

P L Richardson

(c) 2011 Elsevier Inc. All Rights Reserved.

122 130 139 149 162 171

xxxii

Contents

Dynamics of Exploited Marine Fish Populations East Australian Current

M J Fogarty

G Cresswell

179 187

Economics of Sea Level Rise

R S J Tol

197

Ecosystem Effects of Fishing

S J Hall

201

Eels

J D McCleave

208

Effects of Climate Change on Marine Mammals Ekman Transport and Pumping

T K Chereskin, J F Price

El Nin˜o Southern Oscillation (ENSO)

Electrical Properties of Sea Water

Energetics of Ocean Mixing

228

S G Philander

R D Prien

Elemental Distribution: Overview

Y Nozaki

255

A C Naveira Garabato

261

Eutrophication

271

J M Klymak, J D Nash

Estuarine Circulation

288

K Dyer

299

V N de Jonge, M Elliott

Evaporation and Humidity

Fiord Circulation

306

K Katsaros

Exotic Species, Introduction of Expendable Sensors

241 247

w

A V Fedorov, J N Brown

Estimates of Mixing

218 222

K E Trenberth

El Nin˜o Southern Oscillation (ENSO) Models

Equatorial Waves

I Boyd, N Hanson

324

D Minchin

332

J Scott

345

A Stigebrandt

353

Fiordic Ecosystems

K S Tande

359

Fish Ecophysiology

J Davenport

367

Fish Feeding and Foraging Fish Larvae

P J B Hart

E D Houde

Fish Locomotion

381

J J Videler

Fish Migration, Horizontal Fish Migration, Vertical

Fish Reproduction

Fish Vision

392

G P Arnold

402

J D Neilson, R I Perry

Fish Predation and Mortality

Fish Schooling

374

411

K M Bailey, J T Duffy-Anderson

J H S Blaxter

425

T J Pitcher

432

R H Douglas

445

Fish: Demersal Fish (Life Histories, Behavior, Adaptations) Fish: General Review

O A Bergstad

Q Bone

458 467

Fish: Hearing, Lateral Lines (Mechanisms, Role in Behavior, Adaptations to Life Underwater) A N Popper, D M Higgs w

417

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

476

Contents

Fisheries and Climate

K M Brander

Fisheries Economics

483

U R Sumaila, G R Munro

Fisheries Overview

491

M J Fogarty, J S Collie

Fisheries: Multispecies Dynamics Fishery Management

499

J S Collie

505

T P Smith, M P Sissenwine

Fishery Management, Human Dimension

513

D C Wilson, B J McCay

Fishery Manipulation through Stock Enhancement or Restoration Fishing Methods and Fishing Fleets Floc Layers

xxxiii

M D J Sayer

R Fonteyne

522 528 535

R S Lampitt

548

Florida Current, Gulf Stream, and Labrador Current Flow through Deep Ocean Passages Flows in Straits and Channels

P L Richardson

N G Hogg

554 564

D M Farmer

572

Fluid Dynamics, Introduction, and Laboratory Experiments

S A Thorpe

578

Fluorometry for Biological Sensing

D J Suggett, C M Moore

581

Fluorometry for Chemical Sensing

S Draxler, M E Lippitsch

589

Food Webs

A Belgrano, J A Dunne, J Bascompte

Forward Problem in Numerical Models Fossil Turbulence

596

M A Spall

604

C H Gibson

612

Volume 3 Gas Exchange in Estuaries

M I Scranton, M A de Angelis

Gelatinous Zooplankton

L P Madin, G R Harbison

General Circulation Models

Geomorphology

20

C G Langereis, W Krijgsman

C Woodroffe

Geophysical Heat Flow

C A Stein, R P Von Herzen

40 K Lambeck

C C Eriksen A D Mclntyre

Grabs for Shelf Benthic Sampling

67

P F Kingston

70

M McNutt

80

Groundwater Flow to the Coastal Ocean Habitat Modification

49 59

Global Marine Pollution

Gravity

25 33

Glacial Crustal Rebound, Sea Levels, and Shorelines Gliders

9

G R Ierley

Geomagnetic Polarity Timescale

1

A E Mulligan, M A Charette

M J Kaiser

Heat and Momentum Fluxes at the Sea Surface Heat Transport and Climate History of Ocean Sciences

88 99

P K Taylor

H L Bryden H M Rozwadowski

(c) 2011 Elsevier Inc. All Rights Reserved.

105 114 121

xxxiv

Contents

Holocene Climate Variability Hydrothermal Vent Biota

M Maslin, C Stickley, V Ettwein R A Lutz

125 133

Hydrothermal Vent Deposits

R M Haymon

144

Hydrothermal Vent Ecology

C L Van Dover

151

Hydrothermal Vent Fauna, Physiology of

A J Arp

159

Hydrothermal Vent Fluids, Chemistry of

K L Von Damm

164

Hypoxia

N N Rabalais

172

Icebergs

D Diemand

181

Ice-Induced Gouging of the Seafloor Ice–Ocean Interaction

W F Weeks

J H Morison, M G McPhee

191 198

Ice Shelf Stability

C S M Doake

209

Igneous Provinces

M F Coffins, O Eldholm

218

Indian Ocean Equatorial Currents Indonesian Throughflow

M Fieux

J Sprintall

237

Inherent Optical Properties and Irradiance Internal Tidal Mixing Internal Tides

T D Dickey

W Munk

258

C Garrett

266

International Organizations Intertidal Fishes

M R Reeve

274

R N Gibson

Intra-Americas Sea

244 254

R D Ray

Internal Waves

Intrusions

226

280

G A Maul

286

D L Hebert

295

Inverse Modeling of Tracers and Nutrients

R Schlitzer

300

Inverse Models

C Wunsch

312

IR Radiometers

C J Donlon

319

K H Coale

331

Iron Fertilization Island Wakes Krill

E D Barton

343

E J Murphy

349

Kuroshio and Oyashio Currents

B Qiu

Laboratory Studies of Turbulent Mixing Lagoons

358 J A Whitehead

R S K Barnes

Lagrangian Biological Models Land–Sea Global Transfers

377 D B Olson, C Paris, R Cowen F T Mackenzie, L M Ver

Langmuir Circulation and Instability Large Marine Ecosystems

S Leibovich

K Sherman

Laridae, Sternidae, and Rynchopidae Law of the Sea

371

389 394 404 413

J Burger, M Gochfeld

P Hoagland, J Jacoby, M E Schumacher (c) 2011 Elsevier Inc. All Rights Reserved.

420 432

Contents

Leeuwin Current

G Cresswell, C M Domingues

Long-Term Tracer Changes Macrobenthos Magnetics

444

F von Blanckenburg

455

J D Gage

467

F J Vine

478

Manganese Nodules Mangroves

xxxv

D S Cronan

488

M D Spalding

496

Manned Submersibles, Deep Water

H Hotta, H Momma, S Takagawa

Manned Submersibles, Shallow Water

T Askew

505 513

Mariculture Diseases and Health

A E Ellis

519

Mariculture of Aquarium Fishes

N Forteath

524

Mariculture of Mediterranean Species Mariculture Overview

G Barnabe´, F Doumenge

M Phillips

537

Mariculture, Economic and Social Impacts Marine Algal Genomics and Evolution Marine Biotechnology

532

C R Engle

545

A Reyes-Prieto, H S Yoon, D Bhattacharya

H O Halvorson, F Quezada

552 560

Marine Chemical and Medicine Resources

S Ali, C Llewellyn

567

Marine Fishery Resources, Global State of

J Csirke, S M Garcia

576

Marine Mammal Diving Physiology

G L Kooyman

Marine Mammal Evolution and Taxonomy

J E Heyning

Marine Mammal Migrations and Movement Patterns Marine Mammal Overview

582

P J Corkeron, S M Van Parijs

P L Tyack

Marine Mammal Trophic Levels and Interactions Marine Mammals and Ocean Noise

A W Trites

Marine Policy Overview Marine Protected Areas Marine Silica Cycle

635 643

654 G-A Paffenho¨fer

656

P Hoagland, P C Ticco

664

P Hoagland, U R Sumaila, S Farrow D J DeMaster

R S Lampitt

Maritime Archaeology

622

651

J H Steele

Marine Plankton Communities

Mediterranean Sea Circulation

672 678 686

R D Ballard

Meddies and Sub-Surface Eddies

Meiobenthos

S K Hooker

A E S Kemp

Marine Mesocosms

615

628

R R Reeves

Marine Mammals: Sperm Whales and Beaked Whales

Marine Snow

P L Tyack

D Wartzok

Marine Mammals, History of Exploitation

596 605

Marine Mammal Social Organization and Communication

Marine Mats

589

H T Rossby A R Robinson, W G Leslie, A Theocharis, A Lascaratos

B C Coull, G T Chandler (c) 2011 Elsevier Inc. All Rights Reserved.

695 702 710 726

xxxvi

Contents

Mesocosms: Enclosed Experimental Ecosystems in Ocean Science Mesopelagic Fishes

J E Petersen, W M Kemp

A G V Salvanes, J B Kristoffersen

Mesoscale Eddies

748

P B Rhines

Metal Pollution

755

G E Millward, A Turner

Metalloids and Oxyanions

732

768

G A Cutter

776

Methane Hydrates and Climatic Effects

B U Haq

784

Methane Hydrate and Submarine Slides

J Mienert

790

Microbial Loops

M Landry

Microphytobenthos

799

G J C Underwood

807

Mid-Ocean Ridge Geochemistry and Petrology Mid-Ocean Ridge Seismic Structure Mid-Ocean Ridge Seismicity

M R Perfit

815

S M Carbotte

826

D R Bohnenstiehl, R P Dziak

Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology

837 K C Macdonald

Mid-Ocean Ridges: Mantle Convection and Formation of the Lithosphere Millennial-Scale Climate Variability

J T Andrews

Mineral Extraction, Authigenic Minerals Molluskan Fisheries Monsoons, History of Moorings

G Ito, R A Dunn

852 867 881

J C Wiltshire

890

V S Kennedy

899

N Niitsuma, P D Naidu

910

R P Trask, R A Weller

919

Volume 4 Nekton

W G Pearcy, R D Brodeur

Nepheloid Layers

1

I N McCave

Network Analysis of Food Webs

8 J H Steele

Neutral Surfaces and the Equation of State Nitrogen Cycle

19 T J McDougall, D R Jackett

D M Karl, A F Michaels

Nitrogen Isotopes in the Ocean Noble Gases and the Cryosphere Non-Rotating Gravity Currents North Atlantic Oscillation (NAO) North Sea Circulation

25 32

D M Sigman, K L Karsh, K L Casciotti

40

M Hood

55

P G Baines

59

J W Hurrell

65

M J Howarth

73

Nuclear Fuel Reprocessing and Related Discharges

H N Edmonds

82

Ocean Biogeochemistry and Ecology, Modeling of

N Gruber, S C Doney

89

Ocean Carbon System, Modeling of Ocean Circulation

S C Doney, D M Glover

N C Wells

105 115

Ocean Circulation: Meridional Overturning Circulation

J R Toggweiler

(c) 2011 Elsevier Inc. All Rights Reserved.

126

Contents

Ocean Gyre Ecosystems

M P Seki, J J Polovina

Ocean Margin Sediments Ocean Ranching

132

S L Goodbred Jr

138

A G V Salvanes

146

R G Williams

156

Ocean Subduction

Ocean Thermal Energy Conversion (OTEC) Ocean Zoning

S M Masutani, P K Takahashi

M Macleod, M Lynch, P Hoagland

Offshore Sand and Gravel Mining Oil Pollution

Okhotsk Sea Circulation

E Garel, W Bonne, M B Collins

200

H Yamazaki, H Burchard, K Denman, T Nagai

Open Ocean Convection

A Soloviev, B Klinger

Open Ocean Fisheries for Deep-Water Species

Optical Particle Characterization

P H Burkill, C P Gallienne

265 272 274

R Lukas

287

E Thomas

295 W W Hay

Paleoceanography: Orbitally Tuned Timescales Paleoceanography: the Greenhouse World Particle Aggregation Dynamics Past Climate from Corals

T D Herbert

M Huber, E Thomas

A Alldredge

A G Grottoli

K L Denman, J F Dower

Pelagic Biogeography

A Longhurst

D H Cushing

Pelecaniformes

Peru–Chile Current System

C A Paulson, W S Pegau

J Karstensen, O Ulloa

319 330 338 348 356

379 385 393

K C Ruttenberg

Photochemical Processes

311

370

M Rubega

Phosphorus Cycle

303

364

D Siegel-Causey

Penetrating Shortwave Radiation

252 261

I Laing

Paleoceanography, Climate Models in

Phytobenthos

R A Jahnke

K K Turekian

Pacific Ocean Equatorial Currents

Phalaropes

243

G F Lane-Serff

Oysters – Shellfish Farming

Pelagic Fishes

234

K K Turekian

Oxygen Isotopes in the Ocean

Paleoceanography

226

J Joseph

Organic Carbon Cycling in Continental Margin Environments

Overflows and Cascades

208 218

J D M Gordon

Open Ocean Fisheries for Large Pelagic Species

Origin of the Oceans

182 191

L D Talley

One-Dimensional Models

167 174

J M Baker

Patch Dynamics

xxxvii

N V Blough

M Wilkinson

401 414 425

(c) 2011 Elsevier Inc. All Rights Reserved.

xxxviii

Contents

Phytoplankton Blooms

D M Anderson

Phytoplankton Size Structure Plankton

M M Mullin

Plankton and Climate Plankton Viruses

432

E Maran˜o´n

445

w

453

A J Richardson

455

J Fuhrman, I Hewson

465

Platforms: Autonomous Underwater Vehicles Platforms: Benthic Flux Landers

J G Bellingham

R A Jahnke

485

Platinum Group Elements and their Isotopes in the Ocean

G E Ravizza

Plio-Pleistocene Glacial Cycles and Milankovitch Variability Polar Ecosystems

K H Nisancioglu

A Clarke

Pollution, Solids

494 504 514

C M G Vivian, L A Murray

Pollution: Approaches to Pollution Control Pollution: Effects on Marine Communities Polynyas

473

519

J S Grayw, J M Bewers R M Warwick

526 533

S Martin

540

Population Dynamics Models

Francois Carlotti

Population Genetics of Marine Organisms Pore Water Chemistry

546

D Hedgecock

556

D Hammond

Primary Production Distribution

563

S Sathyendranath, T Platt

572

Primary Production Methods

J J Cullen

578

Primary Production Processes

J A Raven

585

Procellariiformes

K C Hamer

590

Propagating Rifts and Microplates

Richard Hey

597

Protozoa, Planktonic Foraminifera

R Schiebel, C Hemleben

606

Protozoa, Radiolarians

O R Anderson

Radiative Transfer in the Ocean Radioactive Wastes Radiocarbon

C D Mobley

619

L Føyn

629

R M Key

637

Rare Earth Elements and their Isotopes in the Ocean Red Sea Circulation Redfield Ratio Refractory Metals

Y Nozaki

w

D Quadfasel

677

K J Orians, C L Merrin

Regime Shifts, Physical Forcing Regime Shifts: Methods of Analysis

653 666

T Tyrrell

Regime Shifts, Ecological Aspects

w

613

L J Shannon, A Jarre, F B Schwing F B Schwing B deYoung, A Jarre

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

687 699 709 717

Contents

Regional and Shelf Sea Models

J J Walsh

Remote Sensing of Coastal Waters

Rigs and offshore Structures River Inputs

722

N Hoepffner, G Zibordi

Remotely Operated Vehicles (ROVs)

xxxix

732

K Shepherd

742

C A Wilson III, J W Heath

748

J D Milliman

754

Rocky Shores

G M Branch

762

Rogue Waves

K Dysthe, H E Krogstad, P Mu¨ller

770

Rossby Waves

P D Killworth

Rotating Gravity Currents

781

J A Whitehead

790

Volume 5 Salmon Fisheries, Atlantic

P Hutchinson, M Windsor

Salmon Fisheries, Pacific Salmonid Farming Salmonids

1

R G Kope

L M Laird

12

w

23

D Mills

29

Salt Marsh Vegetation

C T Roman

Salt Marshes and Mud Flats Sandy Beaches, Biology of Satellite Altimetry

39

J M Teal

43

A C Brown

49

R E Cheney

58

Satellite Oceanography, History, and Introductory Concepts J R Apel w

W S Wilson, E J Lindstrom, 65

Satellite Passive-Microwave Measurements of Sea Ice

C L Parkinson

80

Satellite Remote Sensing of Sea Surface Temperatures

P J Minnett

91

Satellite Remote Sensing SAR

A K Liu, S Y Wu

Satellite Remote Sensing: Ocean Color

C R McClain

Satellite Remote Sensing: Salinity Measurements Science of Ocean Climate Models Sea Ice

103

G S E Lagerloef

S M Griffies

P Wadhams

114 127 133 141

Sea Ice Dynamics

M Leppa¨ranta

159

Sea Ice: Overview

W F Weeks

170

Sea Level Change

J A Church, J M Gregory

179

Sea Level Variations Over Geologic Time Sea Otters

w

M A Kominz

J L Bodkin

185 194

Deceased.

(c) 2011 Elsevier Inc. All Rights Reserved.

xl

Contents

Sea Surface Exchanges of Momentum, Heat, and Fresh Water Determined by Satellite Remote Sensing L Yu

202

Sea Turtles

212

F V Paladino, S J Morreale

Seabird Conservation

J Burger

Seabird Foraging Ecology Seabird Migration

220

L T Balance, D G Ainley, G L Hunt Jr

L B Spear

227 236

Seabird Population Dynamics

G L Hunt Jr

247

Seabird Reproductive Ecology

L S Davis, R J Cuthbert

251

Seabird Responses to Climate Change Seabirds and Fisheries Interactions

David G Ainley, G J Divoky C J Camphuysen

Seabirds as Indicators of Ocean Pollution Seabirds: An Overview Seals

265

W A Montevecchi

274

G L Hunt, Jr

279

I L Boyd

285

Seamounts and Off-Ridge Volcanism Seas of Southeast Asia

R Batiza

292

J T Potemra, T Qu

Seaweeds and their Mariculture Sediment Chronologies

305

T Chopin, M Sawhney

317

J K Cochran

327

Sedimentary Record, Reconstruction of Productivity from the Seiches

Seismic Structure

I Fer, W S Holbrook

L M Dorman

367

K B Katsaros

375

Sensors for Micrometeorological and Flux Measurements Shelf Sea and Shelf Slope Fronts

J B Edson

J Sharples, J H Simpson

H L Kite-Powell

Single Point Current Meters

T I Eglinton, A Pearson

P Collar, G Griffiths

436

Slides, Slumps, Debris Flows, and Turbidity Currents Small Pelagic Species Fisheries Small-Scale Patchiness, Models of

419 428

T J O’Shea, J A Powell G Shanmugam

R L Stephenson, R K Smedbol D J McGillicuddy Jr

Small-Scale Physical Processes and Plankton Biology

J F Dower, K L Denman

M Fieux

447 468 474 488 494

A B Baggeroer

Southern Ocean Fisheries

391

409

Single Compound Radiocarbon Measurements

Sonar Systems

382

401

R P Dinsmore

Somali Current

351 361

Sensors for Mean Meteorology

Shipping and Ports

333 344

A Harding

Seismology Sensors

Sirenians

G Wefer, W H Berger

D C Chapman, G S Giese

Seismic Reflection Methods for Study of the Water Column

Ships

257

504

I Everson

(c) 2011 Elsevier Inc. All Rights Reserved.

513

Contents

Sphenisciformes

L S Davis

520

Stable Carbon Isotope Variations in the Ocean Storm Surges

K K Turekian

529

R A Flather

530

Sub Ice-Shelf Circulation and Processes Submarine Groundwater Discharge Sub-Sea Permafrost Surface Films

xli

K W Nicholls

541

W S Moore

551

T E Osterkamp

559

W Alpers

570

Surface Gravity and Capillary Waves

W K Melville

573

Volume 6 Temporal Variability of Particle Flux Thermal Discharges and Pollution

W Deuser

1

T E L Langford

10

Three-Dimensional (3D) Turbulence Tidal Energy Tides

W D Smyth, J N Moum

A M Gorlov

26

D T Pugh

Tomography

32

P F Worcester

Topographic Eddies Towed Vehicles

40

J H Middleton

57

I Helmond

Trace Element Nutrients

65

W G Sunda

Tracer Release Experiments

75

A J Watson, J R Ledwell

Tracers of Ocean Productivity

Transmissometry and Nephelometry Tritium–Helium Dating

87

W J Jenkins

93

Transition Metals and Heavy Metal Speciation

Tsunami

18

J Donat, C Dryden

100

C Moore

109

W J Jenkins

119

P L-F Liu

127

Turbulence in the Benthic Boundary Layer Turbulence Sensors

R Lueck, L St. Laurrent, J N Moum

N S Oakey

Under-Ice Boundary Layer

148

M G McPhee, J H Morison

Upper Ocean Heat and Freshwater Budgets Upper Ocean Mean Horizontal Structure Upper Ocean Mixing Processes

155

P J Minnett

163

M Tomczak

175

J N Moum, W D Smyth

185

Upper Ocean Structure: Responses to Strong Atmospheric Forcing Events Upper Ocean Time and Space Variability Upper Ocean Vertical Structure Upwelling Ecosystems

141

L K Shay

192

D L Rudnick

211

J Sprintall, M F Cronin

217

R T Barber

Uranium-Thorium Decay Series in the Oceans: Overview

225 M M R van der Loeff

(c) 2011 Elsevier Inc. All Rights Reserved.

233

xlii

Contents

Uranium-Thorium Series Isotopes in Ocean Profiles Vehicles for Deep Sea Exploration

S E Humphris

Viral and Bacterial Contamination of Beaches Volcanic Helium

J Bartram, H Salas, A Dufour

285

Water Types and Water Masses

W J Emery

291

M E McCormick, D R B Kraemer

Waves on Beaches

267 277

E L Kunze

Wave Generation by Wind

244 255

J E Lupton

Vortical Modes

Wave Energy

S Krishnaswami

300

J A T Bye, A V Babanin

304

R A Holman

310

Weddell Sea Circulation

E Fahrbach, A Beckmann

318

Wet Chemical Analyzers

A R J David

326

Whitecaps and Foam

E C Monahan

Wind- and Buoyancy-Forced Upper Ocean Wind Driven Circulation

331 M F Cronin, J Sprintall

P S Bogden, C A Edwards

Zooplankton Sampling with Nets and Trawls

337 346

P H Wiebe, M C Benfield

355

Appendix 1. SI Units and Some Equivalences

373

Appendix 2. Useful Values

376

Appendix 3. Periodic Table of the Elements

377

Appendix 4. The Geologic Time Scale

378

Appendix 5. Properties of Seawater

379

Appendix 6. The Beaufort Wind Scale and Seastate

384

Appendix 7. Estimated Mean Oceanic Concentrations of the Elements

386

Appendix 8. Abbreviations

389

Appendix 9. Taxonomic Outline Of Marine Organisms

L P Madin

401

Appendix 10. Bathymetric Charts of the Oceans

412

Index

421

(c) 2011 Elsevier Inc. All Rights Reserved.

GAS EXCHANGE IN ESTUARIES M. I. Scranton, State University of New York, Stony Brook, NY, USA M. A. de Angelis, Humboldt State University, Arcata, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1113–1119, & 2001, Elsevier Ltd.

the gas in either the gas or liquid phase.) For gases that make up a large fraction of the atmosphere (O2, N2, Ar), pgas does not vary temporally or spatially. For trace atmospheric gases (carbon dioxide (CO2), methane (CH4), hydrogen (H2), nitrous oxide (N2O), and others), pgas may vary considerably geographically or seasonally, and may be affected by anthropogenic activity or local natural sources.

Introduction Many atmospherically important gases are present in estuarine waters in excess over levels that would be predicted from simple equilibrium between the atmosphere and surface waters. Since estuaries are defined as semi-enclosed coastal bodies of water that have free connections with the open sea and within which sea water is measurably diluted with fresh water derived from land drainage, they tend to be supplied with much larger amounts of organic matter and other compounds than other coastal areas. Thus, production of many gases is enhanced in estuaries relative to the rest of the ocean. The geometry of estuaries, which typically have relatively large surface areas compared to their depths, is such that flux of material (including gases) from the sediments, and fluxes of gases across the air/water interface, can have a much greater impact on the water composition than would be the case in the open ocean. Riverine and tidal currents are often quite marked, which also can greatly affect concentrations of biogenic gases.

Gas Solubility The direction and magnitude of the exchange of gases across an air/water interface are determined by the difference between the surface-water concentration of a given gas and its equilibrium concentration or gas solubility with respect to the atmosphere. The concentration of a specific gas in equilibrium with the atmosphere (Ceq) is given by Henry’s Law: Ceq ¼ pgas =KH

½1

where pgas is the partial pressure of the gas in the atmosphere, and KH is the Henry’s Law constant for the gas. Typically as temperature and salinity increase, gas solubility decreases. (Note that Henry’s Law and the Henry’s Law constant also may be commonly expressed in terms of the mole fraction of

Gas Exchange (Flux) Across the Air/ Water Interface The rate of gas exchange across the air/water interface for a specific gas is determined by the degree of disequilibrium between the actual surface concentration of a gas (Csurf) and its equilibrium concentration (Ceq), commonly expressed as R: R ¼ Csurf =Ceq

½2

If R ¼ 1, the dissolved gas is in equilibrium with the atmosphere and no net flux or exchange with the atmosphere occurs. For gases with Ro1, the dissolved gas is undersaturated with respect to the atmosphere and there is a net flux of the gas from the atmosphere to the water. For gases with R41, a net flux of the gas from the water to the atmosphere occurs.

Models of Gas Exchange The magnitude of the flux (F) in units of mass of gas per unit area per unit time across the air/water interface is a function of the magnitude of the difference between the dissolved gas concentration and its equilibrium concentration as given by Fick’s First Law of Diffusion: F ¼ kðCsurf  Ceq Þ

½3

where k is a first order rate constant, which is a function of the specific gas and surface water conditions. The rate constant, k, also known as the transfer coefficient, has units of velocity and is frequently given as k ¼ D=z

½4

where D is the molecular diffusivity (in units of cm2 s1), and z is the thickness of the laminar layer at the

(c) 2011 Elsevier Inc. All Rights Reserved.

1

2

GAS EXCHANGE IN ESTUARIES

air/water interface, which limits the diffusion of gas across the interface. In aquatic systems, Csurf is easily measured by gas chromatographic analysis, and Ceq may be calculated readily if the temperature and salinity of the water are known. In order to determine the flux of gas (F) in or out of the water, the transfer coefficient (k) needs to be determined. The value of k is a function of the surface roughness of the water. In open bodies of water, wind speed is the main determinant of surface roughness. A number of studies have established a relationship between wind speed and either the transfer coefficient, k, or the liquid laminar layer thickness, z (Figure 1). The transfer coefficient is also related to the Schmidt number, Sc, defined as: Sc ¼ n=D

½5

where n is the kinematic viscosity of the water. In calmer waters, corresponding to wind speeds of o5 m s1, k is proportional to Sc2/3. At higher wind speeds, but where breaking waves are rare, k is proportional to S1/2. Therefore, if the transfer coefficient of one gas is known, the k value for any other gas can be determined as: Scn1 =Scn2

k1 =k2 ¼ Smooth surface regime

150

Rough surface regime

_ 2/ 3

Breaking wave (bubble) regime

_ 1/ 2

Kw Sc

Kw Sc

125

½6

where n ¼ the exponent. For short-term steady winds, the transfer coefficient for CO2 has been derived as kCO2 ¼ 0:31ðU10 Þ2 ðSc=600Þ0:5

½7

where U10 is the wind speed at a height of 10 m above the water surface. Eqns [6] and [7] can be used to estimate k for gases other than CO2. In most estuarine studies wind speeds are measured closer to the water surface. In such cases, the wind speed measured at 2 cm above the water surface can be approximated as 0.5U10. In restricted estuaries and tidally influenced rivers, wind speed may not be a good predictor of wind speed due to limited fetch or blockage of prevailing winds by shore vegetation. Instead, streambed-generated turbulence is likely to be more important than wind stress in determining water surface roughness. In such circumstances, the large eddy model may be used to approximate k as: k ¼ 1:46ðDul1 Þ1=2

½8

where u is the current velocity, and l is equivalent to the mean depth in shallow turbulent systems. Much of the reported uncertainty (and study to study variability in fluxes) is caused by differences in assumptions related to the transfer coefficient rather than to large changes in concentration of the gas in the estuary.

Less soluble gas (e.g O2)

Transfer velocity cm h

_1

Direct Gas Exchange Measurements 100

75

More soluble gas (e.g. CO2)

50

25

0.25 0

0

5

0.5 10

_1

0.75 U ms 15 U ms

_1

1 20

Figure 1 Idealized plot of transfer coefficient (Kw) as a function of wind speed (u) and friction velocity (u*). (Adapted with permission from Liss PS and Merlivat L (1985) Air–sea gas exchange rates. In: Buat-Menard P (ed.) The Role of Air–sea Exchange in Geochemical Cycling, p. 117. NATO ASI Series C, vol. 185. Dordrecht: Reidel.)

Gas exchange with the atmosphere for gases for which water column consumption and production processes are known can be estimated using a dissolved gas budget. Through time-series measurements of biological and chemical cycling, gas loss or gain across the air/water interface can be determined by difference. For example, in the case of dissolved O2, the total change in dissolved O2 concentration over time can be attributed to air/water exchange and biological processes. The contribution of biological processes to temporal changes in dissolved O2 may be estimated from concurrent measurements of phosphate and an assumed Redfield stoichiometry, and subtracted from the total change to yield an estimate of air/water O2 exchange. Gas fluxes also may be measured directly using a flux chamber that floats on the water surface. The headspace of the chamber is collected and analyzed over several time points to obtain an estimate of the net amount of gas crossing the air/water interface enclosed by the chamber. If the surface area enclosed

(c) 2011 Elsevier Inc. All Rights Reserved.

GAS EXCHANGE IN ESTUARIES

by the chamber is known, a net gas flux can be determined. Some flux chambers are equipped with small fans that simulate ambient wind conditions. However, most chambers do not use fans and so do not take account of the effects of wind-induced turbulence on gas exchange. Despite this limitation, flux chambers are important tools for measuring gas exchange in environments (such as estuaries or streams) where limited fetch or wind breaks produced by shoreline vegetation make wind less important than current-induced turbulence in shallow systems. While flux chambers may alter the surface roughness and, hence, gas/exchange rates via diffusion, flux chambers or other enclosed gas capturing devices also are the best method for determining loss of gases across the air/water interface due to ebullition of gas bubbles from the sediment. Measurement of radon (Rn) deficiencies in the upper water column can be used to determine gas exchange coefficients and laminar layer thickness, which can then be applied to other gases using eqn [3]. In this method, gaseous 222Rn, produced by radioactive decay of 226Ra, is assumed to be in secular equilibrium within the water column. 222Rn is relatively short-lived and has an atmospheric concentration of essentially zero. Therefore, in nearsurface waters, a 222Rn deficiency is observed, due to flux of 222Rn across the air/water interface. The flux of 222Rn across the air/water interface can be determined by the depth-integrated difference in measured 222Rn and that which should occur based on the 226 Ra inventory. From this flux, the liquid laminar layer thickness, z, can be calculated. Other volatile tracers have been used in estuaries to determine gas exchange coefficients. These tracers, such as chlorofluorocarbons (CFCs) or sulfur hexafluoride (SF6), are synthetic compounds with no known natural source. Unlike 222Rn, these gases are stable in solution. These tracers may be added to the aquatic system and the decrease of the gas due to flux across the air/water interface monitored over time. In some estuaries, point sources of these compounds may exist and the decrease of the tracer with distance downstream may be used to determine k or z values for the estuary.

Individual Gases

plays an important role controlling atmospheric chemistry, including serving as a regulator of tropospheric ozone concentrations and a major sink for hydroxyl radicals in the stratosphere. Methane concentrations have been increasing annually at the rate of approximately 1–2% over the last two centuries. While the contribution of estuaries to the global atmospheric methane budget is small because of the relatively small estuarine global surface area, estuaries have been identified as sources of methane to the atmosphere and coastal ocean and contribute a significant fraction of the marine methane emissions to the atmosphere. Surface methane concentrations reported primarily from estuaries in North America and Europe range from 1 to 42000 nM throughout the tidal portion of the estuaries. Methane in estuarine surface waters is generally observed to be supersaturated (100% saturation B2–3 nM CH4) with R-values ranging from 0.7 to 1600 (Table 1). In general, estuarine methane concentrations are highest at the freshwater end of the estuary and decrease with salinity. This trend reflects riverine input as the major source of methane to most estuaries, with reported riverine methane concentrations ranging from 5 to 10 000 nM. Estuaries with large plumes have been observed to cause elevated methane concentrations in adjacent coastal oceans. In addition to riverine input, sources of methane to estuaries include intertidal flats and marshes, ground-water input, runoff

Table 1 Methane saturation values (R) and estimated fluxes to the atmosphere for US and European estuaries Geographical regiona

Rb

Flux CH4 (mmol m2 h1)

North Pacific coast, USA North Pacific coast, USA Columbia River, USA Hudson River, USA Tomales Bay, CA, USA Baltic Sea, Germany European Atlantic coast Atlantic coast, USA Pettaquamscutt Estuary, USA

3–290 1–550 78 18–376 2–37 10.5–1550 0.7–1580 n.a. 81–111

6.2–41.7c 3.6–8.3c 26.0c 4.7–40.4c 17.4–26.3c 9.4–15.6c 5.5c 102–1107c 0.8–14.2c 541–3375d

a

Methane (CH4)

Atmospheric methane plays an important role in the Earth’s radiative budget as a potent greenhouse gas, which is 3.7 times more effective than carbon dioxide in absorbing infrared radiation. Despite being present in trace quantities, atmospheric methane

3

For studies that report values for a single estuary, the major river feeding the estuary is provided. For studies that report values for more than one estuary, the oceanic area being fed by the estuaries is given. b R ¼ degree of saturation ¼ measured concentration/atmospheric equilibrium concentration. n.a. indicates values not available in reference. c Diffusive flux. d Ebullition (gas bubble) flux.

(c) 2011 Elsevier Inc. All Rights Reserved.

4

GAS EXCHANGE IN ESTUARIES

from agricultural and pasture land, petroleum pollution, lateral input from exposed bank soils, wastewater discharge and emission from organic-rich anaerobic sediments, either diffusively or via ebullition (transport of gas from sediments as bubbles) and subsequent dissolution within the water column. Anthropogenically impacted estuaries or estuaries supplied from impacted rivers tend to be characterized by higher water-column methane concentrations relative to pristine estuarine systems. Seasonally, methane levels in estuaries are higher in summer compared with winter, primarily due to increased bacterial methane production (methanogenesis) in estuarine and riverine sediments. Methane can be removed from estuarine waters by microbial methane oxidation and emission to the atmosphere. Methane oxidation within estuaries can be quite rapid, with methane turnover times of o2 h to several days. Methane oxidation appears to be most rapid at salinities of less than about 6 (on the practical salinity units scale) and is strongly dependent on temperature, with highest oxidation rates occurring during the summer, when water temperatures are highest. Methane oxidation rates decrease rapidly with higher salinities. Methane diffusive fluxes to the atmosphere reported for estuaries (Table 1) fall within a narrow range of 3.6–41.7 mmol m2 h1 (2–16 mg CH4 m2 day1). Using a global surface area for estuaries of 1.4  106 km2 yields an annual emission of methane to the atmosphere from estuaries of 1–8 Tg y1. Because the higher flux estimates given in Table 1 generally were obtained close to the freshwater endmember of the estuary, the global methane estuarine emission is most likely within the range of 1– 3 Tg y1, corresponding to approximately 10% of the total global oceanic methane flux to the atmosphere, despite the much smaller global surface area of estuaries relative to the open ocean. Methane is also released to the atmosphere directly from anaerobic estuarine sediments via bubble formation and injection into the water column. Although small amounts of methane from bubbles may dissolve within the water column, the relatively shallow nature of the estuarine environment results in the majority of methane in bubbles reaching the atmosphere. The quantitative release of methane via this mechanism is difficult to evaluate due to the irregular and sporadic spatial and temporal extent of ebullition. Where ebullition occurs, the flux of methane to the atmosphere is considerably higher than diffusive flux (Table 1), but the areal extent of bubbling is relatively smaller than that of diffusive flux and, except in organic-rich stagnant areas such as tidal marshes, probably does not contribute

significantly to estuarine methane emissions to the atmosphere. Methane emission via ebullition has been observed to be at least partially controlled by tidal changes in hydrostatic pressure, with release of methane occurring at or near low tide when hydrostatic pressure is at a minimum. Nitrous Oxide (N2O)

Nitrous oxide is another important greenhouse gas that is present in elevated concentrations in estuarine environments. At present, N2O is responsible for about 5–6% of the anthropogenic greenhouse effect and is increasing in the atmosphere at a rate of about 0.25% per year. However, the role of estuaries in the global budget of the gas has only been addressed recently. Nitrous oxide is produced primarily as an intermediate during both nitrification (the oxidation of ammonium to nitrate) and denitrification (the reduction of nitrate, via nitrite and N2O, to nitrogen gas), although production by dissimilatory nitrate reduction to ammonium is also possible. In estuaries, nitrification and denitrification are both thought to be important sources. Factors such as the oxygen level in the estuary and the nitrate and ammonium concentrations of the water can influence which pathway is dominant, with denitrification dominating at very low, but non-zero, oxygen concentrations. Nitrous oxide concentrations are typically highest in the portions of the estuary closest to the rivers, and decrease with distance downstream. A number of workers have reported nitrous oxide maxima in estuarine waters at low salinities (o5–10 on the PSU scale), but this is not always the case. The turbidity maximum has been reported to be the site of maximum nitrification (presumably because of increased residence time for bacteria attached to suspended particulate matter, combined with elevated substrate (oxygen and ammonium)). Table 2 presents a summary of the data published for degree of saturation and air–estuary flux of nitrous oxide from a variety of estuaries, all of which are located in Europe and North America. Concentrations are commonly above that predicted from air–sea equilibrium, and estimates of fluxes range from 0.01 mmol m2 h1 to 5 mmol m2 h1. Ebullition is not important for nitrous oxide because it is much more soluble than methane. Researchers have estimated the size of the global estuarine source for N2O based on fluxes from individual estuaries multiplied by the global area occupied by estuaries to range from 0.22 Tg N2O y1 to 5.7 Tg y1 depending on the characteristics of the rivers studied. Independent estimates based on budgets of nitrogen

(c) 2011 Elsevier Inc. All Rights Reserved.

GAS EXCHANGE IN ESTUARIES

Table 2 Nitrous oxide saturation values (R) and estimated fluxes to the atmosphere for US and European estuaries Estuary

R

Flux a N2O (mmol m2 h1)

Europe Gironde River Gironde River Oder River Elbe Scheldt Scheldt

1.1–1.6 E1.0–3.2 0.9–3.1 2.0–16 E1.0–31 E1.2–30

n.a. n.a. 0.014–0.165 n.a. 1.27–4.77 3.56

UK Colne Tamar Humber Tweed Mediterranean Amvraikos Gulf

0.9–13.6 1–3.3 2–40 0.96–1.1

0.9–1.1

1.3 0.41 1.8 E0

5

Table 3 Fluxes of carbon dioxide from estuaries in Europe and eastern USA Estuary

European rivers Northern Europe Scheldt estuary Portugal UK Clyde estuary East coast USA Hudson River (tidal freshwater) Georgia rivers

R

Fluxa CO2 (mmol m2 h1)

0.7–61.1 0.35–26.2 1.6–15.8 1.1–14.4 E0.7–1.8

1.0–31.7 4.2–50 10–31.7 4.4–10.4 n.a.

1.2–5.4

0.67–1.54

Slight supersaturation to 22.9

1.7–23

a n.a., insufficient data were available to permit calculation of this value.

0.043 7 0.0468

North-west USA Yaquina Bay Alsea River

1.0–4.0 0.9–2.4

0.165–0.699 0.047–0.72

East coast USA Chesapeake Bay Merrimack

0.9–1.4 1.2–4.5

n.a. n.a.

a

All fluxes given are for diffusive flux to the atmosphere. n.a. indicates that insufficient data were given to permit calculation of flux.

input to rivers, assumptions about the fraction of inorganic nitrogen species removed by nitrification or denitrification, and the fractional ‘yield’ of nitrous oxide production during these processes indicate that nitrous oxide fluxes to the atmosphere from estuaries is about 0.06–0.34 Tg N2O y1. Carbon Dioxide (CO2) and Oxygen (O2)

Estuaries are typically heterotrophic systems, which means that the amount of organic matter respired within the estuary exceeds the amount of organic matter fixed by primary producers (phytoplankton and macrophytes). Since production of carbon dioxide then exceeds biological removal of carbon dioxide, it follows that estuaries are likely to be sources of the gas to the atmosphere. At the same time, since oxidation of organic matter to CO2 requires oxygen, the heterotrophic nature of estuaries suggests that they represent sinks for atmospheric oxygen. In many estuaries, primary productivity is severely limited by the amount of light that penetrates into the water due to high particulate loadings in the water. In addition, large amounts of organic matter may be supplied to the estuary by runoff from agricultural and forested land, from ground water, from

sewage effluent, and from organic matter in the river itself. There are many reports of estuarine systems with oxygen saturations below 1 (undersaturated with respect to the atmosphere), but few studies in which oxygen flux to the estuary has been reported. However, estuaries are often dramatically supersaturated with respect to saturation with CO2, especially at low salinities, and a number of workers have reported estimates of carbon dioxide flux from these systems (Table 3).

Dimethylsulfide (DMS)

DMS is an atmospheric trace gas that plays important roles in tropospheric chemistry and climate regulation. In the estuarine environment, DMS is produced primarily from the breakdown of the phytoplankton osmoregulator 3-(dimethylsulfonium)-propionate (DMSP). DMS concentrations reported in estuaries are generally supersaturated, ranging from 0.5 to 22 nM, and increase with increasing salinity. DMS levels in the water column represent a balance between tightly coupled production from DMSP and microbial consumption. Only 10% of DMS produced from DMSP in the estuarine water column is believed to escape to the atmosphere from estuarine surface water, since the biological turnover of DMS (turnover time of 3–7 days) is approximately 10 times faster than DMS exchange across the air/water interface. A large part of the estuarine DMS flux to the atmosphere may occur over short time periods on the order of weeks, corresponding to phytoplankton blooms. The DMS flux for an estuary in Florida, USA, was estimated to be on the order of o1 nmoles m2 h1. Insufficient

(c) 2011 Elsevier Inc. All Rights Reserved.

6

GAS EXCHANGE IN ESTUARIES

data are available to determine reliable global DMS air/water exchanges from estuaries. Hydrogen (H2)

Hydrogen is an important intermediate in many microbial catabolic reactions, and the efficiency of hydrogen transfer among microbial organisms within an environment helps determine the pathways of organic matter decomposition. Hydrogen is generally supersaturated in the surface waters of the few estuaries that have been analyzed for dissolved hydrogen with R-values of 1.5–67. Hydrogen flux to the atmosphere from estuaries has been reported to be on the order of 0.06–0.27 nmol m2 h1. The contribution of estuaries to the global atmospheric H2 flux cannot be determined from the few available data. Carbon Monoxide (CO)

Carbon monoxide in surface waters is produced primarily from the photo-oxidation of dissolved organic matter by UV radiation. Since estuarine waters are characterized by high dissolved organic carbon levels, the surface waters of estuaries are highly supersaturated and are a strong source of CO to the atmosphere. Reported R-values for CO range from approximately 10 to 410 000. Because of the highly variable distributions of dissolved CO within surface waters (primarily as the result of the highly variable production of CO), it is impossible to derive a meaningful value for CO emissions to the atmosphere from estuaries. Carbonyl Sulfide (OCS)

Carbonyl sulfide makes up approximately 80% of the total sulfur content of the atmosphere and is the major source of stratospheric aerosols. Carbonyl sulfide is produced within surface waters by photolysis of dissolved organosulfur compounds. Therefore, surface water OCS levels within estuaries exhibit a strong diel trend. Carbonyl sulfide is also added to the water column by diffusion from anoxic sediments, where its production appears to be coupled to microbial sulfate reduction. Diffusion of OCS from the sediment to the water column accounts for B75% of the OCS supplied to the water column and is responsible for the higher OCS concentrations in estuaries relative to the open ocean. While supersaturations of OCS are observed throughout estuarine surface waters, no trends with salinity have been observed. Atmospheric OCS fluxes to the atmosphere from Chesapeake Bay have been reported to range from 10.4 to 56.2 nmol m2 h1. These areal fluxes are over 50 times greater than those determined for the open ocean.

Elemental Mercury (Hg0)

Elemental mercury is produced in estuarine environments by biologically mediated processes. Both algae and bacteria are able to convert dissolved inorganic mercury to volatile forms, which include organic species (monomethyl- and dimethyl-mercury) and Hg0. Under suboxic conditions, elemental mercury also may be the thermodynamically stable form of the metal. In the Scheldt River estuary, Hg0 correlated well with phytoplankton pigments, suggesting that phytoplankton were the dominant factors, at least in that system. Factors that may affect elemental mercury concentrations include the type of phytoplankton present, photo-catalytic reduction of ionic Hg in surface waters, the extent of bacterial activity that removes oxygen from the estuary, and removal of mercury by particulate scavenging and sulfide precipitation. Fluxes of elemental mercury to the atmosphere have been estimated for the Pettaquamscutt estuary in Rhode Island, USA, and for the Scheldt, and range from 4.2–29 pmol m2 h1, although the values are strongly dependent on the model used to estimate gas exchange coefficients. Volatile Organic Compounds (VOCs)

In addition to gases produced naturally in the environment, estuaries tend to be enriched in byproducts of industry and other human activity. A few studies have investigated volatile organic pollutants such as chlorinated hydrocarbons (chloroform, tetrachloromethane, 1,1-dichloroethane, 1,2-dichloroethane, 1,1,1-trichloroethane, trichloroethylene and tetrachloroethylene) and monocyclic aromatic hydrocarbons (benzene, toluene, ethylbenzene, oxylene and m- and p-xylene). Concentrations of VOCs are controlled primarily by the location of the sources, dilution of river water with clean marine water within the estuary, gas exchange, and in some cases, adsorption onto suspended or settling solids. In some cases (for example, chloroform) there also may be natural biotic sources of the gas. Volatilization to the atmosphere can be an important ‘cleansing’ mechanism for the estuary system. Since the only estuaries studied to date are heavily impacted by human activity (the Elbe and the Scheldt), it is not possible to make generalizations about the importance of these systems on a global scale.

Conclusions Many estuaries are supersaturated with a variety of gases, making them locally, and occasionally

(c) 2011 Elsevier Inc. All Rights Reserved.

GAS EXCHANGE IN ESTUARIES

regionally, important sources to the atmosphere. However, estuarine systems are also highly variable in the amount of gases they contain. Since most estuaries studied to date are in Europe or the North American continent, more data are needed before global budgets can be reliably prepared.

Glossary Air–water interface The boundary between the gaseous phase (the atmosphere) and the liquid phase (the water). Catabolic Biochemical process resulting in breakdown of organic molecules into smaller molecules yielding energy. Denitrification Reduction of nitrate via nitrite to gaseous endproducts (nitrous oxide and dinitrogen gas). Ebullition Gas transport by bubbles, usually from sediments. Estuary Semi-enclosed coastal body of water with free connection to the open sea and within which sea water is measurably diluted with fresh water derived from land drainage. Gas solubility The amount of gas that will dissolve in a liquid when the liquid is in equilibrium with the overlying gas phase. Henry’s Law Constant Proportionality constant relating the vapor pressure of a solute to its mole fraction in solution. Liquid laminar thickness The thickness of a layer at the air/water interface where transport of a dissolved species is controlled by molecular (rather than turbulent) diffusion. Molecular diffusivity (D) The molecular diffusion coefficient. Nitrification Oxidation of ammonium to nitrite and nitrate. Practical salinity scale A dimensionless scale for salinity. Redfield stoichiometry Redfield and colleagues noted that organisms in the sea consistently removed nutrient elements from the water in a fixed ratio (C : N : P=106 : 16 : 1). Subsequent workers have found that nutrient concentrations in the sea typically are present in those same ratios. Transfer coefficient The rate constant which determines the rate of transfer of gas from liquid to gas phase.

See also Air–Sea Gas Exchange. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane

7

Hydrocarbons, Organo-Halogens. Air–Sea Transfer: N2O, NO, CH4, CO.

Further Reading Bange HW, Rapsomanikis S, and Andreae MO (1996) Nitrous oxide in coastal waters. Global Biogeochemical Cycles 10: 197--207. Bange HW, Dahlke S, Ramesh R, Meyer-Reil L-A, Rapsomanikis S, and Andreae MO (1998) Seasonal study of methane and nitrous oxide in the coastal waters of the southern Baltic Sea. Estuarine, Coastal and Shelf Science 47: 807--817. Barnes J and Owens NJP (1998) Denitrification and nitrous oxide concentrations in the Humber estuary, UK and adjacent coastal zones. Marine Pollution Bulletin 37: 247--260. Cai W-J and Wang Y (1998) The chemistry, fluxes and sources of carbon dioxide in the estuarine waters of the Satilla and Altamaha Rivers, Georgia. Limnology and Oceanography 43: 657--668. de Angelis MA and Lilley MD (1987) Methane in surface waters of Oregon estuaries and rivers. Limnology and Oceanography 32: 716--722. de Wilde HPJ and de Bie MJM (2000) Nitrous oxide in the Schelde estuary: production by nitrification and emission to the atmosphere. Marine Chemistry 69: 203--216. Elkins JW, Wofsy SC, McElroy MB, Kolb CE, and Kaplan WA (1978) Aquatic sources and sinks for nitrous oxide. Nature 275: 602--606. Frankignoulle M, Abril G, Borges A, et al. (1998) Carbon dioxide emission from European estuaries. Science 282: 434--436. Frost T and Upstill-Goddard RC (1999) Air–sea exchange into the millenium: progress and uncertainties. Oceanography and Marine Biology: An Annual Review 37: 1--45. Law CS, Rees AP, and Owens NJP (1992) Nitrous oxide: estuarine sources and atmospheric flux. Estuarine, Coastal and Shelf Science 35: 301--314. Liss PS and Merlivat L (1986) Air–sea gas exchange rates: introduction and synthesis. In: Baut-Menard P (ed.) The Role of Air–Sea Exchange in Geochemical Cycling, pp. 113--127. Dordrecht: Riedel. Muller FLL, Balls PW, and Tranter M (1995) Processes controlling chemical distributions in the Firth of Clyde (Scotland). Oceanologica Acta 18: 493--509. Raymond PA, Caraco NF, and Cole JJ (1997) Carbon dioxide concentration and atmospheric flux in the Hudson River. Estuaries 20: 381--390. Robinson AD, Nedwell DB, Harrison RM, and Ogilvie BG (1998) Hypernutrified estuaries as sources of N2O emission to the atmosphere: the estuary of the River Colne, Essex, UK. Marine Ecology Progress Series 164: 59--71. Sansone FJ, Rust TM, and Smith SV (1998) Methane distribution and cycling in Tomales Bay, California. Estuaries 21: 66--77.

(c) 2011 Elsevier Inc. All Rights Reserved.

8

GAS EXCHANGE IN ESTUARIES

Sansone FJ, Holmes ME, and Popp BN (1999) Methane stable isotopic ratios and concentrations as indicators of methane dynamics in estuaries. Global Biogeochemical Cycles 463--474. Scranton MI, Crill P, de Angelis MA, Donaghay PL, and Sieburth JM (1993) The importance of episodic events

in controlling the flux of methane from an anoxic basin. Global Biogeochemical Cycles 7: 491--507. Seitzinger SP and Kroeze C (1998) Global distribution of nitrous oxide production and N input in freshwater and coastal marine ecosystems. Global Biogeochemical Cycles 12: 93--113.

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON L. P. Madin and G. R. Harbison, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1120–1130, & 2001, Elsevier Ltd.

Introduction

tissues, which may also deter predation. Large size also permits commensal crustaceans to live on or in the body. Thus, as we look at the diversity of gelatinous zooplankton, we should keep in mind the forces that have led to their remarkable convergence. It is impossible to deal in a short article with the entire range of phyla that have gelatinous representatives, so some of the major groups will be highlighted.

Gelatinous zooplankton comprise a diverse group of organisms with jellylike tissues that contain a high percentage of water. They have representatives from practically all the major, and many of the minor phyla, ranging from protists to chordates. The fact that so many unrelated groups of animals have independently evolved similar body plans suggests that gelatinous organisms reflect the nature of the open ocean environment better than any other group. Whether as predators or grazing herbivores, they seem particularly well adapted to life in the oligotrophic regions of the world oceans, where their diversity and abundance relative to crustacean zooplankton is often greatest. The gelatinous body plan has evolved in a world where physical parameters are relatively constant but food resources are sparse or unpredictable. Gelatinous zooplankton exhibit many common adaptations to this habitat.

Species of polycystine radiolarians form large gelatinous colonies up to several meters in length. Thousands of individual protists are embedded in a common gelatinous matrix from which their pseudopodia extend into the water. The combined efforts of individuals in the colony enable relatively large plankton (such as copepods) to be captured and ingested. In addition to the protistan members of the colony, the matrix also contains symbiotic dinoflagellates (zooxanthellae) that grow on the metabolites of the radiolarians. In turn, the radiolarians digest the zooxanthellae, so that these colonies are planktonic homologues of coral reefs.

•

Medusae

•

•

• •

Transparent tissues provide concealment in the upper layers of the ocean, an environment without physical cover. Transparency is less common below the photic zone. The high water content of gelatinous tissues gives the organisms a density very close to that of sea water. The resulting neutral buoyancy decreases the energy required to maintain depth, but may actually require more energy overall, because of drag. The environment lacks physical barriers, strong turbulence, and current shears, so that gelatinous bodies do not need great structural strength. However, fragility makes many species difficult to sample or handle, and excludes most from more energetic coastal environments. High water content and noncellular jelly permit rapid growth and large body sizes, which can act as, or produce, large surfaces for the collection of food. Relatively large size makes gelatinous animals too big to be attacked by some predators, while their high water content reduces the food value of their

Taxonomic Groups Radiolaria

The phylum Cnidaria has many gelatinous representatives, comprising various groups of medusae and the strictly oceanic siphonophores (see below). What are commonly called jellyfish are medusae belonging to three Classes of the Cnidaria — the Hydrozoa, the Scyphozoa, and the Cubozoa. Since the morphology and life history of all three groups is broadly similar, it is practical to treat them together here. There are perhaps 1000 species of hydro- and scyphomedusae, probably with more to be discovered, especially in deep or polar waters. All are carnivorous, capturing prey with specialized stinging cells, called nematocysts. A wide variety of prey is eaten by different medusae, ranging from larval forms and small crustaceans to other gelatinous animals and large fish. Many epipelagic medusae also harbor zooxanthellae, and presumably they share their resources in the same way as the polycystine radiolarians. Many of these medusae are part of a life history that alternates between a sessile, benthic, asexually reproducing polyp and a sexually

(c) 2011 Elsevier Inc. All Rights Reserved.

9

10

GELATINOUS ZOOPLANKTON

reproducing and dispersing planktonic medusa. However, many oceanic species have lost the polyp stage and evolved instead a variety of sexual and asexual reproductive mechanisms that do not require a benthic habitat. There are several classification schemes for Cnidarians; the group names given here are common usage, but these vary in different taxonomies. Anthomedusae This order of hydromedusae includes small species ranging in size from less than 1 mm to several centimeters. The umbrella is usually shaped like a tall bell, and gonads are almost always found on the sides of the central stomach. There are four radial canals connecting the stomach to a marginal ring canal. Tentacles occur in varying numbers around the umbrella margin and sometimes around the mouth. Anthomedusae alternate with polyp forms, but some also bud medusae directly (Figure 1A). Leptomedusae These medusae are generally flatter than a hemisphere. They usually have four radial canals, but sometimes eight or more, or canals that are branched. Gonads are located on the radial canals, and there may be various sense organs on the margin. The stomach is sometimes flat, and sometimes mounted on a peduncle that can be quite long. There are tentacles around the margin but not the mouth. Leptomedusae also alternate with hydroids, but some species produce new medusae by budding or fission (Figure 1B). Limnomedusae Both high and low umbrella shapes are found in this order. There are usually four radial canals, sometimes branched. Gonads are either on the stomach or on the radial canals, and there is alternation of generations. Species of limnomedusae live in brackish, fresh (one species), or marine environments. Trachymedusae These medusae in the order Trachylina do not alternate generations but develop young medusae directly from planula larvae, or by asexual budding. The umbrella is often high, with stiff mesoglea and well-developed muscle fibers. Most have eight unbranched radial canals and gonads located on them. Many trachymedusae live in deep water and are heavily pigmented (Figure 1D). Narcomedusae Also in the Trachylina, narcomedusae have direct development from planulae, with a larval stage that is often parasitic on other medusae. There are no radial canals, but the flat central stomach is very wide and, in some genera,

extends into radial stomach pouches. Tentacles are solid and stiff, and often extend aborally. Narcomedusae are common in epipelagic and mesopelagic environments; some are strong vertical migrators (Figure 1C). Coronatae This order of scyphomedusae includes mainly deep-water species. The umbrella is divided into a high central part and a thinner marginal part by a coronal groove. The margin of the bell is divided into lappets; sense organs and solid tentacles arise from the cleft between lappets. The mouth has simple lips and the gastrovascular cavity is often deeply pigmented (Figure 1G). Semaeostomae The familiar large jellyfish are mainly in this order of the Scyphozoa. The umbrella margin is divided into lappets, and bears sense organs and hollow tentacles. There is no coronal groove around the umbrella. The mouth opening is surrounded by four long oral arms, often frilled. Gonads are in folds of the subumbrella (Figure 1E). Rhizostomae Medusae in this order of the Scyphozoa are mainly coastal species and can attain large size. They lack tentacles for prey capture, and instead ingest small particles carried into numerous small mouth openings by water currents. Some species in tropical waters host intracellular symbiotic algae. Cubomedusae Medusae in the class Cubozoa also alternate with a benthic polyp form, although details of their life cycles are poorly known. Cubomedusae can be quite large, and have the most virulently toxic nematocysts of any Cnidarians. Some species are responsible for human fatalities. Cubomedusae are also unusual in possessing complex, image-forming eyes, which are not as well developed in other medusae (Figure 1F). Siphonophores

The Order Siphonophora comprises a large and diverse group of predatory Cnidarians in the Class Hydrozoa. Their complex life cycles and colonial morphology are very different from the relatively simple hydromedusae and it is practical to consider the siphonophores as a separate group. The colonial, or polygastric, phase of the life cycle is the largest and most familiar. In this stage, siphonophores consist of an assemblage of medusoid and polypoid zooids, which are budded asexually from a founding larval polyp. The colony may include a gas float, nectophores or swimming bells, and a series of stem

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON

11

Figure 1 Medusae. (A) Pandea conica, an anthomedusa about 2 cm high. (B) Aequorea macrodactyla, a leptomedusa about 10 cm in diameter. (C) Cunina globosa, a narcomedusa about 5 cm in diameter. (D) Benthocodon hyalinus, a trachymedusa about 3 cm in diameter. (E) Cyanea capillata, a semaeostome scyphomedusa which can attain 1 m in diameter. (F) Carybdea alata, a cubomedusa up to 15 cm high. (G) Atolla wyvillei, a coronate scyphomedusa up to 25 cm diameter. (All photographs by L. P. Madin.)

(c) 2011 Elsevier Inc. All Rights Reserved.

12

GELATINOUS ZOOPLANKTON

groups made up of feeding polyps and tentacles. In some siphonophores the stem groups break off as dispersal and sexually reproductive stages called eudoxids. The colony can be thought of as an overgrown, polymorphic juvenile stage that eventually bears the sexually reproductive adults. These are small medusoid zooids called gonophores, which produce gametes. Siphonophores range in size from a few millimeters to over 30 m in length, and occur throughout the water column. All are predators on other small zooplankton, and many genera are known to be luminescent. The colonies are fragile, and usually break up into their various units when collected in plankton nets. For this reason, much of the taxonomy is based on the morphology of the pieces, principally nectophores, and the appearance of the intact colonial stage is not always known. The Order Siphonophora is divided into three suborders and 15 families. Cystonectae This suborder includes siphonophores that possess a float but no swimming bells, so they are at the mercy of ocean currents. The Portuguese manof-war is the most familiar example. It has a float so large that the animal rests on the surface, but most cystonect species have smaller floats and are wholly submerged. Cystonects have virulent nematocysts and capture large, soft-bodied prey such as fish and squids (Figure 2A). Physonectae These siphonophores have more complex colonies, comprising a small apical float, numerous swimming bells that form a nectosome, and a stem containing several groups of gastrozooids, tentacles, bracts, etc. The stem typically contracts when the animal is swimming, and then relaxes so that the stem and tentacles extend to maximum length for fishing. This group is a major contributor to the deep scattering layer in many regions of the ocean. The largest siphonophores (the Apolemiidae, over 30 m long) are found in this group. Physonects prey mainly on small zooplankton, and many species are strong swimmers and vertical migrators. (Figure 2C). Calycophorae In this group, which contains the largest number of species, the float is absent and the nectophores are reduced, usually to two. A sequence of stem groups is budded and breaks free as eudoxids. Calycophorans are the most diverse, widely distributed, and abundant siphonophores. They catch small zooplankton and, when feeding, their tentacles form complex three-dimensional structures in the water, reminiscent of spider webs (Figure 2B).

Ctenophores

Ctenophores are exclusively marine gelatinous animals all but a few of which are holoplanktonic. Although they superficially resemble the Cnidaria, morphological and molecular studies indicate that Cnidarians and ctenophores are not closely related. Ctenophores are predators that use tentacles equipped with ‘glue’ cells or colloblasts to capture prey. The name ‘ctenophore’ is Greek for ‘comb bearer,’ referring to the comb-like plates of fused cilia that are used for propulsion. All ctenophores initially have eight meridional rows of comb plates, although in some groups these are lost or reduced during development. The vast majority of ctenophore species fall into six orders. Cydippida This group contains many species with paired tentacles that exit the body through tentacle sheaths. Species in the family Pleurobrachiidae catch prey ranging from small crustaceans to fish, while members of the Lampeidae feed mainly on large gelatinous animals like salps. The members of one species of cydippid, Haeckelia rubra, eat medusae, and retain the nematocysts of their prey (‘kleptocnidae’) for defensive use in their own tentacles. Before this behavior was known, these nematocysts were considered strong evidence for a close relationship between Cnidarians and ctenophores (Figure 2F). Platyctenida This group is primarily benthic and is distributed widely from the Arctic to the Antarctic. Members of the family Ctenoplanidae have comb rows as adults and are found in the plankton in the Indo-Pacific; all other species in the order lose their comb rows as adults, and live primarily as creeping benthic organisms. Platyctenes have functional tentacles that capture prey. Thalassocalycida This order contains a single species, Thalassocalyce inconstans, which lives in the midwater zone. It superficially resembles a medusa in overall shape, but can easily be distinguished by its eight comb rows and paired tentacles. Lobata Members of this order all have oral lobes and auricles, specialized structures that are used in feeding. Most lobates move through the water with their oral lobes widely spread to form a sort of basket. Small prey, such as crustaceans, are trapped on the mucus-covered oral lobes and tentilla, which stream over the body or extend onto the oral lobes. Ctenophores in the family Ocyropsidae lack

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON

13

Figure 2 Siphonophores and Ctenophores. (A) Rhizophysa filiformis, a cystonect siphonophore up to 2 m long. (B) Sulculeolaria sp. a calycophoran siphonophore up to 1 m long. (C) Physophora hydrostatica, a physonect siphonophore about 10 cm high. (D) Ocyropsis maculata, a lobate ctenophore about 8 cm in diameter. (E) Beroe cucumis, a beroid ctenophore up to 25 cm long. (F) Mertensia ovum, a cydippid ctenophore about 4 cm high. (Photographs A, C, D, F by G. R. Harbison; B, E by L. P. Madin.)

(c) 2011 Elsevier Inc. All Rights Reserved.

14

GELATINOUS ZOOPLANKTON

tentacles, and capture prey by enclosing them in their muscular oral lobes (Figure 2D). Cestida These ctenophores are shaped like long, flat belts. They appear to be related to the Lobata, but lack oral lobes and auricles. There are only two genera (Cestum and Velamen) in one family (Cestidae). The comb rows extend along the aboral edge of the ribbonlike body, propelling the animal with the oral edge forward. Small prey are captured by the fine branches of the tentacles that lie over the flat sides of the body. Cestids are characteristic of oceanic, epipelagic environments. Beroida Beroids lack tentacles altogether. Their large stomodaeum occupies most of the space in the body. All beroids are predators on other ctenophores, and occasionally salps. They capture prey by engulfing them, and can bite off pieces of the prey with specialized macrocilia located immediately behind the mouth (Figure 2E).

others of which (Pseudothecosomata) have gelatinous shells and tissues. There are over 30 species of euthecosome pteropods, in two families, the Limacinidae and the Cavoliniidae. Thecosome pteropods feed by collecting particulate food on the surface of a mucous web or bubble, produced by mucus glands on the wings, and held above the neutrally buoyant and motionless animal. The mucus is periodically retrieved and ingested along with adhering particles, then replaced by a newly secreted web. Some cavoliniids have brightly colored mantle appendages that may aid in maintaining neutral buoyancy or serve as warning devices to predators. When disturbed, animals lose their neutral buoyancy and rapidly sink (Figure 3A). The Pseudothecosomata includes three families, the Peraclididae (one genus), the Cymbuliidae (three genera), and the Desmopteridae (one genus). Pseudothecosomes are larger than euthecosomes, and their mucous webs are correspondingly larger, reaching over a meter across in Gleba cordata (Figure 3C).

Heteropods

The Phylum Mollusca contains many gelatinous representatives, and the gelatinous body plan has apparently arisen independently in several groups. The Heteropoda is a superfamily of prosobranch gastropods that includes the families Atlantidae, Carinariidae, and Pterotracheidae. Heteropods are visual predators with well-developed eyes and a long proboscis containing a radula. Atlantid heteropods have thin, flattened shells into which they can completely withdraw their bodies. They feed on small crustaceans and other mollusks. The family Carinariidae includes eight species in three genera, Carinaria, Pterosoma, and Cardiapoda. These heteropods have a greatly reduced shell, enclosing only a small fraction of the body. Carinariids feed primarily on other gelatinous organisms, such as salps, doliolids and chaetognaths. In the family Pterotracheidae, with two genera — Pterotrachea (four species) and Firoloida (one species) — the shell is completely absent (Figure 3D). Pteropods

This molluskan group comprises two orders in the gastropod subclass Opisthobranchia. The foot in pteropods is modified into two wingshaped paddles responsible for swimming; their fluttering gives rise to the common name for pteropods, sea butterflies. Thecosomata This group contains the shelled pteropods, some of which (Euthocosomata) have calcareous shells and are not truly gelatinous, and

Gymnosomata Members of this order are poorly known, largely because they have no shells and contract into shapeless masses when preserved. Most species live in the deep sea, and only a few of the approximately 50 species have been observed alive. Gymnosomes appear to be highly specialized predators on particular species of thecosome pteropods. The order is divided into two suborders, the Gymnosomata and the Gymnoptera. The four families of the Gymnosomata include the Pneumodermatidae (seven genera and 22 species) with sucker-bearing arms similar to cephalopod tentacles; the Notobranchaeidae (one genus and eight species), with suckerless feeding arms called buccal cones; the Clionidae (eight genera and 16 species); and the Cliopsidae (two genera and three species) (Figure 3B). There are two families in the Gymnoptera, the Hydromylidae (one genus, one species) and the Laginiopsidae (one genus, one species). These groups are very different from each other and from other gymnosome pteropods, and some may not actually belong in the Order Gymnosomata. Cephalopods

Although many cephalopods are active, muscular swimmers, there are several gelatinous and/or transparent genera. The family Cranchiidae is composed entirely of gelatinous species, including the genera Taonius, Megalocranchia, and Teuthowenia. These relatively large, slow-moving squids probably

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON

15

Figure 3 Mollusks and Polychaete. (A) Cuvierina columnella, a euthecosome pteropod about 1 cm high. (B) Clione limacina, a gymnosome pteropod about 2 cm high. (C) Corolla spectabilis, a pseudothecosome pteropod about 10 cm in diameter, with mucous web in background. (D) Carinaria sp. a heteropod about 10 cm long. (E) Teuthowenia megalops, a cranchiid cephalopod about 10 cm long. (F) Alciopid polychaete worm, up to 1 m long. (Photographs A, B, C, E, F by G. R. Harbison; D by L. P. Madin.)

(c) 2011 Elsevier Inc. All Rights Reserved.

16

GELATINOUS ZOOPLANKTON

capture prey through stealth rather than active pursuit. Vitreledonelliid octopods are also gelatinous (Figure 3E). Polychaete Worms

Two major groups of planktonic polychaetes are gelatinous, the Alciopidae and the Tomopteridae. Both are in the order Phyllodocida, although they are probably not closely related. Alciopids are characterized by well-developed eyes with lenses. Many have ink glands along the sides of their bodies, which may function analogously to the ink glands of cephalopods. Their habits are poorly known, but they may feed on gelatinous prey. Alciopids may attain lengths of nearly a meter. Tomopterids do not have well-developed eyes, but probably use chemoreception to locate prey. Some release luminous secretions from glands along their body when disturbed. Deep-sea tomopterids may be 25 cm long, but most shallow species are much smaller. (Figure 3F). Crustaceans

Although arthropods cannot really be considered gelatinous because of their exoskeletons, there are some examples of very transparent bodies, presumably also an adaptation for concealment. The most notable examples are species of the hyperiid amphipods Cystisoma and Phronima. Species of Cystisoma are large and transparent, and the enormous retinas of the compound eye are lightly tinted. Although the retinas of species of Phronima are darkly pigmented, the rest of the body is transparent. It is likely that the transparency of these species is a form of protective coloration, since they live on transparent gelatinous hosts. Holothurians

Although the majority of holuthurian species are rather sedentary benthic deposit feeders, there are several deep-sea genera of swimming or drifting holothurians with gelatinous bodies. Species in the genera Peniagone and Enypniastes feed on bottom deposits, but can swim up into the water column. The genus Pelagothuria appears to be wholly pelagic, with a morphology that suggests it collects and feeds on sinking particulate matter. Few pelagic holuthurians have been observed alive and little is known of their life history or behavior. Pelagic Tunicates

The subphylum Urochordata includes two classes of pelagic tunicates, the Thaliacea and the Larvacea or

Appendicularia. Thaliaceans (including the orders Pyrosomida, Doliolida, and Salpida) are relatively large animals with more or less barrel-shaped bodies. They pump a current of water through their bodies and strain phytoplankton and other small particulates from it with a filter made of mucous strands. The same current provides jet propulsion. Thaliaceans have complex life cycles with alternating generations and multiple zooid types. The class Larvacea comprises a single order of small organisms that filter food particles using an external mucous structure called a house. Both Thaliaceans and Larvaceans are widely distributed in the oceans, and are sometimes extremely abundant. Pyrosomida Pyrosomes form colonies made up of numerous small ascidian-like zooids embedded in a stiff gelatinous matrix or tunic. The colony is tubular, with a single terminal opening. Water is pumped by ciliary action through each zooid, and suspended food particles are retained on the branchial basket within the body. The excurrent water from each zooid passes into the lumen of the colony, forming a single exhalent current that provides jet propulsion. Most pyrosome colonies are a few centimeters to a meter in length, but colonies of at least one species can attain lengths of 20 m. Doliolida This order of the Thaliacea comprises six genera and 23 species of small (2–10 mm), barrel-shaped animals with circumferential muscle bands. The filter feeding mechanism is similar to that of pyrosomes, with currents generated by ciliary beating passing through a mucous net supported on the branchial basket. The life cycle involves five asexual stages and one sexual stage, several of which occur together as parts of large colonies of thousands of zooids. These colonies may attain lengths over 1 m, but are fragile and are rarely collected intact. In most genera of doliolids, the life cycle begins with a sexually produced larva, which becomes the oozooid stage. This stage feeds initially, but then begins budding off the trophozooid and phorozooid stages, thus forming the colony. During this process the oozooid loses its branchial basket and gut, and transforms into the ‘old nurse’ stage, whose function is to swim by jet propulsion and pull the attached colony along behind it. Contractions of the body muscles produce short exhalent pulses that move the colony rapidly. The trophozooids in the colony filter-feed to support themselves and the nurse. The phorozooids grow attached to the colony, but then break free to lead independent lives and produce asexually a

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON

small group of gonozooids. These eventually break free from the phorozooid, and become the sexually reproducing stages (hermaphrodites?) that produce the larvae and begin the whole cycle again (Figure 4B). Salpida This order (with 12 genera and about 40 species) is of larger filter-feeding animals, also with

17

circumferential muscle bands. The salps alternate between two forms, an asexually budding solitary (oozooid) stage and a sexually reproducing aggregate (blastozooid) stage. The aggregate salps usually remain connected together in chains or whorls of various types. Swimming is by jet propulsion, produced by a pulsed water current generated by rhythmic contraction of body muscles.

Figure 4 Pelagic Tunicates. (A) Megalocercus huxleyi, a larvacean of about 5 mm body length, house length about 4 cm. (B) Dolioletta gegenbauri, portion of a colony showing gastrozooids and phorozooids, individuals 2–5 mm long, colonies up to 1 m. (C) Salpa maxima, solitary generation salp, up to 25 cm long. (D) Salpa maxima, chain of aggregate generation salps; orange dots are guts of salps; individuals are to 15 cm, chains up to 10 m long. (E) Pegea socia, aggregate generation salp with attached embryo of solitary generation; aggregate 7 cm, embryo about 1 cm. (F) Traustedtia multitentaculata, solitary generation salp with appendages of uncertain function, about 3 cm long. (All photographs by L. P. Madin.)

(c) 2011 Elsevier Inc. All Rights Reserved.

18

GELATINOUS ZOOPLANKTON

Food particles are strained from the water passing through the body cavity by a mucous filter, which is continuously secreted and ingested. The individual animals range in size from 5 to over 100 mm, and chains can be several meters long (Figures 4C–F). Larvacea This class (also called Appendicularia) is divided into three families (with 15 genera and 70 species) of small (1–10 mm) animals consisting of a trunk and long, flat tail. Larveaceans are also filter feeders on small particulates but are unique among tunicates in the use of an external concentrating and filtering structure called the house. The house surrounds the animal, and contains a complex set of channels and filters made of mucous fibers and sheets. Water is pumped into the house by the oscillation of the larvacean’s tail; the exhalent stream provides slow jet propulsion in some species. Particles are sieved from the flow as it passes through the internal filter; they accumulate and are aspirated at intervals into the pharynx of the larvacean via a mucous tube. The complex house is formed as a mucous secretion on the body of the larvacean, produced by specialized secretory cells. It is inflated with sea water, pumped into it by action of the tail, until it attains its full size, with all the internal structures expanding in proportion. Houses eventually become clogged with particulates and fecal pellets, and are then jettisoned. The larvacean expands a new house (there may be several house rudiments on its body, awaiting expansion) and resumes filter feeding. The abandoned houses can be an important source of marine snow and serve as food for various planktonic scavengers (Figure 4A).

Ecology of Gelatinous Zooplankton Gelatinous zooplankton are found in all of the oceans of the world, from the tropics to polar regions. They also occur at all depths, and many of the largest and most delicate species have been collected in recent years from the mesopelagic and bathypelagic parts of the ocean. The absence of turbulence in the deep sea probably allows these species to attain such large sizes, but there are also robust species that thrive in surface and coastal environments. Examples include the Portuguese man-of-war (Physalia physalis), which lives at the air–water interface and can ride out hurricanes, and the ctenophore Mnemiopsis leidyi, which lives in estuaries with strong tidal currents and turbulence. In general, gelatinous organisms have been rather neglected by zooplankton ecologists, primarily because their delicacy makes them difficult to sample

and study. Most are damaged or destroyed in conventional plankton nets, and many deep-water siphonophores and ctenophores are too delicate to be captured intact even with the most gentle of techniques. Much recent progress in understanding their biology has been based on in situ methods of study using SCUBA diving, submersibles, or remote vehicles. These methods permit observation of undisturbed behavior and collection of intact living specimens. Advances in culture techniques and laboratory measurements have improved our understanding of energetics, reproduction, and life history of some species, but most remain only partially understood. Gelatinous animals occupy every trophic niche, ranging from primary producers (symbiotic colonial radiolaria) to grazers (pteropods and pelagic tunicates) and predators (medusae, siphonophores, and ctenophores). In all these niches, the gelatinous body plan confers advantages of size and low metabolic costs. In addition to attaining large sizes with relatively little food input, gelatinous organisms such as medusa and ctenophores are able to ‘de-grow’ when deprived of food. Metabolic rates remain unchanged, and the animal simply shrinks until higher food levels allow it to resume growth. This energetic flexibility is probably important to the success of gelatinous species in the oligotrophic open ocean and deep sea. Many species of medusae and siphonophores, for example, appear able to survive at low population densities spread over very large areas. In other cases the efficiency of their feeding, growth, and reproduction allows gelatinous species to outcompete other types of zooplankton and form dense populations over large areas, which can have considerable impact on ecosystems. A dramatic recent example was the accidental introduction of the ctenophore Mnemiopsis leidyi into the Black Sea from the eastern seaboard of the Americas. In the late 1980s, this ctenophore reproduced in prodigious quantities, and the resulting predation on zooplankton and larval fishes led to the collapse of pelagic fisheries in the Black Sea. These fisheries have to some extent recovered, but seasonal blooms of this ctenophore continue to occur in the Black Sea, just as they do on the eastern shores of the Americas. Reports of Mnemiopsis in the Caspian Sea suggest that the pattern may be repeated. Many other gelatinous species form dramatic seasonal blooms, such as the medusa Chrysaora quinquecirrha in the Chesapeake Bay, the salp Thalia democratica off Florida, Georgia, and Australia, and the medusa Pelagia noctiluca in the Mediterranean. In the Southern Ocean, immense populations of the salp Salpa thompsoni alternate with those of the

(c) 2011 Elsevier Inc. All Rights Reserved.

GELATINOUS ZOOPLANKTON

Antarctic krill Euphausia superba. The formation of large aggregations through rapid reproduction appears to be a common strategy for taking advantage of favorable conditions. Dense populations are sometimes further concentrated by wind or current action, or are transported close to the coast from their normal habitats farther offshore. The combination of rapid growth and advection can cause the sudden appearance of swarms of medusae, ctenophores, or salps in coastal waters. Although these blooms may sometimes have serious or even catastrophic effects on other organisms, including fisheries or human activities, they are a natural part of the life histories of the species, and not events for which remedial action is needed, or even possible. Gelatinous zooplankton are normal components of virtually all planktonic ecosystems. They are among the most common and typical animals in the oceans, whose biology and ecological roles are now becoming better understood.

19

See also Plankton. Zooplankton Sampling with Nets and Trawls.

Further Reading Bone Q (1998) The Biology of Pelagic Tunicates. Oxford: Oxford University Press. Harbison GR and Madin MP (1982) The Ctenophora. In: Parker SB (ed.) Synopsis and Classification of Living Organisms. New York: McGraw Hill. Lalli CM and Gilmer RW (1989) Pelagic Snails. Stanford, CA: Stanford University Press. Mackie GO, Pugh PR, and Purcell JE (1987) Siphonophore biology. Advances in Marine Biology 24: 98--263. Needler-Arai M (1996) A functional biology of Scyphozoa. London: Chapman and Hall. Wrobel D and Mills CE (1998) Pacific Coast Pelagic Invertebrates: A Guide to the Common Gelatinous Animals. Monterey: Sea Challengers and Monterey Bay Aquarium.

(c) 2011 Elsevier Inc. All Rights Reserved.

GENERAL CIRCULATION MODELS G. R. Ierley, University of California San Diego, La Jolla, CA, USA

•

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1130–1134, & 2001, Elsevier Ltd.

Introduction A general circulation model (GCM) of the ocean is nothing more than that – a numerical model that represents the movement of water in the ocean. Models, and more particularly, numerical models, play an ever-increasing role in all areas of science; in geophysics broadly, and in oceanography specifically. It was perhaps less the early advent of supercomputers than the later appearance of powerful personal workstations (tens of megaflops and megabytes) that effected not only a visible revolution in the range of possible computations but also a more subtle, less often appreciated, revolution in the very nature of the questions that scientists ask, and the answers that result. The range of length scales and timescales in oceanography is considerable. Important dynamics, such as that which creates ‘salt fingers’ and hence influences the dynamically significant profile of density versus depth, takes place on centimeter scales, while the dominant features in the average circulation cascade all the way to basin scales of several thousand kilometers. Timescales for turbulent events, like waves breaking on shore, are small fractions of seconds, while at the opposite end, scientists have reliably identified patterns in the ocean with characteristic evolution times of order several decades.1 Most simply, a ‘model’ is no more than a mathematical description of a physical system. In the case of physical oceanography, that description includes the following elements:

•

The momentum equation (F ¼ mv_ , but expressed in terms appropriate to a continuous medium), or often in its place a derivative form, the ‘vorticity equation’.

1 It is useful to distinguish extrinsic evolution times, which span geologic time, from intrinsic variation, which characterizes an isolated ocean and atmosphere, thus neglecting such secular influences as orbital variation, change in the earth’s rotation rate, variation in the solar constant, etc. Although not all causes of variability have been identified, it is possible that even documented evolution over thousands of years may reflect the latter, intrinsic, variability.

20

• • •

An equation to express the principle of mass conservation. A heat equation, which describes the advection (carrying by the fluid) and diffusion of temperature. A similar ‘advection–diffusion’ equation for salinity. An equation of state, which relates the pressure to the density.

Although it may, after the fact, sound obvious, it took scientists many years to appreciate the full role of rotation – which enters Newton’s law through a latitude dependent Coriolis force – in generating the observed large scale circulation of ocean. Elements of these equations can become quite complicated. For example, momentum in the upper ocean is imparted by a complex, not yet fully understood, process of wind–wave interaction. One has to choose whether to represent the action of the wind kinematically, which means that the spatial and temporal variability of the wind must be given beforehand, or dynamically, in which case we must solve not only the set of equations above, but a similar set that describes the simultaneous evolution of the atmosphere. Clearly the dynamical case is the more ‘realistic’ of the two, but the price is a substantially more involved computation. It is issues such as these that force one to a choice that often pits understanding and intuition on the one hand against verisimilitude and complete dynamical consistency on the other. As the common aim of most large-scale models is to generate results of maximum realism, it usual that the models are frequently corrected, or ‘steered’, on the fly by extensive use of data assimilation. It is not merely a matter of slight refinement: no large-scale model (GCM) is yet sufficiently robust that it can be used for forecasting without considerable input of such observationally derived constraints; in the oceanic component, for example, the need to force model agreement at depth by continuously ‘relaxing’ the solution to the smoothed data of the Levitus atlas (a world-wide compendium of data from many sources, smoothed and interpolated onto a regular grid). How much of this fragility is because of explicit defects (faulty ‘subgrid scale modeling’ and explicit omissions in the physics e.g., neglect of wave breaking) and how much is due to discrete numerical implementation inconsistent with plausible continuous equations remains an open question. Their limitations notwithstanding, large numerical models are one vital means by which we grapple with questions about global warming and a host of other

(c) 2011 Elsevier Inc. All Rights Reserved.

GENERAL CIRCULATION MODELS

environmental issues that affect the way that both we and future generations will live.

instead we reach a tentative conclusion of not proven wrong (yet!). While it may stimulate conjecture, and have other worthy ends, the idea of introducing artificial parameters for the express purpose of manufacturing close agreement with reality is nonetheless formally antithetical to the paradigm of testing independent, quantitative predictions. Backward compatibility. As with confinement of form, the idea of compatibility achieves a purity in mathematics that is not to be expected of the physical sciences. Each new bit of mathematics must fit perfectly into the entire edifice of results already discovered (or invented, as you will). Commonly, though not without exception, in the physical sciences, newer theories are seen to encompass the older theories as special or limiting cases. We speak, for example, of the ‘classical limit’ as a means of recovering prerelativistic or prequantum results. Indeed, it is only in light of Einstein’s theory of special relativity that we can understand the limitations of Newton’s law, which we now understand more fully as not a law, but a limiting approximation. Oceanography has families of theories, each nesting one within the next, like a series of oceanographic Matryoshka dolls, the innermost of which is often the theory of ‘quasigeostrophy,’ which dates from the 1950s.

Models in Theory and Practice Historically, computers were initially so limited that the questions posed were often, in effect, slight extensions of preexisting analytic queries. To that extent, such studies continued to conform (at least in principle) to what we may identity as four basic building blocks of most theories, which, in aiming to describe physical reality become subject to constraints. Classical modes to which physical laws are found to conform generally2 include the following.

•

•

•

Expression in quantity, extent, and duration. The language that offers itself as encompassing all distinct physical conditions and all meaningful physical relations is fundamentally that of mathematics. As others before and since, the great mathematician Eugene Wigner too had a stab at explaining what, in an eponymous essay, he termed ‘the unreasonable effectiveness of mathematics in the physical sciences’. It remains a conundrum. Confinement of form. This attains its purest expression in the Platonic view that mathematical entities are not invented, but discovered, and as such have a prior, if not physical, existence. In extension to physical theories, it corresponds to the belief that there is some true, ultimate, equation association with any natural phenomenon. One important respect in which this idea must be tempered in application to fluids is the mathematical demonstration that a variety of different microscopic laws for interaction may all yield the same generic macroscopic law applicable to behavior at large space scales or timescales. If one’s aim is solely to understand the latter, then although for a given problem we might presume that there is indeed a precise, if complicated, microscopic law, one’s effort might more profitably be spent understanding the passage to the large-scale limit. Falsifiability. Implicit in the progress of science is the idea that one makes and then tests hypotheses. But unlike in mathematics, in the physical sciences we cannot show the hypothesis is correct, only that it is wrong. A hypothesis is never vindicated,

2 We speak gently here, since to insist that those four are always either necessary or sufficient would require that we introduce a theory about theories: a metatheory. But we have no notion of how rigorously to evaluate such ideas!

21

•

The unavoidable adoption and resulting sensitivity of GCMs to ad hoc parameterizations (e.g., subgrid scale modeling) or necessity set them apart from the traditional pursuit of the scientific method. This distinction was (presciently) appreciated at least by the early 1960s and it changes, or ought to change, one’s view of such models as rigorous arbiters of precise truths. And yet while it is true that numerical experiments with GCMs are merely suggestive rather than truly predictive of future evolution of the ocean, the sheer lack of experimental data, to say nothing of the lack of a control, means that theoretical ideas are often assessed on the basis of their success in explaining strictly numerical experiments. As computers became more powerful it was natural to press for the most realistic model runs possible. And, because the growth in computing power was increasingly realized through distributed3 as well as mainframe (super), computing, the school of ‘kitchen sink’ models, which started as a specialized branch off the mainstream of oceanography – largely

3

Both virtually, through high speed and increasingly transparent networks, and physically, through desktop workstations of considerable power.

(c) 2011 Elsevier Inc. All Rights Reserved.

22

GENERAL CIRCULATION MODELS

limited in participation to those in close physical proximity to two or three central machines – became a powerful tributary in its own right: an autonomous discipline within oceanography, which naturally began to evolve its own criteria for relevance.

Limits on Numerical Models We spoke of two sources of error common to large numerical models: difficulties in numerical implementation, and poorly modeled or unrepresented physical processes. In this section we consider specific instances of each, starting with an abstract mathematical point of view. But note that if we could with the wave of hand dispense with these two issues (which one supposes are in principle tractable), the fact that we do not know the exact physical state of the ocean inevitably increases the uncertainty of the results. Moreover, even were we given that exact state at some instant in time, the intrinsic and spontaneous genesis of disorder in such a physical system must forever constrain our predictive power. There are two key mathematical features of the basic momentum (or vorticity) equation which bear comment: conservation and dissipation. The nonlinear term is fundamental to the initiation and sustenance of turbulence and by itself strictly conserves energy. (Other terms introduce explicit dissipation.) In addition, in two dimensions the term conserves ‘enstrophy’ (the square of the vorticity) and in three, both ‘circulation’ and ‘helicity’ (the dot-product of velocity and vorticity). While one might hope for all such conservation properties to be preserved in numerical implementations, some large models, often those based on curvilinear – as opposed to Cartesian – coo¨rdinates – manage only to conserve energy. Beyond the conserved quantities associated with the nonlinear term, which include the energy and, in general, the so-called ‘Noether invariants,’ there is a more subtle property associated with the exact (continuous) equation: its associated ‘multisymplectic geometry.’ Recent mathematical advances make it possible for a discrete numerical model to preserve such structure exactly, though as yet such improvements have not been incorporated into any working GCM. Is it quantitatively important that we do so when basic fluid processes, such as convection, are as yet only crudely modeled? Until the experiment is tried, no one can say. But it is pertinent to note that a similar (that is, ‘symplectic’) refinement is critical for a numerical solution of sufficient accuracy that one can decide whether planetary orbital motions are chaotic on astronomical timescales.4 Finally, strictly speaking, not only ambitious models, but even more confined ‘process’ models, rest

upon a not yet wholly secure foundation: it is still an open research question whether the basic equation of fluid mechanics (the Navier–Stokes equation) is itself ‘globally well-posed’ in three dimensions.5 (It is widely believed to be so, but belief comes cheap. A proof, however, is worth one million dollars(!) – one of seven prizes in a competition recently announced by the Clay Mathematics Institute.) Even overlooking such foundational matters on which mainly mathematicians would cavil about numerical models, the last three of the four principles above are often violated in more apparent ways in the application and development of present-day models. We illustrate this divergence from traditional norms with a few representative examples to emphasize the sometimes causal (not casual!) link between models and ‘reality.’ In delineating the borders of the known, the unknown, and the unknowable, it is important to discriminate between deduction and rationalization as competing processes for exploring and explaining those borders.

•

Although GCM simulations with a viscosity approaching that of water are at present inconceivable, at least as a thought experiment it is worth bearing in mind that those are the numerical results we would in principle compare against observation to assess a given model.6 Short of that, GCMs use various formulations that ostensibly mimic the dynamical effects of the unresolved scales of motions. In the simplest instance, this amounts to choosing a numerical viscosity several orders of magnitude larger than that of sea water. But often the value is dictated by purely heuristic numerical considerations: it is set at a threshold value, any decrease below which leads to rapid

4

On dissipation, a deep, though perhaps insufficiently appreciated, mathematical result is that the solution of a parabolic, dissipative system quickly collapses onto a finite-dimensional ‘attractor’. This is remarkable. If you think of assigning a point to every one of N molecules of water in the ocean, and tracking the velocity and position of each, the associated ‘phase space’ – just a record of that evolution – has dimension 6N. Because the momentum equation is derived on the basis that water is an infinitely divisible continuum, strictly we need to imagine that N approaches infinity. Nonetheless, even in that infinite-dimensional phase space, it remains true that the solution confines itself to only a finite, if quite large, portion. 5 Hadamard introduced the notion of ill-posedness of partial differential equations. A problem is well-posed when a solution exists, is unique, and depends continuously on the initial data. It is ill-posed when it fails to satisfy one or more of these criteria. 6 There is a curious division among physical oceanographers as to whether the large-scale flow we observe is, in the end, actually sensitive to the precise value of the viscosity of sea water. Predictably, there are two camps: yes and no.

(c) 2011 Elsevier Inc. All Rights Reserved.

GENERAL CIRCULATION MODELS

numerical blowup. It cannot be said to be satisfactory feature that a basic parameter is set not by independent dynamical considerations but for stability reasons, and those pertaining solely to the discrete form of the equations. (The solution to the continuous equation would not blow up!) At times, not only the coefficient, but the actual form of the diffusive operator is adjusted. As above, usually the motivation is intrinsically numerical, so it is not surprising that a catalogue of the various choices shows some, for example, that create artificial sources (or sinks) of vorticity in the flow. Others, subject to the given boundary conditions, do not make mathematical sense in a region where the fluid depth tends to zero (like Atlantic City). Oceanographers, unfortunately, do not have the luxury of extensive laboratory measurements from which their dissipative parameterizations can be calibrated. A program of direct observation in particular regions where dissipation is thought to be significant is just getting under way. While such measurements will help reveal deficiencies in present formulations, the largest GCMs will probably rely on a solely heuristic approach for some time to come.

•

An important, but numerically unresolved, process in the ocean is that of convection, which typically occurs at small scales in, for example, localized regions of intense surface cooling. The overall thermal structure of the ocean is sensitive to this, a means by which ‘bottom water’ is formed; a cold, relatively less salty mass that constitutes the deep Atlantic, for example. Because the horizontal resolution is too coarse to encompass the sinking motions, various schemes have been devised to mimic that effect. It has been shown that one of the most common of these leads paradoxically to unacceptable physical (and mathematical) behavior as resolution is improved; it has no verifiable correspondence to a realizable physical process. The temptation with a model that has been extensively tuned to give plausible answers for other observables is to leave well enough alone. Unhappily, a model with one or more such elements whose limits are ill-defined or nonexistent must inevitably produce end results whose errors are typically an opaque mix of effects: some physical, some numerical, some mathematical. In such circumstances, the program of falsifiability of the physical components is apt to be fatally compromised.

The point is not that one should immediately dispense with all ad hoc parametrizations; excepting those that are simply mathematically ill-posed from

23

the outset, theorists generally do not have better alternatives to suggest. But one should always bear in mind the degree to which numerical simulations are sensitive to these components, and seek independent ways in which to constrain their parameters, in isolated settings that test the limits of prediction against known measurements or, failing that, at least against fully converged, adequately resolved simulations of a local or regional character. If, within the acceptable parameter range identified, it is found that the original model no longer gives adequate large-scale predictions, then there are more basic problems to be addressed.

Summary From the numerical side, no computer improvements that can be seen on the horizon seem likely to make reasonably ambitious GCMs accessible to rigorous and extensive parametric and numerical exploration, a prerequisite to their complete understanding. From the mathematical side, it seems to be our fundamental ignorance about turbulence that most severely restricts the range of our grasp, leaving us with an often painfully narrow range of computations to which theoretical remarks can be significantly addressed. For these structural reasons, the gulf between theory and much numerical modeling will probably continue to widen for the foreseeable future, and thus there may grow to be—indeed some would say it already exists—a division akin to C.P. Snow’s ‘Two Cultures’. All the cautions about GCMs notwithstanding, they have become an integral part of the study of physical oceanography. With due regard for the novel capacities and limitations of numerical models, such scientific progress as we do make will more and more often hinge upon judicious computation.

See also Deep Convection. Double-Diffusive Convection. Forward Problem in Numerical Models. Wind Driven Circulation.

Further Reading The literature on ocean modeling is not yet productive of definitive treatises, in large measure because the field is yet young and rapidly evolving. Thus in lieu of textbooks or similar references, the reader is directed to the following series of articles. For some predictions on the perennially intriguing issue of what improvements in large-scale modeling may be

(c) 2011 Elsevier Inc. All Rights Reserved.

24

GENERAL CIRCULATION MODELS

driven by plausible increases in computing speed with massively parallel machines see Semtner A. Ocean and climate modeling Communications of the ACM 43 (43): 81–89. For a look back at the history of one of the single most influential models in physical oceanography, see A.J. Semtner’s Introduction to ‘A numerical method for the study of the circulation of the World Ocean’, which accompanies the reprinting of Kirk Bryan’s now classic 1969 article of the title indicated. This pair appears back-toback, beginning on page 149, in Journal of Computational Physics, (1997) 135 (2).

General readers may wish to consult the following succinct review, accessible to a broad audience: Semtner AJ (1995) Modeling ocean circulation Science 269 (5229): 1379–1385. Finally, for those readers desiring a more in depth appreciation of modeling issues and their implications for specific features of the large scale circulation, consult the careful review. McWilliams JC (1996) Modeling the oceanic general circulation. In: Lumley JL, Van Dyke M. Read HL (eds) Annual Review of Fluid Mechanics Vol. 28, pp. 215–248 Palo Alto, CA: Annual Reviews.

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMAGNETIC POLARITY TIMESCALE C. G. Langereis and W. Krijgsman, University of Utrecht, Utrecht, The Netherlands Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1134–1141, & 2001, Elsevier Ltd.

maintained by convective fluid motion. At the surface of the Earth, the field can conveniently be described as a dipole field, which is equivalent to having a bar magnet at the center of the Earth. Such a dipole accounts for approximately 90% of the observed field; the remaining 10% derives from

Introduction Dating and time control are essential in all geoscientific disciplines, since they allows us to date, and hence correlate, rock sequences from widely different geographical localities and from different (marine and continental) realms. Moreover, accurate time control allows to understand rates of change and thus helps in determining the underlying processes and mechanisms that explain our observations. Biostratigraphy of different faunal and floral systems has been used since the 1840s as a powerful correlation tool giving the geological age of sedimentary rocks. Radiometric dating, originally applied mostly igneous rocks, has provided numerical ages; this method has become increasingly sophisticated and can now — in favorable environments — also be used on various isotopic decay systems in sediments. We are concerned with the application of magnetostratigraphy: the recording of the ancient geomagnetic field that reveals in lava piles and sedimentary sequences, intervals with different polarity. This polarity can either be normal, that is parallel to the present-day magnetic field (north-directed), or reversed (south-directed) (Figure 1). As a rule, it appears that these successive intervals of different polarity show an irregular thickness pattern, caused by the irregular duration of the successive periods of either normal or reversed polarity of the field. This produces a ‘bar code’ in the rock record that often is distinctive. Polarity intervals have a mean duration of some 300 000 y during the last 35 My, but large variations occur, from 20 000 y to several million years. If one can construct a calibrated ‘standard’ or a so-called ‘geomagnetic polarity timescale’ (GPTS), dated by radiometric methods and/or by orbital tuning, one can match the observed pattern with this standard and hence derive the age of the sediments. Magnetostratigraphy with correlation to the GPTS has become a standard tool in ocean sciences.

The Paleomagnetic Signal The Earth’s magnetic field is generated in the liquid outer core through a dynamo process that is

N

Normal

S

Reversed Figure 1 Schematic representation of the geomagnetic field of a geocentric axial dipole (‘bar magnet’). During normal polarity of the field the average magnetic north pole is at the geographic north pole, and a compass aligns along magnetic field lines. Historically, we refer to the north pole as the pole attracting the ‘north-seeking’ needle of a compass, but physically it is a south pole. During normal polarity, the inclination is positive (downward-directed) in the Northern Hemisphere and negative (upward-directed) in the Southern Hemisphere. Conversely, during reversed polarity, the compass needle points south, and the inclination is negative in the Northern and positive in the Southern Hemisphere.

(c) 2011 Elsevier Inc. All Rights Reserved.

25

26

GEOMAGNETIC POLARITY TIMESCALE

higher-order terms: the nondipole field. At any one time, the best-fitting geocentric dipole axis does not coincide exactly with the rotational axis of the Earth, but averaged over a few thousand years we may treat the dipole as both geocentric and axial. The most distinctive property of the Earth’s magnetic field is that it can reverse its polarity: the north and south poles interchange. Paleomagnetic studies of igneous rocks provided the first reliable information on reversals. In 1906, Brunhes observed lava flows magnetized in a direction approximately antiparallel to the present geomagnetic field, and suggested that this was caused by a reversal of the field itself, rather than by a self-reversal mechanism of the rock. In 1929, Matuyama demonstrated that young Quaternary lavas were magnetized in the same direction as the present field (normal polarity), whereas older lavas were magnetized in the opposite direction. His study must be regarded as the first magnetostratigraphic investgation. Initially, it was believed that the field reversed periodically, but as more (K/Ar dating plus paleomagnetic) results of lava flows became available, it became clear that geomagnetic reversals occur randomly. It is this random character that fortuitously provides the distinctive ‘fingerprints’ and gives measured polarity sequences their correlative value. A polarity reversal typically takes several thousands of years, which on geological timescales is short and can be taken as globally synchronous. The field itself is sign invariant: the same configuration of the geodynamo can produce either a normal or reversed polarity. What causes the field to reverse is still the subject of debate, but recent hypotheses suggests that lateral changes in heat flow at the core–mantle boundary play an important role. Although polarity reversals occur at irregular times, over geological time spans the reversal frequency can change considerably. For instance, the polarity reversal frequency has increased from approximately 1 My 1 some 80 My ago to 5 My 1 in more recent times. During part of the Creataceous, no reversals occurred at all from 110 to 80 Ma: the field showed a stable normal polarity during some 30 My. Such long periods of stable polarity are called Superchrons, and the one in the Cretaceous is also recognized as the Cretaceous Normal Quiet Zone in oceanfloor magnetic anomalies. The ancient geomagnetic field can be reconstructed from its recording in rocks during the geological past. Almost every type of rock contains magnetic minerals, usually iron oxide/hydroxides or iron sulfides. During the formation of rocks, these magnetic minerals (or more accurately: their magnetic domains) statistically align with the then ambient field, and will subsequently be ‘locked in,’

preserving the direction of the field as natural remanent magnetization (NRM): the paleomagnetic signal. The type of NRM depends on the mechanism of recording the geomagnetic signal, and we distinguish three basic types: TRM, CRM, and DRM. TRM, or thermoremanent magnetization, is the magnetization acquired when a rock cools through the Curie temperature of its magnetic minerals. Curie temperatures of the most common magnetic minerals are typically in the range 350–7001C. Above this temperature, the magnetic domains align instantaneously with the ambient field. Upon cooling, they are locked: the magnetic minerals carry a remanence that usually is very stable over geological time spans. Any subsequent change of the direction of the ambient field cannot change this remanence. Typically, TRM is acquired in igneous rocks. CRM, or chemical remanent magnetization, is the magnetization acquired when a magnetic mineral grows through a critical ‘blocking diameter’ or grain size. Below this critical grain size, the magnetic domains can still align with the ambient field; above it, the field will be locked and the acquired remanence may again be stable over billions of years. CRM may be acquired under widely different circumstances, e.g., during slow cooling of intrusive rocks, during metamorphosis or (hydrothermal) fluid migration, but also during late diagenetic processes such as weathering processes through formation of new magnetic minerals. A particularly important mechanism of CRM acquisition occurs in marine sediments during early diagenesis: depending on the redox conditions, iron-bearing minerals may dissolve, and iron may become mobile. If the mobilized iron subsequently encounters oxic conditions, it may precipitate again and form new magnetic minerals that then acquire a CRM. DRM, or detrital remanent magnetization, is the magnetization acquired when magnetic grains of detrital origin already carrying TRM or CRM are deposited. The grains statistically align with the ambient field as long as they are in the water column or in the soft water-saturated topmost layer of the sediment (Figure 2). Upon compaction and dewatering, the grains are mechanically locked — somewhere in a ‘lock-in depth zone’ — and will preserve the direction of the ambient field. The Magnetic Signal in Marine Sediments

In sediments, it is often assumed that the natural remanent magnetization is due to DRM. In the past, there has been some debate on the mechanisms of DRM acquisition, and the term pDRM or postdepositional DRM has often been used, because the

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMAGNETIC POLARITY TIMESCALE

27

sedimentary environment. Depending on the role of organic matter, existing magnetic minerals may dissolve and iron may be mobilized and precipitate elsewhere. These geochemical processes may results in the (partial) removal of the original paleomagnetic signal B and in the acquisition of CRM in a particular zone (much) later than the deposition of the sediment in this zone. Clearly, such processes may severely damage the Still water fidelity of the paleomagnetic record, and may offset the position of a reversal boundary by a distance in sediment that can correspond to a time of up to tens of thousands of years. However, in ‘suitable’ sediments Sediment /water interface the damage is usually restricted, and the paleomagnetic signal of such sediments may be considered as Bioturbation reliable and near-depositional. Suitable environments DRM lock-in depth are generally those with a sufficiently high sedimentation rate, a significant detrital input, and a preCompaction dominantly oxic environment. Therefore, it is often Oxic sediment O O O O Oxygen diffusion necessary to check the origin of the NRM using rock magnetic and geochemical methods. Precipitation of iron oxides As a rule, the total NRM is composed of different and CRM formation Anoxic sediment components. Ideally, the primary NRM — that oriDissolution of iron oxides ginating from near the time of deposition — has been conserved, but often this original signal is contaminated with ‘viscous’ remanence components, Fe Fe Fe Fe Iron diffusion (Sub)oxic sediment referred to as VRM. Such a VRM may result from partial realignment of ‘soft’ magnetic domains in the Figure 2 Depositional remanent magnetization (DRM) present-day field or from low-temperature oxidation acquired in sediments involves a continuum of physical and chemical processes. Detrital magnetic minerals (black) will be of magnetic minerals. It is generally easily removed statistically aligned along the ambient geomagnetic field (B) in through magnetic ‘cleaning,’ which consists of a still water and in the soft and bioturbated water-saturated routine laboratory treatment called demagnetization; sediment just below the sediment–water interface. Upon details can be found in any standard texbook on compaction and dewatering, the grains are mechanically paleomagnetism. Despite these pitfalls in sedimentlocked, preserving the direction of the field. Early diagenetic processes such as sulfate reduction may dissolve iron-bearing ary paleomagnetic records, sediments — in contrast minerals. Upon encountering a more oxic environment, iron may to igneous rocks — have in principle the distinctive precipitate as iron oxides, which will acquire a chemical remanent advantage of providing a continuous record of the magnetisation (CRM). The thus-acquired CRM in this layer may geomagnetic field, including the history of geohave a much later age than the depositional age of this layer. magnetic polarity reversals. Paleomagnetic studies of the sediments will then reveal the pattern of geolock-in zone has a certain depth. Sediment at this magnetic reversals during deposition. depth is slightly older than that the sediment–water interface. Hence, the acquisition of NRM is always slightly delayed with respect to sediment age. For The Geomagnetic Polarity Timescale practical purposes (in magnetostratigraphy) this usually has no serious consequences. Nevertheless, the (GPTS) concept of a purely detrital remanent magnetization an Surveys over the ocean basins carried out from the oversimplification of the real world. We therefore 1950s onward found linear magnetic anomalies, prefer to use the term depositional remanent magnet- parallel to mid-ocean ridges, using magnetometers ization (DRM) to refer to a continuum of physical and towed behind research vessels. During the early chemical processes that occur during and shortly after 1960s, it was suggested, and soon confirmed, that deposition. The acquisition of a DRM thus depends these anomalies resulted from the remanent magboth on detrital input of magnetic minerals and the netization of the oceanic crust. This remanence is new formation of such minerals in the sediment acquired during the process of seafloor spreading, through early diagenetic processes (Figure 2). Early when uprising magma beneath the axis of the middiagenesis is widespread and occurs in virtually every ocean ridges cools through its Curie temperature Turbulent water

(c) 2011 Elsevier Inc. All Rights Reserved.

28

GEOMAGNETIC POLARITY TIMESCALE

(Figure 3) in the ambient geomagnetic field, thus acquiring its direction and polarity. The continuous process of rising and cooling of magma at the ridge results in magnetized crust of alternating normal and reversed polarity that produces a slight increase or decrease of the measured field — the marine magnetic anomalies. It was also found that the magnetic anomaly pattern is generally symmetric on both sides of the ridge, and, most importantly, that it provides a wonderfully continuous ‘tape recording’ of the geomagnetic reversal sequence. A major step in constructing a time series of polarity reversals was taken in 1968 by Heirztler and co-workers. They used a long profile from the southern Atlantic Ocean and, extrapolating from a known age for the lower boundary of the Gauss Chron (Figure 4), they constructed a geomagnetic polarity timescale under the assumption of constant 2 Ma

1 Ma

spreading. Subsequent revisions mostly used the anomaly profile of Heirtzler, often adding more detail from other ocean basins, and appending additional calibration points. These calibration points are derived from sections on land, which, first, must contain a clear fingerprint of magnetic reversals that can be correlated to the anomaly profile, and second, contain rocks that can be reliably dated by means of radiometric methods. The GPTS is then derived by linear interpolation of the anomaly pattern between these calibration points, again under the assumption of constant spreading rate between those points. The development of the GPTS (Figure 4) shows increasing detail and gradually improved age control. Periods of a predominant (normal or reversed) polarity are called chrons, and the four youngest ones are named after individuals: Brunhes (normal), who suggested field reversal; Matuyama (reversed),

Present

1 Ma

2 Ma

Model profile

Observed profile

Mid oceanic ridge

Sediment

Uprising magma and ‘locking in’ of magnetic polarity

Crust Lithosphere

Figure 3 Formation of marine magnetic anomalies during seafloor spreading. The oceanic crust is formed at the ridge crest, and while spreading away from the ridge it is covered by an increasing thickness of oceanic sediments. The black (white) blocks of oceanic crust represent the original normal (reversed) polarity thermoremanent magnetization (TRM) acquired upon cooling at the ridge. The black and white blocks in the drill holes represent normal and reversed polarity DRM acquired during deposition of the marine sediments. The model profile (grey) represents computed magnetic anomalies produced by the block model of TRM polarity (top); the observed profile (dark) is the observed sea-level magnetic anomaly profile due to the magnetized oceanic crust.

(c) 2011 Elsevier Inc. All Rights Reserved.

29

APTS, 2000

Cande and Kent, 1992

Berggren et al., 1985

LaBreque et al., 1977

Heirtzler et al., 1968

Cox et al., 1963

GEOMAGNETIC POLARITY TIMESCALE

Brunhes

0

Matuyama

1

Gauss

2

3

Gilbert

Age (Ma)

4

5

6

7

8 Figure 4 Development of the geomagnetic polarity timescale (GPTS) through time. The initial assumption of periodic behavior (in 1963) was soon abandoned as new data became available. The first modern GPTS based on marine magnetic anomaly patterns was established in 1968 by Heirtzler and co-workers. Subsequent revisions show improved age control and increased resolution. A major breakthrough came with the astronomical polarity timescale (APTS), in which each individual reversal is accurately dated.

who proved this; Gauss (normal), who mathematically described the field; and Gilbert (reversed), who discovered that the Earth itself is a huge magnet. Chrons may contain short intervals of opposite

polarity called subchrons, which are named after the locality where they were discovered; for example, the normal Olduvai subchron within the Matuyama reversed chron is named after Olduvai Gorge in

(c) 2011 Elsevier Inc. All Rights Reserved.

30

GEOMAGNETIC POLARITY TIMESCALE

Tanzania, and the Kaena reversed subchron within the Gauss normal chron after Kaena Point on Hawaii. Older chrons were not named but numbered, according to the anomaly numbers earlier given by Heirtzler, which has led to a confusing nomenclature of chrons and subchrons. A major step forward was taken by Cande and Kent in 1992, who thoroughly revised the magnetic anomaly template over the last 110 My. They constructed a synthetic flow line in the South Atlantic, using a set of chosen anomalies that were taken as tie points. The intervals between these tie points were designated as category I intervals. On these intervals they projected (stacks of) the bestquality profiles surveyed in this ocean basin, providing category II intervals. Since the spreading of the Atlantic is slow, they subsequently filled in the category II intervals with high-resolution profiles from fast-spreading ridges (their category III). This enabled them to include much more detail on short polarity intervals (or subchrons); see, for instance, the increase of detail around 7 Ma in Figure 4. Very short and low-intensity anomalies still have an uncertain origin. They may represent very short subchrons of the field, as has been proven for some of them (e.g., the Cobb Mt. subchron at 1.21 Ma, or the Re´union subchron at 2.13–2.15 Ma), or may just represent intensity fluctuations of the geomagnetic field causing the oceanic crust to be less (or more) strongly magnetized. Because of their uncertain or unverified nature, these were called cryptochrons. In addition, Cande and Kent developed a consistent (sub)chron nomenclature that is now used as the standard. In total, they used nine calibration points, but they made a break with tradition by using, for the first time, an astronomically dated age tie point for the youngest one: the Gauss/Matuyama boundary. The correlation of the GPTS to global biostratigraphic zonations is covered extensively by Berggren et al. (1995). The template of magnetic anomaly patterns from the ocean floor has remained central for constructing the GPTS from the late Cretaceous onward (110–0 Ma). Only recently, the younger part of the GPTS has been based on direct dating of each individual reversal through the use of orbitally tuned timescales. In their most recent version of the GPTS, Cande and Kent included the astronomical ages for all reversal boundaries for the past 5.3 My. The CK95 or Cande and Kent (1995) geomagnetic polarity timescale is at present the most widely used standard. Polarity timescales for the Mesozoic rely on some of the oldest magnetic anomaly profiles, down to the late Jurassic, and on dated magnetostratigraphies of sections on land.

The Astronomical Polarity Timescale (APTS) The latest development in constructing a GPTS comes from orbital tuning of the sediment record; for details see the article Orbitally Tuned Timescales. It differs essentially from the conventional GPTS in the sense that each reversal boundary — or any other geological boundary for that matter, e.g. biostratigraphic datum levels or stage and epoch boundaries — is dated individually. This has provided a breakthrough in dating of the geological record and has the inherent promise of increasing understanding of the climate system, since cyclostratigraphy and subsequent orbital tuning rely on decphering and understanding environmental changes driven by climate change, which in turn is orbitally forced. The fact that the age of each reversal is directly determined, rather than interpolated between calibration points, has important consequences for (changes in) spreading rates of plat pairs. Rather than having to assume constant spreading rates between calibration points, one can now accurately determine these rates, and small changes therein. Indeed, Wilson found that the use of astronomical ages resulted in very small and physically realistic spreading rate variations. As a result, the discrepancy in between plate motion rates from the global plate tectonic model (NUVEL-1) and those derived from geodesy has become much smaller. Meanwhile NUVEL-1 has been updated (to NUVEL-1A) to incorporate the new astronomical ages. Another application is the dating of Pleistocene, Pliocene, and Miocene, and older stage boundaries, many of which have been defined in the Mediterranean. The availability of a good astrochronology has effectively become a condition for the definition of a Global Boundary Stratotype Section and Point (GSSP). An example is shown in Figure 5, showing the Tortonian–Messinian GSSP that has recently been defined in the Atlantic margin basin of western Morocco. Perhaps one of the most promising areas of the application of astrochronology is in the bed-to-bed correlation of the two different realms of oceans and continent. Climate forcing may be expected to have a different expression in the different realms because of the different nature of their sedimentary environments. A recently established and refined orbital timescale for the loess sequences of northern China relies upon the correlation of detailed monsoon records to the astronomical solutions and the oceanic oxygen isotope records. An important finding was that the straightforward use of magnetostratigraphy and correlation to the GPTS cannot provide a sufficiently accurate age model for comparison with the

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMAGNETIC POLARITY TIMESCALE

31

Figure 5 Magnetostratigraphy, cyclostratigraphy, and astrochronology of the marine Oued Akrech section from the Atlantic margin of Morocco, which defines the Tortonian–Messinian Global Stratotype Section and Point (GSSP). The sedimentary cycles represent a cyclically changing environment that correlates with variations in insolation. Insolation is strongly related to climatic precession (upper left panel), which induces cyclic changes in seasonal contrast, reflected one-on-one in the sedimentary cycles. (After Hilgen FJ et al. (2000) Episodes 23(3): 172–178.)

ocean record, since the analysis of the astrochronological framework demonstrates considerable downward displacement of reversal boundaries because of delayed lock-in of the NRM. With the new chronology and its direct correlation to the oceanic record, it is now possible to analyse terrestrial paleomonsoon behavior for the past 2.6 My and compare it to climate proxies from the marine realm. This may give important information, for instance, on leads and lags of various systems in response to climate change, on phase relations with insolation, or on the relation between global ice volume and monsoonal climate.

See also Aeolian Inputs. Magnetics. Monsoons, History of. Paleoceanography, Climate Models in. Paleoceanography: Orbitally Tuned Timescales.

Further Reading Berggren WA, Kent DV, Aubry MP, and Hardenbol J (1995) Geochronology, Time Scales and Global Stratigraphic Correlations: A Unified Temporal Framework for an Historical Geology. Society of Economic Paleontologists and Mineralogist, Special Volume 54.

(c) 2011 Elsevier Inc. All Rights Reserved.

32

GEOMAGNETIC POLARITY TIMESCALE

Butler RF (1992) Paleomagnetism, Magnetic Domains to Geologic Terranes. Boston: Blackwell Scientific. Cande SC and Kent DV (1992) A new geomagnetic polarity Time-Scale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research 97: 13917--13951. Cande SC and Kent DV (1995) Revised calibration of the Geomagnetic Polarity Time Scale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research 100: 6093--6095. Gradstein FM, Agterberg FP, Ogg JG, et al. (1995) A Triassic, Jurassic and Cretaceous time scale. In: Berggren WA, Kent DV, Aubry M and Hardenbol J (eds) Geochronology, Time Scales and Stratigraphic Correlation. Society of Economic Paleontologists and Mineralogist, Special Volume 54: 95–128. Hilgen FJ, Krijgsman W, Langereis CG, and Lourens LJ (1997) Breakthrough made in dating of the geological record. EOS, Transactions of the AGU 78: 285--288. Heirtzler JR, Dickson GO, Herron EM, Pitman WC III, and Le Pichon X (1968) Marine magnetic anomalies,

geomagnetic field reversals, and motions of the ocean floor and continents. Journal of Geophysical Research 73: 2119--2136. Heslop D, Langereis CG, and Dekkers MJ (2000) A new astronomical time scale for the loess deposits of Northern China. Earth and Planetory Science Letters 184: 125--139. Maher BA and Thompson R (1999) Quaternary Climates, Environments and Magnetism. Cambridge: Cambridge University Press. Opdyke ND and Channell JET (1996) Magnetic Stratigraphy. San Diego: Academic Press. Shackleton NJ, Crowhurst SJ, Weedon GP, and Laskar J (1999) Astronomical calibration of Oligocene–Miocene time. Philosophical Transactions of the Royal Society of London, A 357: 1907--1929. Wilson DS (1993) Confirmation of the astronomical calibration of the magnetic polarity time scale from seafloor spreading rates. Nature 364: 788--790.

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMORPHOLOGY C. Woodroffe, University of Wollongong, Wollongong, NSW, Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1142–1148, & 2001, Elsevier Ltd.

Introduction Geomorphology is the study of the form of the earth. Coastal geomorphologists study the way that the coastal zone, one of the most dynamic and changeable parts of the earth, evolves, including its profile, plan-form, and the architecture of foreshore, backshore, and nearshore rock and sediment bodies. To understand these it is necessary to examine wave processes and current action, but it may also involve drainage basins that feed to the coast, and the shallow continental shelves which modify oceanographic processes before they impinge upon the shore. Morphodynamics, study of the mutual co-adjustment of form and process, leads to development of conceptual, physical, mathematical, and simulation models, which may help explain the changes that are experienced on the coast. In order to understand coastal variability from place to place there are a number of boundary conditions which need to be considered, including geophysical and geological factors, oceanographic factors, and climatic constraints. At the broadest level plate-tectonic setting is important. Coasts on a plate margin where oceanic plate is subducted under continental crust, such as along the western coast of the Americas, are known as collision coasts. These are typically rocky coasts, parallel to the structural grain, characterized by seismic and volcanic activity and are likely to be uplifting. They contrast with trailing-edge coasts where the continental margin sits mid-plate, which are the locus of large sedimentary basins. Smaller basins are typical of marginal sea coasts, behind a tectonically active island arc. The nature of the material forming the coast is partly a reflection of these broad plate-tectonic factors. Whether the shoreline is rock or unconsolidated sediment is clearly important. Resistant igneous or metamorphic rocks are more likely to give rise to rocky coasts than are those areas composed of broad sedimentary sequences of clays or mudstones. Within sedimentary coasts sandstones may be more resistant than mudstones. Rocky coasts tend to be relatively

resistant to change, whereas coasts composed of sandy sediments are relatively easily reshaped, and muddy coasts, in low-energy environments, accrete slowly. The relative position of the sea with respect to the land has changed, and represents an important boundary condition. The sea may have flooded areas which were previously land (submergence), or formerly submarine areas may now be dry (emergence). The form of the coast may be inherited from previously subaerial landforms. Particularly distinctive are landscapes that have been shaped by glacial processes, thus fiords are glacially eroded valleys, and fields of drumlins deposited by glacial processes form a prominent feature on paraglacial coasts. The oceanographic factors that shape coasts include waves, tides, and currents. Waves occur as a result of wind transferring energy to the ocean surface. Waves vary in size depending on the strength of the wind, the duration for which it has blown and the fetch over which it acts. Wave trains move out of the area of formation and are then known as swell. The swell and wave energy received at a coast may be a complex assemblage generated by specific storms from several areas of origin. Tides represent a largewavelength wave formed as a result of gravitational attractions of the sun and moon. Tides occur as a diurnal or semidiurnal fluctuation of the sea surface that may translate into significant tidal currents particularly in narrow straits and estuaries. In addition storm-generated surges and tsunamis may cause elevated water levels with significant geomorphological consequences. The coast is shaped by these oceanographic processes working on the rocks and unconsolidated sediments of the shoreline. Climate is significant in terms of several factors. First, wind conditions lead to generation of waves and swell, and may blow sand into dunes along the backshore. Climate also influences the rate at which weathering and catchment processes operate. In addition, regional-scale climate factors such as monsoonal wind systems and the El Nin˜o Southern Oscillation phenomenon demonstrate oscillatory behavior and may reshape the coast seasonally or interannually. Gradual climate change may mean that shoreline fluctuations do not revolve around stationary boundary conditions but may exhibit gradual change themselves. In this respect the impact of perceived global climate change during recent decades as a result of human-modified environmental factors is likely to be

(c) 2011 Elsevier Inc. All Rights Reserved.

33

34

GEOMORPHOLOGY

Sea level

Sea level

(B)

(A)

Figure 1 Schematic representation of the original woodcut illustrations by Charles Darwin showing (A) the manner in which he deduced that fringing reefs around a volcanic island would develop into barrier reefs (dotted), and (B) barrier reefs would develop into an atoll, as a result of subsidence and vertical reef growth.

felt in the coastal zone. Of particular concern is anticipated sea-level rise which may have a range of geomorphological effects, as well as other significant socioeconomic impacts.

History Geomorphology has its origins in the nineteenth century with the results of exploration, and the realization that the surface of the earth had been shaped over a long time through the operation of processes that are largely in operation today (uniformitarianism). The observations by Charles Darwin during the voyage of the Beagle extended this view, particularly his remarkable deduction that fringing reefs might become barrier reefs which in turn might form atolls as a result of gradual subsidence of volcanic islands combined with vertical reef growth (Figure 1). In the first part of the twentieth century geomorphology was dominated by the ‘geographical cycle’ of erosion of William Morris Davis who anticipated landscape

denudation through a series of stages culminating in peneplanation, and subsequent rejuvenation by uplift. This highly conceptual model, across landscapes in geological time, was also applied to the coast by Davis who envisaged progressive erosion of the coast reducing shoreline irregularities with time (Figure 2). Such landscape-scale studies were extended by Douglas Johnson who emphasized the role of submergence or emergence as a result of sea-level change (Figure 3). The Davisian view was reassessed in the second half of the twentieth century with a greater emphasis on process geomorphology whereby studies focused on attempting to measure rates of process operation and morphological responses to those processes, reflecting ideas of earlier researchers such as G. K. Gilbert. The concept of the landscape as a system was examined, in which coastal landforms adjusted to equilibrium, perhaps a dynamic equilibrium, in relation to processes at work on them. Studies of

i ii iii (A) Initial

Figure 2 Schematic representation of the planform stages that W. M. Davis conceptualized through which a shoreline would progress from an initial rugged form (1) through maturity (2) to a regularized shoreline (3) that is cliffed (hatched) and infilled with sand (stippled). He envisaged this in parallel to the geographical cycle of erosion by which mountains (like the initial shoreline form, 1) would be reduced to a peneplain (like the solid ultimate shoreline, 3).

(B) Modern

Figure 3 Schematic representation of (A) the initial shoreline form envisaged by Douglas Johnson for an area in Marthas Vineyard, New England, and (B) the modern regularized, form of the shore.

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMORPHOLOGY

sediment movement and the adjustment of beach shape under different wave conditions were typical.

Scales of Study Coastal geomorphology now studies landforms and the processes that operate on them at a range of spatial and temporal scales (Figure 4). At the smallest scale geomorphology is concerned with an ‘instantaneous’ timescale where the principles of fluid dynamics apply. It should be possible to determine details of sediment entrainment, complexities of turbulent flow and processes leading to deposition of individual bedform laminae. The laws of physics apply at these scales, though they may operate stochastically. In theory, behavior of an entire embayment could be understood; in practice studies simply cannot be undertaken at that level of detail and broad extrapolations based on important empirical relationships are made. The next level of study is the ‘event’ timescale, which may cover a single event such as an individual storm or an aggregation of several lesser events over a year or more. The mechanistic relationships from instantaneous time are scaled up in a deterministic or empirical way to understand the operation of coasts at larger spatial and temporal scales. Thus, stripping of a beach during a single storm, and the more gradual reconstruction to its original state under ensuing calmer periods can be observed. Time taken for reaction to the event, and relaxation back to a more ambient state may be known from surveys of beaches, enabling definition of a ‘sweep zone’ within which the beach is regularly active.

35

At a larger scale of operation, the coastal geomorphologist is interested in the way in which coasts change over timescales that are relevant to societies and at which coastal-zone managers need to plan. This is the ‘engineering’ timescale, involving several decades. It is perhaps the most difficult timescale on which to understand coastal geomorphology and to anticipate behavior of the shoreline. The largest timescales are geological timescales. Studies over geological timescales are primarily discursive, conceptual models which recognize that boundary conditions, including climate and the rate of operation of oceanographic processes, change. It is clear that sea level itself has changed dramatically over millennia as a result of expansion and contraction of ice sheets during the Quaternary ice ages, and so the position of the coastline has changed substantially between glaciations and interglaciations.

Models of Coastal Evolution

Space

There has been considerable improvement in understanding the long-term development of coasts as a result of significant advances in paleoenvironmental reconstruction and geochronological techniques (especially radiometric dating). Incomplete records of past coastal conditions may be preserved, either as erosional morphology (notches, marine terraces, etc.) or within sedimentary sequences. Although this record is selective, reconstructions of Quaternary paleoenvironments, together with interpretation of geological sequences in older rocks, have enabled the formulation of geomorphological models based on sedimentary evidence. In the case of deltas, where there may be important hydrocarbon reserves, the complementary development of geological models GEOLOGICAL Global and study of modern deltas has led to better underReef standing of process and response at longer timescales structure ENGINEERING than those for which observations exist. Cliff Unconsolidated sediments are the key to coastal Regional retreat morphodynamics because coasts change through Delta EVENT dynamics erosion, transport, and deposition of sediment. Study of coastal systems has led to many insights, parReef-flat morphology ticularly in terms of the various pathways through Mudflat Local sedimentation which sediment may move within a coastal comBeach states partment or circulation cell. However, it is clear that INSTANTANEOUS a completely reductionist approach to coastal geomm Seconds Years Millennia morphology will not lead to understanding of all Time components and the way in which they interact. Empirical relationships and the presumed determinFigure 4 Representation of space and timescales appropriate istic nature of sediment response to forcing factors for the study of coastal geomorphology and schematic representation of some of the examples discussed in the context of remain incomplete and are all too often formulated instantaneous, event, engineering, and geological scales of on the basis of presumed uniform sediment sizes enquiry. or absence of biotic influence and are ultimately

(c) 2011 Elsevier Inc. All Rights Reserved.

36

GEOMORPHOLOGY

Bore Reflected standing wave

Breaking wave

(B)

geomorphology of each of these behaves differently. Processes operate at different rates, transitions between different states occur over different timescales and the significance of antecedent conditions varies. Rocky Coasts

Bore (A)

Breaking wave

(C)

Figure 5 The morphology of cliffs and shore platforms. Plunging cliffs (A) reflect wave energy and are generally not eroded. Shore platforms dissipate wave energy and may be either (B) ramped, or (C) subhorizontal. Change in these systems tends to occur at geological (hard rock) or engineering (soft rock) timescales.

unrealistic. More recently coastal systems have been investigated as nonlinear dynamical systems, because relationships are not linear and behavior is potentially chaotic. Nonlinear dynamical systems behave in a way that is in part dependent upon antecedent conditions. This is well illustrated by studies of beach state; beach and nearshore sediments are particularly easily reshaped by wave processes, thus a series of characteristics typical of distinct beach states can be recognized (see Figure 6). However, the beach is only partly a response to incident wave conditions, its shape being also dependent on the previous shape of the beach which was in the process of adjusting to wave conditions incident at that time. Coastal systems are inherently unpredictable. However, they may operate within a broad range of conditions with certain states being recurrent. Chaotic systems may tend towards self-organization, for instance patterns of beach cusps characteristic along low-energy beaches which reflect much of the wave energy may adopt a self-organized cuspate morphology in which swash processes, sediment sorting, and form are balanced. Models may be developed based on patterns of change inferred over geological timescales but consistent with mechanistic processes known to operate over lesser event timescales. Simulation modelling is not intended to reconstruct coastal evolution exactly, but becomes a tool for experimentation and extrapolation within which broad scenarios of change can be modeled and sensitivity to parameterization of variables examined. Coasts may be divided into rocky coasts, sandy coasts, and muddy coasts, and coral reefs and deltas and estuaries can be differentiated. The

Rocky coasts are characterized by sea cliffs, especially on tectonically active, plate-margin coasts where there are resistant rocks. Cliffed coastlines in resistant rock appear to adopt one of two forms (states): plunging cliffs where a vertical cliff extends below sea level, and shore platforms where a broad bench occurs at sea level in front of a cliff (Figure 5). Erosion of cliffs occurs where the erosional force of waves exceeds the resistance of the rocks, and where sufficient time has elapsed. Plunging cliffs occur where the rock is too resistant to be eroded. The vertical face results in a standing wave which reflects wave energy, so that there is little force to erode the cliff at water level. Waves exert a greater force if they break, or if they are already broken. This can only occur if the water depth is shallow offshore from the cliff face in which case the increased energy from the breaking waves is also able to entrain sediment from the floor (or rock fragments quarried from the foot of the cliff). This process of erosion accelerates through a positive feedback cutting a shore platform. Shore platforms thus develop in those situations where the erosive force of waves exceeds resistance of the rock. A platform widens as a result of erosion at the foot of the cliff behind the platform. The cliff oversteepens, leading to toppling and fall of detritus onto the rear of the platform, which slows further erosion of the cliff face until that talus has been removed. A series of such negative feedbacks slow the rate at which platforms widen over time, and there is often considerable uniformity of platforms up to a maximum width in any particular lithological setting. Shore platforms adopt either a gradually sloping ramped form, or a subhorizontal form often with a seaward rampart (Figure 5). There has been much discussion as to the relative roles of wave and subaerial processes in the formation of these platforms. Platforms in relatively sheltered locations appear to owe their origin to processes of water-layer leveling (physiochemical processes in pools which persist on platforms at low tide) or wetting and drying and its weakening of the rock. In other cases wave quarrying and abrasion are involved. In many cases both processes may be important. Other platforms may be polygenetic with inheritance from former stands of sea level (reflecting antecedent conditions).

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMORPHOLOGY

(A) Dissipative

(B) Intermediate

37

(C) Reflective

Plunging waves

Spilling waves

Accreted beach

Surging waves

Nearshore bar (D)

Storm

Beach volume

C

A Time

Figure 6 Beach state may be (A) dissipative, where high wave-energy impinges on low-gradient beaches; (B) intermediate (in which several states are possible); or (C) reflective, where low wave-energy surges onto steep or coarse beaches. Change between beach states occurs at event timescales. (D) A beach may react rapidly to a perturbation such as a storm changing from reflective state (C) to more dissipative state (A) but then change much more slowly through intermediate states, readjusting towards state (C) again unless further disrupted.

Coral Reefs

Sandy Coasts

A group of coastlines that are of particular interest to the geomorphologist are those formed by coral reefs. Corals are colonial animals that secrete a limestone exoskeleton that may form the matrix of a reef. Coral reefs flourish in tropical seas in high-energy settings where a significant swell reaches the shoreline. In these circumstances the reef attenuates much of the wave energy such that a relatively quiet-water environment occurs sheltered behind the reef. Not only is this reef ecosystem of enormous geomorphological significance in terms of the solid reef structure that it forms, but in addition, the carbonate reef material breaks down into calcareous gravels, sands and muds which form the sediments that further modify these coastlines. On the one hand reefs can be divided into fringing reefs, barrier reefs, and atolls, distinct morphological states that form part of an evolutionary sequence as a result of gradual subsidence of the volcanic basement upon which the reef established (see Figure 1). This powerful deduction by Darwin relating to reef structure and operating over geological timescales, has been generally supported by drilling and geochronological studies on mid-plate islands in reef-forming seas. On the other hand, the surface morphology of reefs is extremely dynamic with rapid production of skeletal sediments and their redistribution over event timescales, with the landforms of reefs responding to minor sea-level oscillations, storms, El Nin˜o, coral bleaching, and other perturbations.

Beaches form where sandy (or gravel) material is available forming a sediment wedge at the shoreline. The beach is shaped by incident wave energy, and can undergo modification particularly by formation and migration of nearshore bars which in turn modify the wave-energy spectrum. Various beach states can be recognized across a continuum from beaches which predominantly reflect wave energy, and those which dissipate wave energy across the nearshore zone. Reflective beaches are steep, waves surge up the beach and much of the energy is reflected, and the beach face may develop cusps. Dissipative beaches are much flatter, and waves spill before reaching the shoreline (Figure 6). During a storm, sand is generally eroded from the beach face and deposited in the nearshore, often forming shore-parallel or transverse bars. Waves consequently break on the bars and energy is lost, the form modifying the process in a mutual way. It may be possible to recognize a series of beach states intermediate between reflective and dissipative and any one beach may adopt one or several beach states over time (Figure 6). Beach state is clearly modified by incident wave conditions, but the rate of adjustment between states takes time, and a beach is also partly a function of antecedent beach states. Although erosion and redeposition of sand is the way in which the nearshore adjusts, there may also be long-term storage of sediment. In particular broad, flat beaches may develop dunes behind them.

(c) 2011 Elsevier Inc. All Rights Reserved.

38

GEOMORPHOLOGY

In other cases a sequence of beach ridges may develop. Geomorphological changes in state over geological timescales represented by beach-ridge plains may relate to variations in supply of sediment to the system (perhaps by rivers) as well as variations in the processes operating. Deltas and Estuaries

Where rivers bring sediment to the coast, deltas and estuaries can develop. In this case the sediment budget of the shoreline compartment or the receiving basin is augmented and there is generally a positive sediment budget. Deltas are characterized by broad wedges of sediment deposition. Delta morphology tends to reflect the processes that are dominant (Figure 7). Where wave energy is low and tides are minor, it is primarily river flows which account for sedimentation patterns. River flow may be likened to a jet, influenced by inertial forces, friction with the basin floor or buoyancy where there are density differences between outflow and the water of the receiving basin. Wave action tends to smooth the shoreline and a wave-dominated deltaic shoreline will be characterized by shore-parallel bars or ridges. Where riversupplied sediment is relatively low in volume, wave action may form a sandy barrier along the coast and rivers may supply sediment only to lagoons formed behind these barriers. Such is the case for much of the barrier-island shoreline of the eastern shores of the Americas; where barrier islands may be continually reworked landwards during relative sea-level rise. On the other hand if sea level is stable, a stable sand barrier is likely to form closing each

River

Wave

Tide

Figure 7 Deltas adopt a variety of forms, but morphology appears to reflect the relative balance of river, wave, and tide action. The broad morphology of the delta tends to be digitate where river-dominated, shore-parallel where wave-dominated, and tapering where tide-dominated.

embayment or creating a barrier estuary as in the case of coastal lagoons in Asia and Australasia. Estuaries are embayments which are likely to infill incrementally either through the deposition of riverborne sediments or through the influx of sediment from seaward by wave or tidal processes (Figure 8). The sediment from seaward may either be derived from the shelf, or from shoreline erosion. Tidal processes differ from river processes; they are bi-directional, flowing in during flood tide and out during ebb, and the flow is forced by the rising level of the sea. Small embayments are flooded and drained by a tidal prism (the volume of water between low and high tide). Longer estuaries, on the other hand, may have a series of tidal waves which progress up them. Where the tidal range is large this may flow as a tidal bore. Tide-dominated estuarine channels adopt a distinctive tapered form, with width (and depth) decreasing from the mouth upstream (Figure 8). Wave and tidal processes tend to shape deltas and estuaries to varying degrees depending on their relative operation. Many of the deltaic-estuarine processes operate over cycles of change. The hydraulic efficiency of distributaries decreases as they lengthen, until an alternative, shorter course with steeper hydraulic gradient is adopted. The abandoned distributaries of deltas often become tidally dominated with sinuous, tapering tidal creeks dominating what may be a gradually subsiding abandoned delta plain. Wave processes, in wavedominated settings, smooth and rework the abandoned delta shore, often forming barrier islands. Muddy Coasts

Muddy coasts are associated with the lowest energy environments. Mud banks may occur in high-energy, wave-exposed settings where large volumes of mud are supplied to the mouth of large rivers. Thus longshore drift north west of the Amazon, and around Bohai Bay downdrift from the Yellow River, enables mud-shoal deposition in open-water settings. Elsewhere mud flats are typical of sheltered settings, within delta interdistributary bays, around coastal lagoons, etc. These muddy environments are likely to be colonized by halophytic vegetation. Mangrove forests occur in tropical settings, whereas salt marsh occurs in higher latitudes, often extending into tropical areas also. These coastal wetlands further promote retention of fine-grained sediment. The muddy environments are areas of complex hydrodynamics and sedimentation. Sedimentation is likely to occur with a negative feedback such that as the tidal wetlands

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOMORPHOLOGY

(A) Wave-dominated Tidal delta

39

(B) Tide-dominated

Wetland

Wetland

Fluvial delta

Halopytes River Send shoal River

Sand barrier

Saline mudflats

Former shoreline Mud basin

Figure 8 Estuaries are broad embayments which may adopt a wide range of morphologies. They tend to show sand-barrier accumulation where wave-dominated (A), but be prominently tapering where tide-dominated (B). Estuaries are generally sediment sinks with the rates and patterns of infill reflecting the relative dominance of river, wave, and tide processes and sediment sources.

accrete sediment and as the substrate is elevated, they are flooded less frequently and therefore sedimentation decelerates. Boundary conditions, particularly sea level, are likely to vary at rates similar to the rate of sedimentation and prograded coastal plains contain complex sedimentary records of changes in ecological and geomorphological state.

See also

Conclusion

Boyd R, Dalrymple R, and Zaitlin BA (1992) Classification of clastic coastal depositional environments. Sedimentary Geology 80: 139--150. Carter RWG and Woodroffe CD (eds.) (1994) Coastal Evolution: Late Quaternary Shoreline Morphodynamics. Cambridge: Cambridge University Press. Cowell PJ and Thom BG (1994) Morphodynamics of coastal evolution. In: Carter RWG and Woodroffe CD (eds.) Coastal Evolution: Late Quaternary Shoreline Morphodynamics, pp. 33--86. Cambridge: Cambridge University Press. Darwin C (1842) The Structure and Distribution of Coral Reefs. London: Smith Elder. Davis RA (1985) Coastal Sedimentary Environments. New York: Springer-Verlag. Johnson DW (1919) Shore Processes and Shoreline Development. New York: Prentice Hall. Trenhaile AS (1997) Coastal Dynamics and Landforms. Oxford: Clarendon Press. Wright LD and Thom BG (1977) Coastal depositional landforms: a morphodynamic approach. Progress in Physical Geography 1: 412--459.

Coastal geomorphology is the study of the evolving form of the shoreline in response to mutually adjusting processes acting upon it. It spans instantaneous and event timescales, over which beaches respond to wave energy, through engineering and geological timescales, over which deltas build seaward or switch distributaries, sea level fluctuates, and cliff morphology evolves. The morphology (state) of the coast changes in response to perturbations, particularly extreme events such as storms, but also thresholds within the system (as when a cliff oversteepens and falls, or a distributary lengthens and then switches). Human action may also represent a perturbation to the system. In each coastal setting, the influence of human modifications is being felt. Thus there are fewer coasts which are not in some way influenced by society. As anthropogenic modification of climate and sea level occurs at a global scale, the human factor increasingly needs to be given prominence in coastal geomorphology.

Beaches, Physical Processes Affecting. Coral Reefs. Rocky Shores. Salt Marshes and Mud Flats.

Further Reading

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOPHYSICAL HEAT FLOW C. A. Stein, University of Illinois at Chicago, Chicago, IL, USA R. P. Von Herzen, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1149–1157, & 2001, Elsevier Ltd.

Introduction Heat flow is a directly measurable parameter at Earth’s surface that provides important constraints on the thermal state of its interior. When combined with other data, it has proved useful for models of Earth’s origin, structure, composition, and convective motion of at least the upper mantle. Because oceanic crust is mostly thin (about 5–8 km) and deficient in heat-producing radioactive elements compared to continental crust, marine heat flow measurements are particularly illuminating for the cooling and dynamics of the oceanic lithosphere. Indeed, they have provided an important constraint in the development and refinement of the plate tectonics paradigm over the past three decades. Initially influenced by concepts of Sir Edward Bullard, the first successful marine heat flow measurements obtained shortly after the end of World War II were among the first applications of remote electronics in the deep sea. Other than a general understanding of seafloor topography, little information on the nature of the Earth beneath the oceans was then available, so theories to explain the measurements were largely unconstrained. Probably the most important initial finding was that the mean oceanic heat flux was not much different from the mean of the available continental values, contrary to conventional expectations of that era. However, the entire field of marine geophysics (especially seismic reflection and refraction, and geomagnetism) was also being transformed by instrumentation development, and the availability of ships as seagoing platforms gave rise to a ‘golden age’ of ocean exploration between about 1950 and 1970. The large-scale (>1000 km) variations of marine heat flux were shown to be consistent with the plate tectonic paradigm that explained the mean depth and age of much of the ocean floor. Heat flow is now useful for studies of seafloor tectonics and other marine geophysical problems.

40

Marine heat flow measurements now number more than 15 000, and with the exception of high latitude regions, are distributed over most of Earth’s main ocean basins (Figure 1). Values range over about two orders of magnitude (about 10–1000 mW m2), with a general systematic spatial distribution at large (41000 km) scales but also a large variability at small (o50 km) scales caused by pervasive hydrothermal circulation in mostly young (o50– 70 million years (My)) ocean crust. The spatial distribution and magnitudes of marine heat flow values may be used to deduce the geometry and intensity of such circulation.

Measurements and Techniques Instrumentation and Development

Techniques for marine measurements were developed with the realization that the temperature of the deep (42 km) ocean is mostly horizontally stratified, and relatively constant over long time scales (4100 years). In this situation, the heat flux from the Earth’s interior is reflected in a relatively uniform temperature gradient with depth beneath a flat seafloor. Heat flow is the product of the magnitude of this temperature gradient and the thermal conductivity of the material over which it is measured. Gradients are commonly measured using a vertically oriented probe with temperature sensors surmounted by a weight that is lowered from a ship to penetrate the relatively soft sediments that cover most of the seafloor. Thermal conductivity is either measured with in situ sensors during the seafloor penetration by transient heating experiments or aboard ship on sediment cores. For instrumental simplicity, the initial heat flow equipment used only two thermal sensors at either end of a 3–4 m long probe with in situ analog recording techniques on paper or film. Subsequently the number of sensors has been increased, due to a desire to measure non-linear temperature gradients and thermal conductivity that may vary with depth. The wide range of measured thermal gradients and the developments in solid-state electronics have made digital recording the present standard. ‘Pogo’ measurements, i.e., multiple penetrations on the same station, combined with nearly real-time acoustic telemetry of raw data have proven useful for investigation of small-scale (o10–20 km) marine

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved.

0° 60°E

120°E

180°

120°W

Figure 1 Locations of marine heat flow measurements (dots). Thin lines show the location of the major plate boundaries.

60°S

0°

60°N

60°W

0°

GEOPHYSICAL HEAT FLOW 41

42

GEOPHYSICAL HEAT FLOW

heat flow variability, the origins of which are still not fully understood. It was necessary for the ‘Bullard’ probe to remain undisturbed in the seafloor for most of one hour to approach thermal equilibrium with the sediments because of its relatively large diameter (about 2– 3 cm). This probe has been largely supplanted by either much smaller (about 3 mm diameter) individual probes mounted in outrigged fashion on a larger (‘Ewing’) probe to attain deep (5–6 m) penetration, or a ‘violin-bow (Lister)’ probe consisting of a small (about 1 cm) sensor string mounted parallel to, but separate from, the main strength member (Figure 2). Both are capable of in situ thermal conductivity measurements, utilizing either a constant heat source or a calibrated pulse (approximating a delta function) after the gradient measurement. The measurement time in the seafloor is 15–20 min. Depending mostly on the

desired spacing during pogo operations, the mean time between measurement is 1–2 h. Even with acoustic data telemetering, battery life can be 2–3 days, thereby allowing many measurements during a single lowering. Real-time ship location accuracy now approaches a few meters with differential Global Positioning System navigation, although the uncertainty of the probe location during pogo operations in normal ocean depths (4–5 km) is typically 200 m or more. Hence 1–2 km is a useful minimum spacing between pogo penetrations unless seafloor acoustic transponder navigation is employed for higher accuracy (about 10 m) navigation. Instrumentation is also available to determine temperatures to depths up to about 600 m below the seafloor during deep-sea drilling to establish the uniformity of heat flow to greater depths, and may be the only method for reliable measurements in shallow seafloor (o1 km) regions

Cross-section AT A_A′ Weights

Cross-section of outrigger

Instrument housing Strength member

Lister

Bullard

Ewing

Pressure vessel

Instrument housing A

A′ Weight stand

Weight stand Coupling Strength member

Pressure vessel

(A)

Tensioning device (B)

Sensor string Strength member

6 _10 m

4_ 6 m

2m

Outrigger fin with thermistor probe

Nose cone (C)

Figure 2 Diagrams of the three most commonly used marine heat flow probes. (Reproduced with permission from Louden KE and Wright JA (1989) Marine heat flow data: a new compilation of observations and brief review of its analysis. In: Wright JA and Londen KE (eds) Handbook of Seafloor Heat Flow, pp. 2–67, Boca Raton: CRC Press.)

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOPHYSICAL HEAT FLOW

where ocean temperatures are more variable. Geothermal probes have also been developed for use with manned submersibles, allowing visual control of measurements in regions with large lateral gradients of heat flux or over specific geological features. A development is underway to install geothermal instrumentation on an autonomous underwater vehicle (ABE) for measurements in regions that are remote or difficult for normal ship operations. Environmental Corrections

Although most measurements on the deep seafloor with the usual instrumentation do not require corrections, they may be needed for some regions with unusual environmental parameters. As mentioned above, shallow water measurements may be subject to temporal bottom water temperature (BWT) variability, causing non-linear gradients in the seafloor. Using heat conduction theory, corrections may be applied if BWT variability is monitored for a sufficient period (months to years if possible) before the geothermal measurements. Conversely, nonlinear temperature–depth profiles can be inverted to yield BWT history assuming that the initial gradient was linear. This inverse procedure is non-unique, although closely spaced measured temperatures to a sufficient depth below the seafloor and well-determined thermal conductivity can reduce uncertainties. Continental margins and some deep-sea trenches are examples of regions where non-linear gradients have been measured, usually correlated with strong and variable deep currents that are probably focused by the seafloor topography. Topography causes lateral heat flow variability even with uniform and constant BWT, because the seafloor topography distorts isotherms that would otherwise be horizontal below a flat surface. Seafloor sedimentation reduces the heat flow measured because recently deposited sediments modify the seafloor boundary condition to which the equilibrium gradient must adjust. Corrections usually become significant when sedimentation rates exceed a few tens of meters per million years and/or the total sediment thickness exceeds 1 km. Vertical pore water flow in sediments may either enhance (for upward flow) or reduce (for downward flow) temperature gradients because the water advects some of the heat otherwise conducted upwards. The effects become significant for flow rates greater than a few centimeters per year, and nonlinear gradients may be expected if upward flow rates exceed about 10 cm year1. It is unusual for rates on normal deep seafloor to exceed the latter value because the fluid permeability of pelagic sediments is too low, but

43

coarser and sometimes rapidly deposited sediments of continental margins may support relatively rapid flows.

Range of Measured Parameters As discussed later, the plate tectonic paradigm predicts that the highest thermal gradients and heat fluxes should be found in the youngest seafloor, and the lowest values in the oldest. Although this is generally true when measurements are averaged over regions of various seafloor ages, considerable variability occurs over the youngest seafloor as a result of vigorous hydrothermal circulation. The cooling of hot and permeable upper ocean crust supports seawater convection in the crust over lateral circulation scales of at least several to a few tens of kilometers; vertical scales probably do not exceed a few kilometers. The highest gradients and heat fluxes (up to 1001C m1 and 100 W m2, respectively) are measured in sediments near seafloor vents where fluids upwell, and the lowest (o0.0051C m1 and 0.005 W m2, respectively) where fluids downwell. Mean upper basement fluid velocities may be several meters per year, and much greater in conduits that support vigorous seafloor venting. The detailed pore water flow and permeability structure for any system of seafloor hydrothermal circulation have not yet been investigated in detail and are probably very complex. After the seafloor ages to 50–70 My, hydrothermal circulation appears to decrease to a level where modulation of the surface heat flux is small to negligible (Figure 3), probably dependent on the thickness and uniformity of sediment cover. For the oldest seafloor (100–180 My), heat flux is relatively uniform at about 50 mW m2710%. This value probably reflects the secular cooling of the upper oceanic mantle. Thermal conductivity of marine sediments generally varies less than a factor of two, from about 0.7 to o1.4 Wm1 K1. The lowest values are associated with red clays or siliceous oozes, which have up to 70% or more of water by weight and the highest with carbonate oozes or coarse sediments near continental margins with a high proportion of quartz. Since water has a low thermal conductivity (0.6 W m1 K1), sediments with a large percentage of water have low thermal conductivity. Since calcium carbonate and quartz have high thermal conductivity (about 3 W m1 K1 and 4 W m1 K1, respectively), sediments with a large percentage of either of these minerals have high thermal conductivity. Conductivity variations with depth may be caused

(c) 2011 Elsevier Inc. All Rights Reserved.

44

GEOPHYSICAL HEAT FLOW

Depth (km)

3 4 5 GDH1 PSM HS

6 7 0

50

(A)

100 Age (My)

150

250

_

Heat flow (mW m 2)

GDH1 PSM

200

HS 150

Thermal Models 100

Data for Thermal Modeling

50 0

0

150

100

50 Age (My)

(B)

Fraction

1.5 1.0 0.5 0 (C)

climate (e.g., glacial) cycles. When thermal conductivity varies rapidly with depth the temperature–depth profile may not be linear. If thermal conductivity has been measured at closely spaced depth intervals, then the heat flow can be calculated using a method introduced by Bullard. This approach assumes the absence of significant heat sources or sinks and onedimensional, steady-state, conductive heat flow. Thus there is a linear relationship between temperature and the thermal resistance of the sediments (the sum from the surface to the depth of the temperature measurement of the inverse of thermal conductivity times the sediment thickness for that given thermal conductivity). The slope determined from a least-squares fit of this ‘Bullard plot’ gives the appropriate conductive heat flux.

tseal = 65 ± 10 My 0

50

100 Age (My)

150

Figure 3 Data and models for ocean depth (A) and heat flow (B) as a function of age. (Reproduced with permission from Stein and Stein, 1992.) The data are averaged in 2 My bins, and one standard deviation about the mean value for each is shown by the envelope. Shown are the plate model of Parsons and Sclater (1977) (PSM), a cooling halfspace model with the same thermal parameters (HS), and the GDH1 plate model from Stein and Stein (1992). (C) Heat flow fraction (observed heat flow/GDH1 prediction) with age averaged in 2 My bins. The discrepancy for ages o50–70 My presumably indicates the fraction of the heat transported by hydrothermal flow. The fractions for ages o50 My (closed circles), which were not used in deriving GDH1, are fit by a least squares line. The sealing age, where the line reaches the fractional value of one, is 65710 My.

by: (1) turbidity flows that sort grain sizes into layers with different proportions of pore water; (2) ice rafting in higher latitudes that may intersperse higher conductivity sediments derived from the continents and lower conductivity marine pelagic sediments; and (3) variability in the proportion of calcium carbonate in sediments near the equator deposited over

Oceanic lithosphere forms at midocean ridges, where hot magma upwells, and then cools to form plates as the material moves away from the spreading center. As the plate cools, heat flow decreases and the seafloor deepens (Figure 3). However, only shallow (less than 1 km) measurements of lithospheric temperatures are possible. Hence, the two primary data sets used to constrain models for the variation in lithospheric temperature with age are seafloor depths and heat flow. The depth, corrected for sediment load, depends on the temperature integrated over the lithospheric thickness. The heat flow is proportional to the temperature gradient. Initially seafloor depths rapidly increase, with the average increase relative to the ridge crest proportional to the square root of the crustal age. However, for ages greater than about 50–70 My, the average increase in depth is slower and the curve is said to ‘flatten.’ Mean heat flow also decreases rapidly away from the ridge crest, with values approximately proportional to the inverse of the square-root of the age, but after about 50 My this curve also ‘flattens.’ Halfspace and Plate Models

Two different mathematical models are often used to describe the thermal evolution of the oceanic lithosphere, the halfspace (or boundary layer) and plate models. For the halfspace model, the predicted lithospheric thickness increases proportionally with the square root of age. Hence depth and heat flow vary as the square root of age and the reciprocal of the square root of age, respectively. However, the halfspace model cannot explain the ‘flattening’ of the curves. Alternatively, the plate model represents the

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOPHYSICAL HEAT FLOW

lithosphere as a layer with a fixed constant temperature at its base. Initially, the cooling for the plate model is the same as the halfspace model, but for older ages the influence of the lower boundary results in slower cooling, approximately predicting the observed flattening of the depths and heat flow at older ages. The plate base is assumed to represent a depth at which additional heat is supplied from the mantle below to prevent the halfspace cooling at older ages. However, the model does not directly describe how this heat is added.

45

understanding of the physics of processes affecting lithospheric thermal evolution improve, new and better reference models will be developed.

Application to Lithospheric Processes Using a reference model for the expected heat flow, regions can be examined to observe if the measured heat flow differs from that predicted and, if so, to study the causes of the discrepancy. Two primary discrepancies are assumed to reflect hydrothermal circulation and midplate swells.

Reference Models Hydrothermal Circulation

Heat flow measurements for crust of ages 0–65 My are generally lower than the predictions of all commonly used reference models (Figure 3). Because the heat flow measurements primarily reflect conductive heat flow, this difference has been attributed to hydrothermal water flow in the crust and sediments transporting some of the heat assumed in thermal models to be transferred by conduction. The missing heat transported by convection must appear somewhere, either as high conductive heat flow or as advective discharge to the sea. It is estimated that of the predicted global oceanic heat flux of 32  1012 W,

_

Heat flow (W m 1)

1.5

1.0

0.5

0 20°C

Depth (s)

Thermal models are solutions to the inverse problem of finding the temperature as a function of age that best fits the depth and heat flow. The data are used to estimate the primary model parameters (typically plate thickness, basal temperature, and thermal expansion coefficient), subject to other parameters generally specified a priori. The global average depth/heat flow variation with age, calculated for a specific model, data, and parameters is used as a reference model to represent ‘normal’ oceanic lithosphere. ‘Anomalies’, deviations from the reference model, are then investigated to see if they reflect a significant difference in thermal (or other) processes from the global average. It is important to realize that anomalies (defined as those observed minus the model-predicted values) for various reference models may differ significantly, and hence lead to different tectonic inferences. Until recently, a 125 km-thick plate model by Parsons and Sclater (denoted here PSM) was commonly used. Subsequently, as more data became available, it was noted that PSM systematically overpredicts depths and underpredicts heat flow for lithosphere older than 70–100 My, causing widespread ‘anomalies.’ A later joint inversion of the depth and heat flow data by Stein and Stein found that these ‘anomalies’ are reduced significantly by a plate model termed GDH1. GDH1 has a thinner lithosphere (95710 km), and a basal temperature of 145071001C, consistent with the PSM estimate (135072751C). GDH1 predicts heat flow in mW m2 as a function of age, t, in millions of years equal to 510 t1/2 for ages less than about 55 My, and 48 þ 96 exp(  0.0278 t) for older ages. Inversion of the same data, while prescribing a basal temperature of 13501C (a typically assumed temperature for upwelling magma at the ridge based on results from experimental petrology), also yields a thin (100 km) plate, with a somewhat higher value of lithospheric conductivity and hence makes very similar predictions. As the quality of the observed data and our

40°C 60°C 80°C

4.0 0 4.5

W

5 km

Line 88_ 30

E

Figure 4 Portion of the heat flow data from the FlankFlux survey on the Juan de Fuca Ridge sites where bare rock penetrates the sediment column have dramatically higher heat flow than the surrounding areas. Away from these outcrops, heat flow varies inversely with depth to basement rock. Expected heat flow for its age is shown with the dashed line. (Reproduced with permission from Davis EE, Chapman DS, Mottl MJ et al. (1992) FlankFlux: an experiment to study the nature of hydrothermal circulation in young oceanic crust. Canadian Journal of Earth Sciences 29: 925–952.)

(c) 2011 Elsevier Inc. All Rights Reserved.

46

GEOPHYSICAL HEAT FLOW

approximately one third occurs by hydrothermal flow. Characteristic heat flow patterns in young crust appear associated with water flow. Surveys at sites like the Galapagos Spreading Center show high heat flow associated with presumed upwelling zones at basement highs or fault scarps, and low heat flow associated with presumably down-flowing water at basement lows. Heat flow over the highs may be two to ten (or more) times greater than at nearby sites where water may be recharging the system (Figure 4). Closely spaced surveys in young crust show high scatter, presumably in part due to the local variations in sediment distribution, basement relief, and crustal permeability. The most spectacular evidence for hydrothermal circulation are vents near the ridge crest where hydrothermally altered water at temperatures up to 4001C exits the seafloor. Hot rock or magma at shallow depths provides a large heat source for the flow. The high sulfide content of the hot fluids supports unusual local biota from bacteria to tube worms. Vent areas show complex flow patterns with very high heat flow close to low heat flow. Although most vigorous vents are near spreading centers, on older crust isolated crystalline crust outcrops surrounded by a well-sedimented area can also vent measurable amounts of hydrothermal fluid (e.g., the FlankFlux survey area on the Juan de Fuca plate). The persistence of the heat flow discrepancy well away from ridges indicates that hydrothermal heat loss occurs in older crust. The ‘sealing age,’ defined as that beyond which measured heat flow approximately equals that predicted, is presumed to indicate the near cessation of heat transfer by hydrothermal circulation. The sealing age for the entire global data set is about 65710 My. Hence, although there are presumably local variations, water flow in older crust seems not, in general, to transport significant amounts of heat. However, some heat flow surveys (e.g., in the north-west Atlantic Ocean on 80 My crust and the Maderia Abyssal Plain on 90 My crust) indicate that hydrothermal circulation may continue, even if relatively little heat is lost by convection into the sea water. The reasons for the ‘sealing’ are not well understood. An earlier view was that 100– 200 m of sediment would be sufficiently impermeable to ‘seal’ off the crust from the sea, so that heat flow at these sites would yield a ‘reliable’ value, i.e., that predicted by conduction-only thermal models. However, many such sites have lower than expected values. A simple way to reconcile these observations with the present understanding of seafloor hydrology to assume that if water cannot flow vertically through thick sediments at a particular site, it may

flow laterally to a fault or basement outcrop, and then be manifested as either high conductive heat flow or hot water exiting to the sea, such as observed for the FlankFlux area. It has been suggested that the porosity and permeability of the crust decrease with increasing age, thus significantly reducing the water flow. However, recent studies of permeability and seismic velocity suggest that most of the rapid change occurs within the first 5–15 My and that older crust may still retain some relatively permeable pathways. Hydrothermal circulation has profound implications for the chemistry of the oceanic crust and sea water, because sea water reacts with the crust, giving rise to hydrothermal fluid of significantly different composition. The primary geochemical effects are thought to result from the high-temperature water flow observed at ridge axes. Circulation of sea water through hot rock removes magnesium and sulfate from sea water and enriches calcium, potassium, silica, iron, manganese, and other elements within the hydrothermal solution. Estimating the volume of water flowing through the crust depends on the heat capacity of the water, the heat lost by convection, and the assumed temperature of the water, the latter having the largest uncertainties. Although the heat flow anomaly is greatest in young lithosphere, only about 30% of the hydrothermal heat loss and 7% of the water flow occurs in crust younger than 1 My. This effect should be significant for ocean chemistry because a water volume equivalent to that of the total ocean is estimated to circulate through the oceanic crust about once every 0.5–5 My. Hot Spots

Oceanic midplate swells are identified by seafloor depths shallower than expected for their lithospheric age. Thus, models of the processes giving rise to these regions rely on assessments of how their heat flow and other properties differ from unperturbed lithosphere. The origin of these swells is generally thought to be related to upwelling mantle plumes (hot spots) that result in uplift and volcanism. Two basic types of models have been proposed for their origin. In one, the swell is a thermal effect due to the hot spot thinning and heating the lithosphere at depth. In the second, the uplift is primarily due to the dynamic effects of the upwelling plume, which may largely reflect thermal buoyancy forces within the upwelling mantle. The thermal models predict significantly larger heat flow anomalies than do the dynamic models. Detailed heat flow measurements have been made for Hawaii, Cape Verde, Reunion, Bermuda, and Crozet swells. Relative to the PSM reference curve, large heat flow anomalies are suggested.

(c) 2011 Elsevier Inc. All Rights Reserved.

GEOPHYSICAL HEAT FLOW

47

0 1

Mid way 2

5 4

3 4

25

H 7

5 G

4

6

8 20

2000

5

Depth (m)

30°N

E DB

F

5

5

4

4000

8

7

6

5

4

3

2

1

Hawaii

C

GDH1 PSM

A 6000 175

165

170

160

155˚W

200

8

7

400 600 800 1000 1200 Distance (km)

G

F

E C B A

_

H

Heat flow (W m 2)

70

_

Heat flow (mW m 2 )

70

0

60 GDH1 50

PSM

40

4 6

60

5

1 3

2 GDH1

50 PSM 40

30

20 10 Time since hot spot passage (My)

0

0

200

400 600 800 1000 1200 Distance (km)

Figure 5 Heat flow data for transects along (lower left) and across (lower right) the Hawaiian Swell at the locations shown (upper left). The heat flow, though anomalously high with respect to the Parsons and Sclater (PSM) model, is at most slightly above that expected for GDH1. The predicted heat flow values are for 100 My (lower right) and 95–110 My (lower left). Figure modified from Von Herzen et al. (1982, 1989). (Reproduced with permission from Stein CA and Stein S (1993) Constraints of Pacific midplate swells from global depth-age and heat flow-age models. The Mesozoic Pacific, American Geophysical Union Monograph 77: 53–76.)

However, relative to both measured heat flow from off-swell lithosphere and the GDH1 reference curve, smaller heat flow anomalies (less than 5–8 mW m2) are deduced, supporting the idea of a primary dynamic mechanism for hotspots (Figure 5).

Application to Marine Margin Studies Subduction Zones

Many extensive marine heat flow surveys have been done near the Japanese subduction zones. On average, observed heat flow from the trench axis to the forearc area is lower than that characteristic of the crust seaward of the trench. However, heat flow is higher and more variable over the volcanic arc and back arc region compared to the area seaward of the trench. Recent surveys for subduction zones, including Barbados, Nankai, and Cascadia, show that heat flow is highly variable. Within accretionary prisms, high values are often associated with upward advection of pore fluids, typically found along faults and the bottom decollement. Although they have not been

investigated extensively, the active nonaccretionary forearcs seem to have the lowest mean heat fluxes (20– 30 mW m2). Relatively low heat flow values over the subducting lithosphere are the thermal consequence of the subduction of one plate beneath another. Overall, heat flow depends on the age, rate, geometry, and thermal structure of the subducting plate and the sediment thickness and deformation history of the region. Heat flow in marginal basins behind the volcanic arcs is often high. Many back arc basins with high heat flow appear to have formed by back-arc spreading processes similar to those at midocean ridges, within the last 50 My. Passive Continental Margins

Passive continental margins form after continental crust is rifted and seafloor spreading occurs. Because rifting heats the lithosphere, heat flow data can be used to constrain models of rifting. Subsequent to rifting, margins slowly subside as cooling occurs. Simple models for this process suggest that subsequent to rifting the additional heat will almost completely dissipate within 100 My. Although the

(c) 2011 Elsevier Inc. All Rights Reserved.

48

GEOPHYSICAL HEAT FLOW

amount of radioactive heating from continental crust may introduce some variability, older passive margins typically have heat flows similar to old oceanic crust, whereas young margins, such as in the Red Sea, have high heat flow.

Gas Hydrates

Gas hydrates, solid composites of biogenically derived gasses (mainly methane) combined with water ice, may be present in marine sediments, especially at continental margins, under the correct pressure and temperature conditions. Gas hydrates are detected by direct sampling, and inferred from seismic reflection data when a strong bottom-simulating reflector (BSR) is produced by the seismic velocity contrast between the gas hydrate and the sediment below. Given a depth (and thus pressure) of a bottom-simulating reflector, its temperature can be predicted from known relationships and a heat flow calculated. Alternatively, heat flow data can be used to estimate the bottom-simulating reflector temperature. The volume of hydrocarbons contained in marine gas hydrates is large. Changes in eustatic sea level (hence pressure) or bottom seawater temperatures could result in their release and thus increase greenhouse gasses and affect global climate.

Further Reading Humphris S, Mullineaux L, Zierenberg R, and Thomson R (eds.) (1995) Seafloor hydrothermal systems, physical, chemical, biological, and geological interactions, Geophysical Monographs, 91, 425--445. Hyndman RD, Langseth MG, and Von Herzen RP (1987) Deep Sea Drilling Project geothermal measurements: a review. Review of Geophysics 25: 1563--1582. Langseth MG, Jr and Von Herzen RP (1971) Heat Sow through the Soor of the world oceans. In: Maxwell AE (ed.) The Sea, vol. IV, part 1, pp. 299--352. New York: Wiley-Interscience. Lowell RP, Rona PA, and Von Herzen RP (1995) Seafloor hydrothermal systems. Journal of Geophysical Research 100: 327--352. Parsons B and Sclater JG (1977) An analysis of the variation of ocean floor bathymetry and heat flow with age. Journal of Geophysical Research 82: 803--827. Pollack HN, Hurter SJ, et al. (1993) Heat flow from the earth’s interior: analysis of the global data set. Review of Geophysics 31: 267--280. Stein CA and Stein S (1992) A model for the global variation in oceanic depth and heat flow with lithospheric age. Nature 359: 123--129. Von Herzen RP (1987) Measurement of oceanic heat flow. In: Sammis C and Henyey T (eds.) Methods of Experimental Physics – Geophysics, vol. 24, Part B, pp. 227--263. London: Academic Press. Wright JA and Louden KE (eds.) (1989) CRC Handbook of Seafloor Heat Flow. Boca Raton: CRC Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES K. Lambeck, Australian National University, Canberra, ACT, Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1157–1167, & 2001, Elsevier Ltd.

Introduction Geological, geomorphological, and instrumental records point to a complex and changing relation between land and sea surfaces. Elevated coral reefs or wave-cut rock platforms indicate that in some localities sea levels have been higher in the past, while observations elsewhere of submerged forests or flooded sites of human occupation attest to levels having been lower. Such observations are indicators of the relative change in the land and sea levels: raised shorelines are indicative of land having been uplifted or of the ocean volume having decreased, while submerged shorelines are a consequence of land subsidence or of an increase in ocean volume. A major scientific goal of sea-level studies is to separate out these two effects. A number of factors contribute to the instability of the land surfaces, including the tectonic upheavals of the crust emanating from the Earth’s interior and the planet’s inability to support large surface loads of ice or sediments without undergoing deformation. Factors contributing to the ocean volume changes include the removal or addition of water to the oceans as ice sheets wax and wane, as well as addition of water into the oceans from the Earth’s interior through volcanic activity. These various processes operate over a range of timescales and sea level fluctuations can be expected to fluctuate over geological time and are recorded as doing so. The study of such fluctuations is more than a scientific curiosity because its outcome impacts on a number of areas of research. Modern sea level change, for example, must be seen against this background of geologically–climatologically driven change before contributions arising from the actions of man can be securely evaluated. In geophysics, one outcome of the sea level analyses is an estimate of the viscosity of the Earth, a physical property that is essential in any quantification of internal convection and thermal processes. Glaciological modeling of the behavior of large ice sheets during the last cold

period is critically dependent on independent constraints on ice volume, and this can be extracted from the sea level information. Finally, as sea level rises and falls, so the shorelines advance and retreat. As major sea level changes have occurred during critical periods of human development, reconstructions of coastal regions are an important part in assessing the impact of changing sea levels on human movements and settlement.

Tectonics and Sea Level Change Major causes of land movements are the tectonic processes that have shaped the planet’s surface over geological time. Convection within the high-temperature viscous interior of the Earth results in stresses being generated in the upper, cold, and relatively rigid zone known as the lithosphere, a layer some 50–100 km thick that includes the crust. This convection drives plate tectonics — the movement of large parts of the lithosphere over the Earth’s surface — mountain building, volcanism, and earthquakes, all with concomitant vertical displacements of the crust and hence relative sea level changes. The geological record indicates that these processes have been occurring throughout much of the planet’s history. In the Andes, for example, Charles Darwin identified fossil seashells and petrified pine trees trapped in marine sediments at 4000 m elevation. In Papua New Guinea, 120 000-year-old coral reefs occur at elevations of up to 400 m above present sea level (Figure 1). One of the consequences of the global tectonic events is that the ocean basins are being continually reshaped as mid-ocean ridges form or as ocean floor collides with continents. The associated sea level changes are global but their timescale is long, of the order 107 to 108 years, and the rates are small, less than 0.01 mm per year. Figure 2 illustrates the global sea level curve inferred for the past 600 million years from sediment records on continental margins. The long-term trends of rising and falling sea levels on timescales of 50–100 million years are attributed to these major changes in the ocean basin configurations. Superimposed on this are smaller-amplitude and shorter-period oscillations that reflect more regional processes such as large-scale volcanism in an ocean environment or the collision of continents. More locally, land is pushed up or down episodically

(c) 2011 Elsevier Inc. All Rights Reserved.

49

50

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

and, in any discussion of global changes in sea level, information from such localities is best set aside in favor of observations from more stable environments.

Glacial Cycles and Sea Level Change

600 400 200 0 590

500

440 410

360 290 250 210 Time (million years BP)

140

65

35

5

Sea level above present level (m)

Pliocene_Pleistocene

Miocene

Paleocene & Eocene

Oligocene

Cretaceous

Permian

Carboniferous

Devonian

Silurian

Cambrian

Ordovician

in response to the deeper processes. The associated vertical crustal displacements are rapid, resulting in sea level rises or falls that may attain a few meters in amplitude, but which are followed by much longer periods of inactivity or even a relaxation of the original displacements. The raised reefs illustrated in Figure 1, for example, are the result of a large number of episodic uplift events each of typically a meter amplitude. Such displacements are mostly local phenomena, the Papua New Guinea example extending only for some 100–150 km of coastline. The episodic but local tectonic causes of the changing position between land and sea can usually be identified as such because of the associated seismic activity and other tell-tale geological signatures. The development of geophysical models to describe these local vertical movements is still in a state of infancy

Jurassic

Figure 1 Raised coral reefs from the Huon Peninsula, Papua New Guinea. In this section the highest reef indicated (point 1) is about 340 m above sea level and is dated at about 125 000 years old. Elsewhere this reef attains more than 400 m elevation. The top of the present sea cliffs (point 2) is about 7000 years old and lies at about 20 m above sea level. The intermediate reef tops formed at times when the rate of tectonic uplift was about equal to the rate of sea level rise, so that prolonged periods of reef growth were possible. Photograph by Y. Ota.

Triassic

More important for understanding sea level change on human timescales than the tectonic contributions – important in terms of rates of change and in terms of their globality – is the change in ocean volume driven by cyclic global changes in climate from glacial to interglacial conditions. In Quaternary time, about the last two million years, glacial and interglacial conditions have followed each other on timescales of the order of 104–105 years. During interglacials, climate conditions were similar to those of today and sea levels were within a few meters of their present day position. During the major glacials, such as 20 000 years ago, large ice sheets formed in the northern hemisphere and the Antarctic ice sheet expanded, taking enough water out of the oceans to lower sea levels by between 100 and 150 m. Figure 3 illustrates the changes in global sea level over the last 130 000 years, from the last interglacial, the last time that conditions were similar to those of today, through the Last Glacial Maximum and to the present. At the end of the last interglacial, at 120 000 years ago, climate began to fluctuate; increasingly colder conditions were reached, ice sheets over North America and Europe became more or less permanent features of the landscape, sea levels reached progressively lower values, and large parts of today’s coastal shelves were exposed. Soon after the culmination of maximum glaciation, the ice sheets again disappeared, within a period of about

_200

Figure 2 Global sea level variations through the last 600 million years estimated from seismic stratigraphic studies of sediments deposited at continental margins. Redrawn with permission from Hallam A (1984) Pre-Quaternary sea-level changes. Annual Review of Earth and Planetary Science 12: 205–243.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

Relative sea level (m)

20 0 _20 _40 _60 _80 _100 _120

0

20

60 80 100 40 Time (thousand years BP)

120

Figure 3 Global sea level variations (relative to present) since the time of the last interglacial 120 000–130 000 years ago when climate and environmental conditions were last similar to those of the last few thousand years. Redrawn with permission from Chapell J et al. (1996) Reconciliation of Late Quaternary sea level changes derived from coral terraces at Huon Peninsula with deep sea oxygen isotope records. Earth and Planetary Science Letters 141b: 227–236.

10 000 years, and climate returned to interglacial conditions. The global changes illustrated in Figure 3 are only one part of the sea level signal because the actual ice– water mass exchange does not give rise to a spatially uniform response. Under a growing ice sheet, the Earth is stressed; the load stresses are transmitted through the lithosphere to the viscous underlying mantle, which begins to flow away from the stressed area and the crust subsides beneath the ice load. When the ice melts, the crust rebounds. Also, the meltwater added to the oceans loads the seafloor and the additional stresses are transmitted to the mantle, where they tend to dissipate by driving flow to unstressed regions below the continents. Hence the seafloor subsides while the interiors of the continents rise, causing a tilting of the continental margins. The combined adjustments to the changing ice and water loads are called the glacio- and hydro-isostatic effects and together they result in a complex pattern of spatial sea level change each time ice sheets wax and wane.

Observations of Sea Level Change Since the Last Glacial Maximum Evidence for the positions of past shorelines occurs in many forms. Submerged freshwater peats and tree stumps, tidal-dwelling mollusks, and archaeological sites would all point to a rise in relative sea level since the time of growth, deposition, or construction. Raised coral reefs, such as in Figure 1, whale bones cemented in beach deposits, wave-cut rock platforms

51

and notches, or peats formed from saline-loving plants would all be indicative of a falling sea level since the time of formation. To obtain useful sea level measurements from these data requires several steps: an understanding of the relationship between the feature’s elevation and mean sea level at the time of growth or deposition, a measurement of the height or depth with respect to present sea level, and a measurement of the age. All aspects of the observation present their own peculiar problems but over recent years a substantial body of observational evidence has been built up for sea level change since the last glaciation that ended at about 20 000 years ago. Some of this evidence is illustrated in Figure 4, which also indicates the very significant spatial variability that may occur even when, as is the case here, the evidence is from sites that are believed to be tectonically stable. In areas of former major glaciation, raised shorelines occur with ages that are progressively greater with increasing elevation and with the oldest shorelines corresponding to the time when the region first became ice-free. Two examples, from northern Sweden and Hudson Bay in Canada, respectively, are illustrated in Figure 4A. In both cases sea level has been falling from the time the area became ice-free and open to the sea. This occurred at about 9000 years ago in the former case and about 7000 years later in the second case. Sea level curves from localities just within the margins of the former ice sheet are illustrated in Figure 4B, from western Norway and Scotland, respectively. Here, immediately after the area became ice-free, the sea level fell but a time was reached when it again rose. A local maximum was reached at about 6000 years ago, after which the fall continued up until the present. Farther away from the former ice margins the observed sea level pattern changes dramatically, as is illustrated in Figure 4C. Here the sea level initially rose rapidly but the rate decreased for the last 7000 years up to the present. The two examples illustrated, from southern England and the Atlantic coast of the United States, are representative of tectonically stable localities that lie within a few thousand kilometers from the former centers of glaciation. Much farther away again from the former ice sheets the sea level signal undergoes a further small but significant change in that the present level was first reached at about 6000 years ago and then exceeded by a small amount before returning to its present value. The two examples illustrated in Figure 4D are from nearby localities in northern Australia. At both sites present sea level was reached about 6000 years ago after a prolonged period of rapid rise with resulting highstands that are small in amplitude but geographically variable.

(c) 2011 Elsevier Inc. All Rights Reserved.

52

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

Shoreline elevation (m)

250 140

Angermanland, northern Sweden

200

Ottawa Island, Hudson Bay, Canada

120 100

150 80 100

60 40

50 20 0 10

(A)

8

6

4

2

0

Shoreline elevation (m)

50 Andoya, northern Norway

40

7

5

6

4

0

30 20 10 10

0

15

5

10

0

0 15

5

10

0

10

0

0

_ 10 _ 10 _ 20 _ 30 _ 40 10

Shoreline elevation (m)

1

20

_ 20

Dungeness, Kent,

8

6

4

2

Cape Charles, Virginia, USA

0

_ 30 12

10

8

6

4

2

0

3

0

Karumba, Gulf of Carpentaria, North Queensland

_ 10 2 _ 20

2

_ 30

0

1 −1 6 5 4 3 2 1 0 North Queensland, Australia

_ 40 _ 50 12

(D)

2

30

southern England (C)

3

Firth of Forth, Scotland

40

10

Shoreline elevation (m)

8

50

_ 10 20

(B)

0

10

8

6

1

0 0 6 5 4 2 4 Time (1000 radiocarbon years BP)

3

2

1

0

˚ ngermanland, Figure 4 Characteristic sea level curves observed in different localities (note the different scales used). (A) From A Gulf of Bothnia, Sweden, and from Hudson Bay, Canada. Both sites lie close to centers of former ice sheet, over northern Europe and North America respectively. (B) From the Atlantic coast of Norway and the west coast of Scotland. These sites are close to former icesheet margins at the time of the last glaciation. (C) From southern England and Virginia, USA. These sites lie outside the areas of former glaciation and at a distance from the margins where the crust is subsiding in response to the melting of the nearby icesheet. (D) Two sites in northern Australia, one for the Coral Sea coast of Northern Queensland and the second from the Gulf of Carpentaria some 200 km away but on the other side of the Cape York Peninsula. At both localities sea level rose until about 6000 years ago before peaking above present level.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

The examples illustrated in Figure 4 indicate the rich spectrum of variation in sea level that occurred when the last large ice sheets melted. Earlier glacial cycles resulted in a similar spatial variability, but much of the record before the Last Glacial Maximum has been overwritten by the effects of the last deglaciation.

Glacio-Hydro-Isostatic Models If, during the decay of the ice sheets, the meltwater volume was distributed uniformly over the oceans, then the sea level change at time t would be Dze ðtÞ ¼ ðchange in ice volume=ocean surface areaÞ  ri =rw

½1a

where ri and rw are the densities of ice and water, respectively. (See end of article – symbols used.) This term is the ice-volume equivalent sea level and it provides a measure of the change in ice volume through time. Because the ocean area changes with time a more precise definition is Dze ðtÞ ¼

ð dVi ri ri dt dt r0 A0 ðtÞ

½1b

t

where A0 is the area of the ocean surface, excluding areas covered by grounded ice. The sea level curve illustrated in Figure 3 is essentially this function. However, it represents only a zero-order approximation of the actual sea level change because of the changing gravitational field and deformation of the Earth. In the absence of winds or ocean currents, the ocean surface is of constant gravitational potential. A planet of a defined mass distribution has a family of such surfaces outside it, one of which – the geoid – corresponds to mean sea level. If the mass distribution on the surface (the ice and water) or in the interior (the load-forced mass redistribution in the mantle) changes, so will the gravity field, the geoid, and the sea level change. The ice sheet, for example, represents a large mass that exerts a gravitational pull on the ocean and, in the absence of other factors, sea level will rise in the vicinity of the ice sheet and fall farther away. At the same time, the Earth deforms under the changing load, with two consequences: the land surface with respect to which the level of the sea is measured is time-dependent, as is the gravitational attraction of the solid Earth and hence the geoid. The calculation of the change of sea level resulting from the growth or decay of ice sheets therefore

53

involves three steps: the calculation of the amount of water entering into the ocean and the distribution of this meltwater over the globe; the calculation of the deformation of the Earth’s surface; and the calculation of the change in the shape of the gravitational equipotential surfaces. In the absence of vertical tectonic motions of the Earth’s surface, the relative sea level change Dzðj; tÞ at a site j and time t can be written schematically as Dzðj; tÞ ¼ Dze ðtÞ þ DzI ðj; tÞ

½2

where DzI(j, t) is the combined perturbation from the uniform sea level rise term [1]. This is referred to as the isostatic contribution to relative sea level change. In a first approximation, the Earth’s response to a global force is that of an elastic outer spherical layer (the lithosphere) overlying a viscous or viscoelastic mantle that itself contains a fluid core. When subjected to an external force (e.g., gravity) or a surface load (e.g., an ice cap), the planet experiences an instantaneous elastic deformation followed by a timedependent or viscous response with a characteristic timescale(s) that is a function of the viscosity. Such behavior of the Earth is well documented by other geophysical observations: the gravitational attraction of the Sun and Moon raises tides in the solid Earth; ocean tides load the seafloor with a time-dependent water load to which the Earth’s surface responds by further deformation; atmospheric pressure fluctuations over the continents induce deformations in the solid Earth. The displacements, measured with precision scientific instruments, have both an elastic and a viscous component, with the latter becoming increasingly important as the duration of the load or force increases. These loads are much smaller than the ice and water loads associated with the major deglaciation, the half-daily ocean tide amplitudes being only 1% of the glacial sea level change, and they indicate that the Earth will respond to even small changes in the ice sheets and to small additions of meltwater into the oceans. The theory underpinning the formulation of planetary deformation by external forces or surface loads is well developed and has been tested against a range of different geophysical and geological observations. Essentially, the theory is one of formulating the response of the planet to a point load and then integrating this point-load solution over the load through time, calculating at each epoch the surface deformation and the shape of the equipotential surfaces, making sure that the meltwater is appropriately distributed into the oceans and that the total ice–water mass is preserved. Physical inputs into the

(c) 2011 Elsevier Inc. All Rights Reserved.

54

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

formulation are the time–space history of the ice sheets, a description of the ocean basins from which the water is extracted or into which it is added, and a description of the rheology, or response parameters, of the Earth. For the last requirement, the elastic properties of the Earth, as well as the density distribution with depth, have been determined from seismological and geodetic studies. Less well determined are the viscous properties of the mantle, and the usual procedure is to adopt a simple parametrization of the viscosity structure and to estimate the relevant parameters from analyses of the sea level change itself. The formulation is conveniently separated into two parts for schematic reasons: the glacio-isostatic effect representing the crustal and geoid displacements due to the ice load, and the hydro-isostatic effect due to the water load. Thus ½3 Dzðj; tÞ ¼ Dze ðtÞ þ Dzi ðj; tÞ þ Dzw ðj; tÞ where Dzi(j,t) is the glacio-isostatic and Dzw(j,t) the hydro-isostatic contribution to sea level change. (In reality the two are coupled and this is included in the formulation). If Dzi(j,t) is evaluated everywhere, the past water depths and land elevations H(j,t) measured with respect to coeval sea level are given by ½4 H ðj; tÞH0 ðjÞ  Dzðj; tÞ where HðfÞ is the present water depth or land elevation at location j. A frequently encountered concept is eustatic sea level, which is the globally averaged sea level at any time t. Because of the deformation of the seafloor during and after the deglaciation, the isostatic term Dzðj; tÞ is not zero when averaged over the ocean at any time t, so that the eustatic sea level change is Dzeus ðtÞ ¼ Dze ðtÞ þ hDzI fj; tgi0 where the second term on the right-hand side denotes the spatially averaged isostatic term. Note that Dze ðtÞ relates directly to the ice volume, and not Dzeus ðtÞ.

The Anatomy of the Sea Level Function The relative importance of the two isostatic terms in eqn. [3] determines the spatial variability in the sea level signal. Consider an ice sheet of radius that is much larger than the lithospheric thickness: the limiting crustal deflection beneath the center of the load is Iri =Irm where I is the maximum ice thickness and Irm is the upper mantle density. This is the local isostatic approximation and it provides a reasonable approximation of the crustal deflection if the loading time is long compared with the relaxation time of the mantle. Thus for a 3 km thick ice sheet the maximum

deflection of the crust can reach 1 km, compared with a typical ice volume for a large ice sheet that raises sea level globally by 50–100 m. Near the centers of the formerly glaciated regions it is the crustal rebound that dominates and sea level falls with respect to the land. This is indeed observed, as illustrated in Figure 4A, and the sea level curve here consists of essentially the sum of two terms, the major glacio-isostatic term and the minor ice-volume equivalent sea level term Dze ðtÞ (Figure 5A). Of note is that the rebound continues long after all ice has disappeared, and this is evidence that the mantle response includes a memory of the earlier load. It is the decay time of this part of the curve that determines the mantle viscosity. As the ice margin is approached, the local ice thickness becomes less and the crustal rebound is reduced and at some stage is equal, but of opposite sign, to Dze(t). Hence sea level is constant for a period (Figure 5B) before rising again when the rebound becomes the minor term. After global melting has ceased, the dominant signal is the late stage of the crustal rebound and levels fall up to the present. Thus the oscillating sea level curves observed in areas such as Norway and Scotland (Figure 4B) are essentially the sum of two effects of similar magnitude but opposite sign. The early part of the observation contains information on earth rheology as well as on the local ice thickness and the globally averaged rate of addition of meltwater into the oceans. Furthermore, the secondary maximum is indicative of the timing of the end of global glaciation and the latter part of the record is indicative mainly of the mantle response. At the sites beyond the ice margins it is the meltwater term Dze(t) that is important, but it is not the sole factor. When the ice sheet builds up, mantle material flows away from the stressed mantle region and, because flow is confined, the crust around the periphery of the ice sheet is uplifted. When the ice sheet melts, subsidence of the crust occurs in a broad zone peripheral to the original ice sheet and at these locations the isostatic effect is one of an apparent subsidence of the crust. This is illustrated in Figure 5C. Thus, when the ocean volumes have stabilized, sea level continues to rise, further indicating that the planet responds viscously to the changing surface loads. The early part of the observational record (e.g., Figure 4C) is mostly indicative of the rate at which meltwater is added into the ocean, whereas the latter part is more indicative of mantle viscosity. In all of the examples considered so far it is the glacial-rebound that dominates the total isostatic adjustment, the hydro-isostatic term being present but comparatively small. Consider an addition of water that raises sea level by an amount D. The local

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

55

300

300

Relative sea level change (m)

Prediction

ice

200

100

100

0

0 ESL

_ 100 (A)

200

20

Angerman River, Gulf of Bothnia

15

10

5

_ 100

0

20

Relative sea level change (m)

150

10

5

0

15

10

5

0

150

ice

100

100

50

50

0

0 _ 50

_ 50 ESL

_ 100 _ 150 20

15

200

200

(B)

Observation

Upper Firth of Forth, Scotland

15

5

10

_ 100 _ 150

0

20

20

20 water

0

0 Relative sea level change (m)

_ 20

ice

_ 20

_ 40

_ 40

_ 60

_ 60 _ 80

ESL

_ 80

_ 100

_ 100 _ 120

Dungeness, south of England

_ 140 (C)

15

20

Relative sea level change (m)

10

(D)

0 _ 10

10

5

20

10

15

5

0

0 _ 10

ice ESL

Karumba, Gulf of Carpentaria

_ 30 8

_ 140

10

water

_ 20 _ 40 10

0

_ 120

6

4

2

0

_ 20 _ 30 _ 40 10

8

6

4

2

0

Time (1000 years BP)

Figure 5 Schematic representation of the sea level curves in terms of the ice-volume equivalent function Dze ðtÞ (denoted by ESL) and the glacio- and hydro-isostatic contributions Dzi ðf; tÞ, Dzw ðf; tÞ (denoted by ice and water, respectively). The panels on the left indicate the predicted individual components and the panels on the right indicate the total predicted change compared with the observed values (data points). (A) For the sites at the ice center where Dzi ðf; tÞbDze ðtÞbDzw ðf; tÞ. (B) For sites near but within the ice margin where jDzi ðf; tÞjBjDze ðtÞjbjDzw ðf; tÞj, but the first two terms are of opposite sign. (C) For sites beyond the ice margin where jDze ðtÞj > jDzi ðf; tÞjbjDzw ðf; tÞj. (D) For sites at continental margins far from the former ice sheets where jDzw ðf; tÞj > jDzi ðtÞj. Adapted with permission from Lambeck K and Johnston P (1988). The viscosity of the mantle: evidence from analyses of glacial rebound phenomena. In: Jackson ISN (ed.) The Earth’s Mantle – Composition, Structure and Evolution, pp. 461–502. Cambridge: Cambridge University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

Sea Level Change and Shoreline Migration: Some Examples Figure 6 illustrates the ice-volume equivalent sea level function Dze ðtÞ since the time of the last maximum glaciation. This curve is based on sea level

indicators from a number of localities, all far from the former ice sheets, with corrections applied for the isostatic effects. The right-hand axis indicates the corresponding change in ice volume (from the relation [1b]). Much of this ice came from the ice sheets over North America and northern Europe but a not insubstantial part also originated from Antarctica. Much of the melting of these ice sheets occurred within 10 000 years, at times the rise in sea level exceeding 30 mm per year, and by 6000 years ago most of the deglaciation was completed. With this sea level function, individual ice sheet models, the formulation for the isostatic factors, and a knowledge of the topography and bathymetry of the world, it becomes possible to reconstruct the paleo shorelines using the relation [4]. Scandinavia is a well-studied area for glacial rebound and sea level change since the time the ice retreat began about 18 000 years ago. The observational evidence is quite plentiful and a good record of ice margin retreat exists. Figure 7 illustrates examples for two epochs. The first (Figure 7A), at 16 000 years ago, corresponds to a time after onset of deglaciation. A large ice sheet existed over

20

Time (radiocarbon, 1000 years BP) 15 10 5

0

0

0

_20

7.7

_40 _60

23.1

_80 _100

38.6

Ice volume (106 km3 )

isostatic response to this load is Drw =rm , where rw is the density of ocean water. This gives an upper limit to the amount of subsidence of the sea floor of about 30 m for a 100 m sea level rise. This is for the middle of large ocean basins and at the margins the response is about half as great. Thus the hydro-isostatic effect is significant and is the dominant perturbing term at margins far from the former ice sheets. This occurs at the Australian margin, for example, where the sea level signal is essentially determined by Dze ðtÞ and the water-load response (Figure 4D). Up to the end of melting, sea level is dominated by Dze ðtÞ but thereafter it is determined largely by the water-load term Dzw ðj; tÞ, such that small highstands develop at 6000 years. The amplitudes of these highstands turn out to be strongly dependent on the geometry of the waterload distribution around the site: for narrow gulfs, for example, their amplitude increases with distance from the coast and from the water load at rates that are particularly sensitive to the mantle viscosity. While the examples in Figure 5 explain the general characteristics of the global spatial variability of the sea level signal, they also indicate how observations of such variability are used to estimate the physical quantities that enter into the schematic model (3). Thus, observations near the center of the ice sheet partially constrain the mantle viscosity and central ice thickness. Observations from the ice sheet margin partially constrain both the viscosity and local ice thickness and establish the time of termination of global melting. Observations far from the ice margins determine the total volumes of ice that melted into the oceans as well as providing further constraints on the mantle response. By selecting data from different localities and time intervals, it is possible to estimate the various parameters that underpin the sea level eqn. [3] and to use these models to predict sea level and shoreline change for unobserved areas. Analyses of sea-level change from different regions of the world lead to estimates for the lithospheric thickness of between 60–100 km, average upper mantle viscosity (from the base of the lithosphere to a depth of 670 km) of about (1–5)1020 Pa s and an average lower mantle viscosity of about (1–5)1022 Pa s. Some evidence exists that these parameters vary spatially; lithosphere thickness and upper mantle viscosity being lower beneath oceans than beneath continents.

Ice-volume equivalent sea level (m)

56

_120 _140 25

54.0 5 20 15 10 Time (calibrated, 1000 years BP)

0

Figure 6 The ice-volume equivalent function Dze ðtÞ and ice volumes since the time of the last glacial maximum inferred from corals from Barbados, from sediment facies from north-western Australia, and from other sources for the last 7000 years. The actual sea level function lies at the upper limit defined by these observations (continuous line). The upper time scale corresponds to the radiocarbon timescale and the lower one is calibrated to calendar years. Adapted with permission from Fleming K et al. (1998) Refining the eustatic sea-level curve since the LGM using the far- and intermediate-field sites. Earth and Planetary Science Letters 163: 327–342; and Yokoyama Y et al. (2000) Timing of the Last Glacial Maximum from observed sea-level minimum. Nature 406: 713–716.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

57

Figure 8 Shoreline reconstructions for South East Asia and northern Australia at the time of the last glacial maximum, (A) at about 18 000 years ago and (B) at 12 000 years ago. The water depth contours are the same as in Figure 7. The water depths in the inland lakes at 18 000 years ago are contoured at 25 m intervals relative to their maximum levels that can be attained without overflowing.

Figure 7 Shoreline reconstructions for Europe (A) at 16 000 and (B) at 10 500 years ago. The contours are at 400 m intervals for the ice thickness (white) and 25 m for the water depths less than 100 m. The orange, yellow, and red contours are the predicted lines of equal sea level change from the specified epoch to the present and indicate where shorelines of these epochs could be expected if conditions permitted their formation and preservation. The zero contour is in yellow; orange contours, at intervals of 100 m, are above present, and red contours, at 50 m intervals, are below present. At 10 500 years ago the Baltic is isolated from the Atlantic and its level lies about 25 m above that of the latter.

Scandinavia with a smaller one over the British Isles. Globally sea level was about 110 m lower than now and large parts of the present shallow seas were exposed, for example, the North Sea and the English Channel but also the coastal shelf farther south such as the northern Adriatic Sea. The red and orange contours indicate the sea level change between this period and the present. Beneath the ice these rebound contours are positive, indicating that if shorelines could form here they would be above sea level today. Immediately beyond the ice margin, a broad but shallow bulge develops in the topography, which will subside as the ice sheet retreats. At the second epoch selected (Figure 7B) 10 500 years ago, the ice has retreated and reached a temporary halt as the climate briefly returned to colder conditions; the Younger Dryas time of Europe. Much of the Baltic was then ice-free and a freshwater lake developed at some

(c) 2011 Elsevier Inc. All Rights Reserved.

58

GLACIAL CRUSTAL REBOUND, SEA LEVELS, AND SHORELINES

25–30 m above coeval sea level. The flooding of the North Sea had begun in earnest. By 9000 years ago, most of the ice was gone and shorelines began to approach their present configuration. The sea level change around the Australian margin has also been examined in some detail and it has been possible to make detailed reconstructions of the shoreline evolution there. Figure 8 illustrates the reconstructions for northern Australia, the Indonesian islands, and the Malay–IndoChina peninsula. At the time of the Last Glacial Maximum much of the shallow shelves were exposed and deeper depressions within them, such as in the Gulf of Carpentaria or the Gulf of Thailand, would have been isolated from the open sea. Sediments in these depressions will sometimes retain signatures of these pre-marine conditions and such data provide important constraints on the models of sea level change. Part of the information illustrated in Figure 6, for example, comes from the shallow depression on the Northwest Shelf of Australia. By 12 000 years ago the sea has begun its encroachment of the shelves and the inland lakes were replaced by shallow seas.

Symbols used Ice-volume equivalent sea level. Uniform change in sea level produced by an ice volume that is distributed uniformly over the ocean surface. Perturbation in sea level due to glacioiDzI ðtÞ sostatic Dzi(t) and hydro-isostatic Dzw(t) effects. Dzeus ðtÞ Eustatic sea level change. The globally averaged sea level at time t. Dze ðtÞ

See also Beaches, Physical Processes Affecting. Coral Reefs. Fiordic Ecosystems. Geomorphology. Lagoons. Mangroves. Past Climate from Corals. Rocky Shores. Salt Marshes and Mud Flats. Salt Marsh Vegetation. Sandy Beaches, Biology of. Sea Level Change.

Further Reading Lambeck K (1988) Geophysical Geodesy: The Slow Deformations of the Earth. Oxford: Oxford University Press. Lambeck K and Johnston P (1999) The viscosity of the mantle: evidence from analysis of glacial-rebound phenomena. In: Jackson ISN (ed.) The Earth’s Mantle – Composition, Structure and Evolution, pp. 461--502. Cambridge: Cambridge University Press. Lambeck K, Smither C, and Johnston P (1998) Sea-level change, glacial rebound and mantle viscosity for northern Europe. Geophysical Journal International 134: 102--144. Peltier WR (1998) Postglacial variations in the level of the sea: implications for climate dynamics and solid-earth geophysics. Reviews in Geophysics 36: 603--689. Pirazzoli PA (1991) World Atlas of Holocene Sea-Level Changes. Amsterdam: Elsevier. van de Plassche O (ed.) (1986) Sea-Level Research: A Manual for the Collection and Evaluation of Data. Norwich: Geo Books. Sabadini R, Lambeck K, and Boschi E (1991) Glacial Isostasy, Sea Level and Mantle Rheology. Dordrecht: Kluwer.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLIDERS C. C. Eriksen, University of Washington, Seattle, WA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Underwater gliders are a recently developed class of autonomous underwater vehicle (AUV) driven by buoyancy changes to fly along saw-tooth trajectories through the ocean. Gliders quickly have become the AUVs with the highest endurance and longest range. They are able to sample the ocean interior at comparatively low cost because they can operate independently of ships, some for the better part of a year under global remote control. Sampling the ocean on space and timescales as fine as those dominating its variability and over long ranges and durations has been a challenge throughout the history of ocean science. Traditionally, most knowledge of the ocean interior has been collected by ships. Starting in the late twentieth century, moored observations have complemented those from ship-based surveys, and more recently satellite remote sensing has provided images of sea surface conditions globally on ever-finer space and timescales. Were these techniques inexpensive, there would be no particular need for autonomous data collection by moving platforms. Unfortunately, ships stay at sea only for a month or two and rarely are directed to sample the same region for multiple cruises. Moorings sampling the open ocean number in the hundreds globally, typically in the dozens in the deep sea. This means variability of the ocean interior is generally undersampled. Undersampling is arguably the principal impediment to description and understanding of the ocean. Gliders have been developed to address the shortcomings of conventional observing means so that such dominant sources of variability as mesoscale eddies, fronts, and boundary currents can be resolved simultaneously in space and time and over sufficiently long periods and wide domains to allow them to be understood. They are able to make progress in solving the sampling problem simply by the numbers their economy affords. The salient characteristic of gliders is that they cost roughly the equivalent of a few days of research vessel operation to build and can be operated for a few months for the cost of a day of ship time.

The tasks gliders do best are those for which ships are least suited: intensive, regular, and sustained observations of oceanic properties that are readily measured by electronic means. Observations that require large or power-hungry instruments, physical collection of samples, or specialized labor are unsuitable for gliders and continue the need for ship support. Gliders excel at measuring standard properties typically collected by a conductivity–temperature–depth (CTD) package. Among their many advantages is their ability to function well through the most severe seas the ocean has to offer, day and night, around the globe. They are a means to erasing the fair-weather bias of ship-based observations. One may sample the ocean at deliberately chosen locations and times with gliders, examine the data very shortly after it is collected, and alter the sampling plan as often as a glider reaches the sea surface and communicates. The density of observations can readily be scaled to the phenomena of interest by adjusting the number and distribution of platforms to address local, regional, or global variability on diurnal, fortnightly, or seasonal scales, for example. Of course, gliders are not without limitation: long range and high endurance are achieved at the cost of traveling slowly through the ocean. When they cease to communicate, they are lost.

A Short History John Swallow began the era of exploring the ocean interior autonomously in the 1950s with acoustically tracked floats ballasted to follow water motions at depth, famously deployed in the various North Atlantic locales. In following decades, acoustic floattracking technology extended to ocean basin scale. Russ Davis and Doug Webb turned to satellite tracking to follow the motion of Autonomous Lagrangian Circulation Explorer (ALACE) floats that periodically adjusted their buoyancy to reach the sea surface for communication, and then returned to a ‘parking’ depth to drift. ALACE floats were the precursor to Argo floats, platforms used in an ambitious international program to seed the ocean with 3000 profiling floats. While Argo floats continue to track horizontal water movements, the focus of the Argo program is largely on CTD profiles collected periodically to describe large-scale features of ocean circulation. Gliders are the descendants of profiling floats. They can be thought of as profiling floats with wings.

(c) 2011 Elsevier Inc. All Rights Reserved.

59

60

GLIDERS

Their essential difference is that their sampling location is controlled, subject to a limited ability to navigate through currents. Floats randomly sample where currents carry them. While gliders do not offer a Lagrangian view of the ocean, they are commanded to describe oceanic conditions at deliberate times and locations, adding to the Eulerian description of the ocean. Henry Stommel recognized the potential of underwater gliders (the idea was shared with him by his friend, neighbor, and colleague Doug Webb), and published a science fiction article describing them in 1989. Written as the retrospective of a scientist working at a glider network control station in 2021, the article has proved remarkably prescient in describing the gliders that were developed over the next decade and a half. Stommel envisioned a fleet of about 1000 gliders plying the oceans, continuously collecting profiles and communicating them ashore via satellite, while being controlled to variously occupy hydrographic sections or chase features within the ocean. The fundamental vision was of a distributed network of relatively inexpensive platforms. His optimism on the pace of development can understandably be blamed on his enthusiasm, for he predicted the maiden voyage at sea of a glider to have taken place in 1994 and last 198 days. In reality, the first ocean deployment longer than a week took place in 1999 and the longest mission as of this writing lasted 217 days, accomplished in 2005. The Autonomous Oceanographic Sampling Network program set up by the US Office of Naval Research (ONR) in 1995 provided 5 years of support that resulted in gliders becoming a reality. The three operational gliders, Slocum (Webb Research Corp.), Spray (Scripps Institution of Oceanography and Woods Hole Oceanographic Institution), and Seaglider (University of Washington), were developed under this program and have subsequently been supported both by ONR and the US National Science Foundation. Four key technical elements enabled the successful development of underwater gliders: small reliable buoyancy engines, low-power computer microprocessors, Global Positioning System (GPS) navigation, and low-power duplex satellite communication. Seaglider, Slocum, and Spray were designed to fulfill similar missions, hence understandably share many elements (Figure 1). All three are of similar size, roughly 50 kg in mass, in order that both manufacturing and operating costs would be relatively small. The vehicles can be launched and recovered from small vessels by two people without power-assisted equipment. They are designed to carry out missions from a few weeks (Slocum) to

several months (Seaglider and Spray) duration while traveling at B0.5 kt (B0.25 m s  1). All are batterypowered, navigate by dead-reckoning underwater between GPS navigational fixes obtained at the sea surface, and send data and receive commands via a constellation of low Earth-orbit satellites on longrange missions. Derivatives of these three gliders are currently under development, including a version powered in part by extracting thermal energy from ocean stratification, one that operates under ice, and one capable of making open ocean full-depth dives. In addition, parallel development of a vehicle considerably larger, faster, and capable of flying along a shallower glider path is underway.

Design Considerations The three operational gliders were designed to address the sampling deficiencies of infrequent shipbased surveys and widely separated moorings in describing oceanic internal structure. Mesoscale eddies with characteristic space and timescales of O(100 km) and O(1 month) provide the principal noise to ocean general circulation and climate variability. Fronts, eddies, internal waves, diurnal cycles, and tidal fluctuations contribute to variability on smaller scales in problems of regional focus. Measurement platforms capable of resolving these features require sampling on scales O(10 km) and O(10 h) while the need to observe many realizations of random processes calls for sampling to be carried out for several weeks or months over tens or hundreds of kilometers. In order to observe at much lower cost than with ships or moorings, the three operational gliders were designed with endurance and range capability to meet these demands. Gliding is the conversion of a buoyancy force into a body’s forward as well as vertical motion. While aeronautical gliders are heavier than the air around them, hence always fall through it, underwater gliders are designed to alternately be heavier and lighter than the surrounding ocean so that they may glide both when diving and climbing. On average, underwater gliders displace an amount of water equal to their mass. They accomplish gliding by adjusting their volume displacement downward to dive and upward to climb. Gliders pitch themselves downward and are supported by wings to descend along a slanting path (conversely to ascend). Wings are glider propellers. They convert the vertical force of buoyancy into forward motion. The desire to obtain vertical as well as horizontal structure of the ocean is conveniently and efficiently

(c) 2011 Elsevier Inc. All Rights Reserved.

Figure 1 The operational gliders (clockwise from left): Seaglider pitched down at the sea surface with its trailing antenna raised (photo by Keith Van Thiel, University of Washington), top view of Spray at the sea surface (photo by Robert Todd, Scripps Institution of Oceanography), and rear view of Slocum climbing (photo by David Fratantoni, Woods Hole Oceanographic Institution).

GLIDERS

(c) 2011 Elsevier Inc. All Rights Reserved.

61

62

GLIDERS

provided by gliding, provided glide slopes are steep compared to slopes of oceanic property surfaces. The operational gliders variously operate along glide slopes in the range 0.25 to unity (e.g., B14–451), similar to the glide slope of a NASA space shuttle. By contrast, high-performance sail planes attain slopes as gentle as 0.02, not far from the higher slopes of property surfaces within the ocean. Sinking through the ocean is gratis, while climbing through it requires energy input. The force of ambient pressure can be used to partially collapse vehicle volume, raise its mass density over that of seawater surrounding it, and make it sink. Diving requires only the energy to open a valve to allow a swim bladder to drain into a reservoir within the hull. To stop diving and begin ascent, displacement volume must be increased, requiring work against pressure to expand a swim bladder, for example. Work is required not only to provide sufficient potential energy to bring a glider to the sea surface, but also to provide kinetic energy to fly through the ocean. Once made buoyant, gliding eventually brings a vehicle to the sea surface where it can raise an antenna to navigate and communicate. Kinetic energy provided through buoyancy forcing is ultimately dissipated by hydrodynamic drag in gliding. Since drag is generally quadratic with speed, halving speed quadruples endurance and doubles range. Underwater gliders achieve endurance and range 2 orders of magnitude higher than conventional propeller-driven AUVs by virtue of employing speeds an order of magnitude smaller. While a slow propeller-driven AUV might attain considerable endurance and range traveling horizontally, it too must work against the ocean’s potential energy gradient (gravity) to reach the sea surface to navigate or communicate electromagnetically. Slocum, Spray, and Seaglider were each designed to operate at speeds barely faster than typical ocean currents. Since currents are typically surface-intensified due to baroclinicity, deeper dives tend to reduce average current encountered by a glider, hence enhance the ability to navigate effectively. Gliders can navigate more effectively in regions where currents reverse with depth, despite being swift, than in regions of strong barotropic flow. For example, glider navigation in a swift equatorial current system or within a strait with strongly baroclinic flow is expected to be more effective than in weakly stratified coastal jets over the continental shelf. Gliders have operated across the Kuroshio and Gulf Stream by gliding with an upstream component to ferry across them, much as a canoeist paddles both upstream and across stream in order to cross a swift river. The finite spatial extent of ocean currents can be exploited to

make up for downstream drift within them by gliding upstream in weak or opposing flows on their flanks. Navigating the swirling currents of a strong eddy is also possible if glider position within an eddy is recognized. The operational gliders all carry a fixed energy supply in the form of primary batteries. Packs consisting of lithium cells provide the highest energy density, but are classified as hazardous material. Gliders using these are subject to various shipping restrictions, particularly as air freight. Nevertheless, lithium battery packs enable glider missions of half a year or more. Up to about one quarter of the total mass of these gliders can be devoted to carrying batteries, so packs of the equivalent of B100 lithium D-cells can be used, providing roughly 10 MJ of energy, enough for B1000 dive cycles using B10 kJ per cycle. The total average power consumption may be as little as 0.5 W, resulting in mission endurance of several months. Recent improvements suggest that missions longer than a year are feasible for some battery-powered gliders. A buoyancy engine powered by ocean thermal stratification, as envisioned by Stommel and Webb, is under development for Slocum, but will be restricted to regions where sufficient ocean temperature stratification is found. In principle, longer endurance and range can be attained by larger vehicles, since drag is determined by area while energy storage scales with volume. Instrument payload also scales with volume, and also expense. Practically, AUVs larger than B60 kg require power assistance for handling at sea. Both the significant economy of manual handling by a pair of operators and the smaller unit manufacture and service costs associated with small platforms are consistent with the vision expressed by Stommel’s 1989 article and also that guiding the Argo project: that of a network of a large number of relatively small, inexpensive platforms. By keeping gliders small, networks of them are tolerant to occasional vehicle loss.

Glider Dive Cycles Gliders are fully autonomous from the time they leave the sea surface until they return. At the surface, they use a GPS fix to compute a desired heading to a target (or use a preset or communicated heading), then reduce volume displacement to start diving. Retracting a piston accomplishes buoyancy reduction in the 200-m-depth rated Slocum, while in Spray and Seaglider an external bladder is deflated to reduce buoyancy by opening a valve, since subbarometric pressure is maintained inside the pressure

(c) 2011 Elsevier Inc. All Rights Reserved.

GLIDERS

hull. Glider pitch attitude is adjusted by redistributing mass within the vehicles, both by changing buoyancy and moving fixed masses such as battery packs. Gliders turn to attain the desired heading through the use of an active rudder or by redistributing mass to roll the wings. Typical turn radii of the operational gliders are tens of meters. Under water, navigation is by dead reckoning: simply following the chosen compass course. While course through the water varies due to environmental effects such as water-column shear of horizontal current as well as to vehicle control, the overall horizontal displacement due to gliding at various headings and speeds can be measured or inferred. The difference between displacement between surface GPS fixes and gliding displacement gives an estimate of depth-averaged current. The depthaveraged current for each dive cycle can be used to guide the choice of vehicle heading to approach a target through a prediction scheme, either on board or communicated from afar.

Hydrodynamics The existing operational underwater gliders attain effectively steady flight in some tens of seconds due to relatively slow acceleration relative to hydrodynamic drag. The balance of buoyancy by lift and drag forces determines the glide slope. Glider design requires sufficiently small drag that the ratio of drag to lift, equal in steady flight to glide slope, can provide the desired performance. The wings of Seaglider, Slocum, and Spray are somewhat small relative to body vehicle size in comparison to those of typical aeronautical sail planes. Both considerations of how steeply to dive through the ocean’s structure and ease of handling on launch and recovery guided wing sizing. Vehicle drag depends on both shape and surface texture, form, and skin drag, respectively. A significant component of drag is that induced by lift, typically parametrized as proportional to the square of attack angle. Considerations of drag led Spray to approximate a slender ellipsoidal form with a long cylinder capped by ellipsoidal ends and Seaglider to use a long smooth forebody (to maintain laminar boundary layer flow) and short after-section (where turbulence is tolerated). Tow tank, wind tunnel, and open-ocean experience have demonstrated that small appendages can account for disproportionately large contributions to total vehicle drag. A glider with a given set of lift, drag, and induced drag coefficients can be flown over a range of speed and glide slope, the latter limited by stall at the

63

maximum lift to drag ratio. As an example, buoyancy, power, and attack angle dependence on vertical and horizontal speed are shown in Figure 2 for a typical Seaglider. Contours of power (solid curves) reflect the nearly quadratic nature of the drag law for this vehicle: operating at speeds of B0.2 m s  1 costs B0.1 W, while operating at B0.4 m s  1 costs B0.4 W (of delivered power – the buoyancy engines are at most about 50% efficient). The maximum horizontal speed component attainable for a given power use is at a glide slope only slightly steeper than the stall glide slope, as can be seen from the departure of the power curve shape from near-circular about the origin to steeply curved to form a cone of exclusion at the stall slope. The extreme marks the most efficient glide slope for range, B0.28 for speeds of B0.25 m s  1 to B0.20 for speeds of B1 m s  1 for the example given in Figure 2. A given buoyancy force leads to a wide range of speed, depending on glide slope, as illustrated by the dashed curves in Figure 2. For example, a buoyancy of 150 g can provide a horizontal speed of 0.22 m s  1 near the glide slope of maximum efficiency, or 0.33 m s  1 at a glide slope of 0.8 (using about 4 times as much power). The maximum total speed is highest for a given buoyancy if a glider were to be able to rise perfectly vertically (requiring no wing lift), but the maximum horizontal speed is attained at slopes of 0.78–0.84 (glide angles of 30–401 above horizontal) for the particular combination of lift, drag, and induced drag coefficients in the Seaglider example. The buoyancy range plotted in Figure 2 exceeds what the Seaglider buoyancy engine is capable of providing in both positive and negative buoyancy (the end-to-end range is B850 g), but diving through a light near surface layer may produce instantaneously buoyancy as high as is plotted in the figure. Attack angles for gliders (the difference between pitch and glide angle, where the latter is always steeper) range from less than 11 for steep dive slopes to several degrees for glide slopes near stall. For an underwater flying wing, for example, glider pitch can be in the opposite sense to glide angle to produce large lift. For the Seaglider example, typical attack angles are a few degrees for the most efficient glide slope. Attack angle affects flow around the hull, amplifying it as a function of position.

Instrumentation Size, power, hydrodynamic smoothness, and mass invariance are the principal constraints on instrumentation carried on gliders. For the operational

(c) 2011 Elsevier Inc. All Rights Reserved.

64

GLIDERS

Seaglider performance model 1 0

6

750

7 700

0

650

55

0

3

300

0.5

−

45

0.9

5

4

0

0 50

3

600

40

350

2250

0.8 4

2

600 650 700 750 80 0

−1

450

2

500 550

−1

3

0.

8

−1.5

−2

40 350 0

0.3

1 9 0.

7

0.

−1

250

0.3

150

0.4

4

0.6

0.5 100 0.4

−0 .5

0

0.5

0

1 0.9

0.8 0.7

20

Vertical speed (m s−1)

150

.5

−0

400

0

5

25

300

0.6

200

0 35

0.7

3

5

0.

0.2 −2 50 −2.5 0 1.5 0 00.0.1 −3 9 00500 75800 .08 0 0 6 −2.5 5055 3670 0.2 0 2 0 . 0 0 0.05.06 7 −4 −4.5 3.5 45 0 0.0.0341 −5.5 4 3 5 5 4. −5 0.02 −2. 3.5 −78.5 5.5 −6 −6.57.5 10 0.1 0.0 9.5 4 9 8 0.9 1 700 600 650 0 3 550 0500 1 0.2 10 7 0.07 0.02 0.06 0.03 0.01 0.04 0.6 9 .5 45 0.05 98387 5 6.5 9.516350 5−5.5 00.1 .3 −4 4.5 6.5 .0 72.0 80 0.8 3.5 0.7 34.5 15 8.5 50 510 0.4 0.5 8310700 400 .5250 02 8.5 8 7.5 0.00 2.5 5.5550 6 0 00010 50 0535 09.5 10 8 5 9 .5 1 12. 00 47 24150 .2 1 0. 00.4 800 5 0.6 750 63 7 5 5 .0 0.3 450 600 0 5 300 .7 .57 9.5 93049 200 03.5 1.0 6 .0 .2858 0.08 0.09 0.1 002 100 0087.5 7 0.5 0.0 0 0. .5 0 0 0 7 0 5 7 3 .5 0 6 .5 6 0 38 7 6 4 0 8 6 5 .5 8 5 9 350 5 8. 5 0 7 1.5 0 0 50 53 .5 0 10 0.2 5 00 4 0 0 2 5 5200 10 4 01 1 .5 23.5 63 07 1 520.0 5.5 95.0 204 .0 0.0 05 0. 730 .0 40.1 0 80.0 010 0 .5 100 .59785.5 35.56.5 3994.5 79. 36 87 610 1.5 35683 8 52810 5.50.505 4200 0. 0 Buoyancy (g) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Horizontal speed (m s−1) Power (W) Attack angle (°) a = 0.002 758 9 b = 0.009 257 5 c = 9.228 2 e −06 Λ0.25 = 211.863 4 0

30

20

0

2

100

0.6

4

0.

4

−0. 5 0

0.3

Figure 2 Model Seaglider performance based on steady flight with lift, drag, and induced drag coefficients a, b, and c where lift and drag forces follow L ¼ ql 2aa and D ¼ ql 2(bq  1/4 þ c a2), where l ¼ 1.8 m is the Seaglider body length, a is attack angle, and dynamic pressure is given by q ¼ r(u2 þ w2)/2 for horizontal and vertical speed components u and w for water density r. Solid, dashed, and dotted curves contour power, buoyancy, and attack angle, respectively, as functions of horizontal and vertical speed. Contours bunch at the stall limit. Glide slope is the ratio w/u and attack angles are labeled as the difference between vehicle pitch and glide angle, referenced to the horizon (e.g., negative attack for upward glide).

gliders, sensors and their packaging are practically limited to perhaps B500-cc volume and power consumption a fraction of that used for glider propulsion. On long missions, as much as 85% of glider battery energy might be devoted to running the buoyancy engine, limiting the average power consumption available to instrumentation and operating the onboard microprocessor to less than B0.1 W. Average power of this level is achieved by lowering the duty cycle of typical instruments considerably; that is, most of the time sensors are turned off and only briefly are energized for sampling. The low-power sleep current of the microprocessor is a significant constraint on mission endurance. Oceanic fields that are readily measured electronically are well suited to gliders. Temperature, salinity, and pressure are readily measured electronically at little energy cost. In addition to scientific

interest, together with measurements of vehicle pitch, volume, and heading, these can be used to estimate glider speed through the water. Other instrumentation routinely installed on gliders include dissolved oxygen sensors, bio-optical sensors for chlorophyll fluorescence, optical backscatter, and transmissivity, and acoustic sensors for both active and passive measurements. Specially adapted acoustic Doppler current profilers have also been carried on gliders. All of these instruments involve tradeoffs between capability, sampling rate, and mission endurance.

Missions The track of a Seaglider in the Labrador Sea (Figure 3) illustrates the ability of a glider to make transects in regions where depth-averaged flow is

(c) 2011 Elsevier Inc. All Rights Reserved.

GLIDERS

65° W

60° W

55° W

50° W

65

45° W

70

0 60

GPS fix and iridium call

40

05 0

0

400

0

0 00 60 5

20 cm s−1 current 0

300

200

upper 1-km average

10

0

20

0

20

100

0

70

300

800

0

300

90

20

65° N

100

Nuuk

1000

0

Greenland 400

300

200

600

100

65° N

0 40

00 04 0 6080

2000 00

00 5070

5

30

0

0

90 100

30

00

00

10

0

30

600

0

20 00

300

60° N

20

200

60° N 00

000 000 91 60 00080 450

30

700

0

0

200

100

00

30

40

0

0

20

0 30

50

0

20

55° N

55° N 30 0

0

65° W

60° W

55° W

50° W

45° W

Figure 3 Track of a Seaglider in the Labrador Sea from 24 Sep. 2004 to 29 Apr. 2005, the longest AUV mission to date. This vehicle was launched from R/V Knorr in Davis Strait and recovered more than 7 months later offshore Nuuk, Greenland. Pink symbols mark the location of GPS fixes and Iridium calls between dive cycles. Blue arrows indicate depth-averaged currents (to the shallower of bottom depth and 1 km), where the scale length is typical glider speed through the water. Depth contour intervals are 100–1000 m, 500 m thereafter with heavy contours every 1 km.

(c) 2011 Elsevier Inc. All Rights Reserved.

66

GLIDERS

sufficiently weak, but be hampered in its ability to navigate within strong currents. Launched from a research vessel offshore west Greenland, this vehicle made two transects across Davis Strait before being sent south to Hamilton Bank offshore southeast Labrador. It was swept across part of the continental shelf region before escaping the Labrador Current to continue northeastward on its return to Greenland near Kap Desolation. On its way north toward recovery, it was caught in an anticyclonic eddy that drifted offshore. This glider was directed mainly northward as it cycled around the eddy core four times before it finally escaped by being guided radially outward across the eddy. It continued northeastward to Fyllas Bank, offshore Nuuk, Greenland, where it was recovered from a fishing vessel, having completed the longest AUV mission to date as of this writing, lasting 217 days, traveling 3750 km through the water, and making 663 dives, the majority to 1000-m depth. Several different types of glider missions have been undertaken, of which the example in Figure 3 is just one: a solitary glider survey. In other applications, gliders have been used to make repeat surveys along the same track, controlled to maintain position (a ‘virtual mooring’), used in numbers to survey a region intensively, used to interact with one another, to communicate with moored sensors, act as a courier for data recorded by these devices, and to control them.

See also Drifters and Floats. Underwater Vehicles.

Platforms:

Autonomous

Further Reading Davis RE, Eriksen CC, and Jones CP (2003) Autonomous buoyancy-driven underwater gliders. In: Griffiths G (ed.) Technology and Applications of Autonomous Underwater Vehicles, pp. 37--58. New York: Taylor and Francis. Rudnick DL, Davis RE, Eriksen CC, Fratantoni DM, and Perry MJ (2004) Underwater gliders for ocean research. Journal of the Marine Technology Society 38: 73--84. Stommel H (1989) The Slocum mission. Oceanography 2: 22--25.

Relevant Websites http://iop.apl.washington.edu – Operations Summary: Custom View, Seaglider Status, Integrative Observational Platforms, Ocean Physics Department, Applied Physics Laboratory, University of Washington. http://spray.ucsd.edu – SIO IDG Spray Home. http://www.webbresearch.com – Slocum Glider, Webb Research Corporation. http://seaglider.washington.edu – The Seaglider Fabrication Center of the University of Washington.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLOBAL MARINE POLLUTION A. D. Mclntyre, University of Aberdeen, Aberdeen, UK & 2009 Elsevier Ltd. All rights reserved.

This aritcle is designed to provide an entry to the historical and global context of issues related to pollution of the oceans. Until recently, the size and mobility of the oceans encouraged the view that they could not be significantly affected by human activities. Freshwater lakes and rivers had been degraded for centuries by effluents, particularly sewage, but, although from the 1920s coastal oil pollution from shipping discharges was widespread, it was felt that in general the open sea could safely dilute and disperse anything added to it. Erosion of this view began in the 1950s, when fallout from the testing of nuclear weapons in the atmosphere resulted in enhanced levels of artificial radionuclides throughout the world’s oceans. At about the same time, the effluent from a factory at Minamata in Japan caused illness and deaths in the human population from consumption of mercurycontaminated fish, focusing global attention on the potential dangers of toxic metals. In the early 1960s, buildup in the marine environment of residues from synthetic organic pesticides poisoned top predators such as fish-eating birds, and in 1967 the first wreck of a supertanker, the Torrey Canyon, highlighted the threat of oil from shipping accidents, as distinct from operational discharges. It might therefore be said that the decades of the 1950s and 1960s saw the beginnings of marine pollution as a serious concern, and one that demanded widespread control. It attracted the efforts of national and international agencies, not least those of the United Nations. The fear of effects of radioactivity focused early attention, and initiated the establishment of the International Commission on Radiological Protection (ICRP), which produced a set of radiation protection standards, applicable not just to fallout from weapons testing but also to the increasingly more relevant issues of operational discharges from nuclear reactors and reprocessing plants, from disposal of low-level radioactive material from a variety of sources including research and medicine, and from accidents in industrial installations and nuclear-powered ships. Following Minamata, other metals, in particular cadmium and lead, joined mercury on the list of concerns. Since this was seen, like radioactive wastes, as a public health problem, immediate action was taken. Metals

in seafoods were monitored and import regulations were put in place. As a result, metal toxicity in seafoods is no longer a major issue, and since most marine organisms are resilient to metals, this form of pollution affects ecosystems only when metals are in very high concentrations, such as where mine tailings reach the sea. Synthetic organics, either as pesticides and antifoulants (notably tributyl tin (TBT)) or as industrial chemicals, are present in seawater, biota, and sediments, and affect the whole spectrum of marine life, from primary producers to mammals and birds. The more persistent and toxic compounds are now banned or restricted, but since many are resistant to degradation and tend to attach to particles, the seabed sediment acts as a sink, from which they may be recirculated into the water column. Other synthetic compounds include plastics, and the increasing use of these has brought new problems to wildlife and amenities. Oil contaminates the marine environment mainly from shipping and offshore oil production activities. Major incidents can release large quantities of oil over short periods, causing immense local damage; but in the longer term, more oil reaches the sea via operational discharges from ships. These and other threats to the ocean are controlled by the International Convention for the Prevention of Pollution from Ships, administered by the International Maritime Organisation of the UN. Pollution is generated by human activities, and the most ubiquitous item is sewage, which is derived from a variety of sources: as a direct discharge; as a component of urban wastewater; or as sludge to be disposed of after treatment. Sewage in coastal waters is primarily a public health problem, exposing recreational users to pathogens from the local population. The dangers are widely recognized, and many countries have introduced protective legislation. At the global level, the London Dumping Convention controls, among other things, the disposal of sewage sludge. As well as introducing pathogens, sewage also contributes carbon and nutrients to the sea, adding to the substantial quantities of these substances reaching the marine environment from agricultural runoff and industrial effluents. The resulting eutrophication is causing major ecosystem impacts around the world, resulting in excessive, and sometimes harmful, algal blooms. The need for a global approach to ocean affairs was formally brought to the attention of the United

(c) 2011 Elsevier Inc. All Rights Reserved.

67

68

GLOBAL MARINE POLLUTION

Nations in 1967, and over the next 15 years, while sectoral treaties and agreements were being introduced, negotiations for a comprehensive regime led in 1982 to the adoption of the UN Convention on the Law of the Sea (UNCLOS). This provided a framework for the protection and management of marine resources. In the ensuing years, concepts such as sustainable development and the precautionary approach came to the fore and were endorsed in the Declaration of the Rio Summit, while proposals for Integrated Coastal Zone Management (ICZM) are being widely explored nationally and advanced through the Intergovernmental Oceanographic Commission. Early ideas of pollution were focused on chemical inputs, but following the Group of Experts on the Scientific Aspects of Marine Pollution (GESAMP) definition pollution is now seen in a much wider context, encompassing any human activity that damages habitats and amenities and interferes with legitimate users of the sea. Thus, manipulation of terrestrial hydrological cycles, and other hinterland activities including alterations in agriculture, or afforestation, can profoundly influence estuarine regimes, and are seen in the context of pollution. In particular, it is now recognized that excessive fishing can do more widespread damage to marine ecosystems than most chemical pollution, and the need for ecosystem-based fisheries management is widely recognized. Over the years, assessments of pollution effects have altered the priority of concerns, which today are very different from those of the 1950s. Thanks to the rigorous control of radioactivity and metals, these are not now major worries. Also, decades of experience with oil spills have shown that, after the initial damage, oil degrades and the resilience of natural communities leads to their recovery. Today, while the effects of sewage, eutrophication, and harmful algal blooms top the list of pollution concerns, along with the physical destruction of habitats by coastal construction, another item has been added: aquatic invasive species. The particular focus is on the transfer of harmful organisms in ships’ ballast water and sediments, which can cause disruption of fisheries, fouling of coastal industry, and reduction of human amenity. In conclusion, most of the impacts referred to above are on the shallow waters and the shelf, associated with continental inputs and activities. The open ocean, although subject to contamination from the atmosphere and from vessels in shipping lanes, is relatively less polluted United Nations Environment Programme (UNEP).

See also Anthropogenic Trace Elements in the Ocean. Antifouling Materials. Atmospheric Input of Pollutants. Chlorinated Hydrocarbons. Eutrophication. Exotic Species, Introduction of. International Organizations. Marine Policy Overview. Metal Pollution. Oil Pollution. Phytoplankton Blooms. Pollution: Approaches to Pollution Control. Pollution: Effects on Marine Communities. Pollution, Solids. Radioactive Wastes. Thermal Discharges and Pollution. Viral and Bacterial Contamination of Beaches.

Further Reading Brackley P (ed.) (1990) World Guide to Environmental Issues and Organisations. Harlow: Longman. Brune D, Chapman DV, and Gwynne DW (1997) Eutrophication. In: The Global Environment, ch. 30. Weinheim: VCH. Coe JM and Rogers DB (1997) Marine Debris. Berlin: Springer. de Mora SJ (ed.) (1996) Tributyltin Case Study of an Environmental Contaminant. Cambridge, UK: Cambridge University Press. GESAMP (1982) Reports and Studies No. 15: The Review of the Health of the Oceans. Paris: UNESCO. Grubb M, Koch M, and Thomson K (1993) The ‘Earth Summit’ Agreements: A Guide and Assessment. London: Earth Scan Publications. HMSO (1981) Eighth Report of the Royal Commission on Environmental Pollution, Oil Pollution of the Sea, Cmnd 8358. London: HMSO. Hollingworth C (2000) Ecosystem effects of fishing. ICES Journal of Marine Science 57(3): 465--465(1). International Maritime Organization (1991) IMO MARPOL 73/78 Consolidated Edition MARPOL 73/78. London: International Maritime Organization. International Oceanographic Commission (1998) Annual Report of the International Oceanographic Commission. Paris: UNESCO. Kutsuna M (ed.) (1986) Minamata Disease. Kunamoto: Kunammoyo University. Matheickal J and Raaymakers S (eds.) (2004) Second International Ballast Water Treatment R & D Symposium, 21–23 July 2003: Proceedings. GloBallast Monograph Series No. 15. London: International Maritime Organization. National Research Council (1985) Oil in the Sea. Washington, DC: National Academies Press. Park PK, Kester DR, and Duedall IW (eds.) (1983) Radioactive Wastes and the Ocean. New York: Wiley. Pravdic V (1981) GESAMP the First Dozen Years. Nairobi: UNEP. Pritchard SZ (1987) Oil Pollution Control. Wolfeboro, NH: Croom Helm. Sinclair M and Valdimarsson G (eds.) (2003) Responsible Fisheries in the Marine Ecosystem. Oxford, UK: CABI.

(c) 2011 Elsevier Inc. All Rights Reserved.

GLOBAL MARINE POLLUTION

Tolba MK, El-Kholy OA, and El-Hinnawi E (1992) The World Environment 1972–1992: Two Decades of Challenge. London: Chapman and Hall. UNEP (1990) The State of the Marine Environment. UNEP Regional Seas Reports and Studies No. 115. Nairobi: UNEP.

69

UNO (1983) The Law of the Sea: Official Text of UNCLOS. New York: United Nations. World Commission on Environment and Development (1987) Our Common Future. Oxford, UK: Oxford University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING P. F. Kingston, Heriot-Watt University, Edinburgh, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction The sedimentary environment is theoretically one of the easiest to sample quantitatively and one of the most convenient ways to secure such samples is by means of grabs. Grab samplers are used for both faunal samples, when the grab contents are retained in their entirety and then sieved to remove the biota from the sediment, and for chemical/physical samples when a subsample is usually taken from the surface of the sediment obtained. In both cases, the sampling program is reliant on the grab sampler taking consistent and relatively undisturbed sediment samples.

Conventional Grab Samplers The forerunner of the grab samplers used today is the Petersen grab, designed by C.G.J. Petersen to conduct benthic faunal investigations in Danish fiords in the early part of the twentieth century. It consisted of two quadrant buckets that were held in an open

Figure 1 Petersen grab.

70

position and lowered to the seabed (Figure 1). On the bottom, the relaxing of the tension on the lowering warp released the buckets and subsequent hauling caused them to close before they left the bottom. The instrument is still used today but is seriously limited in its range of usefulness, working efficiently only in very soft mud. Petersen’s grab formed the basis for the design of many that came after. One enduring example is the van Veen grab, a sampler that is in common use today (Figure 2). The main improvement over Petersen’s design is the provision of long arms attached to the buckets to provide additional leverage to the closing action. The arms also provided a means by which the complex closing mechanism of the Petersen grab could be simplified with the hauling warp being attached to chains on the ends of the arms. The mechanical advantage of the long arms can be improved further by using an endless warp rig; this has the added advantage of helping to prevent the grab being jerked off the bottom if the ship rolls as the grab is closing. The van Veen grab was designed in 1933 and is still widely used in benthic infaunal studies owing to its simple design, robustness, and digging efficiency. The van Veen grab

Figure 2 Van Veen grab.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING

71

Figure 3 Diagram of a Hunter grab. Reproduced from Hunter B and Simpson AE (1976) A benthic grab designed for easy operation and durability. Journal of the Marine Biological Association 56: 951–957.

typically covers a surface area of 0.1 m2, although instruments of twice this size are sometimes used. A more recent design of frameless grab is the Hunter grab (Figure 3). This is of a more compact design than the van Veen. The jaws are closed by levers attached to the buckets in a parallelogram configuration, giving the mechanism a good overall mechanical advantage. The closing action requires no chains or pulleys and the instrument can be operated by one person. Its disadvantage is that the bucket design does not encourage good initial penetration of the sediment, which is important in hardpacked sediments. A disadvantage of the grab samplers discussed so far is that there is little latitude for horizontal movement of the ship while the sample is being secured: the smallest amount of drift and the sampler is likely to be pulled over. The Smith–McIntyre grab was designed to reduce this problem by mounting the grab buckets in a stabilizing frame (Figure 4). Initial penetration of the leading edge of the buckets is assisted by the use of powerful springs and the buckets are closed by cables pulling on attached short arms in a similar way to that on the van Veen grab. The driving springs are released by two trigger plates, one on either side of the supporting frame to ensure that the sampler is resting flat on the seabed before the sample is taken. In firm sand the Smith–McIntyre grab penetrates to about the same depth of sediment as the van Veen grab. Its main disadvantage is the need to cock the spring mechanism on deck before deploying the sampler, a process that can be quite hazardous in rough weather. The Day grab is a simplified form of the Smith– McIntyre instrument in which the trigger and closing mechanism remains the same, but without spring assistance for initial penetration of the buckets

Figure 4 Smith–McIntyre grab.

Figure 5 Day grab.

(Figure 5). The Day grab is widely used, particularly for monitoring work, despite its poor performance in hard-packed sandy sediments. Most of the grabs thus far discussed have been designed to take samples with a surface area of 0.1 or 0.2 m2. The Baird grab, however, takes samples of 0.5 m2 by means of two inclined digging plates that are pulled together by tension on the warp (Figure 6). The grab is useful where a relatively large surface area needs to be covered, but has the disadvantage of taking a shallow bite and having the surface of the sample exposed while it is being hauled in.

(c) 2011 Elsevier Inc. All Rights Reserved.

72

GRABS FOR SHELF BENTHIC SAMPLING

Warp Activation All the grabs described above use the warp acting against the weight of the sampler to close the jaws. However, direct contact with the vessel on the surface during the closure of the grab mechanism poses several problems.

Warp Heave

As tension is taken up by the warp to close the jaws, there is a tendency for the grab to be pulled up off the bottom, resulting in a shallower bite than might be expected from the geometry of the sampler. This tendency is related to the total weight of the sampler and the speed of hauling and is exacerbated by firm sediments. For example, the theoretical maximum depth of bite of a 120-kg Day grab is 13 cm (based

on direct measurements of the sampler); however, in medium sand, the digging performance is reduced to a maximum depth of only 8 cm (Figure 7(e)). The influence of warp action on the digging efficiency of a grab sampler can also depend on the way in which the sampler is rigged. This is particularly true of the van Veen grab. Figure 7(b) shows the bite profile of the chain-rigged sampler in which the end of each arm is directly connected to the warp by a chain. The vertical sides of the profile represent the initial penetration of the grab while the central rise shows the upward movement of the grab as the jaws close. Figure 7(c) shows the bite profile of a van Veen of similar size and weight (30 kg) rigged with an endless warp in which the arms are closed by a loop of wire passing through a block at the end of each arm (as in Figure 2). The vertical profile of the initial penetration is again apparent; however, in this case, the overall depth of the sampler in the sediment is maintained as the jaws close. The endless warp rig increases the mechanical advantage of the pull of the warp while decreasing the speed at which the jaws are closed. The result is that the sampler is ‘insulated’ from surface conditions to a greater extent than when chain-rigged, giving a better digging efficiency. Grab ‘Bounce’

In calm sea conditions it is relatively easy to control the rate of warp heave and obtain at least some consistency in the volume of sediment secured. However, such conditions are seldom experienced in the open sea where it is more usual to encounter wave action. Few ships used in offshore benthic

Figure 6 Baird grab.

(a)

(b)

(c)

2 4 6 8 10

2 4 6 8 10

2 4 6 8 10

Petersen grab (30 kg)

Chain-rigged van Veen grab (30 kg)

Endless-warp-rigged van Veen grab (short-armed, 30 kg)

(d)

(e)

(f)

2 4 6 8 10 12 14

2 4 6 8 10

2 4 6 8 10

Endless-warp-rigged van Veen grab (long-armed, 70 kg)

Day grab (120 kg)

Smith−mcintyre grab (120 kg)

Figure 7 Digging profiles of a range of commonly used benthic grab samplers obtained in a test tank using a fine sand substratum. (a) Peterson grab (30 kg); (b) chain-rigged van Veen grab (30 kg); (c) endless-warp-rigged van Veen grab (short-armed, 30 kg); (d) endless-warp-rigged van Veen grab (long-armed, 70 kg); (e) Day grab (120 kg); (f) Smith–McIntyre grab (120 kg).

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING

studies are fitted with winches with heave compensators so that the effect of ship’s roll is to introduce an erratic motion to the warp. This may result in the grab ‘bouncing’ off the bottom where the ship rises just as bottom contact is made, or in the grab being snatched off the bottom where the ship rises just as hauling commences. In the former instance, it is unlikely that any sediment is secured; in the latter, the amount of material and its integrity as a sample will vary considerably, depending on the exact circumstances of its retrieval. The intensity of this effect will depend on the severity of the weather conditions. Figure 8 shows the relationship between wind speed and grab failure rate, which is over 60% of hauls at wind force 8. What is of more concern to the scientist attempting to obtain quantitative samples is the dramatic increase in variability with increase in wind speed with a coefficient of variation between 20 and 30 at force 7. The high cost of ship-time places considerable pressure on operators to work in as severe weather conditions as possible and it is not unusual for sampling to continue in wind force 7 conditions with all its disadvantages. Drift

For a warp-activated grab sampler to operate efficiently it should be hauled with the warp positioned vertically above. Where there is a strong wind or current, these conditions may be difficult to achieve. The result is that the grab samplers are pulled on to their sides. This is a particular problem with samplers,

73

such as the van Veen grab, that do not have stabilizing frames. Diver observations have shown, however, that at least in shallow water, where the drift effect is at its greatest on the bottom, even the framed heavily weighted Day and Smith–McIntyre grabs can be toppled. Initial Penetration

It is clear that the weight of the sampler is an important element in determining the volume of the sample secured. Much of the improved digging efficiency of the van Veen grab shown in Figure 7(d) can be attributed to the addition of an extra 40 kg of weight which increased the initial penetration of the sampler on contact with the sediment surface. Initial penetration is one of the most important factors in the sequence of events in grab operation, determining the final volume of sediment secured. Figure 9 shows the relationship between initial penetration and final sample volume obtained for a van Veen grab. Over 70% of the final volume is determined by the initial penetration. Subsequent digging of the sampler is hampered, as already shown, by the pull of the warp. For most benthic faunal studies it is important for the sampler to penetrate at least 5 cm into the sediment (for a 0.1 m2 surface area sample this gives 5 l of sample). In terms of number of species and individuals, over 90% of benthic macrofauna are found in the top 4–5 cm of sediment. Figure 10 shows how the number of individuals relates to average sample 20 18

70

16 Initial penetration depth (cm)

Number of failures due to motion as the % of number of attempts

80

60 50 40 30 20

14 12 10 8 6 4

10

2

0

0 0

1

2

3

4

5

6

7

8

0

2

4

Sea state (Beaufort scale) Figure 8 Relationship between wind speed and grab failure rate.

6

8

10 12 14 16 18 20 22 24

Sample volume (l) Figure 9 Relationship between initial penetration of a van Veen grab sampler and volume of sediment secured.

(c) 2011 Elsevier Inc. All Rights Reserved.

74

GRABS FOR SHELF BENTHIC SAMPLING

1500

Number of individuals per 0.5 m2

1400 1300 1200 1100 1000 900 800 700 600

Figure 11 Shipek grab.

500 4.5

5

5.5

6

6.5

7

7.5

Average volume of grab sample (l) Figure 10 Relationship between number of benthic fauna individuals captured and sample volume for a boreal offshore sand substratum.

volume for 18 stations in the southern North Sea (each liter recorded represents 1 cm of penetration). Although there is a considerable variation in the numbers of individuals between stations, there is no significant trend linking increased abundance with increased sample volume penetration. No sample volumes of less than 4.5 l were taken, indicating that at that level of penetration most of the fauna were being captured. Samplers in which the jaws are held rigidly in a frame have no initial penetration if the edges of the jaw buckets, when held in the open position, are on a level with the base of the frame. The lack of any initial penetration in such instruments has the added disadvantage in benthic fauna work of undersampling at the edges of the bite profile (see Figures 7(e) and 7(f)), although the addition of weight will usually increase the sample volume obtained. Pressure Wave Effect

The descent of the grab necessarily creates a bow wave. Under field conditions, it is usually impracticable to lower the grab at a rate that will eliminate a preceding bow wave, even if the sea were flat calm. There have been several investigations of the effects of ‘downwash’ both theoretical, using artificially placed surface objects, and in situ. The effects of downwash can be reduced by replacing the upper surface of the buckets with an open mesh. Although there is still a considerable effect on the surface flock layer (rendering the samples of dubious

value for chemical contamination studies), the effect on the numbers of benthic fauna is generally very small. Self-Activated Bottom Samplers

There can be little doubt that one of the most important factors responsible for sampler failure or sample variability in heavy seas is the reliance of most presently used instruments on warp-activated closure. The most immediate and obvious answer to this problem is to make the closing action independent of the warp by incorporating a selfpowering mechanism. Spring-Powered Samplers

One solution to the problem is to use a spring to actuate the sampler buckets. Such instruments are in existence, possibly the most widely used being the Shipek grab, a small sampler (0.04 m2) consisting of a spring-loaded scoop (Figure 11). This instrument is widely used where small superficial sediment samples are required for physical or chemical analysis. The use of a pretensioned spring unfortunately sets practical limits on the size of the sampler, since to cock a spring in order to operate a sampler capable of taking a 0.1 m2 sample would require a force that would be impracticable to apply routinely on deck. In addition, in rough weather conditions, a loaded sampler of this size would be very hazardous to deploy. Compressed-Air-Powered Samplers

Another approach has been to use compressed air power. In the 1960s, Flury fitted a compressed air ram to a modified Petersen grab with success. However, the restricted depth range of the instrument and the inconvenience of having to recharge the

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING

air reservoir for each haul limited its potential for routine offshore work. Piston

Hydraulically Powered Samplers

Hydraulically powered grabs are commonly used for large-scale sediment shifting operations such as seabed dredging. The Bedford Institute of Oceanography, Nova Scotia, successfully scaled down this technology to that of a practical benthic sampler. Their instrument is relatively large, standing 2.5 m high and weighing some 1136 kg. It covers a surface area of 0.5 m2 and samples to a maximum sediment depth of 25 cm. At full penetration, the sediment volume taken is about 100 l. The buckets are driven closed by hydraulic rams powered from the surface. The grab is also fitted with an underwater television camera which allows the operator to visually select the precise sampling area on the seabed, close and open the bucket remotely, and verify that the bucket closed properly prior to recovery. The top of the buckets remain open during descent to minimize the effect of downwash and close on retrieval to reduce washout of the sample on ascent. The current operating depth of the instrument is 500 m. The instrument has been successfully used on several major offshore studies, but does require the use of a substantial vessel for its deployment. Hydrostatically Powered Samplers

Hydrostatically powered samplers use the potential energy of the difference in hydrostatic pressure at the sea surface and the seabed. The idea of using this power source is not new. In the early part of the twentieth century, a ‘hydraulic engine’ was in use by marine geologists that harnessed hydrostatic pressure to drive a rock drill. Hydrostatic power has also been used to drive corers largely for geological studies. However, these instruments were principally concerned with deep sediment corers and were not designed to collect macrofauna or material at the sediment–water interface. A more recent development has been that of a grab built by Heriot-Watt University, Edinburgh. The sampler uses water pressure difference to operate a hydraulic ram that is activated when the grab reaches the seabed. Figure 12 shows the general layout of the instrument. Water enters the upper chamber of the cylinder when the sampler is on the seabed, forcing down a piston that is connected to a system of levers that close the jaws. The actuating valve is held shut by the weight of the sampler and there is a delay mechanism to prevent premature closure of the jaws resulting from ‘bounce’. Back on the ship, the jaws are held shut by an overcenter locking mechanism

75

Water inlet valve with release delay mechanism held shut by weight or sampler

Ram operated by hydrostatic pressure Connecting rod

Figure 12 Diagram of a grab sampler using hydrostatic pressure to close the jaws.

and, on release, are drawn open by reversal of the piston motion from air pressure built up on the underside of the piston during its initial power stroke. Since the powering of the grab jaws is independent of the warp, the sampler may be used successfully in a much wider range of surface weather conditions than conventional grabs.

Alternatives to Grab Samplers Ideally a benthic sediment sample for faunal studies should be straight-sided to the maximum depth of its excavation and should retain the original stratification of the sediment. Grab samplers by the very nature of their action will never achieve this end. Suction Samplers

One answer to this problem is to employ some sort of corer designed to take samples of sufficient surface area to satisfy the present approaches to benthic studies. The Knudsen sampler is such a device and is theoretically capable of taking the perfect benthic sample. It uses a suction technique to drive a core tube of 0.1 m2 cross-sectional area 30 cm into the sediment. Water is pumped out of the core tube on the seabed by a pump that is powered by unwinding a cable from a drum. The sample is retrieved by pulling the core out sideways using a wishbone arrangement and returning it to the surface bottomside up (Figures 13 and 14). Under ideal conditions, the device will take a straight-sided sample to a depth

(c) 2011 Elsevier Inc. All Rights Reserved.

76

GRABS FOR SHELF BENTHIC SAMPLING

Figure 13 Knudsen sampler in descent position.

of 30 cm. However, conditions have to be flat calm in order to allow time for the pump to operate on the seabed and evacuate the water from the core. This limits the use of the Knudsen sampler and it is generally not suitable for sampling in unsheltered conditions offshore. Mounting the sampler in a stabilizing frame can improve its success rate and it is used regularly for inshore monitoring work where it is necessary to capture deep burrowing species. Spade Box Samplers

Another approach to the problem is to drive an open-ended box into the sediment, using the weight of the sampler, and arrange for a shutter to close off the bottom end. The most widespread design of such an instrument is that of the spade box sampler, first described by Reineck in the 1950s and later subjected to various modifications. The sampler consists of a removable steel box open at both ends and driven into the sediment by its own weight. The lower end of the box is closed by a shutter supported on an arm pivoted in such a way as to cause it to slide through the sediment and across the mouth of the box (Figure 15). As with the grab samplers previously described, the shutter is driven by the act of hauling on the warp with all the attendant disadvantages. Nevertheless, box corers are very successful and are used widely for obtaining relatively undisturbed samples of up to 0.25 m2 surface area (Figure 16). One big advantage of the box sampler is that the box can usually be removed with the sample and its overlying water left intact, allowing detailed studies of the sediment surface. Furthermore, it is possible to subsample using small-diameter corers for studies of chemical and physical characteristics. Despite their potential of securing the ‘ideal’ sediment sample, box corers are rarely used for routine benthic monitoring work. This is largely because of their size (a box corer capable of taking a 0.1 m2

Figure 14 Knudsen sampler in ascent position.

Lead weights

Box closure (‘spade’) Box holder Box

Figure 15 Diagram of a Reineck spade box sampler.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING

77

Cable drum

Scoops Figure 17 Diagram of a Holme scoop. Reproduced from Holme NA (1953) The biomass of the bottom fauna in the English Channel off Plymouth. Journal of the Marine Biological Association 32: 1–49.

macrobenthos sampling. At present, there is no instrument that fulfils the requirements for a quick turnaround precision multiple corer for offshore sampling.

Sampling Difficult Sediments

Figure 16 A 0.25 m2 spade box sampler.

sample weighs over 750 kg and stands 2 m high) and the difficulty in deployment and recovery in heavy seas. Precision Corers

For chemical monitoring, it is important that the sediment–water interface is maintained intact, for it is the surface flock layer that will contain the most recently deposited material. Unfortunately, such undisturbed samples are rarely obtained using grab samplers or box corers. Precision corers are capable of securing undisturbed surface sediment cores; however, they are unsuitable for routine offshore work because of the time taken to secure a sample on the seabed and dependence on warp-activated closure. Additionally, the cross-sectional area of the core (0.002–0.004 m2) would necessitate the taking of large numbers of replicate samples in order to capture sufficient numbers of benthic macrofauna to be useful. This would be impracticable given the time taken to take a single sample. Large multiple precision corers have been constructed; these are usually too large and difficult to deploy for routine

Most of the samplers so far discussed operate reasonably well in mud or sand substrata. Few operate satisfactorily in gravel or stony mixed ground either because the bottom is too hard for the sampler to penetrate the substratum or because of the increased likelihood of a stone holding the jaws open when they are drawn together. To get around this problem various types of scoops have been devised. The Holme grab has a double scoop action with two buckets rotating in opposite directions to minimize any lateral movement during digging. The scoops are closed by means of a cable and pulley arrangement (Figure 17) and simultaneously take two samples of 0.05 m2 surface area. The Hamon grab, which has proved to be very effective in coarse, loose sediments, takes a single rectangular scoop of the substratum covering a surface area of about 0.29 m2. The scoop is forced into the sediment by a long lever driven by pulleys that are powered by the pull of the warp (Figure 18). Although the samples may not always be as consistent as those from a more conventional grab sampler, the Hamon grab has found widespread use where regular sampling on rough ground is impossible by any other means.

(c) 2011 Elsevier Inc. All Rights Reserved.

78

GRABS FOR SHELF BENTHIC SAMPLING

A generic disadvantage of most samplers presently in use is that they rely on slackening of the warp to trigger the action and the heave of the warp to drive the closing mechanism. In calm conditions, this presents no great problem, but with increasing sea state the vertical movement of the warp decreases reliability dramatically until the variability and finally failure rate of hauls make further sampling effort fruitless.

Specific Problems, Requirements, and Future Developments Scoop

Release Stop plate hook Figure 18 Diagram of a Hamon scoop.

Present State of Technology It is perhaps surprising that given the high state of technology of survey vessels, position fixing, and analytical equipment, the most commonly used samplers are relatively primitive (being designed some 40 or more years ago). Yet the quality of the sample is of fundamental importance to any research or monitoring work. Currently the most popular instruments are grab samplers, probably because of their wide operational weather window and apparent reliability. Samplers such as the Day, van Veen, and Smith–McIntyre grab samplers are still routinely used for sampling sediment for chemical and biological analysis despite their well-documented shortcomings. As discussed earlier, the most important of these are the substantial downwash that precedes the sampler as it descends and the disturbance of the trapped sediment layers by the closing action of the jaws. Both chemical and meiofaunal studies are particularly vulnerable to these. Although box corers go some way to reducing disturbance of the sediment strata, the all-important surface flocculent layer is invariably washed away. A big disadvantage of the box corer for routine offshore work is that it is sensitive to weather conditions; in addition, the larger instruments do not perform well on sand substrata.

Chemical Studies

Studies involving sediment chemistry require precision sampling if undisturbed samples at the sediment/water interface are to be obtained. This can be critical, particularly when the results of recent sedimentation are of interest. The impracticality of using existing hydraulically damped corers such as the Craib corer for offshore work has led to the widespread use of less ‘weather sensitive’ devices such as spade box corers and grabs for routine monitoring purposes. However, studies carried out have shown that these samplers produce a considerable ‘downwash effect’, blowing the surface flocculent layer away before the sample is secured. This can have serious consequences if any meaningful estimation of the surface chemistry of the sediment is desired. Repetitive and accurate sampling is also a prerequisite for determining spatial and temporal change in sediment chemistry.

Meiofauna Studies Meiofauna has increasingly been shown to have potential as an important tool in benthic monitoring. One of the major factors limiting its wider adoption is the lack of a suitable sampler. Although instruments such as the Craib corer and its multicorer derivatives are capable of sampling the critically important superficial sediment layer, these designs provide a poor level of success on harder sediments and in anything but near-perfect weather conditions. They also have a slow turnaround time. Box corers are widely used as an alternative; however, they are known to be unreliable in their sampling of meiofauna. As with the macrobenthos, meiobenthic patchiness results in low levels of precision of abundance estimates unless large numbers of samples are taken.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRABS FOR SHELF BENTHIC SAMPLING

Macrobenthic Studies

The measurement and prediction of spatial and temporal variation in natural populations are of great importance to population biologists, both for fundamental research into population dynamics and productivity and in the characterization of benthic communities for determining change induced by environmental impact. Though always an important consideration, cost-effectiveness of sampling and sample processing is not so crucial in fundamental research, since time and funding may be tailored to fit objectives. This is rarely the case in routine monitoring work where often resolution and timescales have to fit the resources available. Benthic fauna are contagiously distributed and to sample such communities with a precision that will enable distinction between temporal variation and incipient change resulting from pollution effects, it is generally accepted that five replicate 0.1 m2 hauls from each station are necessary (giving a precision at which the standard error is no more than 20% of the mean). This frequency of sampling requires approximately 10–15 man-days of sediment faunal analysis per sample station. While this may be acceptable in community structure studies in which time and manpower (and thus cost) are not a primary consideration, this high cost of analyzing samples is of importance in routine monitoring studies and has led monitoring agencies and offshore operators to reduce sampling frequency on cost grounds to as few as two replicates per station. This reduces the precision with which faunal abundance can be estimated to a level at which only gross change can be demonstrated. However, sampling to an acceptable precision may be achieved from an area equivalent to 0.1 m2 if a smaller sampling unit is used. For example, 50 5-cm core samples (with a similar total surface area) have been shown to give a similar precision to that of 5 to 12 0.1-m2 grab samples. Thus a similar degree of precision may be obtained for around one-fifth to one-twelfth the analytical costs using a conventional approach. The problem is to be able to secure the 50 core samples per site that would be needed for the macrofaunal monitoring in a timescale that would be realistic offshore. At present, there is no instrument capable of supporting such a sampling demand and operating in the range of sediment types that wide-

79

scale monitoring studies demand. Clearly a single core sampler would be impracticable, and one must look to the future development of a multiple corer that is capable of flexibility in its operation, which allows a quick on-deck turnaround between hauls.

See also Benthic Boundary Organisms Overview.

Layer

Effects.

Benthic

Further Reading Ankar S (1977) Digging profile and penetration of the van Veen grab in different sediment types. Contributions from the Asko¨ Laboratory, University of Stockholm, Sweden 16: 1--12. Beukema JJ (1974) The efficiency of the van Veen grab compared with the Reineck box sampler. Journal du Conseil Permanent International pour l’Exploration de la Mer 35: 319--327. Eleftheriou A and McIntyre AD (eds.) (2005) Methods for the Study of the Marine Benthos. Oxford, UK: Blackwell. Flury JA (1967) Modified Petersen grab. Journal of the Fisheries Research Board of Canada 20: 1549--1550. Holme NA (1953) The biomass of the bottom fauna in the English Channel off Plymouth. Journal of the Marine Biological Association 32: 1--49. Hunter B and Simpson AE (1976) A benthic grab designed for easy operation and durability. Journal of the Marine Biological Association 56: 951--957. Riddle MJ (1988) Bite profiles of some benthic grab samplers. Estuarine, Coastal and Shelf Science 29(3): 285--292. Thorsen G (1957) Sampling the benthos. In: Hedgepeth JW (ed.) Treatise on Marine Ecology and Paleoecology, Vol. 1: Ecology. Washington, DC: The Geological Society of America.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRAVITY M. McNutt, MBARI, Moss Landing, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1177–1185, & 2001, Elsevier Ltd.

Introduction The gravity field varies over the oceans on account of lateral variations in density beneath the ocean surface. The most prominent anomalies arise from undulations on density interfaces, such as occur at the water–rock interface at the seafloor or at the crust– mantle interface, also known as the Moho discontinuity. Because marine gravity is relatively easy to measure, it serves as a remote sensing tool for exploring the earth beneath the oceans. The interpretation of marine gravity anomalies in terms of the Earth’s structure is highly nonunique, however, and thus requires simultaneous consideration of other geophysically observed quantities. The most useful auxiliary measurements include depth of the ocean from echo sounders, the shape of buried reflectors from marine seismic reflection data, and/or the density of ocean rocks as determined from dredge samples or inferred from seismic velocities. Depending on the spatial wavelength of the observed variation in the gravity field, marine gravity observations are applied to the solution of a number of important problems in earth structure and dynamics. At the very longest wavelengths of 1000 to 10 000 km, the marine gravity field is usually combined with anomalies over land to infer the dynamics of the entire planet. At medium wavelengths of several tens to hundreds of kilometers, the gravity field contains important information on the thermal and mechanical properties of the lithospheric plates and on the thickness of their sedimentary cover. At even shorter wavelengths, the field reflects local irregularities in density, such as produced by seafloor bathymetric features, magma chambers, and buried ore bodies. On account of the large number of potential contributions to the marine gravity field, modern methods of analysis include spectral filtering to remove signals outside of the waveband of interest and interpretation within the context of models that obey the laws of phyics.

Units Gravity is an acceleration. The acceleration of gravity on the earth’s surface is about 9.81 m s2.

80

Gravity anomalies (observed gravitational acceleration minus an expected value) are typically much smaller, about 0.5% of the total field. The SI-compatible unit for gravity anomaly is the gravity unit (gu): 1 gu ¼ 106 m s2. However, the older c.g.s unit for gravity anomaly, the milligal (mGal), is still in very wide use: 1 mGal ¼ 10 gu. Typical small-scale variations in gravity over the ocean range from a few tens to a few hundreds of gravity units. Lateral variations in gravitational acceleration (gravity gradients) are measured in Eo¨tvo¨s units (E): 1 E ¼ 109 s2. Another quantity useful in gravity interpretation is the density of earth materials, measured in kg m3. In the marine realm, relevant densities range from about 1000 kg m3 for water to more than 3300 kg m3 for mantle rocks. A close relative of the marine gravity field is the marine geoid. The geoid, measured in units of height, is the elevation of the sea level equipotential surface. Geoid anomalies are measured in meters and are the departure of the true equipotential surface from that predicted for an idealized spheroidal Earth whose density structure varies only with radius. Geoid anomalies range from 0 to more than 7100 m. The direction of the force of gravity is everywhere perpendicular to the geoid surface, and the magnitude of the gravitational attraction is the vertical derivative of the geopotential U (eqn [1]). g¼

@U @z

½1

Geoid height N is related to the same equipotential U via Brun’s formula eqn ([2]). N¼

U g0

½2

in which g0 is the acceleration of gravity on the spheroid. For a ship sailing on the sea surface (the equipotential), it is easier to measure gravity. From a satellite in free-fall orbit high over the Earth’s surface, radar altimeters can measure with centimeter precision variations in the elevation of sea level, an excellent approximation to the true geoid that would follow the surface of a motionless ocean. Regardless of whether geoid or gravity is the quantity measured directly, simple formulas in the wavenumber domain allow gravity to be computed from geoid and vice versa. Given the same equipotential, the gravity representation emphasizes the power in

(c) 2011 Elsevier Inc. All Rights Reserved.

GRAVITY

the high-frequency (short-wavelength) part of the spectrum, whereas the geoid representation emphasizes the longer wavelengths. Therefore, for investigations of high-frequency phenomena, gravity is generally the quantity interpreted even if geoid is what was measured. The opposite is true for longwavelength phenomena.

Measurement of Marine Gravity Marine gravity measurements can be and have been acquired with several different sorts of sensors and from a variety of platforms, including ships, submarines, airplanes, and satellites. The ideal combination of sensor system and platform depends upon the needed accuracy, spatial coverage, and available time and funds. Gravimeters

The design for most marine gravimeters is borrowed from their terrestrial counterparts and are either absolute or relative in their measurements. Absolute gravimeters measure the full acceleration of gravity g at the survey site along the direction of the local vertical. Modern marine absolute gravimeters measure precisely the vertical position z of a falling mass (e.g., a corner cube reflector) as a function of time t in a vacuum cylinder using laser interferometry and an atomic clock. The acceleration is then calculated as the second derivative of the position of the falling mass as a function of time (eqn [3]). g¼

d2 zðtÞ dt2

½3

Absolute gravimeters tend to be larger, more difficult to deploy, costlier to build, and more expensive to run than relative gravimeters, and thus are only used when relative gravimeters are inadequate for the problem being addressed. Most gravity measurements at sea are relative measurements, Dg: the instrument measures the difference between gravity at the study site and at another site where absolute gravity is known (e.g., the dock where the expedition originates). Modern relative gravimeters are based on Hooke’s law for the force F required to extend a spring a distance x (eqn [4]), where ks is the spring constant, calculated by extending the spring under a known force. F ¼ ks x ¼ mg

81

at other locations can be calculated by observing how much more or less that same spring is stretched at other locations by the same mass. Although such systems are relatively inexpensive to build and easy to deploy, they suffer from drift: in effect, the spring constant changes with time because no physical spring is perfectly elastic. To first order, the drift can be corrected by returning to the same or another base station with the same instrument, and assuming that the drift was linear with time in between. The accuracy of this linear drift assumption improves with more frequent visits to the base station, but this is usually impractical for marine surveys. Through clever design, the latest generation of marine gravimeters has greatly reduced the drift problem as compared with earlier instruments. The measurement of the gravity gradient tensor was widely used early in the twentieth century for oil exploration, but fell into disfavor in the 1930s as scalar gravimeters became more reliable and easy to use. Gravity gradiometry at sea is currently making a comeback as the result of declassification of military gradiometer technology developed for use in submarines during the Cold War. Gravity gradiometers measure the three-dimensional gradient in the gravity vector using six pairs of aligned gravimeters, with accuracies reaching better than 1 Eo¨tvo¨s. In comparison with measurements of gravity, the gravity gradient has more sensitivity to variations at short wavelenghts (B5 km or less), making it useful for delimiting shallow structures buried beneath the seafloor. Geoid anomalies can be directly measured from orbiting satellites carrying radar altimeters. The altimeters measure the travel time of a radar pulse from the satellite to the ocean surface, from which it is reflected and bounced back to the satellite. Tracking stations on Earth solve for the position of the satellite with respect to the center of the Earth. These two types of information are then combined to calculate the height of the sea level equipotential surface above the center of the Earth. Because the solid land surface does not follow an equipotential, altimeters cannot be used to constrain the terrestrial geoid. Furthermore, it is difficult to extract geoid from ocean areas covered by sea ice. However, in the near future, laser altimeters deployed from satellites hold the promise of extracting geoid information even over ocean surfaces marred with sea ice, on account of their enhanced resolution.

½4

If a mass m is suspended from this spring at a site where gravity g is known (e.g., by deploying an absolute gravimeter at that base station), then gravity

Platforms

Marine gravity data can be acquired either from moving or from stationary platforms. Because the

(c) 2011 Elsevier Inc. All Rights Reserved.

82

GRAVITY

gravity field from variations in the depth of the sea floor is such a large component of the observed signal, most marine gravity surveys have relied on ships or submarines that enable the simultaneous acquisition of depth observations. However, airborne gravity measurements have been acquired successfully over ice-covered areas of the polar oceans, and orbiting satellites have measured the marine geoid from space. A major challenge in acquiring gravity data from a floating platform at the sea surface is in separating the acceleration of the platform in the dynamic ocean from the acceleration of gravity. This problem is overcome by mounting marine gravimeters deployed from ships on inertially stabilized tables. These tables employ gyroscopes to maintain a constant attitude despite the pitching and rolling of the ship beneath the table. The nongravitational acceleration is somewhat mitigated by mounting marine gravimeters deep in the hold and as close to the ship’s center of motion as possible. Special damping mechanisms also prevent the spring in the gravimeter from responding to extremely high-frequency changes in the force on the suspended mass. Instruments deployed in submersibles resting on the bottom of the ocean or in instrument packages lowered to the bottom of the ocean do not suffer from the dynamic accelerations of the moving ocean surface, but bottom currents can also be an important source of noise in submarine gravimetry. Installing instruments in boreholes is the most effective way to counter this problem, but it is also an expensive solution.

Reduction of Marine Data A number of standard corrections must be applied to the raw gravity data (either g or Dg) prior to interpretation. In addition to any drift correction, as mentioned above for relative gravity measurements, a latitude correction is immediately applied to account for the large change in gravity between the poles and the Equator caused by Earth’s rotation. Near the Equator, the centrifugal acceleration from the Earth’s spin is large, and gravity is about 50 000 gu less, on average, than at the poles. Because this effect is 5000 times larger than typical regional gravity signals of interest, it must be removed from the data using a standard formula for the variation of gravity g0 on a spheroid of revolution best fitting the shape of the Earth (eqn [5]; Y ¼ latitude).  g0 ðYÞ ¼ 9:7803185 1 þ 5:278895  103 sin2 Y  þ2:3462  105 sin4 Y ms2 ½5

A second correction that must be made if the gravity is measured from a moving vehicle, such as a ship or airplane, accounts for the effect on gravity of the motion of the vehicle with respect to the Earth’s spin. A ship steaming to the east is, in effect, rotating faster than the Earth. The centrifugal effect of this increased rate of rotation causes gravity to be less than it would be if the ship were stationary. The opposite effect occurs for a ship steaming to the west. This term, called the Eo¨tvo¨s correction, is largest near the Equator and involves only the east–west component vEW of the ship’s velocity vector (eqn [6]), in which o is the angular velocity of the Earth’s rotation. gEOT ¼ 2ovEW cos Y

½6

The free air gravity correction, which accounts for the elevation of the measurement above the Earth’s sea level equipotential surface, is obviously not needed if the measurement is made on the sea surface. The free air correction gFA is required if the measurement is made from a submersible or an airplane: eqn [7], where h is elevation above sea level in meters. gFA ¼ 3:1h

gu

½7

This correction is added to the observation if the sensor is deployed above the Earth’s surface, and subtracted for stations below sea level. For land surveys, the Bouguer correction accounts for the extra mass of the topography between the observation and sea level. For its marine equivalent, it adds in the extra gravitational attraction that would be present if rock rather than water existed between sea level and the bottom of the ocean. Except in areas of rugged bathymetry, the Bouguer correction gB is calculated using the slab formula (eqn [8]). gB ¼ 2pDrGz

½8

Here Dr is the density difference between oceanic crust and sea water, G is Newton’s constant, and z is the depth of the sea floor. This correction is seldom used because it produces very large positive gravity anomalies. Furthermore, there are more accurate corrections for the effect of bathymetry that do not make the unrealistic assumption that the expected state for the oceans should be that the entire depth is filled with crustal rocks displacing the water. The Bouguer correction is necessary, however, when gravity measurements are made from a submarine, in order to combine those data with more conventional observations from the sea surface. In this case, the Bouguer correction is applied twice: once to remove

(c) 2011 Elsevier Inc. All Rights Reserved.

GRAVITY

the upward attraction of the layer of water above the submarine, and once more to add in that layer’s gravitation field below the sensor. Satellite measurements of sea surface height go through a different processing sequence to recover marine geoid anomalies. The most important step is in calculating precise orbits. Information from tracking stations is supplemented with a ‘crossover analysis’ that removes long-wavelength bias in orbit elevation by forcing the height valves to agree wherever orbits cross. Corrections are then made for known physical oceanographic effects such as tides, and wave action is averaged out. The height of the sea level geoid above the Earth’s center, assuming the standard spheroid, is subtracted from the data to create geoid anomalies.

History A principal impediment to the acquisition of useful gravity observations at sea was the difficulty in separating the desired acceleration of gravity from the acceleration of the platform floating on the surface of the moving ocean. For this reason, the first successful gravity measurements to be acquired at sea were taken from a submarine by the Dutch pioneer, Vening Meinesz, in 1923. He used a pendulum gravimeter, which was the state of the art for measuring absolute gravity at that time. By accurately timing the period, T, of the swinging pendulum, the acceleration of gravity, g, can be recovered according to eqn [9], in which l is the length of the pendulum arm. sffiffiffi l T ¼ 2p g

½9

By 1959, five thousand gravity measurements had been acquired from submarines globally. These measurements were instrumental in revealing the large gravity anomalies associated with the great trenches along the western margin of the Pacific. However, these gravity observations were very timeconsuming to acquire because of the long integration times needed to achieve a high-precision estimate of the pendulum’s period, and could not be adapted for use on a surface ship. Gravity measurements at sea became routine and reliable in the late 1950s with the development of gyroscopically stabilized platforms and heavily damped mass-and-spring systems constrained to move only vertically. The new platforms compensated for the pitch and roll of the ship such that simple mass-and-spring gravimeters could collect

83

time series of variations in gravity over the oceans from vessels under way. Without any need to stop the ship on station, a time series of gravity measurements could be obtained at only small incremental cost to ship operations. With the advent of the new instrumentation, the catalogue of marine gravity values has grown in the past 40 years to more than 2.5 million measurements. A new era of precision in marine gravity began with the advent of the Global Positioning System (GPS) in the late 1980s. Prior to this time, the largest source of uncertainty in marine gravity lay in the Eo¨tvo¨s correction. Older navigation systems (dead reckoning, celestial, and even the TRANSIT satellite system) were too imprecise in the absolute position of the ship and too infrequently available to allow accurate velocity estimation from minute to minute, especially if the ship was maneuvering. Typically, gravity data had to be discarded for an hour or so near the time of any change in course. The high positioning accuracy and frequency of GPS fixes now allows such precise calculation of the Eo¨tvo¨s correction that it is no longer the limiting factor in the accuracy of marine gravity data. A breakthough in determining the global marine gravity field was achieved with the launching of the GEOS-3 (1975–1977) and Seasat (1978) satellites, which carried radar altimeters. Altimeters were deployed for the purpose of measuring dynamic sea surface elevation associated with physical oceanographic effects. The Seasat satellite carried a new, high-precision altimeter that characterized the variations in sea surface elevation with unprecedented detail. The satellite failed prematurely, but not before it returned a wealth of data on the marine gravity field from its observations of the marine geoid. The geoid variations at mid- and short-wavelength were so large that the dynamic oceanographic effects motivating the mission could be considered a much smaller noise term. The success of the Seasat mission led to the launch of Geosat, which measured the geoid at even higher precision and resolution. Unfortunately, most of that data remained classified by the US military until the results from a similar European mission were about to be released into the public domain. The declassification of the Geosat data in 1995 fueled a major revolution in our understanding of the deep seafloor (Figure 1). The latest developments in marine gravity stem from the desire to detect the shortest spatial wavelengths of gravity variations by taking gravimeters to the bottom of the ocean. Gravity is one example of a potential field, and as such the amplitude, A, of the signal of interest decays with distance, z, between source and detector as in eqn [10], where k is the

(c) 2011 Elsevier Inc. All Rights Reserved.

84

GRAVITY

Figure 1 Gravity field over the North Pacific. This view is based on satellite altimetry data from the Geosat and other missions. (Data from Sandwell and Smith (1997).)

modulus of the spatial wavenumber, the reciprocal of the spatial wavelength. ABe2pkz

½10

For sensors located on a ship at the sea surface in average ocean depths of 4.5 km, it is extremely difficult to detect short-wavelength variations in gravity of a few kilometers or less. Even lowering the gravimeter to the cruising depth of most submarines (a few hundred meters) does little to overcome the upward attenuation of the signal from localized sources on and beneath the seafloor. The solution to this problem recently has been to take gravimeters to the bottom of the ocean, either in a deep-diving submersible such as Alvin, or as an instrument package lowered on a cable. Most gravity measurements at sea are relative measurements. However, recent advances in instrumentation now allow absolute gravimeters to be deployed on the bottom of the ocean, avoiding the problem of instrument drift that adds error to relative gravity measurements. However, noise associated with the short baselines required for operation in the deep sea remains problematic.

Interpretation of Marine Gravity Short-Wavelength Anomalies

The shortest-wavelength gravity anomalies over the oceans (less than a few tens of kilometers) are the

least ambiguous to interpret since they invariably are of shallow origin. The upward continuation factor guarantees that any spatially localized anomalies with deep sources will be undetectable at the ocean surface. Near-bottom gravity measurements are able to improve somewhat the detection of concentrated density anomalies buried at deeper levels, but most are assumed to lie within the oceanic crust. One of the most useful applications of shortwavelength gravity anomalies has been to predict ocean bathymetry (Figure 2). Radar altimeters deployed on the Seasat and Geosat missions measured with centimeter accuracy the height of the underlying sea surface, an excellent approximation to the marine geoid, over all ice-free marine regions. The accuracy and spatial coverage was far better than had been provided from more than a century of marine surveys from ships. At short wavelengths, undulations of the rock–water interface are the largest contribution to the short-wavelength portion of the geoid spectrum, which opened up the possibility of predicting ocean depth from the excellent geoid data. For example, an undersea volcano, or seamount, represents a mass excess over the water it displaces. The extra mass locally raises the equipotential surface, such that positive geoid anomalies are seen over volcanoes and ridges while geoid lows are seen over narrow deeps and trenches. The prediction of bathymetry from marine geoid or gravity data is tricky: the highest frequencies in the bathymetry cannot be estimated because of the upward attenuation

(c) 2011 Elsevier Inc. All Rights Reserved.

GRAVITY

85

(A)

(B)

Figure 2 Example of bathymetric prediction from gravity anomalies in a largely unexplored region of the South Pacific. (A) The best available bathymetry from sparse echo soundings available in the early 1990s. (B) A diagram shows a dramatic improvement in definition of the bathymetry when satellite gravity observations are used to constrain the short-wavelength component of the bathymetry. (Adapted from McNutt and Bonneville (1996).)

problem, and the longer wavelengths are canceled out in the geoid by their isostatic compensation (see following section). These longer wavelengths in the bathymetry must be introduced into the solution using traditional echo soundings from sparse ship tracks. Nevertheless, the best map we currently have of the depth of the global ocean is courtesy of satellite altimetry. Mid-Wavelength Anomalies

The mid-wavelength part of the gravity spectrum (tens to hundreds of kilometers) is dominated by the effects of isostatic compensation. Isostasy is the process by which the Earth supports variations in topography or bathymetry in order to bring about a condition of hydrostatic equilibrium at depth. The definition of isostasy can be extended to include both static and dynamic compensation mechanisms, but at these wavelengths the static mechanisms are most important. There are a number of different types of isostatic compensation at work in the oceans, and the details of the gravity field can be used to distinguish them and to estimate the thermomechanical behavior of oceanic plates. One of the simplest mechanisms for isostatic compensation is Airy isostasy: the oceanic crust is thickened beneath areas of shallow bathymetry. The thick crustal roots displace denser mantle material, such that the elevated features float on the mantle much like icebergs float in the ocean. Of the various

methods of isostatic compensation, this mechanism predicts the smallest gravity anomalies over a given feature. From analysis of marine gravity, we now know that this sort of compensation mechanism is only found where the oceanic crust is extremely weak, such as on very young lithosphere near a midocean ridge. For example, large plateaus formed when hotspots intersect midocean ridges are largely supported by Airy-type isostasy. Elsewhere the oceanic lithosphere is strong enough to exhibit some lateral strength in supporting superimposed volcanoes and other surface loads. An extremely common form for support of bathymetric features in the oceans is elastic flexure. Oceanic lithosphere has sufficient strength to bend elastically, thus distributing the weight of a topographic feature over an area broader than that of the feature itself (Figure 3). Analysis of marine gravity has been instrumental in establishing that the elastic strength of the oceanic lithosphere increases with increasing age. Young lithosphere near the midocean ridge is quite weak, in some cases hardly distinguished from Airytype isostasy. The oldest oceanic lithosphere displays an effective thickness equivalent to that of a perfectly elastic plate 40 km thick. The fact that this thickness is less than that of the commonly accepted value for the thickness of the mechanical plate that drifts over the asthenosphere indicates that the base of the oceanic lithosphere is not capable of sustaining large deviatoric stresses (of the order of 100 MPa or more) over million-year timescales.

(c) 2011 Elsevier Inc. All Rights Reserved.

86

GRAVITY

Net gravity anomaly Contribution from volcano

Contribution from flexure of plate Volcano

Light crust

Elastic plate Dense mantle

Figure 3 Cartoon showing how the seafloor is warped as an elastic plate under the weight of a small volcano. The gravity anomaly that would be detected by a ship sailing along the sea surface over this feature is the net difference between the positive gravity perturbation from the extra mass of the volcano and the negative gravity perturbation produced when the elastically flexed light crust replaces denser mantle. (Adapted from McNutt and Bonneville (1996).)

Another very important method of isostatic compensation in the ocean is Pratt isostasy. This method of support supposes that the height of a vertical column of bathymetry is inversely proportional to its density. Low-density columns can be higher because they are lighter, whereas heavy columns must be short in order to produce the same integrated mass at some assumed depth of compensation. In the oceans, variations in the temperature of the lithosphere produce elevation changes in the manner of Pratt isostasy. For example, ridges stand 4 km above the deep ocean basins because the underlying lithosphere is hotter when the plate is young. The bathymetric swells around young hotspot volcanoes may also be supported by Pratt-type isostasy, although some combination of crustal thickening and dynamic isostasy may be operating as well. Again, gravity and geoid anomalies have been principal constraints in arguing for the mechanism of support for bathymetric swells. Long-Wavelength Anomalies

At wavelengths from 1000 to several thousand kilometers, gravity anomalies are usually derived from satellite observations and interpreted using equations appropriate for a spherical earth. Geoid is interpreted more commonly than gravity directly, as it emphasizes the longer wavelengths in the geopotential field. Isostatic compensation for smaller-scale bathymetric features, such as seamounts, can usually be ignored in that the gravity anomaly from bathymetry is canceled out by that from its compensation when spatially averaged over longer wavelengths.

The principal signal at these wavelengths arises from the subduction of lithospheric slabs and other sorts of convective overturn within the mantle. Three sorts of gravitational contributions must be considered: (1) the direct effect of mass anomalies within the mantle, either buoyant risers or dense sinkers which drive convection; (2) the warping of the surface caused by viscous coupling of the risers or sinkers to the earth’s surface; and (3) the warping of any deeper density discontinuities (such as the core– mantle boundary) also caused by viscous coupling. In the 1980s, estimates of the locations and densities of mass anomalies in the mantle responsible for the first contribution above began to become available courtesy of seismic tomography. Travel times of earthquake waves constrained the locations of seismically fast and slow regions in the mantle. By assuming that the seismic velocity variations were caused by temperature differences between hot, rising material and cold, sinking material, it was possible to convert velocity to density using standard relations. Knowing the locations of the mass anomalies driving convection inside the Earth led to a breakthrough in understanding the long-wavelength gravity and geoid fields. The amount of deformation on density interfaces above and below the mass anomalies inferred from tomography (contributions (2) and (3) above) depends upon the viscosity structure of Earth’s mantle. Coupling is more efficient with a more viscous mantle, whereas a weaker mantle is able to soften the transmission of the viscous stresses from the risers and sinkers. Therefore, one of the principal uses of marine gravity anomalies at long wavelengths has been to calibrate the viscosity structure of the oceanic upper mantle. This interpretation must be constrained by estimates of the dynamic surface topography over the oceans, which is actually easier to estimate than over the continents because of the relatively uniform thickness of oceanic crust. A fairly common result from this sort of analysis is that the oceanic upper mantle must be relatively inviscid. The geoid shows that there are large mass anomalies within the mantle driving convection that are poorly coupled to variations in the depth of the seafloor. If the upper mantle were more viscous, there should be a stronger positive correlation between marine geoid and depth of the seafloor at long wavelengths.

See also Manned Submersibles, Deep Water. Satellite Altimetry. Satellite Oceanography, History and Introductory Concepts.

(c) 2011 Elsevier Inc. All Rights Reserved.

GRAVITY

Further Reading Bell RE, Anderson R, and Pratson L (1997) Gravity gradiometry resurfaces. The Leading Edge 16: 55--59. Garland GD (1965) The Earth’s Shape and Gravity. New York: Pergamon Press. McNutt MK and Bonneville A (1996) Mapping the seafloor from space. Endeavour 20: 157--161. McNutt MK and Bonneville A (2000) Chemical origin for the Marquesas swell. Geochemistry, Geophysics and Geosystems 1. Sandwell DT and Smith WHF (1997) Marine gravity for Geosat and ERS-1 altimetry. Journal of Geophysical Research 102: 10039--10054. Smith WHF and Sandwell DT (1997) Global seafloor topography from satellite altimetry and ship depth soundings. Science 277: 1956--1962.

87

Turcotte DL and Schubert G (1982) Geodynamics: Applications of Continuum Physics to Geological Problems. New York: Wiley. Watts AB, Bodine JH, and Ribe NM (1980) Observations of flexure and the geological evolution of the Pacific Ocean basin. Nature 283: 532--537. Wessel P and Watts AB (1988) On the accuracy of marine gravity measurements. Journal of Geophysical Research 93: 393--413. Zumberg MA, Hildebrand JA, Stevenson JM, et al. (1991) Submarine measurement of the Newtonian gravitational constant. Physical Review Letters 67: 3051--3054. Zumberg MA, Ridgeway JR, and Hildebrand JA (1997) A towed marine gravity meter for near-bottom surveys. Geophysics 62: 1386--1393.

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN A. E. Mulligan and M. A. Charette, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Water flowing through the terrestrial landscape ultimately delivers fresh water and dissolved solutes to the coastal ocean. Because surface water inputs (e.g., rivers and streams) are easily seen and are typically large point sources to the coast, they have been well studied and their contributions to ocean geochemical budgets are fairly well known. Similarly, the hydrodynamics and geochemical importance of surface estuaries are well known. Only recently has significant attention turned toward the role of groundwater inputs to the ocean. Historically, such inputs were considered insignificant because groundwater flow is so much slower than riverine flow. Recent work however has shown that groundwater flow through coastal sediments and subsequent discharge to the coastal ocean can have a significant impact on geochemical cycling and it is therefore a process that must be better understood. Groundwater discharge into the coastal ocean generally occurs as a slow diffuse flow but can be found as large point sources in certain terrain, such as karst. In addition to typically low flow rates, groundwater discharge is temporally and spatially variable, complicating efforts to characterize site-specific flow regimes. Nonetheless, the importance of submarine groundwater discharge (SGD) as a source of dissolved solids to coastal waters has become increasingly recognized, with recent studies suggesting that SGDderived chemical loading may rival surface water inputs in many coastal areas. So while the volume of water discharged as SGD may be small relative to surface discharge, the input of dissolved solids from SGD can surpass that of surface water inputs. For example, SGD often represents a major source of nutrients in estuaries and embayments. Excess nitrogen loading can result in eutrophication and its associated secondary effects including decreased oxygen content, fish kills, and shifts in the dominant flora. First, let us define ‘groundwater’ in a coastal context. We use the term to refer to any water that resides in the pore spaces of sediments at the land–ocean boundary. Hence, such water can be fresh terrestrially derived water that originates as

88

rainwater infiltrating through the subsurface or it can represent saline oceanic water that flows through the sediments (Figure 1). Therefore, groundwater discharging to coastal waters can have salinity that spans a large range, being some mixture of the two end members. We therefore use the terms fresh SGD and saline SGD to distinguish these sources of fluid and brackish SGD to mean a mixture of the fresh and saline end members.

Basics of Groundwater Flow Groundwater flow in the subsurface is driven by differences in energy – water flows from high energy areas to low energy. The mechanical energy of a unit volume of water is determined by the sum of gravitational potential energy, pressure energy, and kinetic energy: Energy per unit volume ¼ rgz þ P þ

rV 2 2

½1

where r is fluid density, g is gravitational acceleration, z is elevation of the measuring point relative to a datum, P is fluid pressure at the measurement point, and V is fluid velocity. Because groundwater flows very slowly (on the order of 1 m day  1 or less), its kinetic energy is very small relative to its gravitational potential and pressure energies and the kinetic energy term is therefore ignored. By removing the kinetic energy term and rearranging eqn [1] to express energy in terms of mechanical energy per unit weight, the concept of hydraulic head is developed: Energy per unit weight ¼ hydraulic head ¼ z þ

P rg ½2

Groundwater therefore flows from regions of high hydraulic head to areas of low hydraulic head. Because groundwater flows through a porous media, the rate of flow depends on soil properties such as the degree to which pore spaces are interconnected. The property of interest in groundwater flow is the permeability, k, which is a measure of the ease with which a fluid flows through the soil matrix. Groundwater flow rate can then be calculated using Darcy’s law, which says that the flow rate is linearly proportional to the hydraulic gradient:

(c) 2011 Elsevier Inc. All Rights Reserved.

q¼

rgk ðrhÞ m

½3

GROUNDWATER FLOW TO THE COASTAL OCEAN

89

Precipitation Evaporation

Overland flow Infiltratio

n

Septic system

Ocean

Seawater circulation

ate

r

Water table

tw

Groundwater

l sa r− ce e t a fa shw Inter Fre

Figure 1 Simplified water cycle at the coastal margin. Modified from Heath R (1998) Basic ground-water hydrology. US Geological Survey Water Supply Paper 2220. Washington, DC: USGS.

where q is the Darcy flux, or flow rate per unit surface area, and m is fluid viscosity. A more general expression of Darcy’s law is: k q ¼  ðrP þ rgrzÞ m

Runup and waves

z

½4

In inland aquifers, the density of groundwater is constant and eqn [4] is reduced to the simpler form of Darcy’s law (eqn [3]). In coastal aquifers, however, the presence of saline water along the coast means that the assumption of constant density is not valid and so the more inclusive form of Darcy’s law, eqn [4], is required.

Groundwater Flow at the Coast Several forces drive groundwater flow through coastal aquifers, leading to a complex flow regime with significant variability in space and time (Figure 2). The primary driving force of fresh SGD is the hydraulic gradient from the upland region of a watershed to the surface water discharge location at the coast. Freshwater flux is also influenced by several other forces at the coastal boundary that also drive seawater through the sediments. For example, seawater circulates through a coastal aquifer under the force of gravity, from oceanic forces such as

Fresh Tidal pumping 40z

Saline

Dispersive circulation

Figure 2 Simplification of an unconfined coastal groundwater system. Water flow is driven by the inland hydraulic gradient, tides, beach runup and waves, and dispersive circulation. Other forcing mechanisms can drive fluid through coastal sediments, including seasonal changes in recharge to the inland groundwater system and tidal differences across islands.

waves and tides, as a result of dispersive circulation along the freshwater–saltwater boundary within the aquifer, and from changes in upland recharge. Several other forcing mechanisms exist, but they are generally only present in specific settings. For example, tidal height differences across many islands can drive flow through the subsurface. All of these forcing mechanisms affect the rate of fluid flow for

(c) 2011 Elsevier Inc. All Rights Reserved.

90

GROUNDWATER FLOW TO THE COASTAL OCEAN

both fresh and saline groundwater and are ultimately important in controlling the submarine discharge of both fluids. The analysis of coastal groundwater flow must account for the presence of both fresh- and saline water components. When appropriate, such as in regional-scale analysis or for coarse estimation purposes, an assumption can be made that there is a sharp transition, or interface, between the fresh and saline groundwater. While this assumption is not strictly true, it is often appropriate, and invoking it results in simplifying the analysis. For example, we can estimate the position of the freshwater–saltwater interface by assuming a sharp interface, no flow within the saltwater region, and only horizontal flow within the fresh groundwater. Invoking these assumptions means that the pressures at adjacent points along the interface on both the freshwater and saltwater sides are equal. Equating these pressures and rearranging, the depth to the interface can be calculated as follows:

Interface depth ¼

r1 z ¼ 40z r2  r1

½5

where r1 is the density of fresh water (1000 kg m  3) and r2 is the density of seawater (1025 kg m  3). This equation states that the depth of the interface is 40 times the elevation of the water table relative to mean sea level. While eqn [5] is only an approximation of the interface location, it is very helpful in thinking about freshwater and saltwater movement in response to changes in fresh groundwater levels. As recharge to an aquifer increases, water levels increase and the interface is pushed downward. This is also equivalent to pushing the interface seaward and the net effect is to force saltwater out of the subsurface and to replace it with fresh water. The opposite flows occur during times of little to no recharge or if groundwater pumping becomes excessive. While the sharp-interface approach is useful for conceptualizing flow at the coast, particularly in large-scale problems, the reality is more complex. Not only does the saline groundwater flow but also a zone of intermediate salinity extends between the fresh and saline end members, establishing what many refer to as a ‘subterranean estuary’. Like their surface water counterparts, these zones are hotbeds of chemical reactions. Because the water in the interface zone ultimately discharges into coastal waters, the flow and chemical dynamics within the zone are critically important to understand. Research into these issues has only just begun.

Detecting and Quantifying Submarine Groundwater Discharge As a first step in quantifying chemical loads to coastal waters, the amount of water flowing out of the subsurface must be determined. This is particularly challenging because groundwater flow is spatially and temporally variable. A number of qualitative and quantitative techniques have been developed to sample SGD, with each method sampling a particular spatial and temporal scale. Because of limitations with each sampling method, several techniques should be used at any particular site. Physical Approaches

Infrared thermography Infrared imaging has been used to identify the location and spatial variability of SGD by exploiting the temperature difference between surface water and groundwater at certain times of the year. While this technique is quite useful for identifying spatial discharge patterns, it has not yet been applied to estimating flow rates. An example of thermal infrared imagery is shown in Figure 3, an image of the head of Waquoit Bay, a small semi-enclosed estuary on Cape Cod, Massachusetts, USA. In late summer, the groundwater temperature is approximately 13 1C, whereas surface water is about 7–10 1C warmer. Locations of SGD can be seen in the infrared image as locations along the beach face with cooler temperature than the surrounding surface water. The image clearly shows spatial variability in SGD along the beach face, information that is extremely valuable in designing an appropriate field sampling campaign. Hydrologic approaches There are two hydrologic approaches to estimating SGD: the mass balance method and Darcy’s law calculation. Both methods are typically applied to estimating fresh groundwater discharge, although Darcy’s law can also be used to estimate saline flow into and out of the seafloor. To apply Darcy’s law, one must measure the soil permeability and hydraulic head at several locations (at least two) at the field site. Data must also be gathered to determine the cross-sectional flow area. The field data are then used with Darcy’s law to calculate a groundwater flow rate into the coastal ocean. The main disadvantages of this approach include the fact that permeability is highly heterogeneous, often ranging over several orders of magnitude, and so an ‘average’ value to use with Darcy’s law is seldom, if ever, well known. Furthermore, hydraulic head measurements require invasive, typically expensive, well installations. Finally, hydraulic head

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN

91

Temp. (°C) 2.3 2.4−7.4 7.5−8 8.1−9 9.1−9.5 9.6−10 top o con graph tou ic r

10.1−10.5 10.6−11 11.1−11.5

3-m

11.6−12 12.1−12.5 12.6−13 13.1−13.5 13.6−14 14.1−14.5

15

12

18

9

9

11 11

9

11

18

14.6−15 14

15.1−15.5 15.6−16

25

Waquoit Bay

16.1−16.5

Beach/water line

16.6−17 17.1−17.5 0

50

100

200 m

17.6−18 18.1−18.5 18.6−19 19.1−19.5 19.6−20 20.1−20.5

Figure 3 Thermal infrared image of the head of Waquoit Bay, Massachusetts. The beach/water line is shown as a red curve. Light grays imply higher temperatures and darker shades show lower temperatures. Lower temperatures within the bay indicate regions influenced by SGD. The bars show average groundwater seepage rates as measured by manual seep meters from high tide to low tide. Numbers below the bars are average rates in cm d  1. Modified from Mulligan AE and Charette MA (2006) Intercomparison of submarine ground water discharge estimates from a sandy unconfined aquifer. Journal of Hydrology 327: 411–425.

is a point measurement and capturing the spatial variability therefore requires installing many wells. The primary advantage of this approach is that it is well established and easy to implement: head measurements are easy to collect once wells are installed and the flux calculations are simple. The mass balance approach to estimating SGD requires ascertaining all inputs and outputs of water flow, except SGD, through the groundwatershed. Assuming a steady-state condition over a specified time frame, the groundwater discharge rate is calculated as the difference between all inputs and all outputs. Implementing this approach can be quite simple or can result in complex field campaigns, but the quality of the data obviously affect the level of uncertainty. Even with extensive field sampling, water budgets are seldom known with certainty and

so should be used with that in mind. Furthermore, if the spatial and temporal variability of SGD is needed for a particular study, the mass balance approach is not appropriate. Direct measurements: Seepage meters SGD can be measured directly with seepage meters. Manual seepage meters (Figure 4) are constructed using the tops of 55 gallon drums, where one end is open and placed into the sediment. The top of the seepage meter has a valve through which water can flow; a plastic bag prefilled with a known volume of water is attached to the valve, so that inflow to or outflow from the sediments can be determined. After a set length of time, the bag is removed and the volume of water in the bag is measured. The change in water volume over the sampling period is

(c) 2011 Elsevier Inc. All Rights Reserved.

92

GROUNDWATER FLOW TO THE COASTAL OCEAN

then used to determine the average flow rate of fluid across the water–sediment boundary over the length of the sampling period. These meters are very simple to operate, but they are manually intensive and are sensitive to wave disturbance and currents. Furthermore, they only sample a small flow area and so many meters are needed to characterize the spatial variability seen at most sites (Figure 3). Recently, several other technologies have been applied toward developing automated seep meters. Technologies include the heat-pulse method, continuous

Water surface

Chemical Tracers

Collection bag

Connectors

heat, ultrasonic, and dye dilution. These meters can be left in place for days and often weeks and will measure seepage without the manual intervention needed using traditional seepage meters. The trade-off with these meters is that they are expensive and therefore only a limited number are typically employed at any given time. An example of seepage measurements made using the dye dilution meter is shown in Figure 5. Note that the seepage rates are inversely proportional to tidal height. As the tide rises, the hydraulic gradient from land to sea is reduced, seepage slows, and the flow reverses, indicating that seawater is flowing into the aquifer at this location during high tide. Conversely, at low tide, the hydraulic gradient is at its steepest and SGD increases.

Seepage meter

Fluid flow Sediment

Figure 4 Graphic of a manual seepage meter deployed at the sediment–water interface. These meters can be constructed using the top of a 55 gallon drum. The collection bag serves as a fluid reservoir so that both inflow to and outflow from the sediments can be measured. Modified from Lee DR (1977) A device for measuring seepage in lakes and estuaries. Limnology and Oceanography 22: 140–147.

The chemical tracer approach to quantifying SGD has an advantage over seepage meters in that it provides an integrated flux over a wide range of spatial scales from estuaries to continental shelves. The principle of using a chemical tracer is simple: find an element or isotope that is highly enriched (or depleted) in groundwater relative to other sources of water, like rivers or rainfall, to the system under study. If SGD is occurring, then the flux of this element via groundwater will lead to enrichment in the coastal zone that is well above background levels in the open ocean (Figure 6). A simple mass balance/box model for the system under study can be performed, where all sources of the tracer other than groundwater are subtracted from the total inventory of the chemical. The residual inventory, or ‘excess’, is then

75

Flow (cm d−1) Tide (cm)

50 Submarine groundwater discharge

25

0 −25 −50 6 Sep.

Flow Tide 7 Sep.

Seawater flow into sediments

8 Sep.

Date 2002 Figure 5 Seepage rates at Waquoit Bay, Massachusetts, USA, measured using a dye dilution automated seepage meter. Modified from Sholkovitz ER, Herbold C, and Charette MA (2003) An automated dye-dilution based seepage meter for the time-series measurement of submarine groundwater discharge. Limnology and Oceanography: Methods 1: 17–29.

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN

34 Myrtle 20 Beach 23 20 17 15 Inner 13 shelf 12 10 14 15

(dpm 100 l−1) Jul. 1994

226Ra

Winyah Bay

17

25 20 15 Inner 13 shelf 11 10 14 13

Sh

am

re

t lf S

Gu

Open ocean = 8.0 ± 0.5

am

tre

Sh

12 11 10

31 81

ak

re

b elf

8.9

el

32

11 8.9

ak

20 18

15 15 11 9.3 8.7 10 10

re

26

17

Cape Fear

10

17 19 16 13 20 19 10 25 8.2 Charleston 28 24 21 7.8 10 24 20 8.7 20 11 14 11 25 15 10 10 9.7 19 11 11 11 10 9.4 10 11

fb

Sav an Riv nah er

Latitude (°N)

33

25

93

S ulf

0

20

40

G

80

79

78

60 km

80

100

77

Longitude (° W) Figure 6 The distribution of 226Ra offshore South Carolina reveals high activities on the inner shelf that decrease offshore. Moore (1996) used the excess 226Ra to estimate regional SGD fluxes.

divided by the concentration of the tracer in the discharging groundwater to calculate the groundwater flow rate. Naturally occurring radionuclides such as radium isotopes and radon-222 have gained popularity as tracers of SGD due to their enrichment in groundwater relative to other sources and their built-in radioactive ‘clocks’. The enrichment of these tracers is owed to the fact that the water–sediment ratio in aquifers is usually quite small and that aquifer sediments (and sediments in general) are enriched in many U and Th series isotopes; while many of these isotopes are particle reactive and remain bound to the sediments, some like Ra can easily partition into the aqueous phase. Radon-222 (t1/2 ¼ 3.82 days) is the daughter product of 226Ra (t1/2 ¼ 1600 years) and a noble gas; therefore, it is even more enriched in groundwater than radium.

A key issue when comparing different techniques for measuring SGD is the need to define the fluid composition that each method is measuring (i.e., fresh, saline, or brackish SGD). For example, whereas hydrogeological techniques are estimates of fresh SGD, the radium and radon methods include a component of recirculated seawater. Therefore, it is often not possible to directly compare the utility of these techniques. Instead, they should be regarded as complementary. One of the seminal studies that showed how SGD can impact chemical budgets on the scale of an entire coastline was conducted off North and South Carolina, USA. Literature estimates of water residence time, riverine discharge/suspended sediment load, and the activity of desorbable 226Ra on riverine particles were used to determined that only B50% of the 226Ra inventory on the inner continental shelf

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN

222Rn

140

7

120

6

100

5

80

4

60

3

40

2

20

1 18 May

19 May

20 May

40

30

0 21 May

20

10

Seep meter (cm d−1)

8

Water level (cm)

(dpm l−1)

94

0

Date 2002 Tide Seep meter Radon Figure 7 Plot comparing variations in SGD at Shelter Island, NY, between an automated seepage meter and the continuous radon method.

off North and South Carolina could be explained by surface inputs (Figure 6). The remaining inventory enters the system via SGD. Using an estimate of groundwater 226Ra, it is calculated that the groundwater flux to this region of the coastline is on the order of 40% of the river water flux. The approach for quantifying SGD using 222Rn is similar to that for radium (226Ra), except for a few key differences: (1) 222Rn loss to the atmosphere must be accounted for in many situations, (2) there is no significant source from particles in rivers, and (3) decay must be accounted for owing to its relatively short half-life. The first example of 222Rn use to quantify SGD to the coastal zone occurred in a study in the northeastern Gulf of Mexico. The strong pycnocline that develops in the summer time means that fluid that flows from the sediments into the bottom boundary layer is isolated from the atmosphere and therefore no correction for the air–sea loss of 222Rn is needed. Using this approach, diffuse SGD in a 620 km2 area of the inner shelf was estimated to be equivalent to B20 first-magnitude springs (a firstmagnitude spring has a flow equal to or greater than 245 000 m3 water per day, which is equivalent to B60% of the daily water supply for the entire state of Rhode Island, USA). Since 2000, a number of SGD estimation technique intercomparison experiments have been conducted through a project sponsored by the Scientific Committee on Oceanic Research (SCOR) and the Land– Ocean Interaction in the Coastal Zone (LOICZ)

Project. During these intercomparisons, several methods (chemical tracers, different types of seep meters, hydrogeologic approaches, etc.) were run side by side to evaluate their relative strengths and weaknesses. Figure 7 displays a comparison from the Shelter Island, NY, experiment of calculated radon fluxes (based on measurements from a continuous radon monitor), with seepage rates measured directly via the dye dilution seepage meter. During the period (17–20 May) when both devices were operating at the same time, there is a clear and reproducible pattern of higher radon and water fluxes during the low tides. There is also a suggestion that the seepage spikes slightly led the radon fluxes, which is consistent with the notion that the groundwater seepage is the source of the radon (the radon monitor was located offshore of the seepage meter). The excellent agreement in patterns and overlapping calculated advection rates (seepage meter: 2–37 cm day  1, average ¼ 127 8 cm day  1; radon model: 0–34 cm day  1, average ¼ 1177 cm day  1) by these two completely independent assessment tools is reassuring.

Geochemistry of the Subterranean Estuary The magnitude of chemical fluxes carried by SGD is influenced by biogeochemical processes occurring in the subterranean estuary, defined as the mixing zone between groundwater and seawater in a coastal

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN

95

(a)

Fe (ppm) 4500

0

0

1500

Fe (ppm)

3000

Tan/red Tan/gray

100

150

Dark red

150

50

0

50

100 P (ppm)

150

1500

3000

50

100

Core 3 200

4500

150

Core 2 200

0

0

Yellow

Depth (cm)

Orange Red/orange

100

50

4500

Gray

3000

Gray

1500

Gray

0

0

Gray/yellow

Fe (ppm) (b)

200

Core 5 200

0

50

100

150

200

P (ppm)

0

50

100

150

200

P (ppm)

Figure 8 (a) Scientists extruding a sediment core taken through the subterranean estuary of Waquoit Bay, MA. Note the presence of iron oxides within the sediments at the bottom of the core (orange-stained sediments in foreground). (b) Changes in iron and phosphate concentration with depth in three sediment cores similar to the one shown in (a). The red circles indicate Fe concentration (ppm:mg Fe/g dry sediment) while the blue diamonds represent P (ppm:mg P/g dry sediment). Error bars indicate the standard deviation for triplicate leaches performed on a selected number of samples. The dashed lines represent the concentration of Fe and P in ‘off-site’ quartz sand. Also shown is the approximate color stratigraphy for each core. The R2 value for Fe vs. P in cores 2, 3, and 5 is 0.80, 0.91, and 0.16, respectively.

(c) 2011 Elsevier Inc. All Rights Reserved.

96

GROUNDWATER FLOW TO THE COASTAL OCEAN

aquifer. The biogeochemical reactions in such underground estuaries are presumed to be similar to the surface estuary (river water/seawater) counterpart, though few comprehensive studies of chemical cycling in subterranean estuaries have been undertaken. Drivers of biogeochemical reactions in these environments include oxidation–reduction gradients, desorption–sorption processes, and microbially driven diagenesis. In a study of the Waquoit Bay subterranean estuary, a large accumulation of iron (hydr)oxide-coated sediments within the fresh–saline interface was encountered. These iron-oxide-rich sands could act as a geochemical barrier by retaining and accumulating certain dissolved chemical species carried to the subterranean estuary by groundwater and/or coastal seawater. Significant accumulation of phosphorus in the iron oxide zones of the Waquoit cores exemplifies this process (Figure 8). Phosphorous is not the only nutrient that can be retained/removed via reactions in the subterranean estuary. The microbial reduction of nitrate to inert dinitrogen gas, a process known as denitrification, is known to occur in the redox gradients associated with fresh and saline groundwater mixing. Conversely, ammonium, which is more soluble in saline environments, may be released within the subterranean estuary’s mixing zone. While the overall importance of SGD on the ‘global cycle’ of certain chemical species remains to be seen, there is little doubt that SGD is important at the local scale both within the United States and throughout the world.

Summary Groundwater discharge to the coastal ocean can be an important source of fresh water and dissolved solutes. Although a significant amount of research into the role of SGD in solute budgets has only occurred in the past decade or so, there is increasing evidence that solute loading from groundwater can be significant enough to affect solute budgets and even ecosystem health. Proper estimation of solute loads from groundwater requires confident estimates of both the groundwater discharge rate and average solute concentrations in the discharging fluid, neither of which are easily determined. Estimating groundwater discharge rates is complicated by the spatial and temporal variability of groundwater flow. A multitude of time-varying driving mechanisms complicate analysis, as does geologic heterogeneity. Nonetheless, a suite of tools from hydrogeology, geophysics, and geochemistry have been developed for sampling and measuring SGD.

The nature of geochemical reactions within nearshore sediments is not well understood, yet recent studies have shown that important transformations occur over small spatial scales. This is an exciting and critically important area of research that will reveal important process in the near future.

See also Chemical Processes in Estuarine Sediments. Estuarine Circulation. Gas Exchange in Estuaries. Inverse Modeling of Tracers and Nutrients. LongTerm Tracer Changes. Submarine Groundwater Discharge. Trace Element Nutrients. Tracer Release Experiments. Tracers of Ocean Productivity.

Further Reading Burnett WC, Aggarwal PK, Bokuniewicz H, et al. (2006) Quantifying submarine groundwater discharge in the coastal zone via multiple methods. Science of the Total Environment 367: 498--543. Burnett WC, Bokuniewicz H, Huettel M, Moore WS, and Taniguchi M (2003) Groundwater and pore water inputs to the coastal zone. Biogeochemistry 66: 3--33. Cable JE, Burnett WC, Chanton JP, and Weatherly GL (1996) Estimating groundwater discharge into the northeastern Gulf of Mexico using radon-222. Earth and Planetary Science Letters 144: 591--604. Charette MA and Sholkovitz ER (2002) Oxidative precipitation of groundwater-derived ferrous iron in the subterranean estuary of a coastal bay. Geophysical Research Letters 29: 1444. Charette MA and Sholkovitz ER (2006) Trace element cycling in a subterranean estuary. Part 2: Geochemistry of the pore water. Geochimica et Cosmochimica Acta 70: 811--826. Heath R (1998) Basic ground-water hydrology. US Geological Survey Water Supply Paper 2220. Washington, DC: USGS. Kohout F (1960) Cyclic flow of salt water in the Biscayne aquifer of southeastern Florida. Journal of Geophysical Research 65: 2133--2141. Lee DR (1977) A device for measuring seepage in lakes and estuaries. Limnology and Oceanography 22: 140--147. Li L, Barry DA, Stagnitti F, and Parlange J-Y (1999) Submarine groundwater discharge and associated chemical input to a coastal sea. Water Resources Research 35: 3253--3259. Michael HA, Mulligan AE, and Harvey CF (2005) Seasonal water exchange between aquifers and the coastal ocean. Nature 436: 1145--1148 (doi:10.1038/ nature03935). Miller DC and Ullman WJ (2004) Ecological consequences of ground water discharge to Delaware Bay, United States. Ground Water 42: 959--970.

(c) 2011 Elsevier Inc. All Rights Reserved.

GROUNDWATER FLOW TO THE COASTAL OCEAN

Moore WS (1996) Large groundwater inputs to coastal waters revealed by Ra-226 enrichments. Nature 380: 612--614. Moore WS (1999) The subterranean estuary: A reaction zone of ground water and sea water. Marine Chemistry 65: 111--125. Mulligan AE and Charette MA (2006) Intercomparison of submarine ground water discharge estimates from a sandy unconfined aquifer. Journal of Hydrology 327: 411--425. Paulsen RJ, Smith CF, O’Rourke D, and Wong T (2001) Development and evaluation of an ultrasonic ground water seepage meter. Ground Water 39: 904--911. Portnoy JW, Nowicki BL, Roman CT, and Urish DW (1998) The discharge of nitrate-contaminated groundwater from developed shoreline to marsh-fringed estuary. Water Resources Research 34: 3095--3104.

97

Shinn EA, Reich CD, and Hickey TD (2002) Seepage meters and Bernoulli’s revenge. Estuaries 25: 126--132. Sholkovitz ER, Herbold C, and Charette MA (2003) An automated dye-dilution based seepage meter for the timeseries measurement of submarine groundwater discharge. Limnology and Oceanography: Methods 1: 17--29. Slomp CP and Van Cappellen P (2004) Nutrient inputs to the coastal ocean through submarine groundwater discharge: Controls and potential impact. Journal of Hydrology 295: 64--86. Taniguchi M, Burnett WC, Cable JE, and Turner JV (2002) Investigation of submarine groundwater discharge. Hydrological Processes 16: 2115--2129. Valiela I, Foreman K, LaMontagne M, et al. (1992) Couplings of watersheds and coastal waters: Sources and consequences of nutrient enrichment in Waquoit Bay, Massachusetts. Estuaries 15: 443--457.

(c) 2011 Elsevier Inc. All Rights Reserved.

HABITAT MODIFICATION M. J. Kaiser, Bangor University, Bangor, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction This article deals with the causes and consequences of habitat modification in the marine environment, that occur as a result of human activities. The effects of natural forcing (waves, currents, glacial scour, seismic activity, and natural climate change) and how they interact with habitats are dealt with elsewhere. Nevertheless, in order to place the effects of human activities on habitats in a meaningful context, it is necessary to appreciate the different scales at which natural and human-mediated processes occur. Habitat modifications through natural processes can occur as a result of direct physical or chemical processes (e.g., alteration of temperature, seismic activity, and input of freshwater) or through biologically mediated processes (e.g., changes in species composition or dominance). At the largest scale, marine habitats are influenced by the interaction between atmospheric changes and the water column, which in turn affect the biological processes that occur in the habitat. These changes in the watercolumn environment occur over the largest spatial scales (entire oceans) and have periodicities that vary in timescale from millennia to decades (e.g., the North Atlantic Oscillation (NAO)).

Large-Scale Processes: Natural Forcing The NAO is a good example of how physical processes can affect habitat. Decadal changes in the NAO affect wind forcing and precipitation levels according to whether the NAO index is positive or negative. In years when the index is positive, average wind speeds are higher leading to greater physical mixing of the water column, which will affect the timing and persistence of stratification on the continental shelf Northeast Atlantic. When in a negative phase, lower wind speeds are associated with an increased incidence of anoxic events due to prolonged periods of stratification in coastal waters. The latter is particularly pronounced in enclosed bodies of water such as the Baltic Sea. Wind forcing also generates stress at the seabed in shallow water near the coast. These

areas are among the most biologically productive and support important nursery habitats for commercially important fish, and diving and wading birds. Wave stress is a direct controlling factor on benthic production which increases with depth and distance from the shore to a point where it reaches a maximum before decreasing again. Increasing levels of wind stress will decrease near-shore benthic production and lower coastal-carrying capacity for species dependent upon the benthos (Figure 1). Whenever the NAO is positive, increasing precipitation will elevate the amount of sediment discharged from rivers and is associated with increased frequencies of extreme sediment-discharge events linked with flash floods. Such events in large river systems (e.g., the River Amazon) can cause mass mortalities of benthos 10 km2 offshore. Elevated inputs of freshwater discharge will affect the extent of density-driven coastal currents in regions of freshwater influence (ROFIs) leading to extended front systems that run in parallel with the coastline. In addition to these periodic changes in the oceanic regime, we can expect global climate change to greatly exacerbate the extreme nature of many aspects of physical forcing that influences habitat modification in marine systems.

Smaller-Scale Processes: Natural and Biological Forcing The examples given above operate at large scales and influence both water-column and seabed habitat properties. Other more localized natural forcing events can have profound impacts upon the habitat. Examples would include glacial scour that can remove entire habitats on a scale of a few kilometers, localized seismic activity resulting in gas discharge or geothermal activity, localized coastal erosion and daily tidal scour, and carrion falls to the deep seafloor. Apart from the last example these individual processes operate at much larger scales than the modifying processes undertaken by individual biota. Such processes include grazing, predation, and bioturbation. Despite the relatively small scale of individual habitat-modifying events (e.g., an intertidal fiddler crab excavating a burrow), the sum total of many individuals events can equate to a significant disturbance process. However, natural forcing can lead to changes in biological community structure, and habitat changes can arise if ecosystem engineering organisms are affected. Ecosystem engineers

(c) 2011 Elsevier Inc. All Rights Reserved.

99

Number of sightings

(a)

HABITAT MODIFICATION

500

Waves stress−depth

(b)

400

Sea ducks, North Wales

300

Sea ducks, shell flat

200 100 0 1

3

5

7

9

11 13 15 17 19 21 23 25

Wave stress (N m−2)

100

30 25 20 15 10 5 0

Shell flat North Wales Burbo bank

0

5

Depth (m)

20

25

Biomass−Wave stress

(c) log10 bivalve AFDW (g)

10 15 Depth (m)

2 1 0 −0

0

5

10

15

20

25

30

−2 −3 Wave stress (N m−2)

Figure 1 (a) The distribution of shellfish eating diving sea ducks in Liverpool Bay, UK, with mean water depth, (b) the relationship between wave stress and depth, and (c) the relationship between bivalve biomass with depth and wave stress in the same area. Shell flat is more exposed to prevailing winds than the sheltered North Wales coast. Consequently, a peak in bivalve biomass occurs in shallower water off North Wales but in deeper water (where wave stress is lower) on shell flat. The distribution of the diving sea ducks corresponds reasonably well with the observed peak in biomass when it is considered that diving depth limits prey availability. Adapted from Kaiser MJ, Clarke KR, Hinz H, Austen MCV, Somerfield PJ, and Karakassis I (2006) Global analysis and prediction of the response of benthic biota and habitats to fishing. Marine Ecology Progress Series 311: 1–14.

are those taxa that have a strong influence on some component of a marine habitat. A good example of the latter is the fluctuation between kelp and sea urchin-dominated systems in response to periodic natural outbreaks of urchin disease (natural forcing). When urchin population biomass is high, kelp biomass is grazed down, but the latter are released from grazing when urchin populations are reduced by disease. Such systems are also affected by human activities and we will return to this later.

Habitat Modification as an Ecological Process In summary, natural forcing provides a background of habitat modification at generally large scales against which are caste the smaller-scale, more localized natural forcing events and biologically mediated modifying activities. Habitat modification creates disturbance which is a critical process in the maintenance of alpha and beta diversity in terrestrial and aquatic systems. Consequently, the seafloor is a patchwork of communities in different stages of recovery, succession, or climax. It is against this natural background of disturbance and habitat

modification that the impact and significance of human activities needs to be assessed.

Human Forcing The effects of human activities can be split broadly into: direct physical effects on the habitat, and indirect effects through alteration of habitat quality or through changes in biological composition. Direct effects are usually more immediate in terms of their impact and are more readily quantifiable (e.g., aggregate extraction, fishing disturbance, and land reclamation), whereas indirect effects may remain unrecognized for many years or until they interact with some other stressor to induce a habitat change (e.g., increasing levels of sublethal pollutants). Human activities are likely to induce long-lasting habitat changes, if the scale, frequency, or intensity of the human activity exceeds the background natural disturbance regime.

Fishing One of the most widespread agents of direct habitat modification is the use of fishing gears that are towed

(c) 2011 Elsevier Inc. All Rights Reserved.

HABITAT MODIFICATION

Y = loge (1 + (% change from control)/101)

across the seabed. This source of habitat disturbance has been the subject of over 100 experimental manipulations conducted over the last 25 years. In general, as fishing gear is towed across the seabed, parts of the gear dig into the substratum. This causes resuspension of sediments into the water column and, in deeper mud areas, can lead to significant transport of sediment down the continental shelf slope. The consequences of sediment resuspension in more dynamic habitats are likely to be less significant; however, this remains a much understudied topic. What is unequivocal is the direct mortality caused by the direct physical contact of fishing gear with benthic biota. The use of meta-analytic approaches has enabled the general response of benthic biota to different fishing gears in different habitats. What emerges from this analysis is that the direct mortality caused varies among different fishing gears and different habitats. While this result is not surprising, it has important management implications. Scallop (Pectinid) dredges have the most negative effect of all fishing gears across all habitat types, whereas the response of biota to beam trawls varies considerably with habitat type (Figure 2). Perhaps of greater importance is the time taken for disturbed communities to recover to a condition comparable with similar unimpacted areas. These empirical studies have formed the basis for a body-size-based model of the response of benthic communities to fishing disturbance which was validated extensively

with field data. This model has been used to demonstrate that the response of benthic biomass and production to fishing disturbance follows a negative power function such that relatively low levels of fishing intensity cause the greatest proportional loss of virgin biomass and production. Consequently, benthic communities in areas that are fished 2 or 3 times per annum differ little in absolute terms from similar benthic communities that are fished perhaps 10 or more times per annum. This finding is also applicable to other forms of chronic physical disturbance. The preceding observations and predictions apply to seabed habitats on the continental shelf. As yet no one has undertaken experimental manipulations of fishing disturbance in the deep sea. Nevertheless, the life history of deep-sea biota is such that the consequences of physical impacts by fishing gear on the deep-sea benthos and associated habitats will be measured in terms of years to hundreds of years. Recent discoveries of cold-water coral formations on the continental slope in the North Atlantic have coincided with clear evidence of the direct destruction of some of these structures by towed demersal trawl gear. The sensitivity and vulnerability of such longlived biogenic structures subsequently led to the creation of fishing-exclusion areas designed to protect these habitats. Fishing can also modify habitat indirectly through changes in trophic interactions. This is most likely to

Deposit feeders 1

BT

101

Suspension feeders

OT

SD

BT

OT

SD

0

−1

−2

5

4

14

2

38 77

6

28 13

5

25 34

2

35 38

64 11

2

G

S

M

G

S

G

S

G

S

G

S

G

M

M

M

M

M

S

Figure 2 The mean initial response (up to 7 days after impact) of deposit- and suspension-feeding fauna to beam trawling (BT), otter trawling (OT), and scallop dredging (SD) in gravel (G), sand (S), and muddy sand/mud habitats combined (M). Also shown are 95% confidence intervals (dashed lines indicate only two points available for the mean calculation and hence intervals usually extending outside the plotted range). Values above the x-axis denote the number of data points in each mean calculation. An adequate test for a significant initial impact is to note whether the 95% confidence interval crosses the zero-response line (central dashed line). The key point to note from this figure is how the initial response of the biota to different fishing gears varies to a different extent among habitat types. Thus, the response to beam trawling is very variable, whereas scallop dredging appears to have a consistently negative effect in all habitats. Adapted from Kaiser MJ, Galanidi M, Showler DA, et al. (2006) Distribution and behaviour of common scoter Melanitta nigra relative to prey resources and environmental parameters. Ibis 148: 110–128.

(c) 2011 Elsevier Inc. All Rights Reserved.

102

HABITAT MODIFICATION

occur in systems in which keystone species exist that exert a strong influence on other species at a higher or lower trophic level to themselves. Such systems are prone to ‘trophic cascades’. A good example of the latter is the relationship between fishing, sea lions, killer whales, sea otters, sea urchins, and kelp on the west coast of Canada. Fishing activities have depleted populations of forage fish that supported the sea lion population. As sea lions are an important prey species for killer whales, the latter have begun to predate sea otters which in turn feed on sea urchins. The reduction in sea otter population size has released sea urchins from predation pressure such that they have increased in abundance and have grazed down kelp beds such that only urchin barrens dominated by small fleshy algae remain. Similar cascade effects have also been reported for coral reef systems. Cascade effects are manifested relatively quickly due to the strong linkage between each of the components in the system. In other more complex food webs in which the linkages among different species are more complex, the entire removal of one species may have little or no effect on other components. The reason for this lack of response is that multiple species at each trophic level may utilize the same resources; thus, any resources made available after the demise of one species will be utilized by another. This is a phenomenon termed ‘diffuse predation’. Thus, ecosystem complexity provides a buffer against the effects of harvesting specific species. However, a current pressing challenge is to be able to understand when we risk overstepping the threshold beyond which sufficient components of the ecosystem have been removed to alter its functioning. The latter has been thought to occur off the eastern coast of Canada. The removal of demersal fish biomass coupled with a reduction in benthic biomass has resulted in a breakdown in benthopelagic coupling. Consequently, the system has undergone a phase shift such that most of the energy within this ecosystem is recycled within the pelagic compartment.

Aquaculture Wild-capture fishery landings peaked in the early 1990s, while the global contribution of aquaculture to the harvest of marine species has continued to increase at a rate of 10% per year since 1990. Most mariculture practices occur within 10 nautical miles of the coastline due to logistical and technological challenges. Habitat modification can occur through a change in land use through the deforestation of

mangroves for the construction of shrimp ponds or through changes in hydrography and inputs of organic matter by suspended cultivation systems. Most of the important habitat modification processes related to aquaculture relate to pollution from excessive organic inputs from excessive feed and the production of feces. The resulting elevated levels of organic input lead to classic responses in the benthic fauna (Figure 2). Extreme cases of organic-enrichment deoxygenation of the water column and sediment can result in anoxic sediments in the immediate area of cultivation and a benthic community dominated by mats of bacteria. In addition to organic enrichment, the increasing use of antimicrobial products and chemical agents to control sea-lice can alter the microbial composition of the sediment in the immediate vicinity of suspended cages used for fish cultivation. In general, however, the effects of cage cultivation on sediment habitats are relatively localized and can be alleviated by rotating the use of different areas such that areas of the seabed are allowed to recover for a period of time. In enclosed bodies of water, it is possible to modify the water-column habitat by overstocking the biomass of cultivated species such that the carrying capacity of the system is exceeded. Examples of this have occurred in enclosed bays in France (Baie D’Archacon), where oyster culture was allowed to develop in an uncontrolled manner. Overstocking led to poor growth rates (due to competition for food) and changes in benthic community composition. Some bivalve culture systems use direct laying of the stock onto the seabed. This can induce hydrological changes directly above the bivalve bed (e.g., in the case of cockles and mussels) such that turbulence increases. This has the effect of increasing the advection rate of phytoplankton and other particles down to the seabed (and hence enhances food supply) but can also increase the sedimentation rate. These systems are often typified by raised areas of the seabed that are occasionally leveled by the cultivator. The associated community changes that occur with on-bed cultivation are often dramatic, but are confined to a limited area within and around the cultivated plots (Figure 3). Overdevelopment of shrimp culture ponds along the coastal margin has frequently led to degradation of water quality and has exacerbated the spread of disease such that the cultivation activities have become unviable. These environmental problems are well understood and new developments typically involve the use of cultivation practices that are better integrated into natural systems to alleviate and prevent their potential negative effects (e.g., by planting additional mangroves).

(c) 2011 Elsevier Inc. All Rights Reserved.

HABITAT MODIFICATION

Distant control plots (no mussels) Number of species

25

?

20 15 10

Bare mud within mussel bed

5 Mussel covered areas 0

2

4

6

8

10

ln(n + 1) Mussel area cm−2 Figure 3 The response of species richness to the mussel cover in commercially cultivated plots of mussels. At very low mussel cover there is a suggestion that species richness may increase slightly as indicated by the dashed line; however, increasing the extent of mussel cover is associated with a sharp decline in species richness. Nevertheless, these effects are highly restricted to the area of cultivation as indicated by reference to ‘distant control plots’ with no mussel cover. Adapted from Beadman HA, Kaiser MJ, Galanidi M, Shucksmith R, and Willows RI (2004) Changes in species richness with stocking density of marine bivalves. Journal of Applied Ecology 41: 464–475.

Aggregate Extraction and Mining The ecological effects of mining of minerals and aggregates are similar to fishing to the extent that both processes involve the direct removal of the habitat from the seabed. Compared to the spatial area affected by towed bottom fishing gear, the effects of mining and dredging are confined to smaller areas. Nevertheless, the intensity of the impact is much greater (i.e., the depth to which the habitat is affected). Aggregates are often stable substrata that are not easily reworked by natural physical forcing, while mining can occur on the abyssal plain where physical forcing processes are much reduced. Consequently, recovery rates for both the habitats and their associated fauna are likely to exceed those that occur in response to bottom fishing in the same habitat.

River Discharge: Quality and Quantity Agricultural practices such as deforestation have led to a fourfold increase in nutrient loads in river discharge and are associated with phytoplankton blooms that generate mainly diatomaceous phytodetritus. This organic material fuels rapid rates of microbial community respiration and results in depletion of oxygen at the sediment–water interface. Such events are usually periodic and tend to coincide with periods of calm weather when inshore waters undergo stratification. The precise timing and extent of such events will vary interannually according to

103

the exact climatic conditions that prevail and upon the quality and quantity of river discharge. Enclosed water basins such as the Baltic Sea, the Adriatic, and the Gulf of Mexico are well-documented examples of locations where such events occur on a regular basis. These systems are characterized by restricted tidal inundation, poor water exchange, and are highly susceptible to stratification. While anoxic events would appear to be natural events associated with areas of river discharge, the extent and severity of these events would appear to be related to human activities. Nitrogen loading doubled in the River Mississippi catchment between the 1900s and 1980s resulting in eutrophication that has increased the scale of anoxic events in the Gulf of Mexico. This supposition is supported by studies of stable isotope signatures and organic tracers in sediment cores which indicate that nutrient levels started to increase in the 1950s and finally leveled off in the 1980s, and can be linked to the concomitant rise in the use of artificial fertilizers. In addition to concern regarding the effects of elevated nutrient inputs into coastal waters, there are many other contaminants that persist in the marine environment, particularly in sediments. Many of these contaminants, such as radionuclides and organic substances, are derived from industrial and heat-generation sources. Polychlorinated biphenols (PCBs) are particularly persistent and accumulate through the food chain with negative effects expressed in predators at the top of food chains. While the effects of such contaminants on apex predators have been documented in a number of cases, the effects on organisms at lower trophic levels are less well studied. Although invertebrates such as amphipods are used in lethal toxicity tests, the incidence of contaminant mortality associated with bioengineering organisms that affect habitat structure remains unknown. The effects of contaminants may be subtle in that they do not necessarily cause direct mortality but may have negative effects on recruitment processes and larval viability. Sublethal effects may lower the cellular energy activity of organisms such that they are no longer able to cope with more stressful environments. Thus there is a distinct possibility that a combination of factors (increasing temperature, contaminant load) could have lethal effects. The damming of rivers for the purposes of water abstraction has severely lowered the discharge of some river systems with negative consequences for local ecosystems. For example, in the Adriatic Sea, output of the River Po and adjacent river systems has been lowered by 12% in the recent years. The reduction in nutrient inputs has been associated with a

(c) 2011 Elsevier Inc. All Rights Reserved.

104

HABITAT MODIFICATION

decrease in primary production in the local shelf water mass. The structure of many river delta regions is maintained by the supply of suspended river-born sediment. However, human interference with the supply of suspended sediment can lead to alteration in the structure of deltaic regions. For example, the construction of the barrage on the River Nile in 1868 led to the reversal of the accretion of the Nile delta, a situation later worsened by the construction of the Aswan dam. This led to coastal erosion rates of between 5 and 240 m per year. The associated reduction in nutrients transported out to sea was linked with a decline in the catches of sardines. In later years, fishery production increased as the reductions in nutrients supplied through the Nile discharge were replaced by increasing coastal urbanization.

Petroleum Exploration and Production The impacts of oil production on shelf sea ecosystems have been well reviewed. The ecological effects of the contaminants of drilling muds typically lead to a reduction in species diversity, the effects of which ameliorate with increasing distance from the drilling platform. In the North Sea, the chronic use of drilling muds over a period of 6–9 years was associated with community effects at a distance of up to 6 km from the drilling platform, although in cases when water-based muds were used, the ecological effects were reduced. In other areas, the effects may be much more localized as in the case of gas platforms in the Gulf of Mexico where the ecological effects of the drilling muds were confined to a distance of 100 m from the drilling platform. Thus, as with fishing, the impact of hydrocarbon exploration depends upon the nature of the habitat that is affected by this activity.

Discussion The extent of natural and human agents of habitat modification is highly varied both in terms of the temporal and spatial scale at which they affect habitats. Smaller-scale processes should be viewed as being nested with the larger-scale forcing processes that affect global or basin-scale processes. Direct mitigation of human activities that have negative effects on habitat structure that lead to phase shifts or long-term change (such as fishing) are amenable to management at a national or local level. Understanding the interaction between a particular human

activity and the response of different habitat types is a necessary prerequisite for effective spatial management that can lead to sustainable use and conservation of the marine environment.

See also Breaking Waves and Near-Surface Turbulence. Cold-Water Coral Reefs. Groundwater Flow to the Coastal Ocean. Hypoxia. Mangroves. Mariculture Overview.

Further Reading Beadman HA, Kaiser MJ, Galanidi M, Shucksmith R, and Willows RI (2004) Changes in species richness with stocking density of marine bivalves. Journal of Applied Ecology 41: 464--475. Estes JA and Duggins DO (1995) Sea otters and kelp forests in Alaska: Generality and variation in a community ecology paradigm. Ecological Monographs 65: 75--100. Gray JS, Clarke KR, Warwick RM, and Hobbs G (1990) Detection of initial effects of pollution on marine benthos: An examination from the Ekofisk and Eldfisk oilfields, North Sea. Marine Ecology Progress Series 66: 285--299. Hall SJ (1994) Physical disturbance and marine benthic communities: Life in unconsolidated sediments. Oceanography and Marine Biology Annual Review 32: 179--239. Jennings and Kaiser M (1998) The effects of fishing on marine ecosystems. Advances in Marine Biology 34: 201--352. Kaiser MJ, Clarke KR, Hinz H, Austen MCV, Somerfield PJ , and Karakassis I (2006) Global analysis and prediction of the response of benthic biota and habitats to fishing. Marine Ecology Progress Series 311: 1--14. Kaiser MJ, Galanidi M, Showler DA, et al. (2006) Distribution and behaviour of common scoter Melanitta nigra relative to prey resources and environmental parameters. Ibis 148: 110--128. Olsgard F and Gray JF (1995) A comprehensive analysis of the effects of offshore oil and gas exploration and production on the benthic communities of the Norwegian continental shelf. Marine Ecology Progress Series 122: 277--306. Pearson TH and Rosenberg R (1987) Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanography and Marine Biology Annual Review 16: 229--311. Planques A, Guilen J, and Puig P (2001) Impact of bottom trawling on water turbidity and muddy sediment of an unfished continental shelf. Limnology and Oceanography 46: 1100--1110.

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE P. K. Taylor, Southampton Oceanography Centre, Southampton, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1187–1195, & 2001, Elsevier Ltd.

Introduction The maintenance of the earth’s climate depends on a balance between the absorption of heat from the sun and the loss of heat through radiative cooling to space. For each 100 W of the sun’s radiative energy entering the atmosphere nearly 40 W is absorbed by the ocean – about twice that adsorbed in the atmosphere and three times that falling on land surfaces. Much of this oceanic heat is transferred back to the atmosphere by the local sea to air heat flux. The geographical variation of this atmospheric heating drives the weather systems and their associated winds. The wind transfers momentum to the sea causing waves and the wind-driven currents. Major ocean currents transport heat polewards and at higher latitudes the sea to air heat flux significantly ameliorates the climate. Thus the heat and momentum fluxes through the ocean surface form a crucial component of the earth’s climate system. The total heat transfer through the ocean surface, the net heat flux, is a combination of several components. The heat from the sun is the short-wave radiative flux (wavelength 0.3–3 mm). Around noon on a sunny day this flux may reach about 1000 W m2 but, when averaged over 24 h, a typical value is 100–300 W m2 varying with latitude and season. Part of this flux is reflected from the sea surface – about 6% depending on the solar elevation and the sea state. Most of the remaining short-wave flux is absorbed in the upper few meters of the ocean. In calm weather, with winds less than about 3 m s1, a shallow layer may be formed during the day in which the sea is warmed by a few degrees Celsius (a ‘diurnal thermocline’). However, under stronger winds or at night the absorbed heat becomes mixed down through several tens of metres. Thus, in contrast to land areas, the typical day to night variation in sea surface sea and air temperatures is small, o11C. Both the sea and the sky emit and absorb long-wave radiative energy (wavelength 3–50 mm).

Because, under most circumstances, the radiative temperature of the sky is colder than that of the sea, the downward long-wave flux is usually smaller than the upward flux. Hence the net long-wave flux acts to cool the surface, typically by 30–80 W m2 depending on cloud cover. The turbulent fluxes of sensible and latent heat also typically transfer heat from sea to air. The sensible heat flux is the transfer of heat caused by difference in temperature between the sea and the air. Over much of the ocean this flux cools the sea by perhaps 10–20 W m2. However, where cold wintertime continental air flows over warm ocean currents, for example the Gulf Stream region off the eastern seaboard of North America, the sensible heat flux may reach 100 W m2. Conversely warm winds blowing over a colder ocean region may result in a small sensible heat flux into the ocean – a frequent occurrence over the summertime North Pacific Ocean. The evaporation of water vapor from the sea surface causes the latent heat flux. This is the latent heat of vaporization which is carried by the water vapor and only released to warm the atmosphere when the vapor condenses to form clouds. Usually this flux is significantly greater than the sensible heat flux, being on average 100 W m2 or more over large areas of the ocean. Over regions such as the Gulf Stream latent heat fluxes of several hundred W m2 are observed. In foggy conditions with the air warmer than the sea, the latent heat flux can transfer heat from air to sea. In summertime over the infamous fog-shrouded Grand Banks off Newfoundland the mean monthly latent heat transfer is directed into the ocean, but this is an exceptional case.

Measuring the Fluxes The standard instruments for determining the radiative fluxes measure the voltage generated by a thermopile which is exposed to the incident radiation. Typically the incoming short-wave radiation is measured by a pyranometer which is mounted in gimbals for use on a ship or buoy (Figure 1). For better accuracy the direct and scattered components should be determined separately but, apart from at the Baseline Surface Radiation Network stations which are predominantly situated on land, at present this is rarely done. The reflected short-wave

(c) 2011 Elsevier Inc. All Rights Reserved.

105

106

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

Figure 1 A pyranometer used for measuring short-wave radiation. The thermopile is covered by two transparent domes. (Photograph courtesy of Southampton Oceanography Centre.)

radiation is normally determined from the sun’s elevation and lookup tables based on the results of previous experiments. The pyrgeometer used to determine the long-wave radiation is similar to the pyranometer but uses a coated dome to filter out, as far as possible, the effects of the short-wave heating. Because the air close to the sea surface is normally near to the sea temperature, the use of gimbals is less important. However, a clear sky view is required and a number of correction terms have to be calculated for the temperature of the dome and any short-wave leakage. Again, only the downward component is normally measured; the upwards component is calculated from knowledge of the sea temperature and emissivity of the sea surface. The turbulent fluxes may be measured in the nearsurface atmosphere using the eddy correlation method. If upward moving air in an eddy is on average warmer and moister than the downward moving air, then there is an upwards flux of sensible heat and water vapor and hence also an upward latent heat flux. Similarly the momentum flux, or wind stress, may be determined from the correlation between the horizontal and vertical wind fluctuations. Since a large range of eddy sizes may contribute to the flux, fast response sensors capable of sampling at 10 Hz or more must be exposed for periods of the order of 30 min for each flux determination. Three-component ultrasonic anemometers (Figure 2) are relatively robust and, by also determining the speed of sound, can provide an estimate of the sonic temperature flux, a function of the heat and moisture fluxes. The sensors used for determining the fluctuations in temperature and humidity have previously tended to be fragile and prone to contamination by salt particles which are ever-present in the marine atmosphere. However,

Figure 2 The sensing head of a three-component ultrasonic anemometer. The wind components are determined from the different times taken for sound pulses to travel in either direction between the six ceramic transducers. (Photograph courtesy of Southampton Oceanography Centre.)

improved sonic thermometry, and new techniques for water vapor measurement, such as microwave refractometry or differential infrared absorption instruments, are now becoming available. Despite these improvements in instrumentation, obtaining accurate eddy correlation measurements over the sea remains very difficult. If the instrumentation is mounted on a buoy or ship the six components of the wave-induced motion of the measurement platform must be measured and removed from the signal. The distortion both of the turbulence and the mean wind by ship, buoy or fixed tower must be minimized and, as far as possible, corrected for. Thus eddy correlation measurements are not routinely obtained over the ocean, rather they are used in special air–sea interaction experiments to calibrate other less direct methods of flux estimation. For example, in the inertial dissipation method, fluctuations of the wind, temperature, or humidity at a few Hertz are measured and related (through turbulence theory) to the fluxes. This method is less sensitive to flow distortion or platform motion, but relies on various assumptions about the formation and dissipation of turbulent quantities, which may not be valid under some conditions. It has been implemented

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

on a semi-routine basis on some research ships to increase the range of available flux data. The most commonly used method of flux estimation is variously referred to as the bulk (aerodynamic) formulae. These formulae relate the difference between the value of temperature, humidity or wind (‘x’ in [1]) at some measurement height, z, and the value assumed to exist at the sea surface – respectively the sea surface temperature, 98% saturation humidity (to allow for salinity effects), and zero wind (or any nonwind-induced water current). Thus the flux Fx of some quantity x is: Fx ¼ rUz Cxz ðxz  x0 Þ

½1

where r is the air density, and Uz the wind speed at the measurement height. While appearing intuitively correct (for example, blowing over a hot drink will cool it faster) these formulae can also be derived from turbulence theory. The value for the transfer coefficient, Cxz, characterizes both the surface roughness applicable to x and the relationship between Fx and the vertical profile of x. This varies with the atmospheric stability, which itself depends on the momentum, sensible heat, and water vapor fluxes, as well as the measurement height. Thus, although it may appear simple, Eqn [1] must be solved by iteration, initialized using the equivalent neutral value of Cxz at some standard height (normally 10 m), Cx10n. Typical neutral values (determined using eddy correlation or inertial dissipation data) are shown in Table 1. Many research problems remain. For example: CD10n is expected to depend on the state of development of the wave field, but can this be successfully characterized by the ratio of the predominant wave speed to the wind speed (the wave age), or by the wave height and steepness, or is a spectral representation of the wave field required? What are the effects of waves propagating from other regions (i.e., swell waves)? What is the behavior of CD10n in low wind speed conditions? Furthermore CE10n and CH10n are relatively poorly defined by the available experimental data, and Table 1

107

recent bulk algorithms have used theoretical models of the ocean surface (known as surface renewal theory) to predict these quantities from the momentum roughness length.

Sources of Flux Data Until recent years the only source of data for flux calculation routinely available from widespread regions of the world’s oceans was the weather reports from merchant ships. Organized as part of the World Weather Watch system of the World Meteorological Organisation, these ‘Voluntary Observing Ships (VOS)’ are asked to return coded weather messages at 00 00, 06 00, 12 00, and 18 00 h GMT daily, also recording the observation (with further details) in the ship’s weather logbook. The very basic set of instruments provided will normally include a barometer and a means of measuring air temperature and humidity – typically wet and dry bulb thermometers mounted in a hand swung sling psychrometer or a fixed, louvered ‘Stevenson’ screen. Sea temperature is obtained using a thermometer and an insulated bucket, or by reading the temperature gauge for the engine cooling water intake. Depending on which country recruited the VOS an anemometer and wind vane might be provided, or the ship’s officers might be asked to estimate the wind velocity from observations of the sea state using a tabulated ‘Beaufort scale’. Because of the problems of adequately siting an anemometer and maintaining its calibration, these visual estimates are not necessarily inferior to anemometer-based values. Thus the VOS weather reports include all the variables needed for calculating the turbulent fluxes using the bulk formulae. However, in many cases the accuracy of the data is limited both by the instrumentation and its siting. In particular, a large ship can induce significant changes in the local temperature and wind flow, since the VOS are not equipped with radiometers. The short-wave and long-wave fluxes must be estimated from the observer’s estimate

Typical values (with estimated uncertainties) for the transfer coefficientsa

Flux

Transfer coefficients

Typical values

Momentum

Drag coefficient CD10n(  1000) Stanton no., UH10n Dalton no., UE10n

¼ 0.61 (70.05) þ 0.063 (70.005) U10n (U10n43 m s1) ¼ 0.61 þ 0.57/U10no3 m s1 1.1 (70.2)  103 1.2 (70.1)  103

Sensible heat Latent heat

a Neither the low wind speed formula for CD10n, nor the wind speed below which it should be applied, are well defined by the available, very scattered, experimental data. It should be taken simply as an indication that, at low wind speeds, the surface roughness increases as the wind speed decreases due to the dominance of viscous effects.

(c) 2011 Elsevier Inc. All Rights Reserved.

108

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

of the cloud amount plus (as appropriate) the solar elevation, or the sea and air temperature and humidity. The unavoidable observational errors and the crude form of the radiative flux formulae imply that large numbers of reports are needed, and correction schemes must be applied, before satisfactory flux estimates can be obtained. While there are presently nearly 7000 VOS, the ships tend to be concentrated in the main shipping lanes. Thus whilst coverage in most of the North Atlantic and North Pacific is adequate to provide monthly mean flux values, elsewhere data is mainly restricted to relatively narrow, major trade routes. For most of the southern hemisphere the VOS data is only capable of providing useful values if averaged over several years, and reports from the Southern Ocean are very few indeed. These shortcomings of VOS-derived fluxes must be borne in mind when studying the flux distribution maps presented below. Satellite-borne sensors offer the potential to overcome these sampling problems. They are of two types, passive sensors which measure the radiation emitted from the sea surface and the intervening atmosphere at visible, infrared, or microwave frequencies, and active sensors which transmit microwave radiation and measure the returned signal. Unfortunately these remotely sensed data do not allow all of the flux components to be adequately estimated. Sea surface temperature has been routinely determined using visible and infrared radiometers since about 1980. Potential errors due, for example, to changes in atmospheric aerosols following volcanic eruptions, mean that these data must be continually checked against ship and buoy data. Algorithms have been developed to estimate the net surface short-wave radiation from top of the atmosphere values; those for estimating the net surface long wave are less successful. The surface wind velocity can be determined to good accuracy by active scatterometer sensors by measuring the microwave radiation backscattered from the sea surface. Unfortunately scatterometers are relatively costly to operate, since they demand significant power from the spacecraft and, to date, few have been flown. The determination of near-surface air temperature and humidity from satellite is hindered by the relatively coarse vertical resolution of the retrieved data. A problem is that the radiation emitted by the near-surface air is dominated by that originating from the sea surface. Statistically based algorithms for determining the near-surface humidity have been successfully demonstrated. More recently neural network techniques have been applied to retrieving both air temperature and humidity; however, at present there is no routinely available product. Thus the satellite flux products for which useful accuracy has

been demonstrated are presently limited to momentum, short-wave radiation, and latent heat flux. Numerical weather prediction (NWP) models (as used in weather forecasting centers) estimate values of the air–sea fluxes as a necessary part of their calculations. Since these models assimilate most of the available data from the World Weather Watch system, including satellite data, radiosonde profiles, and surface observations, it might be expected that NWP models represent the best source of flux data. However, there are other problems. The vertical resolution of these models is relatively poor and many of the near-surface processes which affect the fluxes have to be represented in terms of larger-scale parameters. Improvements to these models are normally judged on the resulting quality of the weather forecasts, not on the accuracy of the surface fluxes; sometimes these may become worse. Indeed, the continual introduction of model changes results in time discontinuities in the output variables. This makes the determination of interannual variations difficult. Because of this, NWP centres such as the European Centre for Medium Range Weather Forecasting (ECMWF) and the US National Centers for Environmental Prediction (NCEP) have reanalyzed the past weather and have gone back several decades. The surface fluxes from these reanalyses are receiving much study. Those presently available appear less accurate than fluxes derived from VOS data in regions where there are many VOS reports; in sparsely sampled regions the model fluxes may be more accurate. There are particular weaknesses in the shortwave radiation and latent heat fluxes. New reanalyses are planned and efforts are being made to improve the flux estimates; eventually these reanalyses will provide the best source of flux data for many purposes.

Regional and Seasonal Variation of the Momentum Flux The main features of the wind regimes over the global oceans have long been recognized and descriptions are available in many books on marine meteorology (see Further Reading). The major features of the wind stress variability derived from ship observations from the period 1980–93 will be summarized here, using plots for January and July to illustrate the seasonal variation. The distribution of the heat fluxes will be discussed in the next section. In northern hemisphere winter (Figure 3A) large wind stresses due to the strong midlatitude westerly winds are obvious in the North Atlantic and the North Pacific west of Japan. To the south of these

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

109

Figure 3 Monthly vector mean wind stress (N m2) for (A) January and (B) July calculated from Voluntary Observing Ship weather reports for the period 1980–93. (Adapted with permission from Josey SA, Kent EC and Taylor PK (1998) The Southampton Oceanography Centre (SOC) Ocean–Atmosphere Heat, Momentum and Freshwater Flux Atlas. SOC Report no. 6.)

regions the extratropical high pressure zones result in low wind stress values, south of these is the northeast trade wind belt. The Inter-Tropical Convergence Zone (ITCZ) with very light winds is close to the equator in both oceans. In the summertime southern hemisphere the south-east trade wind belt is less well marked. The extratropical high pressure regions are extensive but, despite it being summer, high winds and significant wind stress exist in the midlatitude

southern ocean. The north-east monsoon dominates the wind patterns in the Indian Ocean and the South China Sea (where it is particularly strong). The ITCZ is a diffuse region south of the equator with relatively strong south-east trade winds in the eastern Indian Ocean. In northern hemisphere summer (Figure 3B) the wind stresses in the midlatitude westerlies are very much decreased. Both the north-east and the

(c) 2011 Elsevier Inc. All Rights Reserved.

110

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

south-east trade wind zones are evident respectively to the north and south of the ITCZ. This is predominantly north of the equator. The south-east trades are particularly strong in the Indian Ocean and feed into a very strong south-westerly monsoon flow in the Arabian Sea. The ship data indicate very strong winds in the Southern Ocean south west of Australia. These are also evident in satellite scatterometer data, which suggest that the winds in the Pacific sector of the Southern Ocean, while still strong, are somewhat less than those in the Indian Ocean sector. In contrast the ship data appear to show very light winds. The reason is that in wintertime there are practically no VOS observations in the far south Pacific. The analysis technique used to fill in the data gaps has, for want of other information, spread the light winds of the extratropical high pressure region farther south than is realistic; a good example of the care needed in interpreting the flux maps available in many atlases.

Regional and Seasonal Variation of the Heat Fluxes The global distribution of the mean annual net heat flux is shown in Figure 4A. The accuracy and method of determination of such flux distributions will be discussed further below, here they will be used to give a qualitative description. Averaged over the year the ocean is heated in equatorial regions and loses heat in higher latitudes, particularly in the North Atlantic. However, this mean distribution is somewhat misleading, as the plots for January (Figure 4B) and July (Figure 4C) illustrate. The ocean loses heat over most of the extratropical winter hemisphere and gains heat in the extratropical summer hemisphere and in the tropics throughout the year. The relative magnitude of the individual flux components is illustrated in Figure 5 for three representative sites in the North Atlantic Ocean. At the Gulf Stream site (Figure 5A) the large cooling in winter dominates the incoming solar radiation in the annual mean. However, even at this site the mean monthly short-wave flux in summer is greater than the cooling. Indeed the effect of the longer daylight periods increases the mean short-wave radiation to values similar to or larger than those observed in equatorial regions (Figure 5C). The midlatitude site (Figure 5B) is typical of large areas of the ocean. The ocean cools in winter and warms in summer, in each case by around 100 W m2. The annual mean flux is small – around 10 W m2 – but cannot be neglected because of the very large ocean areas involved. At this site, and generally over the ocean, this annual balance is between the sum of the latent heat flux and net

long-wave flux which cool the ocean, and the net short-wave heating. Only in very cold air flows, as over the Gulf Stream in winter, is the sensible heat flux significant. As regards the interannual variation of the surface fluxes, the major large-scale feature over the global ocean is the El Nin˜o-Southern Oscillation system in the equatorial Pacific Ocean. The changes in the net heat flux under El Nin˜o conditions are around 40 W m2 in the eastern equatorial Pacific. For extratropical and midlatitude regions the interannual variability of the summertime net heat flux is typically about 20–30 W m2, being dominated by the variations in latent heat flux. In winter the typical variability increases to about 30–40 W m2, although in particular areas (such as over the Gulf Stream) variations of up to 100 W m2 can occur. The major spatial pattern of interannual variability in the North Atlantic is known as the North Atlantic Oscillation (NAO). This represents a measure of the degree to which mobile depressions, or alternatively near stationary high pressure systems, occur in the midlatitude westerly zone.

Accuracy of Flux Estimates It has been shown that, although the individual flux components are of the order of hundreds of W m2, the net heat flux and its interannual variability over much of the world ocean is around tens of W m2. Furthermore it can be shown that a flux of 10 W m2 over 1 year would, if stored in the top 500 m of the ocean, heat that entire layer by about 0.151C. Temperature changes on a decadal time scale are at most a few tenths of a degree, so the global mean budget must balance to better than a few W m2. For these various reasons there is a need to measure the flux components, which vary on many time and space scales, to an accuracy of a few W m2. Given the available data sources and methods of determining the fluxes described in the previous sections, it is not surprising that this level of accuracy cannot be achieved at present. To take an example, in calculating the flux maps shown in Figure 4 from VOS data many corrections were applied to the VOS observations to attempt to remove biases caused by the methods of observation. For example, air temperature measurements were corrected for the heat island caused by the ship heating up in sunny, low wind conditions. The wind speeds were adjusted depending on the anemometer heights on different ships. Corrections were applied to sea temperatures calculated from engine room intake data. Despite these and other corrections, the

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

111

Figure 4 Variation of the net heat flux over the ocean, positive values indicate heat entering the ocean: (A) annual mean, (B) January monthly mean, (C) July monthly mean. (Adapted with permission from Josey SA, Kent EC and Taylor PK (1998) The Southampton Oceanography Centre (SOC) Ocean–Atmosphere Heat, Momentum and Freshwater Flux Atlas. SOC Report no. 6.)

global annual mean flux showed about 30 W m2 excess heating of the ocean. Previous climatologies calculated from ship data had shown similar biases and the fluxes had been adjusted to remove the bias,

or to make the fluxes compatible with estimates of the meridional heat transport in the ocean. However, comparison of the unadjusted flux data with accurate data from air–sea interaction buoys showed good

(c) 2011 Elsevier Inc. All Rights Reserved.

112

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

600

600

500

500

400

400

300

300

200

200

100

100

0 Annual (A)

January 40°N, 60°W

July

0 Annual (B)

600

January

July

40°N, 20°W SW radn. LW radn.

500

Latent Heat Sensible Heat

400 300 200 100 0 Annual (C)

January 0°N, 20°W

July

Figure 5 Mean heat fluxes at three typical sites in the North Atlantic for the annual mean, and the January and July monthly means. In each case the left-hand column shows the fluxes which act to cool the ocean while the right-hand column shows the solar heating. (A) Gulf Stream site (401N, 601W), (B) midlatitude site (401N, 201W), (C) equatorial site (01N, 201W).

agreement between the two. This suggests that adjusting the fluxes globally is not correct and that regional flux adjustments are required; however, the exact form of these corrections is presently not shown. In the future, computer models are expected to provide a major advance in flux estimation. Recently coupled numerical models of the ocean and of the atmosphere have been run for many simulated years during which the modeled climate has not drifted. This suggests that the air–sea fluxes calculated by the models are in balance with the simulated oceanic and atmospheric heat transports. However, it does not imply that the presently estimated flux values are realistic. Errors in the short-wave and latent heat fluxes may compensate one another; indeed in a typical simulation the sea surface temperature stabilized to a value which was, over large regions of the ocean, a few degrees different from that which is observed. Nevertheless the estimation of flux values using climate or NWP models is a rapidly developing field and improvements will doubtless occur in the next few years. There will be a continued need for in situ and satellite data for assimilation into the models and for model development and verification. However, it seems very likely that in future the most accurate routine source of the air–sea flux data will

be from numerical models of the coupled ocean–atmosphere system.

See also El Nin˜o Southern Oscillation (ENSO). El Nin˜o Southern Oscillation (ENSO) Models. Evaporation and Humidity. Heat Transport and Climate. IR Radiometers. North Atlantic Oscillation (NAO). Satellite Passive-Microwave Measurements of Sea Ice. Satellite Remote Sensing of Sea Surface Temperatures. Sensors for Mean Meteorology. Turbulence Sensors. Upper Ocean Heat and Freshwater Budgets. Wave Energy. Wave Generation by Wind. Wind- and Buoyancy-Forced Upper Ocean.Wind Driven Circulation

Further Reading Browning KA and Gurney RJ (eds.) (1999) Global Energy and Water Cycles. Cambridge: Cambridge University Press. Dobson F, Hasse L, and Davis R (eds.) (1980) Air–Sea Interaction, Instruments and Methods. New York: Plenum Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT AND MOMENTUM FLUXES AT THE SEA SURFACE

Kraus EB and Businger JA (1994) Atmosphere–Ocean Interaction, 2nd edn. New York: Oxford University Press. Meteorological Office (1978) Meteorology for Mariners, 3rd edn. London: HMSO.

113

Stull RB (1988) An Introduction to Boundary Layer Meteorology. Dordrecht: Kluwer Academic. Wells N (1997) The Atmosphere and Ocean: A Physical Introduction, 2nd edn. London: Taylor and Francis.

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT TRANSPORT AND CLIMATE H. L. Bryden, University of Southampton, Southampton, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction: The Global Heat Budget The Earth receives energy from the sun (Figure 1) principally in the form of short-wave energy (sunlight). The amount of solar radiation is quantified by the ‘solar constant’ which satellite radiometers have measured since about 1985 to have a mean value of 1366 W m  2, an 11-year sunspot cycle of amplitude 1.5 W m  2, and a maximum at maximum sunspot activity. A fraction of the sunlight is reflected directly back into space and this fraction is termed the albedo. Brighter areas like snow in polar regions have high albedo (0.8) reflecting most of the short-wave radiation back to space, while darker areas like the ocean have small albedo (0.05) and small reflection. Averaged over the Earth’s surface, the albedo is about 0.3. Overall, the net incoming short-wave radiation (incoming minus reflected) peaks in equatorial regions and decreases to small values in polar latitudes (Figure 2). The Earth radiates energy back to space in the form of long-wave, black-body radiation proportional to the fourth power of the absolute temperature at the top of the atmosphere. Because the temperature at the top of the atmosphere is relatively uniform with latitude varying only from 200 to 230 K, there is only a small latitudinal variation in outgoing radiation (Figure 1). Over a year, the net

incoming radiation equals the net outgoing radiation within our ability to measure the radiation, thus maintaining the overall heat balance of the Earth. For the radiation budget as a function of latitude, however, there is more incoming short-wave radiation at equatorial and tropical latitudes and more outgoing long-wave radiation at polar latitudes. To maintain this heating–cooling distribution, the atmosphere and ocean must transport energy poleward from the Tropics toward the Pole and the maximum poleward energy transport in each hemisphere occurs at a latitude of c. 351, where there is a change from net incoming to net outgoing radiation, and the maximum has a magnitude of about 5.8 PW (1 petawatt (PW) ¼ 1015 W). As recently as the mid-1990s, it was controversial whether the ocean or the atmosphere was responsible for the majority of the energy transport. Oceanographers found a maximum ocean heat transport of about 2 PW at 25–301 N, while meteorologists reported a maximum atmospheric transport of about 2.5 PW from the analysis of global radiosonde network observations. Thus, there was a missing petawatt of energy transport in the Northern Hemisphere for the combined ocean–atmosphere system. Recent analyses combining observations and models suggest that the atmosphere does carry the additional petawatt that observational analyses alone could not find Radiation balance Incoming radiation (1– )Q 300

300 E

E

Radiation (W m−2)

Outgoing radiation

Q

200

200

100

100

Net radiation R

0

0 90° N 60°

Q

30°

0°

30°

60° 90° S

−100

−100 R = (1−)Q −E Figure 1 Schematic of the Earth’s radiation budget. Q is incoming, short-wave solar radiation; aQ is reflected solar radiation where a is albedo, and E is outgoing, long-wave black-body radiation.

114

Figure 2 Latitudinal profiles of net incoming short-wave radiation, outgoing long-wave radiation, and the net radiative heating of the Earth. Note the latitudinal scale is stretched so that it is proportional to the surface area of the Earth.

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT TRANSPORT AND CLIMATE

due to a sparsity of radiosonde observations over the ocean. Now it is generally accepted that the atmosphere carries the majority of energy transport at 351 N, though some oceanographers point out that half of the maximum atmospheric transport is due to latent heat (water vapor) transport that can be considered to be a joint ocean–atmosphere process (Figure 3). Here we will concentrate on ocean heat transport, especially on the mechanisms of ocean heat transport in the Northern Hemisphere. Conservation of energy in the ocean is effectively expressed as conservation of heat, where heat is defined to be rCpY, where r is seawater density, Cp is specific heat of seawater at constant pressure, and Y is potential temperature, the temperature of a water parcel brought adiabatically (without heat exchange) to the sea surface from depth. Because rCp is nearly constant at about 4.08  106 J m  3 1C  1, heat conservation is essentially expressed as conservation of potential temperature. There are many subtleties to the definitions that are described in entries for density, potential temperature, and heat, but here we use traditional definitions of potential temperature, density, and specific heat based on the internationally recognized equation of state for seawater. Ocean heat transport is then the flow of heat through the ocean, rCpYv, where v is the water

115

velocity. Such definition depends on the temperature scale and has little meaning until it is considered for a given volume of the ocean. Because of mass conservation, there is no net mass transport into or out of a fixed ocean volume over long timescales (neglecting the relatively miniscule contributions from evaporation minus precipitation), so it is the ocean heat transport convergence that is meaningful, that is the amount of heat transport into the volume minus the heat transport out of the volume. Since mass is conserved, the heat transport convergence is proportional to the mass transport times the difference in temperature between the inflow and the outflow across the boundaries of the volume. For a complete latitude band like 251 N, where the Atlantic Ocean and Pacific Ocean volume north of 251 N can be considered to be an enclosed ocean, the heat transport convergence is commonly referred to as the ocean heat transport at 251 N. Individually the Atlantic and Pacific Oceans are nearly enclosed with only a small throughflow connecting them in Bering Straits, so the Atlantic heat transport at 251 N is commonly referred to, even though it is formally the heat transport convergence between 251 N and Bering Straits and similarly for Pacific heat transport at 251 N. Such definitions of heat transport are generally used throughout the Atlantic north of 301 S and the Pacific north of about 101 N, where each ocean

Components of global energy balance 6

Northward energy transport (PW)

4

Dry static atmosphere Latent Ocean Total

2

0

−2

−4

−6 90° S 60°

30° S

15° 0° 15° Latitude scaled for spherical area

30° N

60° 90° N

Figure 3 Components of atmosphere and ocean energy transports required to balance the net radiational heating/cooling of the Earth following Figure 2. The standard atmospheric energy transport is here divided into the dry static atmospheric energy transport and the latent heat transport. Latent heat transport is fundamentally a joint atmosphere–ocean process since the atmospheric water vapor transport is balanced by an opposing oceanic freshwater transport. The ocean heat transport is determined by integrating over the oceans the spatial distribution of atmosphere–surface heat exchange calculated by subtracting the atmospheric energy transport divergence from the radiative heating at the top of the atmosphere.

(c) 2011 Elsevier Inc. All Rights Reserved.

116

HEAT TRANSPORT AND CLIMATE

basin is closed except for the small Bering Straits transport.

Air–Sea Heat Exchange Conservation of heat means that any convergence of ocean heat transport is balanced by heat loss to the atmosphere (the amount of heat gain or loss through the ocean bottom is small in comparison to exchanges with the atmosphere). Thus charts of air–sea energy exchange are primary sources for our understanding of ocean heat transport, where it occurs and how big it is. Estimates of air–sea energy exchange have long been made based on measurements of cloud cover, surface air and water temperatures, wind speed, humidity, and bulk formula exchange coefficients to calculate the size of the radiational heating and latent heat cooling of the ocean surface and the sensible heat exchange between the ocean and atmosphere. Combining such shipboard observations on a global scale produces air–sea flux climatologies giving air–sea exchange by month and region. From such climatologies, the global distribution of annual averaged air–sea energy exchange (Figure 4) shows that the ocean gains heat over much of the equatorial and tropical regions and gives up large amounts of heat over the warm poleward flowing western boundary currents like the Gulf Stream, Kuroshio or Agulhas Current, and over open-water subpolar and polar regions. Ocean heat

transport convergence for any arbitrary ocean volume can technically be estimated by summing up the air–sea energy exchange over the surface area of the ocean volume. There is a problem, however, in that the air–sea energy exchange estimates have an uncertainty of about 30 W m  2. One way to determine this uncertainty is to sum the air–sea exchanges globally and to find that there is on average a heat gain by the ocean of 30 W m  2, in each of the two state-of-the-art air–sea exchange climatologies. It is of course possible to remove this bias, either uniformly, by region or by component (radiative, latent, or sensible heat exchange); despite careful comparison with buoy measurements with bulk formula estimates at several oceanic locations, there is no consensus on how to remove the 30 W m  2 uncertainty in air–sea energy exchange.

Distribution of Ocean Heat Transport Estimating ocean heat transport convergence from in situ oceanographic measurements is a second method to quantify the role of the ocean in the global heat balance. The advantage of this direct approach is that the mechanisms of ocean heat transport are examined rather than just their overall effect in terms of the air–sea exchange. It was first reliably applied at 251 N in the Atlantic, a latitude where the warm northward Gulf Stream flow of about 30 Sv (1 Sv ¼ 1  106 m3 s  1) through Florida Straits is regularly

Net heat flux (W m−2), annual 90

Latitude

75 60

150.00

45

125.00

30

100.00 75.00

15

50.00

0

25.00

−15

0.00 −25.00

−30

−50.00

−45

−75.00

−60

−100.00

−75

−125.00

−90 30° E

90° E

150° E

150° W Longitude

90° W

30° W

30° E

Figure 4 Global distribution of the annual mean net heat gain by the ocean as determined from bulk formula calculations. Positive values indicate a gain of heat by the oceans. SOC climatology – Josey SA, Kent EC, and Taylor PK (1999) New insights into the ocean heat budget closure problem from analysis of the SOC air–sea flux climatology. Journal of Climate 12: 2856–2880.

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT TRANSPORT AND CLIMATE

monitored by submarine telephone cable voltage (the varying flow of conducting seawater through a magnetic field produces varying voltage in the cable which is continuously measured). This is the latitude of the easterly trade winds whose westward wind stress generates a northward surface water transport of about 4 Sv in the Ekman layer. Because the Atlantic is closed to the north except for a small (1 Sv) flow from the Pacific through Bering Straits, a nearly equal amount of southward flow of 35 Sv must cross the mid-ocean section between the Bahamas and Africa. The vertical distribution of this southward return flow (and its temperature) has been estimated from transatlantic hydrographic sections where temperature and salinity profiles are used to derive geostrophic velocity profiles and the reference-level velocity for the geostrophic velocity profiles is set to make the southward geostrophic transport equal to the northward Gulf Stream plus Ekman transport. Analysis of three hydrographic sections along 251 N made in 1957, 1981, and 1992 suggests that the ocean heat transport is 1.3 PW. Estimates of the error in heat transport using this direct method are about 0.3 PW, implying that direct ocean heat transport estimates are more accurate than air–sea flux estimates for areas larger than 301 latitude  301 longitude (approximately, 30 W m  2  3100 km  3100 km ¼ 0.3 PW). In terms of mechanisms, the Gulf Stream carries warm water (transport-weighted temperature of 19 1C) northward and the northward surface wind-driven Ekman transport is also warm (25 1C). The compensating southward mid-ocean flow is about equally divided between a recirculating thermocline flow above 800 m depth with an average temperature close to the 19 1C Gulf Stream flow and a cold deep-water flow with an average temperature less than 3 1C. Overall then, the northward heat transport of 1.3 PW across 251 N is due to a net northward transport of about 18 Sv of warm upper layer waters balanced by a net southward transport of cold deep waters. Deep-water formation in the Nordic and Labrador Seas of the northern Atlantic connects the northward flowing warm waters and southward flowing cold deep waters. This overall vertical circulation is commonly called the Atlantic meridional overturning circulation. During the World Ocean Circulation Experiment 1990–99, transoceanic hydrographic sections were made and ocean heat transport estimated in each ocean basin. The Atlantic heat transport is northward from 401 S to 551 N; it is of the order 0.5 PW at the southern boundary of the Atlantic, increases in the tropical regions as heat is gained at the sea surface, reaches a maximum near 251 N, and then decreases northward as the ocean gives up heat to the

117

atmosphere (Figure 5). At all latitudes, the northward heat transport is due to the meridional overturning circulation where the cold deep-water transport of c. 15–20 Sv persists through the Atlantic and the compensating northward upper water flow changes temperature, warming in tropical regions and then cooling in northern subtropical and subpolar latitudes. The North Pacific, north of 101 N where the basin is essentially closed, exhibits a similar pattern of northward heat transport, but the heat transport achieves a maximum of only 0.8 PW at 251 N, 50% less than the North Atlantic despite the Pacific being more than twice as wide as the Atlantic. The Pacific is different from the Atlantic in that there is no substantial deep meridional overturning circulation, for there is no deep-water formation in the North Pacific. The Pacific heat transport is due to a horizontal circulation where warm waters flow northward in the Kuroshio western boundary current and then recirculate southward over the vast zonal extent of the Pacific, all at depths shallower than 1000 m. The Kuroshio has about the same size and temperature structure as the Gulf Stream, but the zonal temperature distribution in the Pacific exhibits much colder upper water temperatures in mid-ocean and particularly along the eastern boundary of western North America than does the Atlantic along Europe and Africa. Thus, the northward heat transport in the North Pacific is due to the horizontal upper water circulation where warm water flows northward in the Kuroshio, loses heat to the atmosphere at latitudes south of 501 N, and then recirculates southward at colder temperatures in the mid- and eastern Pacific. South of 101 N in the Pacific and throughout the Indian Ocean, ocean heat transports are somewhat ambiguous because of the throughflow from the Pacific to the Indian Ocean through the Indonesian archipelago. The throughflow transport is not yet well defined but there is a substantial observational effort now underway to quantify its mass transport and associated temperature structure. Because mass is not conserved for individual Pacific or Indian zonal sections south of 101 N, heat transports across such zonal sections in the literature generally depend on the temperature scale used as well as on the size of the throughflow assumed. These fluxes should properly be called temperature transports and to be interpreted properly should include both a net mass transport across each section and a transport-weighted temperature. Reported northward temperature transports in the South Pacific are mainly due to a net northward transport (to balance the throughflow) times temperature in degrees Celsius. Similarly, southward temperature transports in the south Indian Ocean

(c) 2011 Elsevier Inc. All Rights Reserved.

118

HEAT TRANSPORT AND CLIMATE

60° N 0.6

30° N

Heat A2

0 P1

1 PW Uncertainty +0.5 PW −0.5 PW

1.3 A5

P3

0.5

A6

0.7

0° 1

A7

0.9

A8

0.8

A10 A11

1.4

1.6

0.2 JADE89

A9

0.3

30° S

I2 I3

P21E

1.8

I4

I10

0.6

1.5

I5

P6

0.7

1.3

60° S

1.7

0.9

1.4

0.7 A21 A12

60° W

0° W

I9S

I6

60° W

P12

120° W

180° W

120° W

Figure 5 Heat transport and temperature transport across hydrographic sections taken during the World Ocean Circulation Experiment. For the Atlantic north of 401 S and the Pacific north of 101 N, heat transport values are presented where there is no net mass transport across the sections. For the remaining sections, temperature transports are presented where there is a net mass transport across the section and the temperature transport includes the net transport multiplied by temperature in degrees Celsius.

tend to be large because they include a substantial net southward mass transport (to balance the throughflow) times an average temperature. Careful ocean heat transport divergence estimates for the Indian Ocean taking into account the throughflow transport and temperature generally suggest that the Indian Ocean gains heat from the atmosphere between 0.4 and 1.2 PW depending on the size and temperature of the throughflow. For the South Pacific, the ocean heat transport convergence or divergence is ambiguous for a normal range of throughflow transport, so it is uncertain whether the South Pacific as a whole gains heat or loses heat to the atmosphere. It is possible to combine the South Pacific and Indian sections to enclose a confined ocean volume north of say 321 S. For such combinations where mass is conserved, the Indo-Pacific Ocean heat transport at 321 S is meaningful and estimates are that it is southward with a magnitude of 0.4–1.2 PW. For the complete latitude band at about 301 S combining Indo-Pacific and Atlantic Ocean heat transports, the ocean heat transport is southward or

poleward but with only a maximum value of about 0.5 PW since the northward Atlantic heat transport partially cancels the southward Indo-Pacific heat transport. Thus, it appears that the ocean contributes much less than the atmosphere to the poleward heat transport in the Southern Hemisphere required to balance the Earth’s radiation budget.

Eddy Heat Transport In addition to heat transport by the steady ocean circulation, temporally and spatially varying currents with scales of 10–100 days and 10–100 km, which are called mesoscale features or eddies, can also transport heat. Correlations between time-varying velocity and temperature fluctuations can transport heat, rCp /Y0 v0 S (where primes denote fluctuations and angular brackets indicate time averages), even when there is no net velocity or mass transport, that is, /v0 S is zero. Such eddy heat transport can be substantial in regions where there are strong lateral temperature gradients like the Gulf Stream or Kuroshio extensions separating

(c) 2011 Elsevier Inc. All Rights Reserved.

HEAT TRANSPORT AND CLIMATE

the subtropical from the subpolar gyre. In the zonally unbounded Southern Ocean where there are several thermal fronts associated with the Antarctic Circumpolar Current, eddy heat transport is the dominant mechanism for transporting heat poleward. Here eddy motions are observed to have colder temperature when they are flowing northward and warmer temperature when following southward though there is no average flow over the eddy scales. The resulting eddy heat transport, typically of the order 4  103 W m  2 for /Y0 v0 S ¼ 0.1 1C cm s  1, is southward, downgradient from high temperature on the northern side toward cold temperature on the southern side of the front, and this downgradient heat flux is a signature of the baroclinic instability process by which eddies form and grow on the potential energy stored in the large-scale lateral temperature distribution. For the 3500 m depth and 20 000 km zonal extent of the Southern Ocean, this poleward eddy heat flux amounts to 0.3 PW across a latitude of 601 S. Individual eddies or rings of isolated water mass properties may also transport heat. For example, Agulhas rings formed with a core of Indian Ocean water properties in the retroflection area south of Africa are observed to transit across the South Atlantic. These eddies have relatively warm water cores and their heat transport is often estimated by multiplying their heat content anomaly by an estimated number of how many such rings are formed each year. Such calculation is somewhat ambiguous because it is not clear how the mass is returned and what its temperature is. Similar estimates have been made with Gulf Stream and Kuroshio rings, both warm core and cold core, and with meddies formed from the outflow of Mediterranean water. While individual rings are impressive, it is not yet clear whether they carry a significant amount of heat compared with the annual averaged air–sea exchange in any region.

Future Developments There is a third method, the residual method, for estimating ocean heat transport that takes the difference between the energy transport required to maintain the Earth’s radiation budget and the atmospheric energy transport to define the ocean heat transport as a residual. In its original implementation, the residual method could only be applied to estimate zonally averaged ocean heat transport into the polar cap north of any given latitude because atmosphere energy transport could only be determined for a complete zonal integral. In a recent development, the atmospheric energy transport divergence is estimated on a grid point basis from a globally consistent model analysis that assimilates atmospheric observations and

119

radiation variables. From such analysis, the surface energy flux can be estimated at each grid point as the difference between radiation input at the top of the atmosphere and atmospheric transport divergence. Presently, the radiation input can only be accurately estimated for the intensive period of Earth Radiation Budget Experiment from 1985 to 1989. Imposing constraints that the annual averaged surface flux over land should be zero and that the net global air–sea flux should be zero leads to realistic charts of air–sea heat exchange from which ocean heat transport divergence can be estimated. Careful comparison of such ocean heat transport convergence with existing air–sea flux climatologies and with direct estimates of ocean heat transport convergence has not yet been done. There are many outstanding questions on how ocean heat transport will change under changing climate conditions. As atmospheric CO2 has increased, the ocean has warmed up by 14  1022 J over the past 40 years. Such warming represents a heat flux of only 0.1 PW or 0.3 W m  2 averaged over the ocean surface area, so it is unlikely to be detectable in local estimates of air–sea exchange that have uncertainties of 30 W m  2 or in direct estimates of ocean heat transport convergence with uncertainties of 0.3 PW. Instead, local estimates of ocean heat content change over decadal timescales provide a sensitive estimate of how the difference between ocean heat transport convergence and air– sea exchange is changing in a changing climate. There may be changes in ocean circulation that will lead to measurable changes in air–sea heat exchange and ocean heat transport. For example, most coupled climate models predict that the Atlantic meridional overturning circulation will slow down by order of 50% over the next century as atmospheric CO2 increases. Because Atlantic heat transport is presently closely related to the strength of the meridional overturning circulation, Atlantic ocean heat transport could reduce measurably. In addition, the absence of a meridional overturning circulation in coupled climate models leads to much colder (10 1C lower) temperatures in the northern Atlantic that greatly reduce the amount of heat given up by the ocean to the atmosphere in northern latitudes. In fact, there has been a recent suggestion that the Atlantic meridional circulation decreased by 30% since 1992. The heat transport decreased by only about 15% from 1.3 to 1.1 PW, however, as the horizontal gyre circulation increased to transport more heat northward. Thus, the Atlantic heat transport may not reduce proportionately with a decreased meridional overturning circulation, because in the absence of a meridional overturning circulation it is possible that the Atlantic will become more like the Pacific with a horizontal

(c) 2011 Elsevier Inc. All Rights Reserved.

120

HEAT TRANSPORT AND CLIMATE

gyre circulation that still transports a substantial amount of heat northward. If Atlantic ocean heat transport reduces under changing climate, will the atmospheric energy transport act to compensate with larger northward energy transport? Presently, the Atlantic Ocean circulation transports more than 20% of the maximum energy transport required to balance the Earth’s radiation budget. The hypothesized Bjerkenes compensation mechanism suggests that a reduction in ocean heat transport will be compensated by increased atmospheric energy transport. For extratropical latitudes, atmospheric transport is primarily effected by eddies, cyclones, and anticyclones. Will a reduction in ocean heat transport then be accompanied by increased mid-latitude storminess and increased atmospheric energy transport? Or will the overall radiation budget for the Earth system be fundamentally altered? Clearly, it is of interest to monitor the changes in ocean circulation and heat transport, most importantly in the Atlantic where there are concerns that increasing atmospheric CO2 may lead relatively quickly to substantial changes in ocean circulation and heat transport. A program to monitor the Atlantic meridional overturning circulation and associated heat transport started in 2004 and such monitoring may provide the first evidence for changes in Atlantic ocean heat transport.

See also Agulhas Current. Ekman Transport and Pumping. Florida Current, Gulf Stream and Labrador Current. Indonesian Throughflow. Kuroshio and Oyashio Currents. Meddies and Sub-Surface Eddies. Mesoscale Eddies. Neutral Surfaces and the Equation of State.

Further Reading Bryden HL (1993) Ocean heat transport across 241 N latitude. In: McBean GA and Hantel M (eds.)

Geophysical Monograph Series, Vol. 75 Interactions between Global Climate Subsystems: The Legacy of Hann, pp. 65--75. Washington, DC: American Geophysical Union. Bryden HL and Beal LM (2001) Role of the Agulhas Current in Indian Ocean circulation and associated heat and freshwater fluxes. Deep-Sea Research I 48: 1821--1845. Bryden HL and Imawaki S (2001) Ocean heat transport. In: Siedler G, Church J, and Gould J (eds.) Ocean Circulation and Climate, pp. 455--474. New York: Academic Press. Bryden HL, Longworth HR, and Cunningham SA (2005) Slowing of the Atlantic meridional overturning circulation at 251 N. Nature 438: 655--657. Cunningham SA, Kanzow T, Rayner D, et al. (2007) Temporal variability of the Atlantic meridional overturning circulation at 26.51 N. Science 317: 935--938. Ganachaud A and Wunsch C (2000) Improved estimates of global ocean circulation, heat transport and mixing from hydrographic data. Nature 408: 453--457. Josey SA, Kent EC, and Taylor PK (1999) New insights into the ocean heat budget closure problem from analysis of the SOC air–sea flux climatology. Journal of Climate 12: 2856--2880. Lavı´n A, Bryden HL, and Parrilla G (1998) Meridional transport and heat flux variations in the subtropical North Atlantic. Global Atmosphere and Ocean System 6: 269--293. Levitus S, Antonov J, and Boyer T (2005) Warming of the World Ocean, 1955–2003. Geophysical Research Letters 32 (doi:10.1029/2004GL021592). Shaffrey L and Sutton R (2006) Bjerknes compensation and the decadal variability of the energy transports in a coupled climate model. Journal of Climate 19(7): 1167--1181. Trenberth KE and Caron JM (2001) Estimates of meridional atmosphere and ocean heat transports. Journal of Climate 14: 3433--3443. Trenberth KE, Caron JM, and Stepanaik DP (2001) The atmospheric energy budget and implications for surface fluxes and ocean heat transports. Climate Dynamics 17: 259--276. Vellinga M and Wood RA (2002) Global climatic impacts of a collapse of the Atlantic thermohaline circulation. Climatic Change 54: 251--267.

(c) 2011 Elsevier Inc. All Rights Reserved.

HISTORY OF OCEAN SCIENCES H. M. Rozwadowski, Georgia Institute of Technology, Atlanta, Georgia, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1206–1210, & 2001, Elsevier Ltd.

Oceanography is the scientific study of the ocean, its inhabitants, and its physical and chemical conditions. Since its emergence as a recognized scientific discipline, oceanography has been characterized less by the intellectual cohesiveness of a traditional academic discipline than by multidisciplinary, often large-scale, investigation of a complex and forbidding environment. The term ‘oceanography’ was not applied until the 1880s, at which point it still competed with such alternatives as ‘thassalography’ and ‘oceanology’. In many countries, ‘oceanography’ now encompasses both biological and physical traditions, but this meaning is not universal. In Russia, for instance, it does not include biological sciences; the umbrella term remains ‘oceanology’. Before scientists studied the ocean as a geographic place with an integrated ecosystem, they addressed individual biological, physical, and chemical questions related to the sea. European expansion promoted the study of marine phenomena, initiating a lasting link between commercial interests and oceanography. Institutional interest in the sea by the Royal Society in London flourished briefly from 1660 to 1675, when members, including Robert Hooke and Robert Boyle, discussed marine research, developed instruments to make observations at sea, and conducted experiments on sea water to discover its physical and chemical properties. Sailors and travelers continued to make scattered observations at sea, working from the research plan and equipment established by the Society. In addition, tidal studies were prosecuted consistently from the late seventeenth century to well into the nineteenth. Following decades of quiescence, a renewal of interest in marine phenomena occurred in the mideighteenth century. As part of that century’s growth of astronomy, geophysics, chemistry, geology, and meteorology, investigators began making observations at sea as part of their scientific pursuit of other fields. Initially, observers were traveling gentlemen, naturalists, or eclipse-expedition astronomers; later they were scientific explorers. Beginning with the voyages of Captain James Cook in the last third of the century, British exploration became

characterized by attention to scientific observations. The amount of energy devoted to marine science depended on the level of interest of expedition leaders or individual members, but, in the last quarter of the century, the volume of experiments and observations made at sea increased. Virtually all ocean investigation during this time focused on temperature and salinity of water, reflecting the growth of chemical sciences. The concept of oceanic circulation sustained by differences in density, which was widely discussed in the early nineteenth century, emerged at this time. The study of waves became more important while marine natural history remained at a modest level. Early in the nineteenth century the emphasis on physical and chemical studies continued. Curtailed exploration due to the American Revolutionary War and the Napoleonic Wars slowed work in marine science until another period of rapid expansion between 1815 and 1830. Work then focused largely on currents and salinity, reflecting the fact that marine science was conducted on Arctic expeditions which searched for sperm whaling grounds and the Northwest Passage. Navigation and the pursuit of whales prompted Arctic explorers to investigate water temperature and pressure at depths. Enthusiastic individuals, most frequently ships’ captains such as William Scorseby, continued to carry on observation programs at sea, but the 1830s saw a decrease of interest in marine science by physical scientists, who turned to rival fields of discovery including meteorology and terrestrial magnetism. At this time, however, British zoologists directed particular attention to marine fauna, embarking on small boats and yachts to collect with modified oyster dredges. Pursuit of new species as well as living relatives of fossilized ones inspired naturalists to reach deeper and deeper into the sea. Beginning in the 1850s, both British and American hydrographic institutions began deep sea sounding experiments which were guided by the promise of submarine telegraphy, although initiated in support of whaling and navigational concerns. The decades after 1840 saw the gradual awareness by scientists, sailors, entrepreneurs, and governments that the ocean’s depths were commercially and intellectually important places to investigate. An unprecedented increase in popular awareness of the sea accompanied this trend. Prefaced by the rage for seaside vacations which was initiated by railroad access to beaches, cultural interest in the ocean was

(c) 2011 Elsevier Inc. All Rights Reserved.

121

122

HISTORY OF OCEAN SCIENCES

manifested by the popularity of marine natural history collecting and the new maritime novels as well as the vogue for yachting and the personal experiences of growing numbers of ocean travelers and emigrants. Submarine telegraphy and public interest in maritime issues helped scientists argue successfully for government funding of oceanographic voyages. Declining fisheries emphasized the need to know more about the biology of the seas. Until 1840, Britain led marine science, although important work was conducted by investigators in other countries, especially Scandinavian countries. In Norway, for example, G. O. Sars carried out important research for fisheries and also studied unusual creatures dredged from very deep water. American marine science began to threaten British dominance after mid-century. Matthew Fontaine Maury was the first investigator to compile wind, current, and whale charts to improve navigation and commerce. He also dispatched the first trans-Atlantic sounding voyages in the early 1850s. Under Alexander Dallas Bache’s tenure, the US Coast Survey began to include, alongside routine charting work, special studies such as detailed surveys of the Gulf Stream, microscopic examinations of bottom sediments, and dredging cruises with the renowned zoologist Louis Agassiz and his son Alexander. Spencer F. Baird, Secretary of the Smithsonian Institution and United States Fish Commissioner, oversaw American efforts to study marine fauna. In Britain, while physical work was undertaken by the Hydrographic Office and Admiralty exploring expeditions, marine biological science centered around the British Association for the Advancement of Science Dredging Committee from 1839 until the mid-1860s. After that time, a Royal Society–Admiralty partnership took the lead and dispatched a series of summer expeditions. These culminated in the famous fouryear circumnavigation of HMS Challenger (1872– 1876), the first expedition sent out with a mandate to study the world’s oceans. The United States, Britain, and other nations as well, quickly followed the example of the Naples Zoological Station and set up coastal marine biological laboratories in the 1880s and 1890s. In the last quarter of the nineteenth century, many nations sponsored oceanographic voyages modeled after that of Challenger. The United States, Russia, Germany, Norway, France, and Italy contributed to the effort to define the limits and contents of the oceans. Late in the nineteenth century, however, the Scandinavian countries promoted a new style of ocean science to replace the great voyage tradition. Mounting concerns about depleted fisheries inspired national efforts in many countries to study the

biology of fish species as well as their migration. Sweden initiated the formation of what became in 1902 the International Council for the Exploration of the Seas (ICES), which coordinated research undertaken by eight northern European member nations. Although not a member, the United States also continued active biological research. Victor Henson’s discovery of plankton and efforts to quantify its distribution and the subsequent realization of how to use physical oceanography to investigate the movements of fish populations gave oceanographers confidence with which to study the ocean as an undivided system. World War I disrupted the international community of oceanographers but provided the impetus for developing echo-sounding technology for submarine detection, which had been pioneered for ice detection partly in response to the Titanic disaster. By the late 1920s, echo-sounders revolutionized the study of underwater topography and helped scientists to recognize the rift valleys of midocean ridges, showing them to be active, unstable regions. Fishermen took up this technology for locating schools of pelagic fish; later scientists adapted echo-sounding to create fishery-independent surveying tools. Although government funding of oceanographic research dropped back almost to pre-war levels, the late 1920s saw the appearance of the first oceanographic institutions, sponsored mostly by foundations and private individuals. In the United States, the Scripps Institution changed its mission from biological research to oceanography in 1925 and, five years later, the Woods Hole Oceanographic Institution was established on the Atlantic. The availability of oceangoing vessels in the 1930s spurred development of American oceanography on a new scale. Fisheries research, which blossomed in Europe in the 1930s, included important theoretical advances which provided the foundation for later fish population dynamics modeling as well as open ocean research such as Sir Alister Hardy’s Continuous Plankton Recorder surveys and Johannes Schmidt’s publicly acclaimed search for the mid-Atlantic spawning ground of the European eel. Economic depression affected practical government science, such as fisheries research, as well as private projects, such as several of Schmidt’s voyages, which were sponsored by the Carlsberg Foundation. As a consequence, oceanographic work, which was particularly expensive, slowed down dramatically until preparations began for World War II. As with other sciences, World War II partnerships helped forge a new relationship between governments and oceanography. Oceanographic work during and after the war carried the imprint of wartime

(c) 2011 Elsevier Inc. All Rights Reserved.

HISTORY OF OCEAN SCIENCES

government support and policy in both its problem selection and its scale. Physical studies gained and maintained precedence over biological ones. The areas of inquiry promoted by wartime efforts related to submarine and antisubmarine tactics. New research began on underwater acoustics, and ocean floor sediment charts were compiled from existing data. Wave studies also received precedence for their value in predicting surf conditions for landings. After the war, oceanographers and academic institutions, newly accustomed to generous funding, learned to accept and even encourage government support. The foundation of the National Science Foundation, which became the major federal supporter of marine science in the USA, exemplified the new high levels of support for this technology-intensive science. Oceanography became characterized by large-scale, expensive research projects such as the 1960s deepsea drilling by proponents of the theory of seafloor spreading. Major international projects also became an integral part of ocean sciences, in conjunction with postwar internationalization of science and other sectors. The International Indian Ocean Expedition, for example, targeted the least well-studied of the world’s oceans. By 1950 physical oceanographers were aware of the wanderings of the Gulf Stream and pressed urgently for systematic exploration of ocean variability. Development of deep ocean mooring technology provided the opportunity to study currents and temperatures continuously. This work showed that mesoscale variability was important and was incorporated into general circulation models. Progress in postwar physical oceanography in general proceeded in step with technological improvements, particularly the development of computing power. Another example, that of radioactive tracers and ‘experiments’ of atomic bomb-testing at sea, led to an intensification of research into biogeochemical cycling from the 1950s. Through programs such as the Geochemical Ocean Sections Study (GEOSECS), tracer use permitted estimates of mixing and circulation times by the late 1960s and led to programs such as the World Ocean Circulation Experiment (WOCE). Circulation and diffusion studies took on renewed importance as public concern about pollution rose. The idea that it was necessary to understand how the ocean functioned in order to avoid inadvertently destroying it fanned development of biological and chemical oceanography from the late 1960s onward. Much effort was funneled into baseline surveying and monitoring programs, initially focusing on potential threats to human health through eating poisoned seafood but soon broadening to investigations of biological effects of

123

contaminants. The fast growth of mariculture from this period also encouraged these studies. Although many oceanographers became interested in open ocean research and theoretical modeling in the postwar period, some in Europe resuscitated efforts, begun within ICES in the 1930s, to integrate hydrographic knowledge with fisheries biology. This work encompassed the practical attempt to guide fishermen to catch more fish as well as the scientific project of trying to relate recruitment fluctuations to environmental factors. Parallel efforts in the United States bore fruit in the California Cooperative Oceanic Fisheries Investigations (CalCOFI), a cooperative government and academic program which investigated causes for the decline of the California coastal sardine fishery. CalCOFI, though, was an exceptional project; in general through the 1960s little significant intellectual exchange occurred between biological oceanography and fisheries science. Within biological oceanography, the legacy of E. Steeman-Nielsen’s radioactive carbon tracer method led to active work measuring primary productivity, with global estimates by J. H. Ryther and others in the late 1960s. Discovery of high biological diversity in the deep sea in the late 1960s led to work which was considerably enhanced by deep-diving submersibles, whose use dramatically changed our perception of the ocean’s depths. In the area of marine geology and geophysics, the two greatest achievements since World War II have been the development of the theory of plate tectonics and the deciphering of Earth’s paleoclimate record from deep-sea sediments. The first of these was set in motion by the Vine and Matthews hypothesis of 1963, which rocked the scientific community but was accepted with remarkable speed. Research submersibles played a central role in proving the theory of seafloor spreading. The 1979 discovery of hydrothermal vents, made in the process of these geophysical investigations, led to the surprise discovery of chemosynthetic ecosystems. Reconstruction of Earth’s paleoclimates, which became possible due to the large accumulation of data and the availability of deep sea cores from the ocean drilling projects, was fueled by growing scientific concerns about global climate change. This work began during the 1960s with C. David Keeling’s launch of time-series measurements of carbon dioxide, which provided the data documenting the increase of atmospheric carbon dioxide attributable to human activities and prompted inquiry into the magnitude of exchange of carbon dioxide between the atmosphere and oceans. Studies of the ocean’s role in global warming and weather production have grown in importance in recent years.

(c) 2011 Elsevier Inc. All Rights Reserved.

124

HISTORY OF OCEAN SCIENCES

Accompanying the rise of the environmental movement, oceanography’s multi-faceted attempt to understand the oceans as integrated biological and physical environments made it a compelling discipline to ecologically and environmentally minded scientists from the 1970s onward. Food web investigations became prominent, including those which used mesoscale enclosures such as one designed and run at Canada’s Pacific Biological Station in Nanaimo. Discovery of the microbial character of the pelagic food web added a new dimension to biological oceanography. New understanding of production, including a distinction between recycled nutrients and new production made by Dugdale and Goering in 1967, provided biological oceanography with the mathematical formalism for rigorous, quantitative modeling of ocean productivity and biogeochemical fluxes. Modeling capability encouraged a link to be forged between physical and biological oceanography, because new concepts required that physical processes of mixing and upwelling be integrated into ecosystem models dealing with new production, fish production, or export of organic material from the surface layer. The international Global Ocean Ecosystem Dynamics program and the recent resurgence of fisheries oceanography exemplify relatively successful efforts to bridge trophic levels, from plankton to marine mammals and sea birds, and thereby study the ecosystem as a whole.

Distribution. Primary Production Primary Production Processes.

Methods.

Further Reading Deacon M (1971) Scientists and the Sea, 1650–1900: A Study of Marine Science. London: Academic Press. Herdman W (1923) Founders of Oceanography and their Work: An Introduction to the Science of the Sea. London: Edward Arnold. Idyll CP (1969) Exploring the Ocean World: A History of Oceanography. New York: Thomas Y. Crowell. Lee AJ (1992) The Directorate of Fisheries Research: Its Origins and Development. London: Ministry of Agriculture, Fisheries, and Food. Mills EL (1989) Biological Oceanography: An Early History, 1870–1960. Ithaca: Cornell University Press. Schlee S (1973) The Edge of an Unfamiliar World: A History of Oceanography. New York: E.P. Dutton. Smith TD (1994) Scaling Fisheries: The Science of Measuring the Effects of Fishing, 1855–1955. Cambridge: Cambridge University Press. Went AEJ (1972) Seventy years agrowing: a history of the International Council for the Exploration of the Sea, 1902–1972. Rapport et Proce`s-Verbaux des Re´unions, Conseil Permanent International pour l’Exploration de la Mer 165: 1--252. National Research Council 50 (2000) Years of Ocean Discovery: National Science Foundation, 1950–2000. Washington, DC: National Academy Press, 2000.

See also Continuous Plankton Recorders. Fisheries and Climate. Indonesian Throughflow. International Organizations. Law of the Sea. Leeuwin Current. Maritime Archaeology. Primary Production

(c) 2011 Elsevier Inc. All Rights Reserved.

HOLOCENE CLIMATE VARIABILITY M. Maslin, C. Stickley, and V. Ettwein, University College London, London, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1210–1217, & 2001, Elsevier Ltd.

Introduction Until a few decades ago it was generally thought that significant large-scale global and regional climate changes occurred at a gradual pace within a timescale of many centuries or millennia. Climate change was assumed to be scarcely perceptible during a human lifetime. The tendency for climate to change abruptly has been one of the most surprising outcomes of the study of Earth history. In particular, paleoceanographic records demonstrate that our present interglacial, the Holocene (the last B10 000 years), has not been as climatically stable as first thought. It has been suggested that Holocene climate is dominated by millennial-scale variability, with some authorities suggesting that this is a 1500 year cyclicity. These pronounced Holocene climate changes can occur extremely rapidly, within a few centuries or even within a few decades, and involve regional-scale changes in mean annual temperature of several degrees Celsius. In addition, many of these Holocene climate changes are stepwise in nature and may be due to thresholds in the climate system. Holocene decadal-scale transitions would presumably have been quite noticeable to ancient civilizations. For instance, the emergence of crop agriculture in the Middle East corresponds very closely with a sudden warming event marking the beginning of the Holocene, and the widespread collapse of the first urban civilizations, such as the Old Kingdom in Egypt and the Akkadian Empire, coincided with a cooling event at around 4300 BP. In addition, paleo-records from the late Holocene demonstrate the possible influence of climate change on the collapse of the Mayan civilization (Classic Period), while Andean ice core records suggests that alternating wet and dry periods influenced the rise and fall of coastal and highland cultures of Ecuador and Peru. It would be foolhardy not to bear in mind such sudden stepwise climate transitions when considering the effects that humans might have upon the present climate system, via the rapid generation of greenhouse gases for instance. Judging by what we

have already learnt from Holocene records, it is not improbable that the system may gradually build up over hundreds of years to a ‘breaking point’ or threshold, after which some dramatic change in the system occurs over just a decade or two. At the threshold point, the climate system is in a delicate and somewhat critical state. It may take only a relatively minor ‘adjustment’ to trigger the transition and tip the system into abrupt change. This article summarizes the current paleoceanographic records of Holocene climate variability and the current theories for their causes. Concentrating on records that cover a significant portion of the Holocene. The discussion is limited to centennial– millennial-scale variations. Figure 1 illustrates the Holocene and its climate variability in context of the major global climatic changes that have occurred during the last 2.5 million years. Short-term variations such as the North Atlantic Oscillation and the El Nin˜o-Southern Oscillation will not be discussed.

The Importance of the Oceans and Holocene Paleoceanography Climate is created from the effects of differential latitudinal solar heating. Energy is constantly transferred from the equator (relatively hot) toward the poles (relatively cold). There are two transporters of such energy – the atmosphere and the oceans. The atmosphere responds to an internal or external change in a matter of days, months, or may be a few years. The oceans, however, have a longer response time. The surface ocean can change over months to a few years, but the deep ocean takes decades to centuries. From a physical point of view, in terms of volume, heat capacity and inertia, the deep ocean is the only viable contender for driving and sustaining long-term climate change on centennial to millennial timescales. Since the process of oceanic heat transfer largely regulates climate change on longer time-scales and historic records are too short to provide any record of the ocean system prior to human intervention, we turn to marine sediment archives to provide information about ocean-driven climate change. Such archives can often provide a continuous record on a variety of timescales. They are the primary means for the study and reconstruction of the stability and natural variability of the ocean system prior to anthropogenic influences.

(c) 2011 Elsevier Inc. All Rights Reserved.

125

126

HOLOCENE CLIMATE VARIABILITY

Future 0

+100 a

Quaternary Period Holocene Pleistocene

Recent 100 a

1 ka

6 NOW Little Ice Age

El Niño and the North Atlantic Oscillation

Dansgaard _ Oeschger cycles

10 Ma

2 3

Deglaciation and the Younger Dryas

Global warming

1 Ma

4

5

7 8

100 ka

10 ka

Plio-Miocene

Heinrich events and Dansgaard _ Oeschger cycles

Development of NADW Mid-Pleistocene revolution

Interglacials: 410 ka and 125 ka

1

Onset of Northern Hemisphere glaciation

Figure 1 Log time scale cartoon, illustrating the most important climate events in the Quaternary Period. (a, ka, Ma, refer to years ago.) (1) Onset of Northern Hemisphere Glaciation (3.2–2.5 Ma), ushering in the strong glacial–interglacial cycles which are characteristic of the Quaternary Period. (2) Mid-Pleistocene Revolution when the dominant periodicity of glacial–interglacial cycles switched from 41 000 y, to every 100 000 y. The external forcing of the climate did not change; thus, the internal climate feedback’s must have altered. (3) The two closest analogues to the present climate are the interglacial periods at 420 000 to 390 000 years ago (oxygen isotope stage 11) and 130 000 to 115 000 years ago (oxygen isotope stage 5e, also known as the Eemian). (4) Heinrich events and Dansgaard–Oeschger cycles (see text). (5) Deglaciation and the Younger Dryas events. (6) Holocene Dansgaard– Oeschger cycles (see text). (7) Little Ice Age (AD 1700), the most recent climate event that seems to have occurred throughout the Northern Hemisphere. (8) Anthropogenic global warming.

One advantage of marine sediments is that they can provide long, continuous records of Holocene climate at annual (sometimes intraannual) to centennial time-resolutions. However, there is commonly a trade-off between temporal and spatial resolution. Deep-ocean sediments usually represent a large spatial area, but sedimentation rates in the deep-ocean are on average between 0.002 and 0.005 cm y 1, with very productive areas producing a maximum of 0.02 cm y 1. This limits the temporal resolution to a maximum of 200 years per cm (50 y cm 1 for productive areas). Mixing by the process of bioturbation will reduce the resolution further. On continental shelves and in bays and other specialized sediment traps such as anoxic basins and fiords, sedimentation rates can exceed 1 cm y 1 providing temporal resolution of over 1 y cm 1. More local conditions are recorded in laminated marine sediments formed in anoxic environments, where biological activity can not disturb the sediments. For example, Pike and Kemp (1997) analysed annual and intraannual variability within the Gulf of California from laminated sediments containing a record of diatom-mat accumulation. Time series analysis highlighted a decadal-scale variability in mat-deposition associated with Pacific-wide changes in surface water circulation, suggested to be influenced by solar-cycles. In addition, anoxic sediments from the Mediterranean Ridge (ODP Site 971) reveal

seasonal-scale variability during the late Quaternary from a laminated diatom-ooze sapropel. Pearce et al., (1998) inferred changes in the monsoon-related nutrient input to the Mediterranean Basin via the Nile River as the main cause of the variations in the laminated sediments, which suggests a wide influence of changes in seasonality. Other potentially extremely high-resolution studies will come from Saanich Inlet, a Canadian fiord and Prydz Bay in Antarctica, sites recently drilled by the Ocean Drilling Program. However, the main drawback to such high-resolution locations is that they contain highly localized environmental and climate information. An additional problem associated mainly with continental margins is reworking, erosion, and redistribution of the sediment by mass density flows such as turbidities and slumps. Hence we concentrate on wider-scale records of Holocene climate change.

Holocene Climatic Variability Initial studies of the Greenland ice core records concluded the absence of major climate variation within the Holocene. This view is being progressively eroded, particular in the light of new information being obtained from marine sediments (Figure 2). Long-term trends indicate an early to mid-Holocene climatic optimum with a cooling trend in the late

(c) 2011 Elsevier Inc. All Rights Reserved.

127

HOLOCENE CLIMATE VARIABILITY

Northwest African Climate

North Atlantic Climate

ODP 658C

V29-191

Summer radiation _ Terrigenous (%) SST anomaly (˚C) (W m 2 ) 450

470

490

40

50

60

_8_

6 _4 _2 0

2

_1

(lithics g ) 1000 100 10

1

NEAP 15K Sortable silt mean size (10_63 µm) 11 12 13 14 15

0

Calendar age (ka BP)

_ 34

1

0

2

2

2

3

3

3 4

4

4 5

5

5

6

6

10000

7 8

7

5000

Calendar age (BP)

African Humid Period

10000

_ 35

Holocene

5000

δ18O (ppt) _ 36

LIA 1

LIA 1

GISP2

FASTER

8

COLD

WARM YD

(A)

(B)

(C)

SSTwarm SSTcold

(D)

(E)

(F)

Figure 2 Comparison of summer insolation for 651N with north-west African climate (deMenocal et al., 2000) and North Atlantic climate (V29-191, Bond et al., 1997; NEAP 15 K, Bianchi and McCave, 1999; GISP2, O’Brien et al., 1996). Note the similarity of events labeled 1 to 8 and the Little Ice Age (LIA).

Holocene. Superimposed on this trend are several distinct oscillations or climatic cooling steps that appear to be of widespread significance (see Figure 2), the most dramatic of which occurred 8200, 5500, and 4400 years ago and between AD 1200 and AD 1650. The event 8200 years ago is the most striking and abrupt, leading to widespread cool and dry conditions lasting perhaps 200 years, before a rapid return to climates warmer and generally moister than at present. This event is noticeably present in the GISP2 Greenland ice cores, from which it appears to have been about half as severe as the Younger Dryas to Holocene transition. Marine records of North African to Southern Asian climate suggest more arid conditions involving a failure of the summer monsoon rains. Cold and/or arid conditions also seem to have occurred in northernmost South America, eastern North America and parts of north-west Europe. In the middle Holocene approximately 5500–5300 years ago there was a sudden and widespread shift in precipitation, causing many regions to become either noticeably drier or moister. The dust and sea surface temperature records off north-west Africa show that the African Humid Period, when much of subtropical West Africa was vegetated, lasted from 14 800 to 5500 years ago and was followed by a 300year transition to much drier conditions (de Menocal et al., 2000). This shift also corresponds to the decline of the elm (Ulmus) in Europe about 5700, and of hemlock (Tsuga) in North America about 5300

years ago. Both vegetation changes were initially attributed to specific pathogen attacks, but it is now thought they may have been related to climate deterioration. The step to colder and drier conditions in the middle of an interglacial period is analogous to a similar change that is observed in records of the last interglacial period referred to as Marine Oxygen Isotope Stage 5e (Eemian). There is also evidence for a strong cold and arid event occurring about 4400 years ago across the North Atlantic, northern Africa, and southern Asia. This cold, and arid event coincides with the collapse of a large number of major urban civilizations, including the Old Kingdom in Egypt, the Akkadian Empire in Mesopotamia, the Early Bronze Age societies of Anatolia, Greece, Israel, the Indus Valley civilization in India, the Hilmand civilization in Afganistan, and the Hongshan culture of China.

Little Ice Age (LIA) The most recent Holocene cold event is the Little Ice Age (see Figures 2 and 3). This event really consists of two cold periods, the first of which followed the Medieval Warm Period (MWP) that ended B1000 years ago. This first cold period is often referred to as the Medieval Cold Period (MCP) or LIAb. The MCP played a role in extinguishing Norse colonies on Greenland and caused famine and mass migration in Europe. It started gradually before AD 1200 and ended at about AD 1650. This second cold period,

(c) 2011 Elsevier Inc. All Rights Reserved.

128

HOLOCENE CLIMATE VARIABILITY

Greenland temperature

Bermuda Rise SST

(˚C)

(˚C)

_ 32

_ 31

21

22

23

24

West African SST 25

_6

(˚C, anomaly) _4 _2 0

2

0 a 500

LIA

Age (BP)

b 1000 MWP 1500 Mid-Neoglacial

2000 2500

Figure 3 Comparison of Greenland temperatures, the Bermuda Rise sea surface temperatures (SST) (Keigwin, 1996), and west African and a sea surface temperature (deMenocal et al., 2000) for the last 2500 years. LIALittle Ice Age; MWPMedieval Warm Period. Solid triangles indicate radiocarbon dates.

may have been the most rapid and the largest change in the North Atlantic during the Holocene, as suggested from ice-core and deep-sea sediment records. The Little Ice Age events are characterized by a drop in temperature of 0.5–11C in Greenland and a sea surface temperature falls of 41C off the coast of west Africa and 21C off the Bermuda Rise (see Figure 3).

Holocene Dansgaard–Oeschger Cycles The above events are now regarded as part of the millennial-scale quasiperiodic climate changes characteristic of the Holocene (see Figure 2) and are thought to be similar to glacial Dansgaard–Oeschger (D/O) cycles. The periodicity of these Holocene D/O cycles is a subject of much debate. Initial analysis of the GISP2 Greenland ice core and North Atlantic sediment records revealed cycles at approximately the same 1500 (7500)-year rhythm as that found within the last glacial period. Subsequent analyses have also found a strong 1000-year cycle and a 550year cycle. These shorter cycles have also been recorded in the residual d14 C data derived from dendrochronologically calibrated bidecadal tree-ring measurements spanning the last 11 500 years. In general, during the coldest point of each of the millennial-scale cycles shown in Figure 2, surface water temperatures of the North Atlantic were about 2– 41C cooler than during the warmest part. One cautionary note is that Wunsch has suggested a more radical explanation for the pervasive 1500year cycle seen in both deep-sea and ice core, glacial and interglacial records. Wunsch suggests that the

extremely narrow spectral lines (less than two bandwidths) that have been found at about 1500 years in many paleo-records may be due to aliasing. The 1500-year peak appears precisely at the period predicted for a simple alias of the seasonal cycle sampled inadequately (under the Nyquist criterion) at integer multiples of the common year. When Wunsch removes this peak from the Greenland ice core data and deep-sea spectral records, the climate variability appears as expected to be a continuum process in the millennial band. This work suggests that finding a cyclicity of 1500 years in a dataset may not represent the true periodicity of the millennialscale events. The Holocene Dansgaard–Oeschger events are quasi periodic, with different and possibly stochastic influences.

Causes of Millennial Climate Fluctuation during the Holocene As we have already suggested, deep water circulation plays a key role in the regulation of global climate. In the North Atlantic, the north-east-trending Gulf Stream carries warm and relatively salty surface water from the Gulf of Mexico up to the Nordic seas. Upon reaching this region, the surface water has cooled sufficiently that it becomes dense enough to sink, forming the North Atlantic Deep Water (NADW). The ‘pull’ exerted by this dense sinking maintains the strength of the warm Gulf Stream, ensuring a current of warm tropical water into the North Atlantic that sends mild air masses across to the European continent. Formation of the NADW can be weakened by two processes. (1) The presence

(c) 2011 Elsevier Inc. All Rights Reserved.

HOLOCENE CLIMATE VARIABILITY

of huge ice sheets over North America and Europe changes the position of the atmospheric polar front, preventing the Gulf Stream from traveling so far north. This reduces the amount of cooling and the capacity of the surface water to sink. Such a reduction of formation occurred during the last glacial period. (2) The input of fresh water forms a lens of less-dense water, preventing sinking. If NADW formation is reduced, the weakening of the warm Gulf Stream causes colder conditions within the entire North Atlantic region and has a major impact on global climate. Bianchi and McCave, using deep-sea sediments from the North Atlantic, have shown that during the Holocene there have been regular reductions in the intensity of NADW (Figure 2E), which they link to the 1500-year D/O cycles identified by O’Brien and by Bond (1997). There are two possible causes for the millennial-scale changes observed in the intensity of the NADW: (1) instability in the North Atlantic region caused by varying

129

freshwater input into the surface waters; and (2) the ‘bipolar seesaw’. There are a number of possible reasons for the instability in the North Atlantic region caused by varying fresh water input into the surface waters:

• • • •

Internal instability of the Greenland ice sheet, causing increased meltwater in the Nordic Seas that reduces deep water formation. Cyclic changes in sea ice formation forced by solar variations. Increased precipitation in the Nordic Seas due to more northerly penetration of North Atlantic storm tracks. Changes in surface currents, allowing a larger import of fresher water from the Pacific, possibly due to reduction in sea ice in the Arctic Ocean.

The other possible cause for the millennial-scale changes is an extension of the suggested glacial intrinsic millennial-scale ‘bipolar seesaw’ to the

0.8

Northward heat flux (1015 W)

0.4 N. Atlantic heat piracy

0.0 (1) Modern

S. Atlantic counter heat piracy

(2) CLIMAP LGM (3) Cold tropics LGM (4) Heinrich event (HL)

_ 0.4

(5) Heinrich event (HB) (6) D/O event

_ 0.8 _ 40 N ˚

0˚N

40˚S Latitude

80˚S

120˚S

Figure 4 Atlantic Ocean poleward heat transport (positive indicates a northward movement) as given by the ocean circulation model (Seidov and Maslin, 1999) for the following scenarios: (1) present-day (warm interglacial) climate; (2) last glacial maximum (LGM) with generic CLIMAP data; (3) ‘Cold tropics’ LGM scenario; (4) a Heinrich-type event driven by the meltwater delivered by icebergs from decaying Laurentide ice sheet; (5) a Heinrich-type event driven by meltwater delivered by icebergs from decaying Barents Shelf ice sheet or Scandinavian ice sheet; (6) a general Holocene or glacial Dansgaard–Oeschger (D/O) meltwater confined to the Nordic Seas. Note that the total meridional heat transport can only be correctly mathematically computed in the cases of cyclic boundary conditions (as in Drake Passage for the global ocean) or between meridional boundaries, as in the Atlantic Ocean to the north of the tip of Africa. Therefore the northward heat transport in the Atlantic ocean is shown to the north of 301S only.

(c) 2011 Elsevier Inc. All Rights Reserved.

130

HOLOCENE CLIMATE VARIABILITY

Holocene. One of the most important finds in the study of glacial millennial-scale events is the apparent out-of-phase climate response of the two hemispheres seen in the ice core climate records from Greenland and Antarctica. It has been suggested that this bipolar seesaw can be explained by variations in the relative amount of deep water formation in the two hemispheres and heat piracy (Figure 4). This mechanism of altering dominance of the NADW and the Antarctic Bottom Water (AABW) can also be applied to the Holocene. The important difference with this theory is that the trigger for a sudden ‘switching off’ or a strong decrease in rate of deep water formation could occur either in the North Atlantic or in the Southern Ocean. AABW forms in a different way than NADW, in two general areas around the Antarctic continent: (1) near-shore at the shelf–ice, sea-ice interface and (2) in open ocean areas. In near-shore areas, coastal polynya are formed where katabatic winds push sea ice away from the shelf edge, creating further opportunity for sea ice formation. As ice forms, the surface water becomes saltier (owing to salt rejection by the ice) and colder (owing to loss of heat via latent heat of freezing). This density instability causes sinking of surface waters to form AABW, the coldest and

saltiest water in the world. AABW can also form in open-ocean Antarctic waters; particularly in the Weddell and Ross seas; AABW flows around Antarctica and penetrates the North Atlantic, flowing under the less dense NADW. It also flows into the Indian and Pacific Oceans, but the most significant gateway to deep ocean flow is in the south-west Pacific, where 40% of the world’s deep water enters the Pacific. Interestingly, Seidov and colleagues have shown that the Southern Ocean is twice as sensitive to meltwater input as is the North Atlantic, and that the Southern Ocean can no longer be seen as a passive player in global climate change. The bipolar seesaw model may also be self sustaining, with meltwater events in either hemisphere, triggering a train of climate changes that causes a meltwater event in the opposite hemisphere, thus switching the direction of heat piracy (Figure 5).

Conclusion The Holocene, or the last 10 000 years, was once thought to be climatically stable. Recent evidence, including that from marine sediments, have altered this view, showing that there are millennial-scale

Internal ice sheet oscillation ?

North Atlantic

Heinrich Events

Reduced NADW

Antarctic Warm Events Antartica melting or surging?

Southern Ocean warms up

S. Hemisphere heat piracy

1500 year D/O cycle

Meltwater

Warmer southern ocean

S. Hemisphere heat piracy

Reduced NADW

Southern Ocean

Laurentide ice sheet collapse

Sea level

Meltwater

Deep water oscillating system

Greenland / Iceland melting or surging?

Warmer N. Atlantic

N. Hemisphere heat piracy

Increased NADW

Reduced AABW

? Internal ice sheet oscillation

Figure 5 Possible deep water oscillatory system explaining the glacial and interglacial Dansgaard–Oeschger cycles. Additional loop demonstrates the possible link between interglacial Dansgaard–Oeschger cycles and Heinrich events.

(c) 2011 Elsevier Inc. All Rights Reserved.

HOLOCENE CLIMATE VARIABILITY

Solar cycles North Atlantic Oscillation

100

D/O cycles

Heinrich events

0.1

Obliquity

1

Precession

10 Eccentricity

Relative variance

Annual cycle

Millennial-scale quasiperiodicities?

0.01

El Nino_Southern Oscillation

Astronomical forcing

1000

131

? T h e _6

10

_5

10

_

‘ g a p ’ _

_

_1

10 4 10 3 10 2 10 _ Frequency (cycles y 1)

100

Figure 6 Spectrum of climate variance showing the climatic cycles for which we have good understanding and the ‘gap’ between hundreds and thousands of years for which we still do not have adequate understanding of the causes.

climate cycles throughout the Holocene. In fact we are still in a period of recovery from the last of these cycles, the Little Ice Age. It is still widely debated whether these cycles are quasiperiodic or have a regular cyclicity of 1500 years. It is also still widely debated whether these Holocene Dansgaard–Oeschger cycles are similar in time and characteristic to those observed during the last glacial period. A number of different theories have been put forward for the causes of these Holocene climate cycles, most suggesting variations in the deep water circulation system. One suggestion is that these cycles are caused by the oscillating relative dominance of North Atlantic Deep Water and Antarctic Bottom Water. Holocene climate variability still has no adequate explanation and falls in the ‘gap’ of our knowledge between Milankovitch forcing of ice ages and rapid variations such as El Nin˜o and the North Atlantic Oscillation (Figure 6). Future research is essential for understanding these climate cycles so that we can better predict the climate response to anthropogenic ‘global warming.’

See also Antarctic Circumpolar Current. Calcium Carbonates. Cenozoic Oceans – Carbon Cycle Models. Current Systems in the Atlantic Ocean. Deep Convection. Deep-Sea Drilling Methodology. Deep-Sea Drilling Results. MillennialScale Climate Variability. Ocean Circulation. Ocean Subduction. Ocean Margin Sediments. Paleoceanography. Sediment Chronologies. Water Types and Water Masses.

Further Reading Adams J, Maslin MA, and Thomas E (1999) Sudden climate transitions during the Quaternary. Progress in Physical Geography 23(1): 1--36. Alley RB and Clark PU (1999) The deglaciation of the Northern hemisphere: A global perspective. Annual Review of Earth and Planetary Science 27: 149--182. Bianchi GG and McCave IN (1999) Holocene periodicity in North Atlantic climate and deep-ocean flow south of Iceland. Nature 397: 515--523. Broecker W (1998) Paleocean circulation during the last deglaciation: a bipolar seesaw? Paleoceanography 13: 119--121. Bond G, Showers W, Cheseby M, et al. (1997) A pervasive millenial-scale cycle in North Atlantic Holocene and glacial climates. Science 278: 1257--1265. Chapman MR and Shackleton NJ (2000) Evidence of 550 year and 1000 years cyclicities in North Atlantic pattern during the Holocene. Holocene 10: 287--291. Cullen HM, et al. (2000) Climate change and the collapse of the Akkadian Empire: evidence from the deep sea. Geology 28: 379--382. Dansgaard W, Johnson SJ, and Clausen HB (1993) Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364: 218--220. deMenocal P, Ortiz J, Guilderson T, and Sarnthein M (2000) Coherent high- and low-latitude climate variability during the Holocene warm period. Science 288: 2198--2202. Keigwin LD (1996) The Little Ice Age and Medieval warm period in the Sargasso sea. Science 274: 1504--1507. Maslin MA, Seidov D, and Lowe J (2001) Synthesis of the nature and causes of sudden climate transitions during the Quaternary. AGU Monograph: Oceans and Rapid Past and Future Climate Changes: North–South Connections. Washington DC: American Geophysical. Union No. 119.

(c) 2011 Elsevier Inc. All Rights Reserved.

132

HOLOCENE CLIMATE VARIABILITY

O’Brien SR, Mayewski A, and Meeker LD (1996) Complexity of Holocene climate as reconstructed from a Greenland ice core. Science 270: 1962--1964. Pearce RB, Kemp AES, Koizumi I, Pike J, Cramp A, and Rowland SJ (1998) A lamina-scale, SEM-based study of a late quaternary diatom-ooze sapropel from the Mediterranean ridge, Site 971. In: Robertson AHF, Emeis K-C, Richter C, and Camerlenghi A (eds.) Proceedings of the Ocean Drilling Program, Scientific Results 160, 349--363. Peiser BJ (1998) Comparative analysis of late Holocene environmental and social upheaval: evidence for a disaster around 4000 BP. In: Peiser BJ, Palmer T, and Bailey M (eds.) Natural Catastrophes during Bronze Age Civilisations, 117--139.

Pike J and Kemp AES (1997) Early Holocene decadal-scale ocean variability recorded in Gulf of California laminated sediments. Paleoceanography 12: 227--238. Seidov D and Maslin M (1999) North Atlantic Deep Water circulation collapse during the Heinrich events. Geology 27: 23--26. Seidov D, Barron E, Haupt BJ, and Maslin MA (2001) Meltwater and the ocean conveyor: past, present and future of the ocean bi-polar seesaw. AGU Monograph: Oceans and Rapid Past and Future Climate Changes: North–South Connections. Washington DC: American Geophysical. Union No. 119.. Wunsch C (2000) On sharp spectral lines in the climate record and millennial peak. Paleoceanography 15: 417--424.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA R. A. Lutz, Rutgers University, New Brunswick, NJ, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1217–1227, & 2001, Elsevier Ltd.

On 17 February 1977, the deep-submergence vehicle Alvin descended 2500 m to the crest of the Galapagos Rift spreading center to first visit an ecosystem that would forever change our view of life in the deep sea. Cracks and crevices in the ocean floor were emanating fluids with temperatures up to 171C. None of the bizarre organisms clustering around these ‘hydrothermal vents’ had ever been encountered; they comprised new species, genera, families, superfamilies, and bizarre ‘tubeworms,’ up to 2 m long, which were subsequently placed in a new phylum (Vestimentifera) (Figure 1). Since the Galapagos Rift discovery, numerous hydrothermal vent sites have been found throughout the world’s oceans and over 500 new species have been described from these regions. Figure 2 depicts many of the major hydrothermal systems from which organisms have been collected to date. Fluids with temperatures as high as 4031C exit from polymetallic sulfide chimneys in many of these regions (Figure 3). Most ecosystems on earth ultimately rely on photosynthesis, with the energy source being solar. In marked contrast, deep-sea hydrothermal ecosystems

are based predominantly on chemosynthesis, with the energy source being geothermal. Many of the chemosynthetic microbes are fueled by hydrogen sulfide, which is present at low-temperature vents in concentrations up to several hundred micromoles per liter and at high-temperature vents in concentrations up to 100 milimoles per liter. These microbial organisms can be either ‘free-living’ (in the water or on the surface of various substrates) or symbiotic in association with certain vent organisms. The vestimentiferan tubeworms Riftia pachyptila and Tevnia jerichonana (Figures 1, 4, 5, and 6), for example, each have a specialized ‘tissue,’ known as the trophosome, which is comprised entirely of chemosynthetic bacteria. The tubeworms have no mouth, no digestive system, and no anus; in short, no opening to the external environment. Hydrogen sulfide diffuses across cell membranes and is transported via the hemoglobin-containing circulatory system to the trophosome, where it is utilized by the associated symbionts. Mussels (Bathymodiolus thermophilus) (Figures 7 and 8) and vesicomyid clams (Calyptogena magnifica) (Figure 9), common along both the Galapagos Rift and East Pacific Rise (EPR), represent two of the other dominant members of the vent megafauna that house chemosynthetic symbionts. In the case of each of these bivalves, the symbionts are associated with the gills and both species have modified feeding apparatuses relative to those of shallow-water related species (likely a result of their

Figure 1 A cluster of vestimentiferan tubeworms (Riftia pachyptila and Tevnia jerichonana), together with a zoarcid fish (Thermarces andersoni) and brachyuran crabs (Bythograea thermydron) inhabiting low-temperature hydrothermal vents at 91500 N along the East Pacific Rise (depth 2500 m).

(c) 2011 Elsevier Inc. All Rights Reserved.

133

134

HYDROTHERMAL VENT BIOTA

Lucky Strike

Southern Explorer Ridge (SEXP) Juan de Fuca Ridge (JDFR) Gorda Ridge (GORD) Sunrise Okinawa Trough

Mariana Trough Manus Basin

Rainbow Broken Spur

Guaymas Basin

TAG

Galapagos′ Rift (GAR)

Carlsberg Ridge Central Indian Ridge

Ri

se

e

dg

A R)

Ch

ile

Red Sea

Snake Pit Logatchev

ic Ridge (M

21° 13° Northern East 11° Pacific Rise (NEPR) 9° 7° Lau Basin 11° 14° 16° 17° 18° 20° 21° Southern East 22° Pacific Rise (SEPR) 31° 32°

Mid Atlant

North Fiji Basin

Menez Gwen

n dia

Ri

t In

as

e uth

Southeast Indian Ridge So

ge

arctic Rid

nt Pacific A

Figure 2 Deep-sea hydrothermal vent sites along mid-oceanic and back-arc ridge systems from which vent organisms have been collected to date. Numbers indicate approximate latitude of site along the East Pacific Rise.

predominant reliance on the associated symbionts for nutrition). Closely related mussels and clams within the families Mytilidae and Vesicomyidae are common constituents of the fauna associated with vents along mid-oceanic ridge and back-arc spreading centers (as well as at many cold-water hydrocarbon seeps) throughout the world’s oceans. All of

Figure 3 A ‘black smoker’ polymetallic sulfide chimney at 201500 N along the East Pacific Rise (depth 2615 m). Temperatures as high as 4031C have been recorded at the orifice of such edifices from which mineral-rich fluids emanate violently in many deep-sea hydrothermal systems.

these bivalve mollusks retrieved to date appear to contain thiotrophic (‘sulfur-feeding’) or methanotrophic (‘methane-feeding’) chemosynthetic symbionts. It should be mentioned that, while chemosynthetic bacteria play a critical role in food chains associated with many vent systems, there are numerous other microbes at the vents belonging to two other recognized kingdoms. Both eukaryotes and Archaea occupy specialized niches within various vent ecosystems and certain Archaea, which have been isolated from environments associated with hightemperature sulfide chimneys, have been reported to occupy the most primitive ‘node’ on the phylogenetic tree of life. Such reports have led to considerable speculation as to whether or not life itself may have originated in hydrothermal vent environments. At various high-temperature vent sites, numerous organisms colonize the sides of active sulfide edifices. Figure 4 depicts a sulfide edifice (named ‘Tubeworm Pillar’) that is 11 m high, the sides of which are covered with tubeworms (both R. pachyptila and T. jerichonana), crabs, zoarcid fish, and a spectrum of smaller, associated vent fauna. At other sites, the sides of ‘black smoker’ chimneys are frequently covered with the tubes of polychaetes, such as Alvinella pompejana, a ‘Pompeii’ worm named after the

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA

135

Figure 4 A portion of an 11 m high polymetallic sulfide edifice known as Tubeworm Pillar at 91500 N along the East Pacific Rise (depth 2500 m). The top two-thirds of the edifice is covered with vestimentiferan tubeworms, both Riftia pachyptila (larger organisms) and Tevnia jerichonana (smaller organisms), as well as numerous brachyuran crabs (Bythograea thermydron) and zoarcid fish (Thermarces andersoni).

submersible Alvin (Figure 10). This organism has been reported to routinely withstand long-term exposure to temperatures ranging from 21C to 351C and short-term exposure to temperatures in excess of 1001C, rendering this annelid perhaps the most eurythermal organism on the planet. In addition to the numerous species of bivalves, such as mussels and clams mentioned earlier, there are myriad other common mollusks at vents, including numerous gastropods, particularly archeogastropod limpets. Thirteen different gastropod species have been reported from vents along the narrow range of the East Pacific Rise from 91170 N to

91540 N. These organisms can achieve high population densities and are found on a wide variety of substrates including the tubes of Riftia pachyptila and the shells of Bathymodiolus thermophilus (Figures 8 and 11). Limpets graze on free-living microbes that coat the majority of surfaces associated with both low-temperature and high-temperature (e.g., sulfide chimney) vents. Over 100 species of mollusks have been collected to date from the mid-oceanic and back-arc spreading centers visited to date. Virtually all of the invertebrates inhabiting deepsea hydrothermal vents have planktonic larval stages. These free-swimming stages serve as the primary

Figure 5 Higher magnification of the side of Tubeworm Pillar (depicted in Figure 4), showing tubeworms (Riftia pachyptila and Tevnia jerichonana) and scavenging brachyuran crabs (Bythograea thermydron).

(c) 2011 Elsevier Inc. All Rights Reserved.

136

HYDROTHERMAL VENT BIOTA

Figure 6 Close-up image of a cluster of Tevnia jerichonana, together with a brachyuran crab (Bythograea thermydron) and a zoarcid fish (Thermarces andersoni). Note the ‘accordion-like’ morphology of the tube of this species of vestimentiferan tubeworm.

Figure 7 A dense population of mussels (Bathymodiolus thermophilus) inhabiting a low-temperature hydrothermal vent field along the East Pacific Rise. Associated fauna in the field of view include tubeworms (Riftia pachyptila), brachyuran crabs (Bythograea thermydron), zoarcid fish (Thermarces andersoni), and a galatheid crab (Munidopsis subsquamosa) (lower left).

means of dispersal between isolated vent systems, which can be separated by hundreds of kilometers. The larval stages can be either planktotrophic (feeding within the water column) or lecithotrophic (utilizing yolk reserves for nutrition). In either case, it appears that the early life history stages of the vast majority of vent organisms are capable of staying in the water column for considerable lengths of time (likely months in the case of many vent species). Such extended planktonic durations facilitate passive transport over vast distances via ocean currents (velocities of ocean currents between 15 and 30 cm s1 are commonly encountered along the crest of ridge systems). Numerous crustaceans inhabit the vent environment and represent perhaps the dominant scavengers

Figure 8 Close-up of mussels (Bathymodiolus thermophilus) attached to the tubes of the tubeworm Riftia pachyptila. Limpets (Lepetodrilus elevatus) are seen attached to the external surfaces of both the mussel shells and tubeworm tubes.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA

137

Figure 9 Vesicomyid clams (Calyptogena magnifica) line cracks and crevices from which low-temperature fluids are venting in an area known as Clam Acres at 201500 N along the East Pacific Rise (depth 2615 m). This species is common at many vent sites that are believed to be in relatively late stages of succession along the Galapagos Rift and the entire stretch of the East Pacific Rise from 211N to 181S.

of the ecosystem. Brachyuran crabs (e.g., Bythograea thermydron on the Galapagos Rift and the East Pacific Rise) (Figures 1, 5, 6, and 7) appear to be one of the earliest colonizers of hydrothermal vent environments. When new vents were formed during the April, 1991 volcanic eruption along the crest of the East Pacific Rise at 91500 N, several regions were referred to as ‘crab nurseries’ owing to the relatively large abundance of crab larvae (megalopae) in areas

of low-temperature discharge. Eleven months after the eruptive event, the region was populated by tremendous numbers of large crabs. These were frequently observed to be holding with their claws various substrates, such as empty tubeworm tubes or pieces of basalt covered with microbial mats, and voraciously scraping the surfaces with their mouth parts, suggesting that microbes and/or their products may represent an important source of nutrition for

Figure 10 The polychaete Alvinella pompejana emerging from its tube on the side of a black smoker chimney at 91500 N along the East Pacific Rise. It has been suggested that the organism may be the most eurythermal invertebrate on the planet, capable of withstanding short-term exposures to temperatures ranging from less than 21C to in excess of 1001C.

(c) 2011 Elsevier Inc. All Rights Reserved.

138

HYDROTHERMAL VENT BIOTA

Figure 11 Limpets (Lepetodrilus elevatus) coating the surfaces of the tubes of Riftia pachyptila and the shell of Bathymodiolus thermophilus (left). A zoarcid fish (Thermarces andersoni) is seen emerging from the tubeworm tubes.

these organisms. They are also frequently seen extending their claws into the tubes of Riftia pachyptila and will readily devour the tissues of any tubeworm or mussel damaged during routine submersible activities. They are easily captured in ‘crab traps’ baited with a wide variety of dead fish. Intraspecific attacks appear to be relatively common occurrences and crabs with only one claw or a missing leg are frequently seen on the bottom, presumably reflecting an encounter with another crab. Galatheid crabs (e.g., Munidopsis subsquamosa) (Figure 12) inhabit peripheral areas of the vents during earlier stages of succession and, as hydrogen sulfide levels gradually decrease over time, they are

commonly encountered in central areas as well, often among tubeworms and mussels. While these ‘squat lobsters’ may also be scavengers, they are seldom, if ever, caught in traps. One chance encounter with a relatively large dead octopod in the peripheral area of a vent revealed a large quantity of galatheid crabs on the carcass and a noticeable absence of brachyuran crabs or other vent organisms (Figure 13). Close-up imagery revealed that the crabs were actively feeding on the dead tissues of the carcass. Galatheids that have been collected to date have generally been ‘transported’ back to the surface in random regions of the submersible itself. The carapaces of this species are frequently covered with fine, filamentous bacteria (Figure 12).

Figure 12 A galatheid crab (Munidopsis subsquamosa) with filamentous bacteria on its carapace. This species is a common inhabitant of peripheral vent environments and is also frequently observed in ambient deep-sea environments.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA

Figure 13 Numerous galatheid crabs (Munidopsis subsquamosa) on and in the immediate vicinity of the carcass of a cirrate octopod approximately 50 m from an active hydrothermal vent at 91500 N along the East Pacific Rise. Closeup imagery revealed that the crabs were actively feeding on the dead tissues of the carcass.

Numerous species of shrimp (Figure 14) have been encountered at vent sites visited to date along ridge axes in all ocean basins. At several sites (e.g., the TAG hydrothermal vent field) along the Mid-Atlantic Ridge, thousands of shrimp are frequently seen ‘swarming’ on top of one another, completely carpeting the sides of large sulfide chimneys. It has been suggested that a large ‘eye spot’ on the back of the carapace of the shrimp is capable of sensing longwavelength light emitted by high-temperature smokers. Shrimp are also relatively common inhabitants

139

of vents visited to date in the western Pacific (e.g., at the Alice Springs vent field along the back-arc spreading center in the Mariana Trough). Other smaller, common vent-endemic crustaceans include numerous species of amphipods, copepods, and leptostracans. Halice hesmonectes, an amphipod common at various vent sites along the East Pacific Rise is frequently seen ‘swarming’ above mussel and tubeworm colonies in regions of active low-temperature venting. It has been reported that these dense swarms represent the highest concentration of planktonic invertebrates in the ocean. Serpulid polychaetes (Figure 15) are common inhabitants of the peripheral area of many vent fields. Commonly referred to as ‘feather dusters,’ they were seen extending over large expanses of lava during the early expeditions to the Galapagos Rift in 1977 and 1979 and were subsequently reported at numerous vents along the East Pacific Rise. Their small tubes, which generally reach lengths of about 5 cm, consist of calcite. When the tentacular plumes are withdrawn into the tube, a small ‘plug’ seals the tube from the external environment. Numerous species of apparently vent-endemic fish have been reported from hydrothermal systems along ridge systems throughout the world’s oceans. Bythites hollisi is a bythitid that was encountered on the first dive to the Galapagos Rift vent field in 1977. Bythitids have been observed in large numbers at several other vent sites, such as the hydrothermal field at 91500 N along the East Pacific Rise (Figure 16). One large ‘pit,’ with a diameter of several meters, from which cloudy, shimmering water was emanating, had a concentration of over 20 bythitids,

Figure 14 The shrimp Alvinocaris lusca perched atop a tube of Riftia pachyptila at a low temperature vent along the East Pacific Rise.

(c) 2011 Elsevier Inc. All Rights Reserved.

140

HYDROTHERMAL VENT BIOTA

Figure 15 Serpulid polychaetes in the peripheral area of a vent field at 91500 N along the East Pacific Rise.

which were frequently observed with their heads projecting downward into the cloudiest portions of the water at the base of the pit. Zoarcids (eel pouts) are common members of the vent fauna at various sites along the East Pacific Rise and the Mid-Atlantic Ridge, with several different species having been encountered at the various vent fields visited to date. The common species that inhabits many East Pacific Rise vents is Thermarces andersoni (Figures 1, 4, 6, and 11). From analyses of extensive video footage and stomach contents of retrieved specimens of this species, it appears to commonly feed on a wide range of organisms, including bacteria, amphipods, leptostracans, and shrimp. It has also been observed scavenging on the plumes of specimens of vestimentiferan tubeworms that have

been damaged in the process of sampling or maneuvering with the submersible. Enteroptneusts, commonly referred to as ‘spaghetti worms’ were first observed in 1977 draped over pillow lava in peripheral regions of the Galapagos Rift vent field. Few individuals were observed at any of the other many vent fields visited over the next 20 years throughout the world. In 1997, six years after the volcanic eruption at 91500 N along the East Pacific Rise, numerous enteroptneusts were observed at distances ranging from a few meters to a few hundred meters from active vent sites within the region. No individual organisms were observed living in direct association with venting fluids or at distances in excess of a kilometer from active hydrothermal systems. This unusual, soft-bodied

Figure 16 The vent fish Bythites hollisi emerging from a high-temperature vent region known as Hole-to-Hell at 91500 N along the East Pacific Rise. The species is common at various vent fields along the Galapagos Rift and East Pacific Rise.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA

Figure 17 The enteroptneust Saxipendium coronatum (commonly known as a ‘spaghetti worm’) on the surface of basalt about 30 m from an active low-temperature vent field at 91500 N along the East Pacific Rise. The image was taken with a prototype high-resolution video camera system equipped with a macro-lens.

invertebrate may represent an organism that is uniquely adapted to an ecotone between active vents and the ambient deep sea. Analyses of video images of numerous individuals taken with a high-resolution video camera system equipped with a macro-lens (Figure 17) have revealed behavioral patterns suggesting that the organism may be ‘grazing’ directly on basaltic surfaces, potentially consuming microbes or organic substances ultimately originating in the vent environment. One of the most unusual and spectacular organisms inhabiting vent ecosystems is the vestimentiferan tubeworm Riftia pachyptila which thrives at numerous vent fields along the Galapagos Rift and East Pacific Rise (Figures 1, 4, and 8). As mentioned earlier, it lives in a symbiotic relationship with chemosynthetic bacteria concentrated within its body. Such a relationship provides an internal, hydrogen sulfide-nourished ‘garden’ that, in turn, nourishes the tubeworm. Although the mechanism by which the host obtains energy from the bacteria is unclear, the energy transfer appears remarkably efficient. An unique opportunity to determine the growth rate potential of R. pachyptila arose as a result of the April, 1991 volcanic eruptive event at 91500 N along the East Pacific Rise mentioned above. In March,

Figure 18 (Right)Temporal sequence of vent community development at a low-temperature vent in a region known as Hole-to-Hell at 91500 N along the East Pacific Rise. This was the site of a volcanic eruption in April, 1991 that entirely decimated previously existing communities within the region. The field of view of each image and the heading of the camera system utilized are approximately the same for each image taken at the following times: (A) April, 1991; (B) March, 1992; (C) December, 1993; (D) October, 1994; and (E) November, 1995. (Shank et al., 1998.)

(c) 2011 Elsevier Inc. All Rights Reserved.

141

142

HYDROTHERMAL VENT BIOTA

1992, no R. pachyptila were present within the region, which had been devastated by the eruption. By December, 1993, less than 2 years later, huge colonies of this tubeworm had colonized the active lowtemperature vents; the tube lengths of many individuals were in excess of 1.5 m (Figure 18 and 19). Such growth rates of more than 85 cm y1 increase in tube length represent the fastest rates of growth documented for any marine invertebrate. It is interesting to note that one of the other fastest-growing marine invertebrates is the giant clam Tridacna squamosa, a bivalve that has a symbiotic association with photosynthetic algae (zooxanthellae). The efficiency with which energy is transferred from the symbiont to the host in certain invertebrates may well be a contributing factor to the remarkable growth rates of these organisms. The rapid succession of a tubeworm-dominated vent community over the 5-year period following the April, 1991 eruption is dramatically illustrated in Figures 18 and 19. Inhabitants of deep-sea hydrothermal vents are among the most spectacular and unusual organisms on the planet. Given the relatively small number of vent ecosystems found to date, many questions have been raised concerning whether conservation measures need to be taken to protect vent communities from anthropogenic disturbances. Any assessment of the potential consequences of anthropogenic impacts needs to consider that the organisms inhabiting deepsea hydrothermal vents thrive in an environment that is constantly being altered radically by geological process. Periodic devastation of entire biological communities is a relatively common occurrence as volcanic and tectonic processes proceed along active ridge axes. Over the past two decades we have learned that many vent communities are remarkably resilient. Populations of essentially all vent organisms indigenous to regions decimated by massive volcanic eruptions have, like a Phoenix rising from the ashes, reestablished themselves in less than a decade. This remarkable resilience in the face of huge natural disasters has profound implications as one considers the potential impacts of exploitation of

Figure 19 (Left)Temporal sequence of vent community development at a low-temperature vent located approximately 500 m from the community depicted in Figure 18. This was also a region that was buried with fresh lava by the April, 1991 volcanic eruptive event, decimating previously existing communities within the area. The field of view of each image and the heading of the camera system utilized are approximately the same for each image taken at the following times: (A) April, 1991; (B) March, 1992; (C) December, 1993; (D) October, 1994; and (E) November, 1995. (Shank et al., 1998.)

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT BIOTA

precious mineral and biological resources associated with active hydrothermal systems throughout the world’s oceans.

Acknowledgments I thank the pilots and crew of the DSV Alvin and the R/V Atlantis for their expertise, assistance, and patience over the years; W. Lange and the Woods Hole Oceanographic Institution for technical expertise and the provision of the camera and recording systems critical to the generation of the majority of images presented in this article; Emory Kristof, Stephen Low, and Michael V. DeGruy for inspiration and assistance with the procurement of video images using a variety of camera systems; and Matt Tieger for assistance in the generation of video prints. Supported by National Science Foundation Grants OCE-95-29819 and OCE-96-33131.

See also Deep-Sea Ridges, Microbiology. Hydrothermal Vent Ecology Hydrothermal Vent Fauna, Physiology of

143

Further Reading Childress JJ and Fisher CR (1992) The biology of hydrothermal vent animals: physiology, biochemistry, and autotrophic symbioses. Oceanography and Marine Biology Annual Review 30: 337--441. Gage JD and Tyler PA (1991) Deep-Sea Biology: A Natural History of Organisms at the Deep-Sea Floor. Cambridge: Cambridge University Press. Lutz RA (2000) Deep sea vents. National Geographic 198(4): 116--127. Shank TM, Fornari DJ, and Von Damm KL (1998) Temporal and spatial patterns of biological community development at nascent deep-sea hydrothermal vents (91N, East Pacific Rise). Deep-Sea Research. 45: 465– 515. Jones ML (ed.) (1985) The Hydrothermal Vents of the Eastern Pacific: An Overview. Bulletin of the Biological Society of Washington, vol. 6. Washington, DC: Biological Society of Washington. Rona PA, Bostrom K, Laubier L, and Smith KL Jr (eds.) (1983) Hydrothermal Processes at Seafloor Spreading Centers. New York: Plenum Press. Tunnicliffe V (1991) The biology of hydrothermal vents: ecology and evolution. Oceanography and Marine Biology Annual Review 29: 319--417. Van Dover CL (2000) The Ecology of Deep-Sea Hydrothermal Vents. Princeton, NJ: Princeton University Press, Princeton.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT DEPOSITS R. M. Haymon, University of California, CA, USA Copyright & 2001 Elsevier Ltd. All rights reserved. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1228 –1234, & 2001, Elsevier Ltd.

Where Deposits Form: Geologic Controls

Introduction In April 1979, submersible divers exploring the midocean ridge crest at latitude 211N on the East Pacific Rise discovered superheated (3807301C) fluids, blackened by tiny metal-sulfide mineral crystals, spewing from the seafloor through tall mineral conduits (see Hydrothermal Vent Biota, Hydrothermal Vent Fluids, Chemistry of). The crystalline conduits at these ‘black smoker’ hydrothermal vents were made of minerals rich in copper, iron, zinc, and other metals. Since 1979, hundreds of similar hydrothermal deposits have been located along the midocean ridge. It is now clear that deposition of hydrothermal mineral deposits is a common process, and is integrally linked to cracking, magmatism, and cooling of new seafloor as it accretes and spreads away from the ridge (see Propagating Rifts and Microplates, Mid-Ocean Ridge Geochemistry and Petrology, Seamounts and Off-Ridge Volcanism, Mid-Ocean Ridge Seismic Structure). For thousands of years before mid-ocean ridge hot springs were discovered in the oceans, people mined copper from mineral deposits that were originally formed on oceanic spreading ridges. These fossil deposits are embedded in old fragments of seafloor called ‘ophiolites’ that have been uplifted and emplaced onto land by fault movements. The copperrich mineral deposits in the Troodos ophiolite of Cyprus are well-known examples of fossil oceanridge deposits that have been mined for at least 2500 years; in fact, the word ‘copper’ is derived from the Latin word ‘cyprium’ which means ‘from Cyprus.’ The mineral deposits accumulating today at hot springs along the mid-ocean ridge are habitats for a variety of remarkable organisms ranging in size from tiny microbes to large worms (see Hydrothermal Vent Biota, Deep-Sea Ridges, Microbiology, Hydrothermal Vent Fauna, Physiology of). The properties of the mineral deposits are inextricably linked to the organisms that inhabit them. The mineral deposits contain important clues about the physical–chemical environments in which some of these organisms live, and also preserve fossils of some organisms, creating a geologic record of their existence.

144

Hydrothermal vent deposits are thus a renewable source of metals and a record of the physical, chemical, biological, and geological processes at modern and ancient submarine vents.

Less than 2% of the total area of the mid-ocean ridge crest has been studied at a resolution sufficient to reveal the spatial distribution of hydrothermal vents, mineral deposits, and other significant small-scale geologic features. Nevertheless, because study areas have been carefully selected and strategically surveyed, much has been learned about where vents and deposits form, and about the geologic controls on their distribution. The basic requirements for hydrothermal systems include heat to drive fluid circulation, and high-permeability pathways to facilitate fluid flow through crustal rocks. On midocean ridges, vents and deposits are forming at sites where ascending magma intrusions introduce heat into the permeable shallow crust, and at sites where deep cracks provide permeability and fluid access to heat sources at depth.

Fast-spreading Ridges Near- and on-bottom studies along the fast-spreading East Pacific Rise suggest that most hydrothermal mineral deposits form along the summit of the ridge crest within a narrow ‘axial zone’ less than 500 m wide. Only a few active sites of mineral deposition have been located outside this zone; however, more exploration of the vast area outside the axial zone is needed to establish unequivocally whether or not mineral deposition is uncommon in this region. The overall spatial distribution of hydrothermal vents and mineral deposits along fast-spreading ridges traces the segmented configuration of cracks and magma sources along the ridge crest (see Propagating Rifts and Microplates, Mid-Ocean Ridge Geochemistry and Petrology, Seamounts and Off-Ridge Volcanism, Mid-Ocean Ridge Seismic Structure). Within the axial zone, mineral deposition is concentrated along the floors and walls of axial troughs created by volcanic collapse and/or faulting along the summit of the ridge crest. The majority of the deposits are located along fissures that have opened above magmatic dike intrusions, and along collapsed

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT DEPOSITS

lava ponds formed above these fissures by pooling and drainage of erupted lava. Where fault-bounded troughs have formed along the summit of the ridge crest, mineral deposition is focused along the bounding faults and also along fissures and collapsed lava ponds in the trough floor. Hydrothermal vents appear to be most abundant along magmatically inflated segments of fast-spreading ridges; however, the mineral deposits precipitated on the seafloor on magmatically active segments are often buried beneath frequent eruptions of new lava flows. The greatest number of deposits, therefore, are observed on inflated ridge segments that are surfaced by somewhat older flows, i.e., along segments where: (1) much heat is available to power hydrothermal vents; and (2) mineral deposits have had time to develop but have not yet been buried by renewed eruptions.

Intermediate- and Slow-spreading Ridges Most hydrothermal deposits that have been found on intermediate- and slow-spreading ridge crests are focused along faults, fissures, and volcanic structures within large rift valleys that are several kilometers wide. The fault scarps along the margins of rift valleys are common sites for hydrothermal venting and mineral deposition. Fault intersections are thought to be particularly favorable sites for hydrothermal mineral deposition because they are zones of high permeability that can focus fluid flow. Mineral deposition on rift valley floors is observed along fissures above dike intrusions, along eruptive fissures and

Dead edifice

Flanges

volcanic collapse troughs, and on top of volcanic mounds, cones and other constructions. In general at slower-spreading ridges, faults appear to play a greater role in controlling the distribution of hydrothermal vents and mineral deposits than they do at fast-spreading ridges, where magmatic fissures are clearly a dominant geologic control on where vents and deposits are forming.

Structures, Morphologies, and Sizes of Deposits A typical hydrothermal mineral deposit on an unsedimented mid-ocean ridge accumulates directly on top of the volcanic flows covering the ridge crest. On sedimented ridges, minerals are deposited within and on top of the sediments. Beneath seafloor mineral deposits are networks of feeder cracks through which fluids travel to the seafloor. Precipitation of hydrothermal minerals in these cracks and in the surrounding rocks or sediments creates a subseafloor zone of mineralization called a ‘stockwork’. In hydrothermal systems where fluid flow is weak, unfocused, or where the fluids mix extensively with sea water beneath the seafloor, most of the minerals will precipitate in the stockwork rather than on the seafloor. Hydrothermal deposits on mid-ocean ridges are composed of: (1) vertical structures, including individual conduits known as ‘chimneys’ (Figure 1) and larger structures of coalesced conduits that are often called ‘edifices’; (2) horizontal ‘flange’ structures that extend outward from chimneys and edifices

Black smoke plume

Fossil worm tubes

145

Black smoker chimneys

T decreasing, mixing increasing

White smoke plume White smoker edifices Live alvinelline worm Basal mound

Basalt Figure 1 Composite sketch of the mineral structures and zones in hydrothermal mineral deposits on unsedimented ridge crests (modified after Haymon, 1989). Although mound interiors are seldom observed on the seafloor, the simplified sketch of mineral zoning within the mound is predicted by analogy with chimneys and massive sulfide deposits exposed in ophiolites. An outer peripheral zone (unshaded) of anhydrite þ amorphous silica þ Zn-rich sulfide, dominantly ZnS þ FeS2, is replaced in the interior by an inner zone (hatched) of Cu-rich sulfide (CuFeS2 þ FeS2) þ minor anhydrite and amorphous silica. The inner zone may be replaced by a basal zone (cross-pattern) of Cu-rich sulfide (CuFeS2 þ FeS2) þ quartz. Zones migrate as thermochemical conditions within the mound evolve. Although not shown here, it is expected that zoning around individual fractures cutting through the mound will be superimposed on the simplified zone structure in this sketch.

(c) 2011 Elsevier Inc. All Rights Reserved.

146

HYDROTHERMAL VENT DEPOSITS

at a particular site, which depends on the nature of the heat source and plumbing system, and the rate of seafloor spreading; frequency with which deposits are buried beneath lava flows; and the compositions of the vent fluids and minerals. The large deposits found on slower-spreading ridge crests are located on faults that have moved slowly away from the ridge axis and have experienced repeated episodes of venting and accumulated mineral deposition over thousands of years, without being buried by lava flows. The tall Endeavor Segment edifices are formed because ammonia-enriched fluid compositions favor precipitation of silica in the edifice walls. The silica is strong enough to stabilize these structures so that they do not collapse as they grow taller.

How Do Chimneys Grow? A relatively simple two-stage inorganic growth model has been advanced to explain the basic characteristics of black smoker chimneys (Figure 2). In this model, a chimney wall composed largely of anhydrite (calcium sulfate) precipitates initially from sea water that is heated around discharging jets of hydrothermal fluid. The anhydrite-rich chimney wall precipitated during stage I contains only a small Cross-section chimney wall +

+

+

+

+

+

+ +

Hydrothermal fluid

+ +

+ +

+ +

+ +

+ +

+

Pillow basalt (A)

Arrows indicate directions of growth

+ + +

Anhydrite ± (caminite) + FeS2 and Zn(Fe)S Fine-grained pyrrhotite + pyrite + sphalerite intergrowths in an anhydrite matrix +

+

+

+

+

+

+

+

Hydrothermal fluid

+

+

+ +

+ +

(B)

+

+

Pillow basalt + + +

(Figure 1); (3) mounds of accumulated mineral precipitates (Figure 1); and (4) horizontal layers of hydrothermal sediments, debris, and encrustations. Chimneys are initially built directly on top of the seabed around focused jets of high-temperature effluents. Chimneys and edifices are physically unstable and often break or collapse into pieces that accumulate into piles of debris. The debris piles are cemented into consolidated mounds by precipitation of minerals from solutions percolating through the piles. New chimneys are constructed on top of the mounds as the mounds grow in size. Hydrothermal plume particles and particulate debris from chimneys settle around the periphery of the mounds to form layers of hydrothermal sediment. Diffuse seepage of fluids also precipitates mineral encrustations on mound surfaces, on volcanic flows and sediments, and on biological substrates, such as microbial mats or the shells and tubes of sessile macrofauna. The morphologies of chimneys are highly variable and evolve as the chimneys grow, becoming more complex with time. Black smoker chimneys are often colonized by organisms and evolve into ‘white smokers’ that emit diffuse, diluted vent fluids through a porous carapace of worm tubes (Figure 1). Fluid compositions and temperatures, flow dynamics, and biota are all factors that influence the development of chimney morphology. The complexity of the interactions between these factors, and the high degree of spatial–temporal heterogeneity in the physical, chemical, biological and geological conditions influencing chimney growth, account for the diverse morphologies exhibited by chimneys, and present a challenge to researchers attempting to unravel the processes producing these morphologies. The sizes of hydrothermal mineral deposits on ridges also vary widely. It has been suggested that the largest deposits are accumulating on sedimented ridges, where almost all of the metals in the fluids are deposited within the sediments rather than being dispersed into the oceans by hydrothermal plumes. On unsedimented ridges, the structures deposited on the seabed at fast spreading rates are usually relatively small in dimension (mounds are typically less than a few meters in thickness and less than tens of meters in length, and vertical structures are o15 m high). On intermediate- and slow-spreading ridges, mounds are sometimes much larger (up to tens of meters in thickness, and up to 300 m in length). On the Endeavour Segment of the Juan de Fuca Ridge, vertical structures reach heights of 45 m. The size of a deposit depends on many factors, including: magnitude of the heat source, which influences the duration of venting and mineral deposition; tendency of venting and mineral deposition to recur episodically

Chalcopyrite, or cubanite + (pyrrhotite) Cu_Fe sulfides in an anhydrite matrix As in stage I:higher sulfide:sulfate ratio

Figure 2 Two-stage model of black smoker chimney growth. (A) Stage I, sulfate-dominated stage; (B) stage II, sulfide replacement stage. During stage II, several different sulfide mineral zonation sequences develop, depending on permeability and thickness of chimney walls, hydrodynamic variables, and hydrothermal fluid composition.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT DEPOSITS

component of metal sulfide mineral particles that crystallize because of rapid chilling of the hydrothermal fluids. In stage II, the anhydrite-rich wall continues to grow upward and to thicken radially,

Table 1 Minerals occurring in ocean ridge hydrothermal mineral deposits

147

protecting the fluid flowing through the chimney from very rapid chilling and dilution by sea water. This allows metal sulfide minerals to precipitate into the central conduit of the chimney from the hydrothermal fluid. The hydrothermal fluid percolates outward through the chimney wall, gradually replacing anhydrite and filling voids with metal sulfide minerals. During stage II, the chimney increases in height, girth and wall thickness, and both the calcium sulfate/metal sulfide ratio and permeability of the walls decrease. Equilibration of minerals with pore fluid in the walls occurs continuously along steep, time–variant temperature and chemical gradients between fluids in the central conduit and sea water surrounding the chimney. This equilibration produces sequences of concentric mineral zones across chimney walls that evolve with changes in thermal and chemical gradients and wall permeability. The model of chimney growth described above is accurate but incomplete, as it does not include the effects on chimney development of fluid phase separation, biological activity, or variations in fluid composition. Augmented models that address these

Table 2 Ranges of elemental compositions in bulk midocean ridge hydrothermal mineral deposits

(c) 2011 Elsevier Inc. All Rights Reserved.

148

HYDROTHERMAL VENT DEPOSITS

Morphological and Mineralogical Evolution of Chimneys on the East Pacific Rise at 9˚-10˚N Stage 2

1991 "Proto-chimney" Anhydrite-dominated T = 389˚-403˚C

Stage 2

1992-1995 CuFe-sulfide-dominated T = 340˚-392˚C

1992-1995 Zn-sulfide-dominated T = 264˚-340˚C

KEY an _ anhydrite cp _ chalcopyrite po _ pyrrhotite py _ pyrite

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT DEPOSITS

complexities are needed to fully characterize the processes governing chimney growth.

Elemental and Mineral Compositions of Deposits Ridge crest hydrothermal deposits are composed predominantly of iron-, copper- and zinc sulfide minerals, calcium- and barium-sulfate minerals, iron oxide and iron oxyhydroxide minerals, and silicate minerals (Table 1). These minerals precipitate from diverse processes, including: heating of sea water; cooling of hydrothermal fluid; mixing between sea water and hydrothermal fluid; reaction of hydrothermal minerals with fluid, sea water, or fluid–sea water mixtures; reaction between hydrothermal fluid and seafloor rocks and sediments; and reactions that are mediated or catalyzed biologically. This diversity in the processes and environments of mineral precipitation results in the deposition of many different minerals and elements (Tables 1 and 2). High concentrations of strategic and precious metals are found in some deposits (Table 2). The deposits are potentially valuable, if economic and environmentally safe methods of mining them can be developed. Chimneys can be classified broadly by composition into four groups: sulfate-rich, copper-rich, zinc-rich and silica-rich structures. Copper-rich chimney compositions are indicative of formation at temperatures above 3001C. Sulfate-rich compositions are characteristic of active and immature chimneys. Many chimneys are mineralogically zoned, with hot interior regions enriched in copper, and cooler exterior zones enriched in iron, zinc, and sulfate (Figures 1 and 2). Mounds exhibit a similar gross mineral zoning, and those which are exposed by erosion in ophiolites often have silicified (quartzrich) interiors (Figure 2). Seafloor weathering of deposits after active venting ceases results in dissolution of anhydrite, and oxidation and dissolution of metal-sulfide minerals. Small deposits that are not sealed by silicification or buried by lava flows will

149

not be well preserved in the geologic record (Figure 3).

Chimneys as Habitats Chimney and mound surfaces are substrates populated by microbial colonies and sessile organisms such as vestimentiferan and polychaete worms, limpets, mussels, and clams. It is likely that pore spaces in exterior regions of chimney walls are also inhabited by microbes. All of these organisms that are dependent on chemosynthesis benefit from the seepage of hydrothermal fluid through active mineral structures, and from the thermal and chemical gradients across mineral structures. The structures provide an interface between sea water and hydrothermal fluid that maintains tolerable temperatures for biota, and allows organisms simultaneous access to the chemical constituents in both sea water and hydrothermal fluid. However, organisms attached to active mineral structures must cope with changes in fluid flow across chimney walls (which sometimes occur rapidly), and with ongoing engulfment by mineral precipitation. Some organisms actively participate in the precipitation of minerals; for example, sulfide-oxidizing microbes mediate the crystallization of native sulfur crystals, and microbes are also thought to participate in the precipitation of marcasite and iron oxide minerals. Additionally, the surfaces of organisms provide favorable sites for nucleation and growth of amorphous silica, metal sulfide and metal oxide crystals, and this facilitates mineral precipitation and fossilization of vent fauna (Figure 3).

Fossil Record of Hydrothermal Vent Organisms Fossil molds and casts of worm tubes, mollusc shells, and microbial filaments have been identified in both modern ridge hydrothermal deposits and in Cretaceous, Jurassic, Devonian, and Silurian deposits. This fossil record establishes the antiquity of vent

Figure 3 (Left)On left: a time series of seafloor photographs showing the morphological development of a chimney that grew on top of lava flows erupted in 1991 on the crest of the East Pacific Rise near 9150.30 N (Haymon et al., 1993). Within a few days-to-weeks after the eruption, anhydrite-rich ‘Stage 1 Protochimneys’ a few cm high had formed where hot fluids emerged from volcanic outcrops covered with white microbial mats (top left). Eleven months later, the chimney consisted of cylindrical ‘Stage 2’ anhydrite-sulfide mineral spires approximately one meter in height, and as-yet unpopulated by macrofauna (middle left). Three and a half years after the eruption, the cylindrical conduits had coalesced into a 7 m-high chimneys structure that was covered with inhabited Alvinelline worms tubes (bottom left). On right: photomicrographs of chimney samples from the eruption area that show how the chimneys evolved from Stage 1 (anhydrite-dominated; top right) to Stage 2 (metal-sulfide dominated) mineral compositions (see text). As the fluids passing through the chimneys cooled below B3301C during Stage 2, the CuFe-sulfide minerals in the chimney walls (middle right) were replaced by Zn- and Fe-sulfide minerals (bottom right).

(c) 2011 Elsevier Inc. All Rights Reserved.

150

HYDROTHERMAL VENT DEPOSITS

communities and the long evolutionary history of specific faunal groups. The singular Jurassic fossil assemblage preserved in a small ophiolite-hosted deposit in central California is particularly interesting because it contains fossils of vestimentiferan worms, gastropods and brachiopods, but no clam or mussel fossils. In contrast, modern and Paleozoic faunal assemblages described thus far include clams, mussels and gastropods, but no brachiopods. Does this mean that brachiopods have competed with molluscs for ecological niches at vents, and have moved in and out of the hydrothermal vent environment over time? Fossilization of organisms is a selective process that does not preserve all the fauna that are present at vents. Identification of fossils at the species level is often difficult, especially where microbes are concerned. Notwithstanding, it is important to search for more examples of ancient fossil assemblages and to trace the fossil record of life at hydrothermal vents back as far as possible to shed light on how vent communities have evolved, and whether life on earth might have originated at submarine hydrothermal vents.

Summary Formation of hydrothermal deposits is an integral aspect of seafloor accretion at mid-ocean ridges. These deposits are valuable for their metals, for the role that they play in fostering hydrothermal vent ecosystems, for the clues that they hold to understanding spatial–temporal variability in hydrothermal vent systems, and as geologic records of how life at hydrothermal vents has evolved. From these deposits we may gain insights about biogeochemical processes at high temperatures and pressures that can be applied to understanding life in inaccessible realms within the earth’s crust or on other planetary bodies. We are only beginning to unravel the complexities of ridge hydrothermal vent deposits. Much exploration and interdisciplinary study remains to be done to obtain the valuable information that they contain.

See also Deep-Sea Ridges, Microbiology. Hydrothermal Vent Fluids, Chemistry of. Hydrothermal Vent Biota. Hydrothermal Vent Fauna, Physiology of. Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Propagating Rifts and Microplates. Seamounts and Off-Ridge Volcanism.

Further Reading Dilek Y, Moores E, Elthon D, and Nicolas A (eds.) (2000) Ophiolites and Oceanic Crust: New Insights from Field Studies and the Ocean Drilling Program. Geological Society of America Memoir. Boulder: Geological Society of America. Haymon RM (1989) Hydrothermal processes and products on the Galapagos Rift and East Pacific Rise, 1989. In: Winterer EL, Hussong DM, and Decker RW (eds.) The Geology of North America: The Eastern Pacific Ocean and Hawaii, vol. N, pp. 125--144. Boulder: Geological Society of America. Haymon RM (1996) The response of ridge crest hydrothermal systems to segmented, episodic magma supply. In: MacLeod CJ, Tyler P, and Walker CL (eds.) Tectonic, Magmatic, Hydrothermal, and Biological Segmentation of Mid-Ocean Ridges, vol. Special Publication 118, pp. 157--168. London: Geological Society. Humphris SE, Zierenberg RA, Mullineaux LS, and Thomson RE (eds.) (1995) Seafloor Hydrothermal Systems: Physical, Chemical, Biological, and Geological Interactions, Geophysical Monograph, vol. 91. Washington, DC: American Geophysical Union. Little CTS, Herrington RJ, Haymon RM, and Danelian T (1999) Early Jurassic hydrothermal vent community from the Franciscan Complex, San Rafael Mountains, California. Geology 27: 167--170. Tivey MK, Stakes DS, Cook TL, Hannington MD, and Petersen S (1999) A model for growth of steep-sided vent structures on the Endeavour Segment of the Juan de Fuca Ridge: results of a petrological and geochemical study. Journal of Geophysical Research 104: 22859--22883.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT ECOLOGY C. L. Van Dover, The College of William and Mary, Williamsburg, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1234–1241, & 2001, Elsevier Ltd.

Introduction Most of the ocean floor is covered with a thick layer of sediment and is populated by sparse and minute, mud-dwelling and mud-consuming invertebrates. In striking contrast, the volcanic basalt pavement of mid-ocean ridges hosts hydrothermal vents and their attendant lush communities of large invertebrates that ultimately rely on inorganic chemicals for their nutrition. Vents themselves are sustained by tectonic forces that fracture the basalt and allow sea water to penetrate deep within the ocean crust, and by volcanism, which generates the hot rock at depth that strips sea water of oxygen and magnesium. The hot rock gives up to the nascent vent fluid a variety of metals, especially copper, iron, and zinc, as well as reduced compounds such as hydrogen sulfide and methane. The vent fluid, thermally buoyant, rises to exit as hot springs on the seafloor. Discovered first by geologists in 1977 along a stretch of mountain range known as the Galapagos Rift, near the equator in the eastern Pacific Ocean, hydrothermal vents are now known to occur along every major ridge system on the planet. Several of these ridge systems – in the Arctic and Antarctic, in the Indian Ocean, in the southern Atlantic – are only just beginning to be explored. Regional species composition of the vent fauna differs between ocean basins and, sometimes, even within a basin. Most of the species that occur at vents have never been found in the adjacent, non-vent deep sea and are considered to be endemic, adapted to the chemical milieu of the vent environment. Reduced compounds carried in hydrothermal fluids, together with oxygen from the surrounding sea water, fuel the microbial fixation of inorganic carbon into organic carbon that forms the chemosynthetic base of the vent food web. Hydrothermal vents on midocean ridges are thus globally distributed, insular ecosystems that support endemic faunas through chemosynthetic processes rather than through photosynthesis. They are effectively decoupled by depth (typically 41000 m) from climatic variations and anthropogenic activities, but are tightly coupled to geophysical processes

of tectonism and volcanism. Vents thus offer unique opportunities for biologists to study adaptations that allow life to persist in this extreme environment and to explore planetary controls on biodiversity and biogeography along submarine, hydrothermal ‘archipelagoes’, where propagules are water-borne and subject to dispersal in an open system. Further, because vents are thought to have been a primordial component of the oceans and because there is increasing speculation that early life on this planet may have thrived in hot environments and on chemicals rather than on an organic soup for nourishment, extant hydrothermal systems are thought by many to be analogues for sites where early life may have evolved on this and other planets or planetary bodies in our solar system.

Microorganisms and the Chemosynthetic Basis for Life at Vents The terms ‘chemosynthesis’ and ‘photosynthesis’ are imprecise. While a voluminous nomenclature is available to differentiate among variations in these processes, for simplicity, chemosynthesis and photosynthesis are used here. In photosynthesis, sunlight captured by proteins provides energy for the conversion of inorganic carbon (carbon dioxide, CO2) and water (H2O) into organic carbon (carbohydrates, [CH2O] and oxygen (O2) (eqn [1]). light

CO2 þ H2 O - ½CH2 O þ O2

½1

Photosynthesis by plants is the basis for consumer and degradative food webs both on land and, as a rain of organic detritus derived from surface phytoplankton productivity, on the seabed. In the deep sea, detrital inputs of organic carbon are exceedingly small, accounting for the paucity of consumer biomass in abyssal muds. At hydrothermal vents, the supply of surface-derived organic material is overwhelmed by the supply of new organic carbon generated through chemical oxidation of hydrogen sulfide (H2S) (eqn [2]). CO2 þ H2 O þ H2 S þ O2 -½CH2 O þ H2 SO4

½2

Metabolic fixation pathways for carbon can be identical in photosynthetic plants and chemosynthetic

(c) 2011 Elsevier Inc. All Rights Reserved.

151

152

HYDROTHERMAL VENT ECOLOGY

microorganisms, namely the Calvin–Benson cycle, but the energy-yielding processes that fuel the Calvin– Benson cycle (photon capture versus chemical oxidation) are distinctive. High biomass at hydrothermal vents is in part a consequence of the aerobic nature of the process described in eqn [2]. Oxygen is used to oxidize the hydrogen sulfide, generating a large energy yield that in turn can fuel the production of large

amounts of organic carbon (Figure 1). Nonaerobic chemical reactions, such as oxidation of vent-supplied hydrogen (H2) by carbon dioxide (CO2), can also support chemosynthesis at vents, but energy yields under such anaerobic conditions are much lower than from aerobic oxidation. Microorganisms using these anaerobic reactions cannot by themselves support complex food webs and large invertebrates.

Symbiosis and the Host–Symbiont Relationship One of the hallmarks of many hydrothermal vent communities is the dominance of the biomass by invertebrate species that host chemosynthetic microorganisms within their tissues. Giant, redplumed, vestimentiferan tubeworms (Riftia pachyptila; Figure 2) so far provide the ultimate in host accommodation of endosymbiotic bacteria. These worms live in white, chitinous tubes, with their plumes extended into the zone of turbulent mixing of warm (B201C), sulfide-rich, hydrothermal fluid and cold (21C), oxygenated sea water. When discovered

CO2+H2O solar energy [CH2O] + O2 Photosynthesis

Chemical energy CO2 + H2O + H2S + O2 [CH2O] + H2SO4

Chemosynthesis

Figure 1 Photosynthetic and chemosynthetic processes in the ocean. Sunlight fuels the generation of organic material (CH2O) from inorganic carbon dioxide (CO2) and water (H2O) by phytoplankton in surface, illuminated waters. At depths where hydrothermal vents exist (typically>2000 m), no sunlight penetrates. In place of sunlight, the chemical oxidation of sulfide (H2S) by oxygen (O2) fuels the conversion of carbon dioxide to organic carbon by chemosynthetic bacteria.

Figure 2 The giant tubeworm Riftia pachyptila. The red plume of the tubeworm acts as a gill for uptake of dissolved gases. The trunk of the worm is found inside the white, chitinous tube. (Photograph by Dudley Foster, Woods Hole Oceanographic Institution.)

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT ECOLOGY

in 1977, vestimentiferan tubeworms were remarkable for their size (up to several meters in length) and the complete absence of a digestive system in adults. In fact, the digestive system has been replaced by the trophosome, which is a specialized, paired organ derived from the larval gut. The trophosome is richly infiltrated with blood capillaries and each of its lobes lies within a blood-filled body cavity. The blood itself is rich in hemoglobin. Host bacteriocyte cells in the trophosome house chemosynthetic bacteria that use hydrogen sulfide and oxygen to fuel the production of organic carbon, as described above. The metabolic requirements of the tubeworm endosymbiotic bacteria place some remarkable burdens on the host. First, there is a novel requirement for delivery of sulfide to the bacteria, which reside at a location remote from the site of gas exchange (the plume). Sulfide is normally a potent toxin to animals, poisoning the cellular enzyme system that generates ATP (adenosine triphosphate), the currency of metabolism. Sulfide also competes with oxygen for binding sites on hemoglobin. Tubeworm hemoglobin has separate binding sites for oxygen and sulfide, so that both can be transported throughout the worm in the circulatory system without competition. When bound to hemoglobin, the sulfide is not reactive and so enzyme systems remain unchallenged. Once delivered to the bacteria in the trophosome, the sulfide is quickly oxidized and loses its toxic potential. Novel requirements for carbon dioxide are also found in tubeworms. The usual flow of CO2 is out of an animal, as the end-product of metabolism, but the resident bacteria of the trophosome require a net uptake of CO2. Maintenance of high concentrations of inorganic carbon in the blood of the tubeworm is facilitated by the high partial pressure of CO2 in the water surrounding the site of uptake (the plume) and by the alkaline internal pH of the blood (7.3–7.4), which favors the bicarbonate form (HCO 3 ) of carbon dioxide and thus maintains a steep concentration gradient for diffusion of CO2 from the environment into the blood. As described above, the anatomy of the tubeworm is well adapted for life in sulfide- and CO2-rich vent fluids and for supporting its endosymbiotic, chemosynthetic bacteria. The bacteria provide nearly all of the nutrition for the host, with the exception, perhaps, of small amounts of dissolved organic materials taken up across the tissues of the plume. In turn, the bacteria are provided with a chemically rich and stable environment for growth. Other large invertebrates at vents also derive much of their nutrition from endosymbiotic, chemosynthetic bacteria, including 20 to 30-cm long to vesicomyid clams and bathymodiolid mussels. While

153

Figure 3 Swarming shrimp (Rimicaris exoculata) at a hydrothermal vent on the Mid-Atlantic Ridge. (Photograph by C.L. Van Dover.)

vent mussels have a fairly normal digestive system and are capable of filter-feeding just as shallow-water mussels do, vent clams lack a functional digestive system. Both types of bivalves have enlarged gills and it is within these gills that the endosymbiotic bacteria are found. Clams are thought to take up hydrogen sulfide via their highly vascularized foot, with which they probe cracks in the basalt where vent fluids emanate. Not all chemosynthetic bacteria that nourish vent invertebrates are endosymbiotic. Shrimp that dominate vents in the Atlantic (Figure 3) host chemosynthetic bacteria on their carapace (i.e., the bacteria are episymbiotic) and seem to depend on these bacteria for a significant portion of their diet. Other chemosynthetic bacteria are free-living, suspended in the water column, providing nourishment to suspension-feeding invertebrates such as barnacles, or grow as mats or films on surfaces, where grazers such as limpets and polychaetes forage. Heterotrophic bacteria (using organic rather than inorganic compounds) may also be important for consumers within the vent invertebrate food web, but this has yet to be examined carefully.

Thermal Adaptations While hydrothermal vent communities live at temperatures slightly elevated above ambient sea water temperature, the existence of highly productive communities at vents is a consequence of fluid chemistry rather than thermal input. Nevertheless, there are some invertebrates, notably the large, thumb-sized polychaetes in the family Alvinellidae (Figure 4), that are especially tolerant of high temperatures and that compete with desert ants for the

(c) 2011 Elsevier Inc. All Rights Reserved.

154

HYDROTHERMAL VENT ECOLOGY

Guinness Book of World Records in the category of multicellular animal living at a thermal extreme. Alvinellids live on the sides of black smoker sulfide chimneys, and are reported to survive brief exposures to temperatures as high as 1051C. They routinely experience a thermal gradient of 50–601C over the length of their bodies. Temperature tolerance in these worms is not completely understood, but thermal stability of enzymes has been shown to increase in alvinellid species that occupy increasingly warmer habitats. Membranes may also be adapted for thermal stability through an increase in the degree of double bonding in fatty acids in species living in warmer portions of the vent habitat. Shrimp (Rimicaris exoculata) that swarm on black smoker chimneys in close proximity to extremely high temperature (3501C) fluids are not bathed in excessively hot water, but they do have paired photoreceptive organs that may function to detect the glow emitted by the hot water. The organs are derived from ordinary shrimp eyes, but the photoreceptive surfaces are hypertrophied and rich in visual pigment (rhodopsin). The eyestalks are lost and the derived eyes extend back along the dorsal surface of the shrimp, beneath the transparent carapace. These ‘eyes’ have also lost their optic (lens) systems, so they cannot form an image, but they are

Figure 4 The alvinellid polychaete Alvinella caudata, beside its fragile tube. (Photograph by J. Porteous, Woods Hole Oceanographic Institution.)

optimized to detect gradients of dim light. While light from black smokers has now been well documented, the behavioral response of the shrimp to this light has not been studied.

Community Dynamics There could hardly be a greater dynamic contrast than that found between the cold, food-limited, relentlessly stable and vast deep sea environment, and the thermally complex, trophically rich, ephemeral, and insular deep-sea hydrothermal vent fields. Since the discovery of vents, ecologists have attempted to predict the cycle of community development over the life span of a vent field by interpolation and extrapolation from snapshot observations. The 1991 seafloor volcanic eruption at the Venture Hydrothermal Field on the East Pacific Rise was witnessed within days to weeks of the event, providing biologists with the first submarine equivalent of Krakatau. At the time of the Venture eruption, existing vent communities were obliterated by fresh lava flows, pervasive warm-water venting was observed along B1.5 km of ridge axis, and a dense ‘bloom’ of flocculent material poured from cracks between lobes of lava, obscuring visual navigation (Figure 5). The flocs were determined to be the mineral by-product of microbial production. Within one year, venting became focused at numerous sites along the ridge axis and the first colonists had arrived, including a small tubeworm species (Tevnia jerichonana) and dense aggregations of several species of limpets. Mobile vent organisms, including vent crabs and squat lobsters, zoarcid fish, and swarming amphipods were also well represented. After 2.5 years, some vents had shut off but, where venting persisted, mature colonies of the giant tubeworm (Riftia

Figure 5 Flocculent material suspended in the water column 1 m above new ocean crust. The floc is a mineral by-product of microbial production and emanated from cracks in the seafloor. (Photograph by Alvin, Woods Hole Oceanographic Institution.)

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT ECOLOGY

pachyptila) were established, along with a variety of smaller invertebrates that live among the tubeworm tubes, including shrimp, limpets, amphipods, and polychaetes. Growth rates of R. pachyptila were measured to be among the most rapid of any aquatic invertebrate. Within 5 years, 75% of the regional species pool could be found at the new vents and mussel beds were well-established and beginning to overwhelm the R. pachyptila thickets. In this example, vesicoymid clams were the last of the big megafaunal species to arrive at the site, despite the presence of adult populations within several kilometers of the area overrun by fresh lava. Mussels appear to have a competitive edge among the larger taxa at vents, in part because they are mobile and can relocate as necessary to cope with changing flow patterns of vent fluids, while tubeworms are stationary and have few options for tracking vent flow. In addition, because mussels can filter-feed as well as derive nutrition from their endosymbiotic bacteria, they are among the last species to disappear as a vent shuts down. Ultimately, it is the mobile scavengers – the crabs and fish and octopus – that witness the final demise of a vent field. The eruption at the Venture Hydrothermal Field was not entirely unexpected – geologists were studying this region of the ridge axis because it was so inflated and appeared to be ripe for an eruptive event. Localization of seafloor eruptions on the Juan de Fuca Ridge in the northeast Pacific (off the coast of Vancouver Island) now takes place in real-time, facilitated by a legacy of the cold war era, namely through acoustic signals received by underwater sound-surveillance systems originally designed to track enemy submarines. As eruptions take place, T-waves (low-frequency, tertiary waves characteristic of lava in motion) are transmitted into the water column. Navy hydrophone networks allow the T-waves to be placed in a geographical context and traced from start to finish. One migration of lava in June 1993 was traced for about 40 km below the ocean crust over a two-day period before it finally erupted onto the seafloor. As at the Venture Field, fresh lava on the seafloor was observed, along with venting of flocculent material derived from microbial production. Sites of persistent venting were colonized by populations of vent invertebrates within one year, some of which were reproductively mature. The two examples of community dynamics cited here suggest that the species composition of all vent communities is constantly changing. But some vent sites are long-lived, and repeat visits to these longlived sites over a 15-year period document essentially no change in the nature of the fauna. The best

155

example of this to date is the TAG site on the MidAtlantic Ridge. TAG is a large sulfide mound (100 m diameter) that has occupied the junction of crosscutting faults and fissures for more than 100 000 years. Swarming shrimp (Rimicaris exoculata) and anemones (Maractis rimicarivora) dominated the site when it was first discovered in 1985 and continued to dominate through the most recent set of observations in 1998, despite massive disruptions of hydrothermal flow caused by drilling in 1994. There have been shifts in the precise location of the primary masses of shrimp as they track local natural and anthropogenic changes in vent flow on the mound, and we assume that there has been replacement of generations by recruitment. But there has been no succession observed, no invasion by other taxa. The absence of change does little to attract ecologists, who thrive on dynamic systems, but the stability of the species composition at TAG and other sites has profound implications regarding the selective pressures encountered by the species that inhabit these sites compared to sites that are constantly threatened by lava overruns or tectonic shifts in plumbing.

Origins of Vent Faunas, Biogeography, and Biodiversity Evolutionary paths that brought invertebrate taxa to vents are varied. Several major taxa, including the galatheid squat lobsters, pycnogonid sea spiders, and echinoderms are likely to be immigrants from the surrounding deep sea. Some species are closely related to (and presumably derived from) shallowwater genera. Many of the most familiar vent taxa – the vesicomyid clams, bathymodiolid mussels, vestimentiferan tubeworms, alvinocarid shrimp – are allied to genera and families found in a variety of deepsea chemosynthetic ecosystems (i.e., seeps and whale-falls as well as vents). The direction of invasion (seep-to-vent or vice versa) can be inferred using molecular techniques. For example, molecular phylogenetics suggests that the bathymodiolid mussel group invaded vents from seeps, with several seep species resulting from reinvasion of the seep habitat by vent ancestors. There are also specialized taxa so far known only from hydrothermal vents. The most conspicuous of these is the alvinellid polychaete family, whose members often occupy the warmest habitable waters of a vent site. Still other taxa appear to be relicts of ancient lineages that have found refuge in the vent environment. These relict taxa include the stalked

(c) 2011 Elsevier Inc. All Rights Reserved.

156

HYDROTHERMAL VENT ECOLOGY

1 cm

Figure 6 The ancient barnacle Neolepas zevinae. (Photograph by C.L. Van Dover.)

Present

barnacle, Neolepas zevinae (Figure 6) and the archaeogastropod limpet, Neomphalus fretterae. Fossil vent communities found on land provide a glimpse of Silurian vent assemblages. Vestimentiferans are reported from some of the oldest vent deposits known, accompanied by brachiopods and monoplacophorans. Brachiopods and monoplacophorans are so far poorly represented in modern vent communities, if at all. When vents were first discovered, some biologists hypothesized that the vent fauna would be globally cosmopolitan. Subsequent explorations have shown this hypothesis to be false, and mechanisms that allow isolation and differentiation of faunas have been postulated. One of the best examples comes from a comparison of faunas from the Juan de Fuca Ridge and East Pacific Rise. At one time (56 Ma), these ridges were one continuous ridge system, but around 37 Ma the North American Plate began to

56 Ma

Kula North America

Pacific Pacific

65 Ma

Farallon

110 Ma

Kula Farallon

Izanagi Farallon Pacific

Pacific

Figure 7 Bisection of a mid-ocean ridge by the North American Plate. At one time (56 Ma), there was a continuous mid-ocean ridge in the eastern Pacific basin but, as the North American Plate overrode the ridge system, it was bisected to form the northeast Pacific ridge system (Juan de Fuca, Explorer, Endeavour, and Gorda Ridges) and the East Pacific Rise (emerging in the Gulf of California and running south toward Antarctica). (Reproduced from Tunnicliffe et al. 1996.)

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT ECOLOGY

override this ridge (Figure 7), splitting the parent faunal assemblage into two daughter assemblages that are distinct yet closely allied at the generic level. Similarities among faunas at the species level may be reduced to nearly zero at vent sites that are in separate ocean basins. This observation suggests that major additions to the global vent faunal inventory await us in the unexplored ocean basins; the potential for discovery of novel adaptations to the vent ecosystem is extremely high. Vent biologists are just beginning to examine global patterns in biodiversity. Preliminary measures show a strong correlation between spreading rate and species richness, raising the hypothesis that patterns of volcanism may be an ultimate control on species diversity in hydrothermal vent ecosystems. Where ridges are fast-spreading, the magma budget is high and the temporal and spatial frequency of vents along the ridge axis is high. In contrast, on slow-spreading ridges, the magma budget is low and vents are far apart. Because distances between vents are short on fast-spreading ridges, species are less likely to go extinct. At slow-spreading ridges, allopatric speciation by isolation may be favored, but species are more likely to go extinct because of the distance between vents. Many of the species that dominate vents on slow-spreading systems have mobile rather than sessile adults, suggesting that distance may act as a selective filter for dispersal capability in these systems.

Vent Systems and the Origin of Life Without doubt, the most provocative consequence of the discovery of seafloor hydrothermal vents is the suggestion that vents may have been the site where life originated on Earth. In contrast to the heterotrophic hypothesis of the origin of life, with its nourishing ‘organic soup’, the vent theory suggests that the earliest life was chemosynthetic, taking biochemical advantage of the large degree of chemical disequilibrium associated with mixing zones of low- and high-temperature portions of hydrothermal systems. A case has been made that the thermophilic nature of the most ancient known lineages of life is indicative of an origin of life at hot springs, but this argument finds limited support, since these most ancient cell types are still very complex and far removed from the progenitors of life. One model for the origin of life at vents suggests the following acellular precursor: a monomolecular, negatively-charged organic layer bonded to positively charged mineral surfaces at the interface of hot water. In this model, pyrite, which forms exergonically from

157

iron monosulfide and hydrogen sulfide (both components of vent fluids), serves as the mineral surface. The pyrite-forming reaction yields two free electrons that can be used for building biochemical constituents; simple organic molecules interacting with the pyrite could be reduced to more complex organic molecules. The simple organics also derive from the vent fluids, through abiogenic synthesis. In a secondary stage of development, the precursor evolves to being semicellular, still supported by minerals, but with a lipid membrane and internal broth, with increasing metabolic capabilities. In the final stage of origin, the pyrite support is abandoned and true cellular organisms arise. The elegance of this model is that it uses an energetically realistic inorganic chemical reaction to create a cationic substrate that can bind with organic compounds, all within a single setting. Discovery of vent ecosystems and the appreciation of their chemosynthetic basis has influenced the search for life elsewhere in the solar system. When the Viking Mission to Mars took place, the emphasis was on a search for photoautotrophic processes, but now the search for evidence of past or extant life on other planets highlights environments where chemosynthetic processes may take place, including hydrothermal areas.

Closing Remarks Hydrothermal vent ecology remains a field ripe for discovery of novel faunas and adaptations, as vents in new ocean basins are explored. Because access to deep-sea ecosystems is continually improving, biologists can now undertake quantitative sampling and time-series investigations that are certain to reshape our understanding of the physiological ecology, population biology, community dynamics, and biogeography of vent faunas in the near future.

See also Hydrothermal Vent Biota. Hydrothermal Vent Fluids, Chemistry of. Hydrothermal Vent Deposits. Hydrothermal Vent Fauna, Physiology of. MidOcean Ridge Tectonics, Volcanism, and Geomorphology.

Further Reading Bock GR and Goode JA (eds.) (1996) Evolution of Hydrothermal Ecosystems on Earth (and Mars?). New York: Ciba Foundation.

(c) 2011 Elsevier Inc. All Rights Reserved.

158

HYDROTHERMAL VENT ECOLOGY

Childress JJ and Fisher CR (1992) The biology of hydrothermal vent animals: physiology, biochemistry, and autotrophic symbioses. Oceanography and Marine Biology Annual Review 30: 337--441. Desbruye`res D and Segonzac M (eds.) (1997) Handbook of Deep-Sea Hydrothermal Vent Fauna. Brest: IFREMER. Fisher CR (1990) Chemoautotrophic and methanotrophic symbioses in marine invertebrates. Critical Reviews in Aquatic Science 2: 399--436. Humphris SE, Zierenberg RA, Mullineaux LS, and Thomson RE (eds.) (1995) Seafloor Hydrothermal Systems: Physical, Chemical, Biological and Geological Interactions. Washington, DC: American Geophysical Union. Karl DM (ed.) (1995) The Microbiology of Deep-Sea Hydrothermal Vents. New York: CRC Press. Shank TM, Fornari DJ, Von Damm KL et al. (1998) Temporal and spatial patterns of biological community

development at nascent deep-sea hydrothermal vents (91N, East Pacific Rise). Deep-Sea Research 45: 465--516. Tunnicliffe V (1991) The biology of hydrothermal vents: ecology and evolution. Ocenography and Marine Biology Annual Review 29: 319--407. Tunnicliffe V, Fowler CMR, and McArthur AG (1996) Plate tectonic history and hot vent biogeography. In: MacLeod CJ, Tyler PA, and Walker CL (eds.) Tectonic, Magmatic, Hydrothermal and Biological Segmentation of Mid-Ocean Ridges, Geological Society Special Publication 118, pp. 225–238. Tyler PA and Young CM (1999) Reproduction and dispersal at vents and cold seeps: a review. Journal of the Marine Biology Association of the UK 79: 193--208. Van Dover CL (2000) The Ecology of Deep-Sea Hydrothermal Vents. Princeton: Princeton University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FAUNA, PHYSIOLOGY OF A. J. Arp, Romberg Tiburon Center for Environment Studies, Tiburon, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1242–1246, & 2001, Elsevier Ltd.

Hydrothermal Vent Environments are Dynamic, Hot, and Toxic The hydrothermal vent environments, lying at the bottom of the ocean at depths of 2.5 km or more, were discovered in 1977 by a group of geologists exploring spreading centers at midocean ridges on the sea floor. As fissures open up in the earth’s surface, lava is extruded onto the ocean floor and sea water is pulled towards the center of the earth so deeply that it comes into contact with hot, molten magma. The sea water is superheated and then discharged back into the environment through fissures in the ocean floor. As sea water moves from the center of the earth and into the vent habitat on the seafloor, it becomes laden with inorganic chemicals. In particular, hydrogen sulfide, an essential chemical in this unique environment, leaches into the water at depth, and water discharging into the environment can be highly enriched in this toxic but energy-rich chemical. There are two types of hydrothermal vents. In those characterized by diffuse venting, sea water percolates out at a moderate rate and is approximately 10–201C in temperature. As the bottom water temperature of the majority of the earth’s deep ocean is about 21C, these hydrothermal fluids are elevated in temperature, but rapidly mix with the surrounding sea water. One of the first hydrothermal vent environments discovered, Rose Garden at the hydrothermal vent environment near the Galapagos Islands, is an example of a diffuse vent habitat. In the black smoker environment of the hydrothermal vents, things are a lot hotter, such as at those on the Juan de Fuca Ridge off the coast from the state of Washington. This is a very dynamic, high-temperature environment where water issuing forth from large chimneylike structures can be as hot as 4001C in temperature. In spite of the hydrogen sulfide-enriched environment, elevated temperatures, and the dynamic volcanic activity, numerous and varied animals cluster around these sites, taking advantage of the hard substrate provided by the extruded pillow lava. A

typical vent environment begins with a stretch of pillow lava in the foreground, with clams wedged into the numerous fissures. At the heart of the vent environment are diffuse venting hydrothermal fluids or actively spewing white smoker chimneys and black smoker chimneys, teeming with life. Typical inhabitants include dense clusters of tubeworms and many free-ranging animals roaming in and out of the vent environment such as brachyuran crabs, galatheid crabs, numerous amphipods, a few species of fish, and a host of other smaller animals.

Chemosynthesis — The Basis of All Life in the Vent Environment Possibly the most revolutionary outcome of the discovery of hydrothermal vents is the story of how life exists in this challenging habitat and the unique nature of the food chain and the source of basic energy in this remote location. Prior to the discovery of the hydrothermal vents, most biologists believed that all life depended upon the energy of sunlight and that the basis of all food chains was photosynthesis. When the hydrothermal vents were discovered, it was immediately clear that they represent a very enriched biological environment very remote from the surface sunlight. It was difficult to imagine that organic material could drift down in large enough quantities to provide the energy to fuel this environment. Other interesting data materializing rapidly after the discovery of the vents indicated that most of the animals, especially the large invertebrates, have no digestive systems. For example, the large tubeworm, Riftia pachyptila, has no mouth and no intestine; however, it is a large animal approximately 1.5 m in length and up to 2 cm in diameter, and large colonies of these animals flourish in the remote vent habitat. The question of how animal life is supported in the deep sea communities of the hydrothermal vents became, and remains to this day, a focus of intense research. Several lines of evidence lead to the realization that some of the major invertebrates endemic to the vent environment harbor bacteria within their body cavities. Free-living bacteria in this environment and in our own backyard have been known for years to be able to use chemical energy as a basis of their metabolism. In the case of the free-living bacteria, there are many hydrogen sulfide-oxidizing bacteria that can use this chemical as the basis of

(c) 2011 Elsevier Inc. All Rights Reserved.

159

160

HYDROTHERMAL VENT FAUNA, PHYSIOLOGY OF

their metabolic pathways and produce organic compounds that form the basis of their nutrition. Hydrothermal vent animals harboring these chemical-utilizing bacteria within their body tissues as symbionts include the large tubeworm Riftia pachyptila, which has dense aggregations of bacteria in a residual gutlike organ, and the clam Calyptogena magnifica, which harbors bacteria in the gills. These symbiotic bacteria are able to utilize the inorganic chemical hydrogen sulfide, so plentiful in this environment, in a manner analogous to what plants do with energy from the sun. This process is therefore termed chemosynthesis rather than photosynthesis. The symbiotic bacteria living within the bodies of the larger invertebrate animals have been demonstrated to oxidize hydrogen sulfide, and the energy released from this biochemical process is used to power the fixation of carbon dioxide into small organic compounds – just as in free-living bacterial sulfide oxidation. The Calvin–Benson cycle employed in both cases is the same metabolic pathway that is utilized by plants in photosynthesis to transform inorganic carbon dioxide into organic compounds that are then utilized as food higher up in the food chain. The critical difference with chemosynthetic metabolism is that, rather than using sunlight, these animals and bacteria utilize chemical energy to power that reaction and are completely independent of sunlight (Figure 1). The net result is that freeliving bacteria in the environment and symbiotic bacteria living within animal tissues are able to live independently of sunlight by utilizing chemicals from the core of the earth, thus forming a very different basis for the food chain in the hydrothermal vent environment. The discovery of the hydrothermal vent environment was a fundamental discovery of a well-defined ecosystem that is completely independent of sunlight at any level of the food chain.

Figure 1 Chemosynthetic pathways in Riftia pachyptila.

The Ecophysiology of the Giant Tubeworm Riftia pachyptila One of the most dramatic and best-known of the animals endemic to the hydrothermal vent environment is the giant tubeworm Riftia pachyptila. Colonies of these worms are clumped together around effluent points in the hydrothermal vent habitat, growing toward and into the water that is percolating out from the seafloor. An individual animal lives inside a single, unbranched chitinous tube and the red structure protruding out of the end of the tube is the respiratory plume. The animal can retract the plume back into the tube if disturbed by a roaming predator. There is a collarlike vestimentum organ that positions the animal within the tube, and a large trunk region of the animal is filled with an organ termed the trophosome. This organ is believed to be the vestigial gut of the worm and is composed, literally, of masses of bacteria. A pool of coelomic fluid bathes the trophosome and contains a largemolecular-weight extracellular respiratory hemoglobin (Figure 2). Riftia pachyptila also has a separate pool of blood that contains high concentrations of an extracellular hemoglobin that circulates in an elaborate closed circulatory system powered by a heartlike structure in the vestimentum region. The blood is pumped in a complete circuit

Figure 2 The external anatomy of Riftia pachyptila.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FAUNA, PHYSIOLOGY OF

161

from the respiratory plume to body tissues, and on to elaborate capillary beds in the region of the trophosome and bacteria (Figure 3). The red color of the blood is due to the high concentration of hemoglobin and gives the characteristic red color to the plumes. The respiratory plume and the circulating hemoglobin are essential for the transport of the key metabolites oxygen, hydrogen sulfide, and carbon dioxide, which are the principal components of the metabolism of the symbiotic bacteria. The respiratory hemoglobins present in the plume and the coelomic fluid of the animal bind oxygen with a very high affinity. The binding is reversible and cooperative, such that oxygen uptake is enhanced at the respiratory plume, and oxygen delivery is augmented at the tissues and trophosome organ. Hydrogen sulfide is a highly toxic molecule that typically acts in a similar manner to cyanide by binding at the iron center of cytochrome molecules and hemoglobin molecules, thus arresting aerobic metabolism. Although Riftia pachyptila and other hydrothermal vent animals utilize hydrogen sulfide for their metabolism, they also have tissues that are highly sensitive to sulfide poisoning. Detoxification of hydrogen sulfide is essential for aerobic life in this

dynamic, chemically enriched environment. The key to the simultaneous needs for transportation and detoxification is the respiratory hemoglobin present in the plume and coelomic fluid. Riftia pachyptila hemoglobin binds hydrogen sulfide with a very high affinity. The toxic hydrogen sulfide is transported to the trophosome region in the center of the worm’s body as a tightly bound molecule that cannot chemically interact with sulfide-sensitive tissues. Oxygen and sulfide are simultaneously bound to the hemoglobin at separate binding sites and are transported to the trophosome, where they are believed to be delivered to the symbiotic bacteria for metabolism. In this way hydrogen sulfide is taken up from the surrounding sea water and transported to the site of bacterial metabolism while interaction is prevented with other tissues, such as the body wall, that are highly aerobic and sensitive to the toxic effects of hydrogen sulfide. These unusual adaptations function for respiratory gas transport and metabolism as well as for detoxification and tolerance of toxic chemicals in what would be a very inhospitable environment for most animals. Dense colonies of Riftia pachyptila flourish in a specialized microhabitat within the vent environment. The worms anchor themselves on the rocks where the hydrothermal vent fluid is issuing out into the seafloor. The base of the tube is bathed in hydrothermal fluid enriched in hydrogen sulfide and carbon dioxide, but devoid of oxygen. Temperatures are relatively elevated here, and a gradient develops along the length of the tube. The respiratory plume extends into ocean-bottom sea water that is 21C in temperature, devoid of hydrogen sulfide and enriched in oxygen (Figure 4). By occupying the interface between the hydrothermal fluids and the surrounding bottom water, animals are exposed to both of these essential metabolites which are then taken up by the circulating hemoglobin and

Figure 3 The internal anatomy of Riftia pachyptila.

Figure 4 The microhabitat of Riftia pachyptila.

(c) 2011 Elsevier Inc. All Rights Reserved.

162

HYDROTHERMAL VENT FAUNA, PHYSIOLOGY OF

transported to internal tissues and symbiotic chemosynthetic bacteria.

The Ecophysiology of the Giant Clam Calyptogena magnifica The vesicomyid clam Calyptogena magnifica is a common inhabitant of hydrothermal vents that orients in the fissures of the pillow lava near the periphery of the vent environment. Individual clams can reach up to 20 cm in length. The clams position themselves in cracks in the pillow lava and wedge their muscular foot down into the region where the hydrothermal plume is percolating out (Figure 5). This water is enriched in hydrogen sulfide and is elevated in temperature. This position orients the siphon end-up into surrounding sea water and enables the uptake of oxygen and carbon dioxide. Calyptogena magnifica possess a high concentration of an intracellular, circulating hemoglobin that functions for oxygen binding and transport. The clam’s blood also contains a separate component that binds hydrogen sulfide and transports this essential metabolite to internal symbionts in the gill region. In this manner, the clams are able to accumulate hydrogen sulfide from the hydrothermal fluids bathing the foot, and oxygen via binding to the circulating hemoglobin through gill ventilation at the siphon region. These essential metabolites are transported as bound substances to symbiotic chemosynthetic bacteria in the gill via the circulatory system of the clam, providing both essential respiratory gas transport and detoxification of toxic hydrogen sulfide. The vent clams, like the tubeworms, seek out, exploit, and flourish in a unique microhabitat in the hydrothermal vent community. Other hydrothermal vent animals include the mytilid mussel Bathymodiolus thermophilus that also harbors chemosynthetic bacteria in its gill tissues. There are many free-ranging animals that are

Figure 5 The microhabitat of Calyptogena magnifica.

not fueled by chemosynthesis but may feed on the larger invertebrates that benefit from that symbiotic, chemoautotrophic metabolism. Animals such as the brachyuran crabs, Bythograea thermydron, that wander through the environment scavenging dead or dying material, numerous swarming amphipods, as well as slow-moving fishes, are plentiful. Along with the giant tubeworms, clams, and mussels, these animals benefit directly or indirectly from the chemicalbased metabolism that supports this dynamic and robust deep-sea community. It is the specialized physiological adaptations for transport and detoxification of hydrogen sulfide and other processes essential for life that provide the underlying mechanisms that make this possible.

Summary A fascinating variety of marine invertebrates occur in dense assemblages in organically enriched deep-sea hydrothermal vent environments. Intensive studies on the hydrothermal vent fauna have been conducted since their discovery in the late 1970s. Many investigations have focused on the fact that these organisms, including vestimentiferan tubeworms and vesicomyid clams emphasized here, are nutritionally dependent upon the chemical-based metabolism of large populations of symbiotic bacteria that they harbor internally in dense concentrations. These bacteria utilize hydrogen sulfide as an energy source to fix inorganic carbon into nutrients. Hydrogen sulfide is extremely toxic to aerobic organisms in nanomolar to micromolar concentrations. However, uptake and transport of sulfide to internal symbionts is essential for the host animal’s metabolism and survival. Chemical-based metabolism, or chemoautotrophy, and detoxification of sulfide through binding to blood-borne components occur in vent tubeworms and clams, and are particularly well-characterized for the tubeworm Riftia pachyptila. These chemosynthetic endosymbiont-harboring worms simultaneously transport sulfide bound to the respiratory hemoglobin, providing an electron donor for the bacterial symbiont metabolism and protection against sulfide toxicity at the tissues. How animals living in sulfide-rich environments like hydrothermal vent communities transport, metabolize, and detoxify hydrogen sulfide has been one of the major questions posed by hydrothermal vent researchers. The clarification of both the phylogenetic importance and the ecological status of these animals requires knowledge of their physiological adaptations to low oxygen conditions and high

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FAUNA, PHYSIOLOGY OF

concentrations of sulfide, and provides answers to the puzzle of how these animals flourish in such seemly hostile environments.

See also Deep-Sea Ridges, Microbiology. Hydrothermal Vent Fluids, Chemistry of. Hydrothermal Vent Biota. Hydrothermal Vent Deposits. Hydrothermal Vent Ecology.

Further Reading Arp AJ, Childress JJ, and Fisher CR (1985) Blood gas transport in Riftia pachyptila. In: Jones ML (ed.) The Hydrothermal Vents of the Eastern Pacific: An Overview, Bulletin of the Biological Society of Washington, No. 6, p. 289. Washington DC: Biological Society of Washington. Childress JJ and Fisher CR (1992) The biology of hydrothermal vent animals: physiology, biochemistry and autotrophic symbioses. Oceanography and Marine Biology Annual Review 30: 337. Delaney JR (1998) Life on the seafloor and elsewhere in the solar system. Oceanus 41(2): 10. Feldman RA, Shank TB, Black MB, et al. (1998) Vestimentiferan on a whale fall. Biology Bulletin 194: 116.

163

Fisher CR (1990) Chemoautotrophic and methanotrophic symbioses in marine invertebrates. Review of Aquatic Science 2: 399. Lutz RA (2000) Deep sea vents: science at the extreme. National Geographic (October): 116. MacDonald IR and Fisher CR (1996) Life without light. National Geographic (October): 313. Macdonald KC (1998) Exploring the global mid-ocean ridge. Oceanus 41(1): 2--9. Mullineaux L and Manahan D (1998) The LARVE Project explores how species migrate from vent to vent. Oceanus 41(2) Nelson K and Fisher CR (2000) Speciation of the bacterial symbionts of deep-sea vestimentiferan tube worms. Symbiosis 28: 1--15. Somero GN, Childress JJ, and Anderson AE (1989) Transport, metabolism, and detoxification of hydrogen sulfide in animals from sulfide-rich environments. Review of Aquatic Science 1: 591--614. Shank T, Fornari DJ, Von Damm KL et al. (1998) Temporal and spatial patterns of biological community development at nascent deep-sea hydrothermal vents (91500 N, East Pacific Rise). Deep-Sea Research II 45: 465. Tivey MK (1998) How to build a smoker chimney. Oceanus 41(2): 22. Van Dover CL (2000) The Ecology of Deep-Sea Hydrothermal Vents. Princeton, NJ: Princeton University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF K. L. Von Damm, University of New Hampshire, Durham, NH, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 2, pp 1246–1253, & 2001, Elsevier Ltd.

Introduction It was not until 1977 that we knew that fluids exit from the seafloor along the global midocean ridge system. With the first discovery of hydrothermal venting at the Galapagos Spreading Center, our ideas on how elements cycle through the oceans and lithosphere, and even how and where life on our planet may have originated were fundamentally and irrevocably changed. Although these fluids were only a few tens of degrees hotter than ambient sea water (o301C vs. 21C), on the basis of their chemical compositions it was immediately clear that these fluids were derived from reactions at much higher temperatures between sea water and the oceanic crust. Less than two years later, spectacular jets of hot (X3501C) and black water were discovered several thousand kilometers away on the northern East Pacific Rise, and ‘black smokers’ and ‘chimneys’ (Figure 1) entered the oceanographic lexicon. From a chemical oceanography perspective, hydrothermal venting and hydrothermal vent fluids provide both new source and sink mechanisms for elemental cycling in the ocean, and therefore possible resolutions to a number of the outstanding chemical flux imbalances. They therefore play a fundamental role in regulating the chemistry of the oceans through geological time. From a biological oceanography perspective, hydrothermal vents provide us with new ecosystems based on chemosynthetic, rather than photosynthetic energy. Of the 4500 new species discovered at these sites, the archea and other microbiological components are attracting increasing interest for both biotechnological applications and ‘origin of life’ questions on our own and other planets. From a physical oceanographic perspective, hydrothermal vents provide an input of both heat and materials into the oceanic mid-depth circulation. From a geological oceanographic perspective they provide an efficient means of removing heat from newly formed oceanic crust, as well as a means of altering the elements recycled into the mantle when the oceanic crust formed at spreading centers is later subducted back into the Earth’s interior.

164

Where Are Hydrothermal Vents Found? Hydrothermal vents are now known to exist at approximately 30 locations on the global midocean ridge system (Figure 2). Initially some people had speculated that venting would only be found on intermediate- or faster-spreading ridges (i.e., ridges with full spreading rates of at least 60 mm yr1); we now know of numerous locations on slow-spreading ridges (e.g., Mid-Atlantic Ridge) where they occur. Known sites occur at depths from 800 to 43600 m, with spreading rates from o20 to 4150 mm y1 (full rate), on both bare basalt and sedimented-covered ridges, as well as on seafloor where ultramafic rock types are known to outcrop, and at temperatures up to 4051C. If one looks at the global distribution of known vent sites, it is obvious that many of the sites are in relatively close proximity to nations that operate submersibles, and are clustered disproportionately in the north Atlantic and eastern Pacific. Although not strictly part of the midocean ridge system, venting associated with back arc spreading centers is also known from a number of sites in the western Pacific. No sites have yet been discovered in the Indian Ocean, although cruises to this ocean are now planned. Similarly, no sites are known in the south Atlantic, or at high latitudes. This is an exploration issue, not a lack of their existence in these areas. While initially vents were thought to occur at the mid-point of ridge segments, this was a largely selffulfilling prophesy, as this is where exploration for them was focused. There is increasing evidence that more venting occurs on the magmatically robust portions of the ridge, rather than on those areas that are deemed to be magma-starved, on the basis of their morphological characteristics.

How Are They Found? Vent fields have been discovered in numerous ways, but surveys of the overlying water column and camera tows are the most common systematic approaches employed today. The venting of hot/warm water often forms a plume with unique temperature, salinity, reduced light transmittance, and other specific chemical signals several hundred meters above the ridge. The presence of such a plume is often the first indication that a given section of ridge is hydrothermally active. Cameras, or other

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

(B)

(A)

(C)

165

(

D)

(E) Figure 1 Pictures of black smokers. (A)–(C) Brandon Vent a black smoker with 4051C fluid temperatures on the Southern East Pacific Rise at 211330 S at a depth of 2834 m; in (A) and (B), shown instrumented with a Hobo recording temperature probe and in (C) with one of Alvin’s manipulators. (D) Nadir vent at 171260 S at a depth of 2562 m and fluid temperatures of 3431C shown prior to sampling in 1998. In the foreground is Alvin’s basket with water sampling equipment. (E) Water sampling of Nadir vent fluids.

(c) 2011 Elsevier Inc. All Rights Reserved.

166

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

Figure 2 Global distribution of known (sampled) hydrothermal systems. (Modified from Baker et al. (1995).)

deep-towed survey vehicles may also find visual evidence for communities of vent-specific organisms, sulfide structures (both active and extinct), and murky water. These surface ship surveys may then lead to the use of an ROV (remotely operated vehicle) or submersible, in what has been called a ‘nested survey strategy.’ Sites have also been found by the fortuitous dredging up of both pieces of hydrothermal chimneys and vent-specific animals.

Controls on Fluid Compositions Hydrothermal vent fluids are primarily, if not entirely, sea water that has been altered due to reaction

within the oceanic crust at high temperatures (at least 3501C, and more likely 44001C) and elevated pressures (at least 80 bar, and more typically at least 250 bar). The two primary controls on vent fluid compositions are phase separation and water–rock reaction (Figure 3). Unlike pure water, sea water is a two-component system, containing both water and salt (primarily NaCl, especially for hydrothermal fluids from which the magnesium and sulfate have been essentially quantitatively removed). The critical point for sea water is higher than that for distilled water – 4071C and 298 bar rather than 3741C and 220 bar – and most importantly the two-phase curve therefore continues beyond the critical point

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

Sea water Oceanic crust 4

CaSo 2+

Mg

W_R rxn 2+

+

Ca , H W_R rxn n

io

t ra pa

Vapor e

s ha

Brine

se

P

W_R rxn

Gas input Heat source Magma lens or dike

Figure 3 Schematic of a hydrothermal flow cell. W– Rrxn ¼ water–rock reaction. (Von Damm 1995 in Humphris et al.)

(Figure 4). If fluids intersect the two-phase curve at pressure and temperature conditions less than the critical point, they will undergo subcritical phase separation or boiling. The vapor phase formed by

0 50

Vapor + halite

100

Pressure (bar)

150 200 250 300

Liquid

CP Liquid + vapor

350

167

this process will contain some salt, the exact amount and composition of which will depend on the pressure and temperature conditions at which the phase separation occurred, and will be enriched in dissolved gases. If the fluid intersects the two phase curve at pressure and temperature conditions higher than the critical point, supercritical phase separation or condensation will occur, and, rather than a lowsalinity vapor being formed, a high-salinity brine or liquid phase will condense. In most, if not all, hydrothermal fluids that have been sampled, phase separation has occurred as evidenced by the chlorinities of these fluids, which can be from B6% to 200% of those in sea water. While the change in the absolute chlorinity of seafloor hydrothermal fluids is primarily the result of phase separation, changes in the other dissolved ions in sea water also reflect substantial reaction of the fluids with the rock, mostly basalt, but also ultramafics and sometimes sediments. Seafloor hydrothermal fluids therefore do not show the ‘constancy of composition’ and elemental ratios known to characterize the major chemical species in sea water. The two primary mechanisms for determining the composition of hydrothermal fluids are therefore phase separation and water–rock interaction. This does not mean that other mechanisms may not be important. Two that have been identified but whose global importance is not yet known are magmatic degassing and biological uptake/removal. On the East Pacific Rise at 91500 N latitude unusually high gas concentrations and gas/heat ratios have been observed in the hydrothermal fluids that suggest that we are seeing the degassing of a recently resupplied magma chamber. This has not yet been observed at any other sites, although it has been observed for B7 years at this site. The potential biological effects on fluid compositions have long been speculated on, but there are few quantitative data at this time that can directly address this question. As the ‘black smoker’ fluids are at temperatures well beyond the known bounds of life on this planet, if biological effects are present they are most likely to be found in fluids with temperatures less than B1101C.

400

The Division of Fluids by Temperature and Styles of Venting

450 500 250

300

350

400

450

500

Temperature (°C) Figure 4 Two-phase curve for sea water, showing the location of the critical point (CP) for sea water (4071C, 298 bar), and the relative locations of the liquid, liquid þ vapor, and vapor þ halite stability fields. Note that, unlike for pure water, the two-phase curve continues beyond the critical point.

Hydrothermal fluids are often subdivided into two categories: high-temperature or focused flow, and lowtemperature or diffuse flow. These terms are often poorly defined in the contexts in which they are used, and may mean different things to different authors. High-temperature fluids, generally 4 200–2501C, are usually focused jets of water that are exiting from

(c) 2011 Elsevier Inc. All Rights Reserved.

168

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

constructional features that either are precipitated from the hydrothermal fluids themselves when they encounter cold, alkaline sea water or are formed by precipitation due to the mixing of hydrothermal fluids with sea water, or, and more usually, both of these processes. Although these chimneys comprise metal sulfides and sulfates, other minerals are also present, usually in lesser quantities, although on sedimentcovered ridges carbonates, for example, may also be abundant in these chimney or mound structures. Lowtemperature fluids, generally p351C, but sometimes approaching 1001C, have usually deposited (if they ever contained) much of their metal load below the seafloor and hence usually neither smoke nor have specific constructional features associated with their fluid flow. Because of this lack of specific structure, the fluid flow is often less organized, and hence is referred to as ‘diffuse.’ Some authors use this term to refer to fluids exiting directly from the basaltic substrate. These sites are usually well-colonized by various types of vent megafauna, making it difficult, if not impossible, to identify a specific orifice. Other authors use this same term to refer to fluids that are oozing out of large sulfide structures. At some sites, only hightemperature ‘black smoker’ vents are found, while at other sites only the lower-temperature diffuse venting is found (as at the Galapagos Spreading Center where venting was first discovered). At other sites, both focused and diffuse flow may occur right next to one another, and, at least at one site, we have now observed low-temperature diffuse flow evolving into high-temperature focused flow over several years. For all types of ‘low temperature and/or diffuse flow,’ all of the fluids that have been sampled to date appear to contain some fluid that has reacted at significantly greater temperatures than are measured directly in these fluids. Hence one could argue that, at least at the ridge axes, all fluids are ultimately high-temperature fluids and that some of them have undergone significant mixing with sea water at some depth, likely relatively shallow, within the oceanic crust. With the continued inability to drill these systems, our knowledge of their subsurface hydrology is rudimentary at best. One might note that fluids with temperatures in the B100–2001C range have been left out of the above classification. This is because there are few fluids that have been sampled in this temperature range, which may reflect a true lack of abundance of fluids at such temperatures, or simply a sampling gap.

were chemically stable for years at a time, yet each individual vent had a distinctive chemistry, and rapid changes in the hydrothermal plumes above these sites had been noted. In 1991 some of this puzzle was resolved with the first opportunity to sample a midocean ridge hydrothermal site immediately following a volcanic eruption and to study the evolution of this site on timescales now approaching decadal. The discovery of an eruption on the East Pacific Rise at 9145–520 N was fortuitous, but it could be sampled and documented with the DSV Alvin within weeks of the volcanic event. Since that time using the US Navy’s SOSUS system, it has been possible to determine real-time when some of these events are occurring on the Juan de Fuca and Gorda Ridges, but immediate response to these events have been limited to surface ship observations due to logistical constraints. The events studied so far have been characterized by a volcanic eruption at the seafloor, the intrusion of a dike, or both. The knowledge gained from responding to these events has revolutionized our understanding of these systems, especially their pronounced temporal variability on very short timescales (much less than days to weeks). Presumably we will be able to study more of these events and to gain a sense of their frequency, perhaps as a function of spreading rate, as we begin to instrument more of the ridge crest with hydrophones to detect the T-phase signals associated with these events. The Influence of Tectonic/Cracking Events

Magmatic events have provided new insights into the processes that drive hydrothermal systems and their fluid compositions on intermediate- to fast-spreading ridges, but presumably tectonic events are more important (or at least more frequent than magmatic events) on slow-spreading ridges. As we have not yet been able to observe one of these events, we cannot assess their importance. In a relatively small cracking event on a fast-spreading ridge observed with a seismic array, changes in fluid temperatures were marked, and changes in fluid compositions were profound, leading to major changes in the biological communities existing at this site. Presumably the observation of one or more tectonic events on slowspreading ridges will also provide critical new insights into how hydrothermal systems on these types of ridges function and evolve. On-Axis versus Off-Axis

The Influence of Volcanic Events

In the late 1980s there was a real dichotomy in our understanding of the controls on seafloor hydrothermal systems. We knew of individual vents that

The discussion above has focused on the axial component of midocean ridge hydrothermal systems. Many debates have focused on the importance of these axial systems compared to hydrothermal

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

systems on the ridge flanks. The net amount of heat transported and the total geographical extent of the flanks is much larger than that of the axial parts of the ridges, the division often being drawn at 1 My, resulting in a partitioning of 70%:30% relative heat loss. Hence, the relative chemical anomaly and volume of fluids expelled on the axis versus the flanks of the ridge will determine which of these parts of the midocean ridge hydrothermal system is most important for the cycles of individual chemical elements and species. It has been difficult to acquire quantitative data on the fluids exiting from the flanks, one of the reasons being that sites of flank fluid venting are difficult to identify because water column plumes are not present, nor are the distinctive animal communities (distinctive both for their animal types as well as their white color on the black basalt) found at the axial sites of venting. In the last several years drilling by the Ocean Drilling Program on the flanks of the Juan de Fuca Ridge has provided important new information on the chemistry and hydrology of these systems, as well as demonstrating a viable method for further approaching the flank question.

Table 1

Observed Fluid Compositions Compositions of hydrothermal fluids not only vary widely but also are almost always very different from those of sea water (Table 1). While some of the low temperature diffusely venting fluids may be close to sea water in their major element compositions, they will often have very different compositions of dissolved gases (e.g., H2S, CO2, CH4, H2, and He) and will usually be highly enriched in iron and manganese compared to local ambient sea water. Compared to sea water, hydrothermal fluids have lost essentially all of their magnesium and sulfate, and are highly enriched in H2S, CH4, H2, He, Si, Li, Fe, and Mn. As hydrothermal fluids are very acid, they also have no alkalinity, and with the loss of sulfate, chloride becomes the major, and almost only anion (bromide is present in much lower concentrations). The behavior of the cations is more variable. As the amount of chloride present is a result of the phase separation history of the fluids, and the fluids must maintain electroneutrality, to determine whether a particular cation has been added to or removed from the fluid,

Range of physical parameters and chemical compositions for hydrothermal vent fluidsa

Parameter

Unitsb

Vents

Seawater

Temperature Depth (pressure) pH 251C, 1 atm Alkalinitytotal Cl SO24 H2S Si Li Na K Ca Mg Sr Fe Mn Cu Zn Cd Co Pb B Al Br CO2 CH4 H2

1C m (bars)

4 2–405 800–3600 (80–360) 2.5–7.8  2.7–10.6 31–1245 0 0–110 2.7b–24.0 o0.012–2.35 o15–924 o1–58.7 o0.2–109 0 o1–348 0–18.7 0–4.48 0–310 0–900 o10–1000 o5–2570 50–2200 416–1630 0–20.0 29–1880 2.3–375 0.0003–6800 o0.001–38

B1–5 (bottom water)

a b

meq kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 nmol kg1 nmol kg1 nmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1 mmol kg1

169

Only includes mid-ocean ridge systems, both bare rock and sediment-covered. From high-temperature vents.

(c) 2011 Elsevier Inc. All Rights Reserved.

7.8 2.3 545 28 0 0.032–0.180 0.026 465 10 10 53 87 o0.001 o0.001 0.007 0.01 1 0.03 0.01 416 0.02 840 2.3 0.0003 0.0000003

170

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

not only the absolute concentration but also its ratio to chloride must be compared to the local ambient sea water value. Hydrothermal fluids are also very reducing, often containing large amounts of H2S, H2, and CH4. In most, but not all hydrothermal fluids (fluids that are formed immediately after a volcanic eruption being a major exception), Li, K, Rb, Cs, Ca, Sr, Si, the transition metals in reduced forms, including Fe, Mn, Cu, and Zn, are enriched in hydrothermal fluids, the cause being water–rock interaction. Sodium may be either enriched or depleted with respect to the chloride content of the fluids, the loss being due to albitization, with a concomitant gain in the calcium-content of the fluids. It is the loss of magnesium to form magnesiumhydroxy silicates that, along with other aluminosilicate reactions, generates and then maintains the acidity of the fluids. In the case of sulfate, some is lost as anhydrite (CaSO4) in the downflow zone, while some is reduced to sulfide (as H2S). The compositions of fluids exiting from a single hydrothermal vent may vary widely over time. The cases in which this has been observed are increasing, and are usually associated with vents where a known magmatic event has occurred. The variation in the composition of a single vent can vary from vapor to brine, and may encompass almost the entire range of known compositions. In contrast, some sites of venting are known where the fluid compositions have been stable during the time interval over which they have been sampled. None of these vents with constant compositions have known ‘magmatic’ or ‘tectonic’ events associated with them, although several of these sites have now been sampled over times of B15 years. Presumably, these vents are in a period of steady-state venting, although we do not have adequate constraints to determine how long after an eruptive event, or at what spreading rates, this may occur. While our data has increased on the temperature and chemical characteristics of vents during their early histories, few data exist for their waning stage(s). Presumably all this variability – or lack thereof – and the timescale(s) on which it occurs can ultimately be tied to the nature of the heat source at a given site. Aside from the most general characteristics related to the presence or absence of a seismic low-velocity zone, and the depth at which it occurs, little is known about the specifics of the heat sources at sites that are hydrothermally active, especially in contrast to those that are not. The composition of a hydrothermal fluid cannot be correlated to, or predicted by, such known physical parameters as the depth of the seafloor on which it occurs, the spreading rate of the ridge on which it occurs, and so on. As the fluid compositions in most

cases are probably due to the achievement of steadystate, if not true thermodynamic equilibrium, of the fluids with the rock substrate, some of the measured compositions can be tied to either the measured exit temperature or the presumed in situ conditions within the hydrothermal system itself. While there has been some success, especially recently, with understanding the chemical controls on these systems using thermodynamic modeling, a major limitation in many cases remains the proximity to the critical point of both pure water and sea water.

The Flux Question One of the driving questions for the study of seafloor hydrothermal systems is to understand their net flux to the ocean in terms of energy (thermal and chemical) and mass. The thermal, or heat, energy they carry is believed to be relatively well constrained, as various independent ways of estimating this flux provide similar values. The mass of chemicals they add and/or remove remains problematic. In some cases whether hydrothermal activity is a net source or sink for particular elements remains unresolved as well. In addition to absolute concentrations, or concentrations normalized to the chloride content, the isotopic signature of various species can also be used to constrain the source and sink terms, as well as helping to identify the important processes occurring within the hydrothermal circulation cell.

Summary and Conclusions Hydrothermal venting along the global midocean ridge system is a process that is widespread throughout the ocean basins and impacts all of the oceanographic disciplines. Our studies of these systems remain in their infancy, however, and we do not yet completely understand the controls on the chemistry of these systems, the controls on the locations of individual vent sites, their overall importance to ocean chemistry, productivity, and circulation, and their net effects on the structure and composition of the oceanic crust.

See also Hydrothermal Vent Deposits. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. MidOcean Ridge Geochemistry and Petrology. MidOcean Ridge Seismic Structure. Seamounts and Off-Ridge Volcanism.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYDROTHERMAL VENT FLUIDS, CHEMISTRY OF

Further Reading Cowen JP and Baker ET (1998) Topical studies in oceanography: detection of and response to mid-ocean ridge magmatic events. Deep-Sea Research II 45(12): 2503--2766. Elderfield H and Schultz A (1996) Mid-ocean ridge hydrothermal fluxes and the chemical composition of the ocean. Annual Review of Earth and Planetary Science 24: 191--224. Humphris SE, Zierenberg RA, Mullineaux LS, and Thomson RE (eds.) (1995) Seafloor Hydrothermal Systems: Physical, Chemical, Biological, and Geological

171

Interactions. AGU Monograph 91. Washington, DC: American Geophysical Union. Parson LM, Walker CL, and Dixon DR (eds.) (1995) Hydrothermal Vents and Processes. London: Geological Society: Geological Society Special Publication 87. Seyfried WE Jr (1987) Experimental and theoretical constraints on hydrothermal alteration processes at mid-ocean ridges. Annual Review of Earth and Planetary Science 15: 317--335. Von Damm KL (1990) Seafloor hydrothermal activity: black smoker chemistry and chimneys. Annual Review of Earth and Planetary Science 18: 173--204.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYPOXIA N. N. Rabalais, Louisiana Universities Marine Consortium, Chauvin, LA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: Definitions Hypoxic (low-oxygen) and anoxic (no-oxygen) waters have existed throughout geologic time. Presently, hypoxia occurs in many of the ocean’s deeper environs, open-ocean oxygen-minimum zones (OMZs), enclosed seas and basins, below western boundary current upwelling zones, and in fjord. Hypoxia also occurs in shallow coastal seas and estuaries, where their occurrence and severity appear to be increasing, most likely accelerated by human activities (Figure 1). A familiar term used in the popular press and literature, ‘dead zone’, used for coastal and estuarine hypoxia, refers to the fish and shellfish killed by the suffocating conditions or the failure to catch these animals in bottom waters when the oxygen concentration in the water covering the seabed is below a critical level.

Based on laboratory or field observations or both, the level of oxygen stress and related responses of invertebrate and fish faunas vary. The units are often determined by oxygen conditions that are physiologically stressful, but these levels also differ depending on the organisms considered, and the pressure, temperature, and salinity of the ambient waters. The numerical definition of hypoxia varies as do the units used, but hypoxia has mostly been defined as dissolved oxygen levels lower than a range of 3–2 ml l  1, with the consensus being in favor of 1.4 ml l  1 ( ¼ 2 mg l  1 or ppm). This value is approximately equivalent to 30% oxygen saturation at 25 1C and 35 salinity (psu). Below this concentration, bottom-dragging trawl nets fail to capture fish, shrimp, and swimming crabs. Other fishes, such as rays and sharks, are affected by oxygen levels below 3 mg l  1, which prompts a behavioral response to evacuate the area, up into the water column and shoreward. Water-quality standards in the coastal waters of Long Island Sound, New York, and Connecticut, USA, consider that dissolved oxygen conditions below 5 mg l  1 result in behavioral effects in marine organisms and fail to support living resources at sustainable levels.

Oxygen depletion Annual Episodic Periodic Persistent Unknown

Figure 1 Distribution of coastal ocean hypoxic areas; excludes naturally occurring hypoxia, such as upwelling zones and OMZs. Reproduced from Dı´az RJ, Nestlerode J, and Dı´az ML (2004) A global perspective on the effects of eutrophication and hypoxia on aquatic biota. In: Rupp GL and White MD (eds.) Proceedings of the 7th International Symposium on Fish Physiology, Toxicology and Water Quality, Tallinn, Estonia, May 12-15, 2003. EPA 600/R-04/049, pp. 1–33. Athens, GA: Ecosystems Research Division, US Environmental Protection Agency, with permission from Robert J. Dı´az.

172

(c) 2011 Elsevier Inc. All Rights Reserved.

HYPOXIA

The most commonly used definition for oceanic waters is dissolved oxygen content less than 1 ml l  1 (or 0.7 mg l  1). Disoxyic or disaerobic refers to oxygen levels between 0.1 and 1.0 ml l  1. OMZs are usually defined as waters less than 0.5 ml l  1 dissolved oxygen.

173

shifts in trophic interactions and food webs, and impacts on living resources.

Hypoxic Systems Oxygen-minimum Zones

Causes Hypoxia occurs where the consumption of oxygen through respiratory or chemical processes exceeds the rate of supply from oxygen production via photosynthesis, diffusion through the water column, advection, or mixing. The biological and physical water-column characteristics that support the development and maintenance of hypoxia include (1) the production, flux, and accumulation of organic-rich matter from the upper water column; and (2) watercolumn stability resulting from stratification or long residence time. Dead and senescent algae, zooplankton fecal pellets, and marine aggregates contribute significant amounts of organic detritus to the lower water column and seabed. Aerobic bacteria consume oxygen during the decay of the carbon and deplete the oxygen, particularly when stratification prevents diffusion of oxygen. Stratification is the division of the water column into layers with different densities caused by differences in temperature or salinity or both. Hypoxia will persist as long as oxygen consumption rates exceed those of supply. Oxygen depletion occurs more frequently in estuaries or coastal areas with longer water residence times, with higher nutrient loads and with stratified water columns. Hypoxia is a natural feature of many oceanic waters, such as OMZs and enclosed seas, or forms in coastal waters as a result of the decomposition of high carbon loading stimulated by upwelled nutrientrich waters. Hypoxia in many coastal and estuarine waters, however, is but one of the symptoms of eutrophication, an increase in the rate of production and accumulation of carbon in aquatic systems. Eutrophication very often results from an increase in nutrient loading, particularly by forms of nitrogen and phosphorus. Nutrient over-enrichment from anthropogenic sources is one of the major stressors impacting estuarine and coastal ecosystems, and there is increasing concern in many areas around the world that an oversupply of nutrients is having pervasive ecological effects on shallow coastal waters. These effects include reduced light penetration, increased abundance of nuisance macroalgae, loss of aquatic habitat such as seagrass or macroalgal beds, noxious and toxic algal blooms, hypoxia and anoxia,

Persistent hypoxia is evident in mid-water OMZs, which are widespread in the world oceans where the oxygen concentrations are less than 0.5 ml l  1 (or about 7.5% oxygen saturation, o22 mM). They occur at different depths from the continental shelf to upper bathyal zones (down to 1300 m). Many of the OMZs form as a result of high primary production associated with coastal upwelled nutrient-rich waters. Their formation also requires stagnant circulation, long residence times, and the presence of oxygen-depleted source waters. The extensive OMZ development in the eastern Pacific Ocean is attributed to the fact that intermediate depth waters of the region are older and have overall oxygen concentrations lower than other water masses. The largest OMZs are at bathyal depths in the eastern Pacific Ocean, the Arabian Sea, the Bay of Bengal, and off southwest Africa. The upper boundary of an OMZ may come to within 10 or 50 m of the sea surface off Central America, Peru, and Chile. The OMZ is more than 1000-m thick off Mexico and in the Arabian Sea, but off Chile, the OMZ is o400-m thick. Along continental margins, minimum oxygen concentrations occur typically between 200 and 700 m. The area of the ocean floor where oceanic waters permanently less than 0.5 ml l  1 impinge on continental margins covers 106 km2 of shelf and bathyal seafloor, with over half occurring in the northern Indian Ocean. These permanently hypoxic waters account for 2.3% of the ocean’s continental margin. These hypoxic areas are not related to eutrophication, but longer-term shifts in meteorological conditions and ocean currents may increase their prevalence in the future with global climate change. Shifts in ocean currents have been implicated in the increased frequency of continental shelf hypoxia along the northwestern US Pacific coast of Oregon. Deep Basins, Enclosed Seas, and Fjord

Many of the existing permanent or periodic anoxic ocean environments occur in enclosed or semienclosed waters where a mass of deep water is bathymetrically isolated from main shelf or oceanic water masses by surrounding landmasses or one or more shallow sills. In conjunction with a pycnocline, the bottom water volume is restricted from exchange with deep open water. Examples of hypoxic and

(c) 2011 Elsevier Inc. All Rights Reserved.

174

HYPOXIA

anoxic basins include anoxic deep water fjord, such as Saanich Inlet, the deeper basins of the Baltic, the basin of the Black Sea, the Japanese Seto Inland Sea, deep waters of the Sea of Cortez, Baja California, and Santa Barbara Basin in the southern California borderland. Coastal Seas and Estuaries

Periodic hypoxia or anoxia also occurs on open continental shelves, for example, the northern Gulf of Mexico and the Namibian and Peruvian shelves where upwelling occurs. More enclosed shelves such as the northern Adriatic and the northwestern shelf of the Black Sea also have periodic hypoxia or anoxia. In these instances, there is minimal exchange of shelf-slope water and/or high oxygen demand on the shallow shelf. Estuaries, embayments, and lagoons are susceptible to the formation of hypoxia and anoxia if the water residence time is sufficiently long, especially where the water column is stratified. Light conditions are also important in these coastal habitats as a limiting factor on phytoplankton growth, which, if excessive, contributes to high organic loading within the confined waters. Coastal ecosystems that have been substantially changed as a result of eutrophication exhibit a series of identifiable symptoms, such as reduced water clarity, excessive, noxious, and, sometimes, harmful algal blooms, loss of critical macroalgal or seagrass habitat, development or intensification of oxygen depletion in the form of hypoxia or anoxia, and, in some cases, loss of fishery resources. More subtle responses of coastal ecosystems to eutrophication include shifts in phytoplankton and zooplankton communities, shifts in the food webs that they support, loss of biodiversity, changes in trophic interactions, and changes in ecosystem functions and biogeochemical processes. In a review of anthropogenic hypoxic zones in 1995, Dı´az and Rosenberg noted that no other environmental variable of such ecological importance to estuarine and coastal marine ecosystems around the world has changed so drastically, in such a short period of time, as dissolved oxygen. For those reviewed, there was a consistent trend of increasing severity (either in duration, intensity, or size) where hypoxia occurred historically, or hypoxia existed presently when it did not occur before. While hypoxic environments have existed through geologic time and are common features of the deep ocean or adjacent to areas of upwelling, their occurrence in estuarine and coastal areas is increasing, and the trend is consistent with the increase in human activities that result in nutrient over-enrichment.

The largest human-caused hypoxic zone is in the aggregated coastal areas of the Baltic Sea, reaching 84 000 km2. Hypoxia existed on the northwestern Black Sea shelf historically, but anoxic events became more frequent and widespread in the 1970s and 1980s, reaching over areas of the seafloor up to 40 000 km2 in depths of 8–40 m. There is also evidence that the suboxic zone of the open Black Sea enlarged toward the surface by about 10 m since 1970. The condition of the northwestern shelf of the Black Sea, in which hypoxia covered up to 40 000 km2, improved over the period 1990–2000 when nutrient loads from the Danube River decreased, but may be experiencing a worsening of hypoxic conditions more recently. Similar declines in bottom water dissolved oxygen have occurred elsewhere as a result of increasing nutrient loads and cultural eutrophication, for example, the northern Adriatic Sea, the Kattegat and Skaggerak, Chesapeake Bay, Albemarle-Pamlico Sound, Tampa Bay, Long Island Sound, New York Bight, the German Bight, and the North Sea. In the United States, over half of the estuaries experience hypoxia at some time over an annual period and many experience hypoxia over extensive areas for extended periods on a perennial basis. The number of estuaries with hypoxia or anoxia continues to rise. Historic data on Secchi disk depth in the northern Adriatic Sea in 1911 through the present, with few interruptions of data collection, provide a measure of water transparency that could be interpreted to depict surface water productivity. These data coupled with surface and bottom water dissolved oxygen content determined by Winkler titrations and nutrient loads outline the sequence of eutrophication in the northern Adriatic Sea. Similar historical data from other coastal areas around the world demonstrate a decrease in water clarity due to phytoplankton production in response to increased nutrient loads that are paralleled by declines in water column oxygen levels. There are strong relationships between river flow and nutrient flux into the Chesapeake Bay and northern Gulf of Mexico and phytoplankton production and biomass and the subsequent fate of that production in spring deposition of chlorophyll a. Further there is a strong relationship between the deposited chlorophyll a and the seasonal decline of deep-water dissolved oxygen. Excess nutrients in many watersheds are driven by agricultural activities and atmospheric deposition from burning of fossil fuels. The link with excess nutrients in more urban areas, such as Long Island Sound, is with the flux of nutrients associated from numerous wastewater outfalls.

(c) 2011 Elsevier Inc. All Rights Reserved.

HYPOXIA

Swift currents that move materials away from a river delta and that do not permit the development of stratification are not conducive to the accumulation of biomass or depletion of oxygen, for example in the Amazon and Orinoco plumes. Similar processes off the Changjiang (Yantze River) and high turbidity in the plume of the Huanghe (Yellow River) were once thought to be reasons why hypoxia did not develop in those coastal systems. Incipient indications of the beginning of symptoms of cultural eutrophication were becoming evident at the terminus of both these systems as nutrient loads increased. The severely reduced, almost minimal, flow of the Huanghe has prevented the formation of hypoxia, but other coastal ecosystem problems remain. There is, however, now a hypoxic area off the Changjiang Estuary and harmful algal blooms are more frequent in the East China Sea. The likelihood that more and more coastal systems, especially in developing countries, where the physical conditions are appropriate will become eutrophic with accompanying hypoxia is worrisome. Northern Gulf of Mexico

The hypoxic zone on the continental shelf of the northern Gulf of Mexico is one of the largest hypoxic zones in the world’s coastal oceans, and is representative of hypoxia resulting from anthropogenic activities over the last half of the twentieth century (Figure 2). Every spring, the dissolved oxygen levels in the coastal waters of the northern Gulf of Mexico decline and result in a vast region of oxygen-starved water that stretches from the Mississippi River westward along the Louisiana shore and onto the Texas coast. The area of bottom covered by hypoxic water can reach 22 000 km2, and the volume of hypoxic waters may be as much as 1011 m3. Hypoxia

L.Calcasieu

30 Sabine L.

175

in the Gulf of Mexico results from a combination of natural and human-influenced factors. The Mississippi River, one of the 10 largest in the world, drains 41% of the land area of the lower 48 states of the US and delivers fresh water, sediments, and nutrients to the Gulf of Mexico. The fresh water, when it enters the Gulf, floats over the denser saltier water, resulting in stratification, or a two-layered system. The stratification, driven primarily by salinity, begins in the spring, intensifies in the summer as surface waters warm and winds that normally mix the water subside, and dissipates in the fall with tropical storms or cold fronts. Hypoxic waters are found at shallow depths near the shore (4–5 m) to as deep as 60 m. The more typical depth distribution is between 5 and 35 m. The hypoxic water is not just located near the seabed, but may rise well up into the water column, often occupying the lower half of a 20-m water column (Figure 3). The inshore/offshore distribution of hypoxia on the Louisiana shelf is dictated by winds and currents. During typical winds from the southeast, downwelling favorable conditions force the hypoxic bottom waters farther offshore. When the wind comes from the north, an upwelling favorable current regime promotes the movement of the hypoxic bottom waters close to shore. When the hypoxic waters move onto the shore, fish, shrimp, and crabs are trapped along the beach, resulting sometimes in a ‘jubilee’ when the stunned animals are easily harvested by beachgoers. A more negative result is a massive fish kill of all the sea life trapped without sufficient oxygen. Hypoxia occurs on the Louisiana coast west of the Mississippi River delta from February through November, and nearly continuously from mid-May through mid-September. In March and April, hypoxic water masses are patchy and ephemeral. The hypoxic

Atchafalaya.R

LA

Mississippi R

TX 2

2

Terrebonne Bay

2

2

29

2

2

2

2

2

2

Dissolved oxygen less than 2.0 (mg l−1) −94

−93

2001 Area −92

2 2

0

−91

−90

25

(km)

50

−89

Figure 2 Similar size and expanse of bottom water hypoxia in mid-July 2002 (shaded area) and in mid-July 2001 (outlined with dashed line). Data source: N. N. Rabalais, Louisiana Universities Marine Consortium.

(c) 2011 Elsevier Inc. All Rights Reserved.

176

HYPOXIA 7

7 6 5

−5

4 3

1

2 1

6 5

4 1 3 2

6 5 4 3

6

6

5

5

7 6 5 4

4 5

−15

5

5

43 2

1

Depth (m)

−10

7

1

−20

5

4 3 2

4 3

5 −25

34 2

8/16/1990 −30 0

5

15

20

6

−5

25

30

35

40

45

50

55

8 7 6 5 43 2

7

6 6 6

6

5 4

−10

Depth (m)

10 7

1

3 2

−15

4

5 4

5

3 2

1

5 4

3 2

3 2

6 45

3 2

1 2

1

1

−20

4 3 21

3 2

−25 9/17/1990 −30 0

5

10

15

20

25

30

35

40

45

50

55

Distance (km) Figure 3 Contours of dissolved oxygen (mg l  1) across the continental shelf of Louisiana approximately 200 km west of the Mississippi River delta in summer. The distribution across the shelf in August is a response to an upwelling favorable oceanographic regime and that of September to a downwelling favorable oceanographic regime. These contours also illustrate the height above the seabed that hypoxia can reach, i.e., over half the water column. Data source: N. N. Rabalais, Louisiana Universities Marine Consortium.

zone is most widespread, persistent, and severe in June, July, and August, and often well into September, depending on whether tropical storm activity begins to disrupt the stratification and hypoxia. Anoxic waters occur periodically in midsummer. The midsummer size of the hypoxic zone varies annually, and is most closely related to the nitrate load of the Mississippi River in the 2 months prior to the typically late-July mapping exercise. The load of nitrate is determined by the discharge of the Mississippi River multiplied by the concentration of the nitrate, so that the amount of water coming into the Gulf of Mexico is also a factor. The relationship of the size of hypoxia, however, is stronger with the load of nitrate than with the total river water discharge or any other nutrient or combination of

nutrients. Changes in the severity of hypoxia over time are related mostly to the change in nitrate concentration in the Mississippi River (80%), the remainder to changes in increased discharge (20%).

Historical Change in Oxygen Conditions Historical dissolved oxygen data such as those for the northern Adriatic Sea beginning in 1911 are not commonly available. A solution is to turn to the sediment record for paleoindicators of long-term transitions related to eutrophication and oxygen deficiency. Biological, mineral, or chemical indicators of plant communities, level of productivity, or

(c) 2011 Elsevier Inc. All Rights Reserved.

HYPOXIA

conditions of hypoxia preserved in sediments, where sediments accumulate, provide clues to prior hydrographic and biological conditions. Data from sediment cores taken from the Louisiana Bight adjacent to the Mississippi River where sediments accumulate with their biological and chemical indicators document increased recent eutrophication and increased organic sedimentation in bottom waters, with the changes being more apparent in areas of chronic hypoxia and coincident with the increasing nitrogen loads from the Mississippi River system beginning in the 1950s. This evidence comes as an increased accumulation of diatom remains and marine-origin carbon accumulation in the sediments. Benthic microfauna and chemical conditions provide several surrogates for oxygen conditions. The mineral glauconite forms under reducing conditions in sediments, and its abundance is an indication of low-oxygen conditions. (Note that glauconite also forms in reducing sediments whose overlying waters are 42 mg l  1 dissolved oxygen.) The average glauconite abundance in the coarse fraction of sediments in the Louisiana Bight was B5.8% from 1900 to a transition period between 1940 and 1950, when it increased to B13.4%, suggesting that hypoxia ‘may’ have existed at some level before the 1940–50 time period, but that it worsened since then. Benthic foraminiferans and ostracods are also useful indicators of reduced oxygen levels because oxygen stress decreases their overall diversity as measured by the Shannon–Wiener diversity index (SWDI) and shifts community composition. Foraminiferan and ostracod diversity decreased since the 1940s and early 1950s, respectively. While presentday foraminiferan diversity is generally low in the Louisiana Bight, comparisons among assemblages from areas of different oxygen depletion indicate that the dominance of Ammonia parkinsoniana over Elphidium spp. (A–E index) was much more pronounced in oxygen-depleted compared to welloxygenated waters. The A–E index has also proven to be a strong, consistent oxygen-stress signal in other coastal areas, for example, Chesapeake Bay and Long Island Sound. The A–E index from sediment cores increased significantly after the 1950s, suggesting increased oxygen stress (in intensity or duration) in the last half century. Buliminella morgani, a hypoxia-tolerant species, known only from the Gulf of Mexico, dominates the present-day population (450%) within areas of chronic seasonal hypoxia, and has also increased markedly in recent decades. Quinqueloculina sp., a hypoxia-intolerant foraminiferan, was a conspicuous member of the fauna from 1700 to 1900, indicating that oxygen

177

stress was not a problem prior to 1900, but this species is no longer present on northern Gulf of Mexico shelf in the Louisiana Bight. Multiple lines of evidence from sediment cores indicate an overall increase in phytoplankton productivity and continental shelf oxygen stress (in intensity or duration) in the northern Gulf of Mexico adjacent to the plume of the Mississippi River, especially in the last half of the twentieth century. The changes in these indicators are consistent with the increases in river nitrate-N loading during that same period. OMZ intensity and distribution vary over geological timescales as a result of shifts in productivity or circulation over a few thousands to 10 ky. These changes affect expansions and contractions of the oxygen-depleted waters both vertically and horizontally. Paleoindicators, including foraminiferans, organic carbon preservation, carbonate dissolution, nitrogen isotopes, and Cd:Ca ratios that reflect productivity maxima and shallow winter mixing of the water column, are used to trace longer-term changes in OMZs, similar to studies of continental shelf sediment indicators.

Consequences Direct Effects

The obvious effects of hypoxia/anoxia are displacement of pelagic organisms and selective loss of demersal and benthic organisms. These impacts may be aperiodic so that recovery occurs; may occur on a seasonal basis with differing rates of recovery; or may be permanent so that a shift occurs in long-term ecosystem structure and function. As the oxygen concentration falls from saturated or optimal levels toward depletion, a variety of behavioral and physiological impairments affect the animals that reside in the water column or in the sediments or that are attached to hard substrates (Figure 4). Hypoxia acts as an endocrine disruptor with adverse effects on reproductive performance of fishes, and loss of secondary production may therefore be a widespread environmental consequence of hypoxia. Mobile animals, such as shrimp, fish, and some crabs, flee waters where the oxygen concentration falls below 3–2 mg l  1. As dissolved oxygen concentrations continue to fall, less mobile organisms become stressed and move up out of the sediments, attempt to leave the seabed, and often die (Figure 5). As oxygen levels fall from 0.5 toward 0 mg l  1, there is a fairly linear decrease in benthic infaunal diversity, abundance, and biomass. Losses of entire higher taxa are features of the

(c) 2011 Elsevier Inc. All Rights Reserved.

Nekton

Demersal invertebrates

Epibenthic invertebrates

Most present

Most present

Present

Absent

Few mantis and penaeid shrimp

2.0

Larger infauna

Present and burrowed

Macroinfauna

Present and burrowed

(mg O2 l−1)

HYPOXIA

(mg O2 l−1)

178

2.0

1.5

1.5 Absent

Stressed starfish, brittle stars) 1.0

1.0

Dead

Stressed (anemones, gastropods)

Moribund polychaetes

0.5

0.5 Dead

Dead

(Fish kills) 0.2

Bacterial mats

0.05

0.2 0.05

Anoxic sediment, H2S in sediment and water Figure 4 Progressive changes in fish and invertebrate fauna as oxygen concentration decreases from 2 mg l  1 to anoxia. From Rabalais NN, Harper DE, Jr., and Tuner RE (2001) Responses of nekton and demersal and benthic fauna to decreasing oxygen concentrations. In: Rabalais NN and Turner RE (eds.) Coastal and Estuarine Studies 58: Coastal Hypoxia: Consequences for Living Resources and Ecosystems, pp. 115–128. Washington, DC: American Geophysical Union.

depauperate benthic fauna in the severely stressed seasonal hypoxic/anoxic zone of the Louisiana inner shelf in the northern Gulf of Mexico. Larger, longerlived burrowing infauna are replaced by short-lived, smaller surface deposit-feeding polychaetes, and certain typical marine invertebrates are absent from the fauna, for example, pericaridean crustaceans, bivalves, gastropods, and ophiuroids. Long-term trends for the Skagerrak coast of western Sweden in semi-enclosed fiordic areas experiencing increased oxygen stress showed declines in the total abundance and biomass of macroinfauna, abundance and biomass of mollusks, and abundance of suspension feeders and carnivores. These changes in benthic communities result in an impoverished diet for bottom-feeding fish and crustaceans and contribute, along with low dissolved oxygen, to altered sediment biogeochemical cycles. In waters of Scandinavia and the Baltic, there was a reduction of 3 million t in benthic macrofaunal biomass during the worst years of hypoxia occurrence. This loss, however, may be

partly compensated by the biomass increase that occurred in well-flushed organically enriched coastal areas not subject to hypoxia. Where oxygen minimum zones impinge on continental margins or sea mounts, they have considerable effects on benthic assemblages. The benthic fauna of OMZs consist mainly of smaller-sized protozoan and meiofaunal organisms, with few or no macrofauna or megafauna. The few eukaryotic organisms are nematodes and foraminiferans. Meiofauna appear to be more broadly tolerant of oxygen depletion than are macrofauna. The numbers of metazoan meiofaunal organisms, primarily nematodes, are not reduced in OMZs, presumably due to abundant particulate food and reduced predation pressure. In hypoxic waters of the northern Gulf of Mexico, harpacticoid copepod meiofauna are reduced at low oxygen levels, but the nematodes maintain their densities. Benthic macrofauna are found in all hypoxic sediments of the northern Gulf of Mexico, although the density is severely reduced

(c) 2011 Elsevier Inc. All Rights Reserved.

HYPOXIA

•

Direct mortality

•

Altered migration

•

Reduction in suitable habitat

•

Increased susceptibility to predation

•

Changes in food resources

•

Susceptibility of early life stages

179

Figure 5 Effects of hypoxia on fishery resources and the benthic communities that support them. Upper right: Dead demersal and bottom-dwelling fishes killed by the encroachment of near-anoxic waters onto a Grand Isle, Louisiana, beach in August 1990. Photo provided by K. M. St. Pe´. Lower right: dead spider crab (family Majidae) at sediment surface. Photo provided by Franklin Viola. Lower left: dead polychaete (family Spionidae) and filamentous sulfur bacteria. Photo provided by Franklin Viola.

below 0.5 mg l  1, and the few remaining organisms are polychaetes of the families Ampharetidae and Magelonidae and some sipunculans. While permanent deep-water hypoxia that impinges on 2.3% of the ocean’s continental margin may be inhospitable to most commercially valuable marine resources, they support the largest, most continuous reducing ecosystems in the world oceans. Large filamentous sulfur bacteria, Thioploca and Beggiatoa, thrive in hypoxic conditions of 0.1 ml l  1. OMZ sediments characteristically support large bacteria, both filamentous sulfur bacteria and giant spherical sulfur bacteria with diameters of 100–300 mm. The filamentous sulfur bacteria are also characteristic of severely oxygen depleted waters in the northern Gulf of Mexico. Secondary Production

An increase in nutrient availability results in an increase of fisheries yield to a maximal point; then there are declines in various compartments of the fishery as further increases in nutrients lead to seasonal hypoxia and permanent anoxia in semienclosed seas. Documenting loss of fisheries related

to the secondary effects of eutrophication, such as the loss of seabed vegetation and extensive bottom water oxygen depletion, is complicated by poor fisheries data, inadequate economic indicators, increase in overharvesting that occurred at the time that habitat degradation progressed, natural variability of fish populations, shifts in harvestable populations, and climatic variability. Eutrophication often leads to the loss of habitat (rooted vegetation or macroalgae) or low dissolved oxygen, both of which may lead to loss of fisheries production. In the deepest bottoms of the Baltic proper, animals have long been scarce or absent because of low oxygen availability. This area was 20 000 km2 until the 1940s. Since then, about a third of the Baltic bottom area has intermittent oxygen depletion. Lowered oxygen concentrations and increased sedimentation have changed the benthic fauna in the deeper parts of the Baltic, resulting in an impoverished diet for bottom fish. Above the halocline in areas not influenced by local pollution, benthic biomass has increased due mostly to an increase in mollusks. On the other hand, many reports document instances where local pollution resulting in severely depressed oxygen levels has greatly

(c) 2011 Elsevier Inc. All Rights Reserved.

180

HYPOXIA

impoverished or even annihilated the soft-bottom macrofauna. Eutrophication of surface waters accompanied by oxygen-deficient bottom waters can lead to a shift in dominance from demersal fishes to pelagic fishes. In the Baltic Sea and Kattegatt where eutrophicationrelated ecological changes occurred mainly after World War II, changes in fish stocks have been both positive (due to increased food supply; e.g., pike perch in Baltic archipelagos) and negative (e.g., oxygen deficiency reducing Baltic cod recruitment and eventual harvest). Similar shifts are inferred with limited data on the Mississippi River-influenced shelf with the increase in two pelagic species in bycatch from shrimp trawls and a decrease in some demersal species. Commercial fisheries in the Black Sea declined as eutrophication led to the loss of macroalgal habitat and oxygen deficiency, amid the possibility of overfishing. After the mid-1970s, benthic fish populations (e.g., turbot) collapsed, and pelagic fish populations (small pelagic fish, such as anchovy and sprat) started to increase. The commercial fisheries diversity declined from about 25 fished species to about five in 20 years (1960s to 1980s), while anchovy stocks and fisheries increased rapidly. The point on the continuum of increasing nutrients versus fishery yields remains vague as to where benefits are subsumed by environmental problems that lead to decreased landings or reduced quality of production and biomass.

Future Expectations The continued and accelerated export of nitrogen and phosphorus to the world’s coastal ocean is the trajectory to be expected unless societal intervention takes place (in the form of controls or changes in culture). The largest increases are predicted for southern and eastern Asia, associated with predicted large increases in population, increased fertilizer use to grow food to meet the dietary demands of that population, and increased industrialization. The implications for coastal eutrophication and subsequent ecosystem changes such as worsening conditions of oxygen depletion are significant.

Ecosystems. Fishery Management, Dimension. Upwelling Ecosystems.

Human

Further Reading Dı´az RJ, Nestlerode J, and Dı´az ML (2004) A global perspective on the effects of eutrophication and hypoxia on aquatic biota. In Rupp GL and White MD (eds.) Proceedings of the 7th International Symposium on Fish Physiology, Toxicology and Water Quality, EPA 600/ R-04/049, pp. 1–33. Tallinn, Estonia, 12–15 May 2003. Athens, GA: Ecosystems Research Division, US EPA. Dı´az RJ and Rosenberg R (1995) Marine benthic hypoxia: A review of its ecological effects and the behavioural responses of benthic macrofauna. Oceanography and Marine Biology Annual Review 33: 245--303. Gray JS, Wu RS, and Or YY (2002) Review. Effects of hypoxia and organic enrichment on the coastal marine environment. Marine Ecology Progress Series 238: 249--279. Hagy JD, Boynton WR, and Keefe CW (2004) Hypoxia in Chesapeake Bay, 1950–2001: Long-term change in relation to nutrient loading and river flow. Estuaries 27: 634--658. Helly J and Levin LA (2004) Global distributions of naturally occurring marine hypoxia on continental margins. Deep-Sea Research 51: 1159--1168. Mee LD, Friedrich JJ, and Gomoiu MT (2005) Restoring the Black Sea in times of uncertainty. Oceanography 18: 100--111. Rabalais NN and Turner RE (eds.) (2001) Coastal and Estuarine Studies 58: Coastal Hypoxia – Consequences for Living Resources and Ecosystems. Washington, DC: American Geophysical Union. Rabalais NN, Turner RE, and Scavia D (2002) Beyond science into policy: Gulf of Mexico hypoxia and the Mississippi River. BioScience 52: 129--142. Rabalais NN, Turner RE, Sen Gupta BK, Boesch DF, Chapman P, and Murrell MC (2007) Characterization and long-term trends of hypoxia in the northern Gulf of Mexico: Does the science support the Action Plan? Estuaries and Coasts 30(supplement 5): 753--772. Turner RE, Rabalais NN, and Justic´ D (2006) Predicting summer hypoxia in the northern Gulf of Mexico: Riverine N, P and Si loading. Marine Pollution Bulletin 52: 139--148. Tyson RV and Pearson TH (eds.) Geological Society Special Publication No. 58: Modern and Ancient Continental Shelf Anoxia, 470pp. London: The Geological Society.

See also

Relevant Websites

Coastal Topography, Human Impact on. Ecosystem Effects of Fishing. Eutrophication. Fiordic

http://www.gulfhypoxia.net – Hypoxia in the Northern Gulf of Mexico.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICEBERGS D. Diemand, Coriolis, Shoreham, VT, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1255–1264, & 2001, Elsevier Ltd.

Introduction Icebergs are large blocks of freshwater ice that break away from marine glaciers and floating ice shelves of glacial origin. Although they originate on land, they are often included in discussions of sea ice because they are commonly found surrounded by it. However, unlike sea ice, they are composed of fresh water; therefore their origin, crystal structure, and chemical composition, as well as the hazards they pose, are different. They are found in both polar regions, their sizes and numbers generally being greater at higher latitudes. They pose a hazard both to shipping and to seabed structures.

Origins and Spatial Distribution The great ice sheets of Greenland and Antarctica, which produce by far the greatest number of the world’s icebergs, flow off the land and into the sea through numerous outlet glaciers. In many cases, especially in Antarctica, the ice spreads out on the sea surface, staying connected to land and forming a floating ice shelf of greater or lesser extent. There are two major differences between the calving fronts in Greenland and those in Antarctica. First, most of the Greenland icebergs are calved directly from the parent glaciers into the sea, while Antarctic icebergs are mostly calved from the edges of the huge ice shelves that fringe much of the continent. The result is that southern icebergs at the time of calving tend to be very large and tabular, while the northern ones are not so large and have a more compact configuration. Second, the equilibrium line of the Greenland ice sheet is above 1000 m. Therefore, the entire volume of a Greenland iceberg is composed of ice. In Antarctica, on the other hand, the equilibrium line is at or near the edge of the ice shelves, so that icebergs are commonly calved with an upper layer of permeable firn (see Ice Properties below) of varying thickness that influences later deterioration rates and complicates estimates of draft and mass.

The drift of icebergs is largely governed by ocean currents, although wind may exert some influence. Since ocean currents at depth may differ in speed and direction from surface currents, a large iceberg may move in a direction different from that of the surrounding sea ice, creating a patch of open water behind it. Since its speed and direction are heavily dependent on the depth and shape of the keel, which is usually unknown, trajectory predictions are seldom reliable, even when local current profiles are known. Because the Antarctic continent is surrounded by oceans while the Arctic is an ocean surrounded by continents, the drift patterns of the icebergs from these areas are very different. Baffin Bay to North Atlantic Region

About 95% of icebergs in northern latitudes originate on Greenland. Most of these are from western Greenland where they calve directly into Baffin Bay, but a few are produced in eastern Greenland. Many of these remain trapped in the fiords where they originated, deteriorating to a great degree before they reach the sea. Those from eastern Greenland that do reach the sea drift south in the East Greenland Current, a small number continuing south into the North Atlantic where they rapidly dwindle, others being carried around the southern tip of Greenland and then north in the warm West Greenland Current, where the few that survive the long trip join the great numbers of bergs originating from Disko Bay north. Figure 1 shows some of the most active glaciers on Greenland and the drift paths generally followed by icebergs. Icebergs may remain in Baffin Bay for several years, circulating north along the Greenland coast and then south along the Canadian arctic islands. Since the water temperature in Baffin Bay remains consistently low throughout the year, little deterioration takes place. However, many do escape southward through the Davis Strait and drift down the Labrador coast in the cold Labrador Current until they break free of the annual pack ice and reach the Grand Banks off Newfoundland. Icebergs have been sighted as far south as Bermuda and the Azores. Arctic Ocean

The remaining 5% of northern icebergs are calved from numerous glaciers on Ellesmere Island in the

(c) 2011 Elsevier Inc. All Rights Reserved.

181

ICEBERGS

Baffin Bay

cir cle cti c Ar

60

Di sk o

Ba ffin Is.

˚N

Ba y

Is. Devon

Sc Su ores nd by

p tr u ns e e St

sh av n

Gade

Greenland

Ja ko b

boldt Hum

Up er Ri nav nk ik s

80˚N

Ellesmere Is.

70

˚N

80

˚N

182

Davis strait

70˚N

Labrador Sea Arctic circle

Labrador 60˚N

Figure 1 Sources and drift paths of North Atlantic icebergs.

Canadian Arctic, the many islands in or bordering the Barents, Kara, and Laptev Seas, and Alaska (see Figure 2). Many of these, especially those calved from ice shelves on Ellesmere Island, are tabular in form. When they were first discovered drifting among the Arctic pack, they were referred to as ‘ice islands’, and the name stays with them. Once they have become incorporated into the pack, they tend to stay there indefinitely, although occasionally one may escape and join the southbound flux through Davis Strait or the east coast of Greenland. The sources and trajectories of these icebergs and ice islands are shown in Figure 2. Southern Regions

Since there is no significant runoff from Antarctica, iceberg production accounts for most mass loss from

the continent. Most of these icebergs are calved from the massive ice shelves, such as the Ross, Filchner, Ronne, Larsen, and Amery. About 60–80% by volume are calved from the ice shelves, the remainder from outlet glaciers that empty directly into the sea or from active ice tongues. Once free of the ice front, they drift with the prevailing current along the coast, in some places westward, in others eastward as shown in Figure 3. They may remain close to the coast, where their concentration is the greatest, for periods up to 4 years, protected by the sea ice and the cold water. There are several localized places around the coast where icebergs turn north away from the continent. Once they drift beyond the northern limit of the pack ice, about 601S, they are carried east and north into ever warmer waters until they deteriorate. The most northerly reported sighting was at 261S near the Tropic of Capricorn. Few pass 551S.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICEBERGS

183

Figure 2 Sources and drift paths of icebergs in the Arctic Ocean.

Numbers and Size Distribution Our knowledge of the numbers and size distribution of icebergs is based on visual observations from ships and aircraft; from radar data from ships, shore, and aircraft; and from satellite imagery. Each method has its advantages and shortcomings. For example, satellite imagery covers very large areas and all times of the year, but will not detect small bergs; ship’s radar will pick up most icebergs within its range but may miss rounded bergs or small bergs in heavy seas; visual observation will catch all sizes of bergs, but

only within a limited area in good weather when someone is looking. Thus, any iceberg census will be slanted toward the size and shape categories favored by the observation method used. The most detailed records of iceberg numbers and sizes in a single location have been kept by the US Coast Guard’s International Ice Patrol (IIP) which was formed in the aftermath of the sinking of the Titanic. The IIP began patrolling the Grand Banks in 1914 and reporting iceberg locations to ships in the area. Since that time the IIP has kept a detailed record of all icebergs crossing 481N. These numbers

(c) 2011 Elsevier Inc. All Rights Reserved.

184

ICEBERGS

0°E / W 60°S

Antarctic circ le Larsen Ice Shelf

90°W 70°S

Filchner Ice Shelf Ronne Ice Shelf South Pole

Amery Ice Shelf 60°S 90°E

Ross Ice Shelf

70°S 180°E/W

Figure 3 Sources and drift paths of Antarctic icebergs. The shaded box shows the area covered by the satellite image in Figure 5

are highly variable from year to year as is apparent from Figure 4. The reason for this variability is not clear. Both the numbers and sizes are greater at higher latitudes, since the bergs gradually disintegrate as they drift south into warmer waters. No such long-standing record exists for the southern oceans, so estimates of numbers here may be less reliable. However, the National Ice Center, using satellite imagery, does identify and track icebergs whose longest dimension is greater than 10 nautical miles (18.5 km) when first sighted. They also continue to track fragments smaller than this that may break away but are still detectable by satellite radar. At the same time, Norway’s Norsk Polarinstitutt has kept a record of all icebergs sighted in Antarctic waters by ‘ships of opportunity’, which is most ships in the area, since 1981. This data set

includes icebergs of all sizes, but the coverage is restricted to those times and areas where ships are present. Northern Regions

The total estimated volume of ice calved annually from Greenland is about 225765 km3. Estimated numbers of icebergs calved from Greenland’s glaciers range from about 10 000 to 30 000 per year. The greatest numbers in northern oceans are found in Baffin Bay. Icebergs may also be seasonally very numerous along the coast of eastern Canada, especially before the pack ice melts. Of those that drift south into the North Atlantic, the annual numbers crossing 481N, just north of the Grand Banks of Newfoundland, according to the IIP are shown in Figure 4.

Figure 4 Total numbers of icebergs crossing 481N each year from 1900 through 1999. Note: The figures for the years of World War I and II are incomplete. (Data courtesy of the US Coast Guard International Ice Patrol.)

(c) 2011 Elsevier Inc. All Rights Reserved.

ICEBERGS

Once they encounter the warm Gulf Stream waters, they rapidly deteriorate. Southern Regions

The total estimated volume of ice calved annually from Antarctica ranges from 750 to 3000 km3 per year. The occasional release of extremely large icebergs has a major impact on annual estimates of Antarctic mass loss. The iceberg shown in Figure 5. represents as much ice as the total annual mass loss from a ‘normal’ year. The shaded box shown in Figure 3 is the area covered by this image. Estimated numbers of icebergs calved range from 5000 to 10 000 each year.

Shapes and Sizes The range of sizes of icebergs is enormous, spanning about eight orders of magnitude, from small

fragments with a mass around 1000 tonnes to the immense Antarctic tabular begs with masses in excess of 1010 t. Table 1 shows the normal range of iceberg sizes in the Labrador Sea. In terms of shape, no two icebergs are the same. However, as bergs deteriorate they do tend to assume characteristic forms (Figures 6–10). The shape classification in common use is given in Table 2. Specialized terms used in classification and description are defined in the Glossary. It should be borne in mind that the shape or extent of the ‘sail’ does not necessarily reflect the shape of the entire iceberg. Figure 11A and B show a photograph of an iceberg and a computer-generated image of its underwater configuration. The nearly spherical shape of this medium-sized berg suggests that it had rolled, probably recently and probably several times. Any horizontal tongues of ice, or ‘rams’, would have broken away during this energetic process. Figure 12A and B show a larger iceberg that had only tilted from its original in the water. Extensive rams are visible extending outward underwater far beyond the extent of the sail, remnants of a far greater mass that has been lost since the berg moved into relatively warm waters. While these two icebergs have roughly similar shapes above water, their underwater configurations are very different.

Table 1

Figure 5 Satellite image showing an extremely large iceberg breaking away from the Ronne Ice Shelf in 1998. The area covered by this image is indicated in Figure 3. A and B are the remnants of two other very large icebergs, both of which broke free in 1986 and were grounded at the time of this image. C is a rapidly moving stream of ice moving off the continent through the Filchner Ice Shelf and past Berkner Island (D) into the Weddell Sea. (Radarsat data ^ 1998 Canadian Space Agency/Agence spatiale canadienne. Received by the Canada Centre for Remote Sensing (CCRS). Processed by Radarsat International (RSI) and the Alaska SAR facility (ASF). Image enhancement and interpretation by CCRS. Provided courtesy of RSI, CCRS, ASF, and the National Ice Center.)

185

Iceberg size categories

Designation

Height (m)

Length (m)

Approximate mass (Mt)

Growler Bergy Bit Small Medium Large Very Large

o1 1–5 5–15 16–45 46–75 475

o5 5–15 15–60 61–120 121–200 4200

0.001 0.01 0.1 2 10 410

Figure 6 Tabular iceberg. (Photograph by Deborah Diemand.)

(c) 2011 Elsevier Inc. All Rights Reserved.

186

ICEBERGS

Figure 7 Wedge iceberg. (Photograph by Deborah Diemand.) Figure 10 Domed iceberg. (Photograph by Deborah Diemand.)

Table 2

Iceberg shape categories

Designation

Description

Tabular types Tabular Blocky

Nontabular types Wedge

Figure 8 Pinnacle iceberg. (Photograph by Deborah Diemand.) Pinnacle Drydock Dome

Steep sides with a flat top; length to height ratio >5:1 (Figure 6) Similar to tabular, but length to height ratio o5:1

One flat side sloping gradually to the water; the opposite side sloping steeply, the two meeting at the peak as a spine (Figure 7) With one or more sharp peaks (Figure 8) With two or more peaks separated by a water-filled channel (Figure 9) Small with rounded top (Figure 10)

To approximate the mass in tonnes, that of eqn [2] can be used. Mass ¼ 3:01  ½ðlongest sail dimensionðmÞÞ ðorthogonal widthðmÞÞ ðmaximum heightðmÞÞ Figure 9 Drydock iceberg. (Photograph by Deborah Diemand.)

Because of this uncertainty in the underwater shape, it is impossible to calculate iceberg draft and mass accurately using the sail dimensions. However, certain rules of thumb have emerged from empirical studies. To approximate the draft in meters, the relationship of eqn [1] can be used.   Draft ¼ 49:4  height0:2

½1

½2

These calculations apply only to icebergs composed entirely of ice, with no firn layer. Northern Regions

In general, the mean size of icebergs in Baffin Bay is about 60 m height, 100 m width and 100 m draft. Mean mass is about 5–10 Mt. The sizes of icebergs in this area are constrained by the water depth near the calving fronts, which is less than 200 m. Icebergs with a mass greater than 20 Mt are extremely rare, and for those found south of 601N, a mass greater

(c) 2011 Elsevier Inc. All Rights Reserved.

ICEBERGS

Figure 11 (A) Medium-sized iceberg grounded in 107 m of water. (B) Computer-generated image showing the underwater profile of this berg. The shape and size of the sail were established through stereophotography; those of the keel were determined using acoustic profiling techniques. These two datasets were joined electronically to create this image as well as Figure 12B. WL, water line. ((A) Photograph by Deborah Diemand; (B) courtesy of Dr. James H. Lever.)

than 10 Mt is seldom found. The maximum sail height on record for an iceberg in the North Atlantic is 168 m. The ice islands in the Arctic Ocean extend about 5 m above sea level. They have a thickness of 30–50 m and an area from a few thousand square metres to 500 km2 or more.

Southern Regions

In general, the thickness of the ice shelves at the calving fronts is about 200–250 m, increasing away from the front. Since the ice edge is very long, and often seaward of seabed obstructions, Antarctic icebergs may be extremely large. The iceberg shown in Figure 5 is more than 5800 km2, larger than the US state of Rhode Island. The

187

Figure 12 (A) Large iceberg grounded in 134 m of water. (B) Computer-generated image showing the underwater profile of this berg. WL, water line. ((A) Photograph by Deborah Diemand; (B) courtesy of Dr. James H. Lever.)

largest iceberg ever reported was about 180 km long, with an estimated volume of 1000 km3.

Deterioration Deterioration begins as soon as an iceberg calves from its parent glacier. Even icebergs locked into sea ice over the winter show signs of mass loss. However, significant deterioration does not usually begin until after the berg breaks free from the pack ice and is exposed to warmer surface water and wave action. The major causes of ice loss are melting, calving, and splitting and ram loss. Melting

Ice loss due to melting alone is hard to quantify, but is thought to fall within the shaded region shown in Figure 13. It is highly dependent on water temperature, but is also influenced by wave action, water currents, and bubble release. Melting rate at the

(c) 2011 Elsevier Inc. All Rights Reserved.

188

ICEBERGS

Antarctic tabular bergs, and the resulting fragments can still be extremely large. In this case, probably the main cause of breakup is flexure due to ocean waves, 160 although grounding or collision, or both, may con140 tribute. It is likely that grounding is the main cause of 120 splitting for bergs smaller than 1 km2. 100 While undercut cliffs on the sides of icebergs tend 80 to shed many small fragments, the corresponding 60 underwater rams remain intact until they are suf40 ficiently large that buoyancy forces alone cause them 20 to break away, or the berg grounds. These rounded 0 fragments, which may be of considerable size, 2 4 8 10 0 12 6 probably represent a large proportion of the domed Seawater temperature (˚C above freezing point) icebergs common in warmer waters. For example, Figure 13 Dependence of the melt rate of icebergs on the the calving of the ram extending to the right in temperature above the freezing point of sea water. Figure 12B would create a new iceberg weighing roughly 100 000 tonnes. Splitting and ram loss are the major cause of size water line is far greater than that over the rest of the ice surface. This causes a groove to form, undercut- reduction in extremely large bergs. ting the ice cliffs and creating sometimes extensive underwater rams. The importance of melting in the overall mass loss Ice Properties depends on the surface/volume ratio, being more Glacial ice is formed by the gradual accumulation of significant for small bergs and growlers than for large snow over many centuries. As the snow compacts ones. and recrystallizes, it forms firn, a granular, permeable One of the side effects of the rapid side melting of material. The firn layer may reach as deep as 100 m icebergs is the vertical mixing of the surrounding in very cold places, but is seldom deeper than 50 m. seawater. Driven mostly by the release of air from the This firn layer is not present on the calving fronts of bubbles in the ice and partly by the lower density of Greenland, but is present on Antarctica’s ice shelves, the fresh water of the melted ice, water flows upward and in the icebergs calved from them. When the firn near the berg, drawing deep water to the surface. The reaches a density of 830 kg m3 the pores close off, combination of nutrients brought up from depth by trapping any air that remains. At this point the ice this process and the decreased salinity of the melt- contains about 10% air by volume. Further densifiwater surrounding the berg results in a specialized cation is a result of compression of the air in the community of plankton and fish in the vicinity of bubbles. The bubbles become smaller and may beicebergs. come incorporated into the crystals through recrystallization. The pressure inside them may be as Calving of Cliff Faces high as 2 MPa (20 bars). 200

_

Melt rate (m y 1)

180

Small pieces of ice are constantly breaking off the sides of icebergs, mostly owing to waterline undercutting. Such calving events may produce only a few small pieces, or a great number, especially in warm water. Usually the individual pieces are quite small and are quickly melted, but the total mass loss can be considerable and the resulting imbalance can cause the berg to roll, causing further ice loss. Once the berg has rolled and stabilized, waterline erosion begins anew. This is probably the major cause of mass loss in medium-sized bergs.

Acoustics

Melting icebergs in the open ocean make a characteristic sound sometimes referred to as ‘bergy seltzer’. This is probably created by the explosion or implosion of bubbles as the ice melts. The frequency range of audible sound produced is quite wide, and is largely masked by ambient ocean noise at frequencies below 6 kHz. The sound seems to vary from berg to berg and is undoubtedly influenced by environmental conditions. Estimated detection distances at frequencies above 6 kHz range from 2 to 150 km.

Splitting and Ram Loss

Splitting occurs when a large iceberg breaks into two or more pieces, each of which is an iceberg in its own right. This is a common occurrence for very large

Ice Temperature

Since ice is a good insulator, the original temperature of a large berg at the time of calving will be retained

(c) 2011 Elsevier Inc. All Rights Reserved.

ICEBERGS

in its central core, and may be as low as  221C. After a year or more in cold water, where little or no ablation takes place, the surface ice will warm to about 01C. In relatively warm waters, however, these outer layers of ice are removed more rapidly than the inner cold core warms up, leaving much colder ice near the iceberg’s surface. Since the strength of ice is greater at lower temperatures, the result of a collision with such a berg could be more severe than with one that had not undergone significant melting. Color

In small quantities, ice appears both colorless and transparent. However, because ice selectively transmits light in the blue portion of the visible spectrum while it absorbs light of other frequencies, sufficiently large pieces of clear, bubble-free ice can appear blue. However, this color is frequently masked in glacial ice by the scattering of light of all wavelengths by the bubbles included in the ice, causing the ice to appear white. Blue bands, commonly present within the greater white mass, are caused when cracks form on the parent glacier or later on the iceberg itself and are filled with meltwater that then freezes relatively bubble free. These blue bands range in size from hairline cracks to a meter or more in width. Green icebergs are fairly common in certain regions. This has variously been attributed to copper or iron compounds, the incorporation of dissolved organic compounds, or to an optical trick caused by red light of the sun near the horizon causing the apparent green color. It is likely that there is no single cause and that all of these factors may make the ice appear green. In some cases the trace substances may originate on land, as in the blue cracks mentioned above. In others they may result from sea water freezing to the underside of an ice shelf. Unlike in ice formed at the water surface, most salts and bubbles are rejected, but certain compounds may be trapped in trace amounts, causing the green appearance of otherwise clear ice. Icebergs may also have bands of brown or black; these are caused by morainal or volcanic material deposited while the ice was still part of the parent glacier.

189

much smaller bergs. These ‘small’ bergs may weigh in excess of 100 000 t, but owing to their small above-water size and frequently rounded shape they may not be detected by radar until they are dangerously close to the ship, especially in storm conditions when the radar return from the rough sea surface (sea clutter) will tend to mask the weak radar return from the iceberg. In such conditions the peril is further increased because these relatively small ice masses, tossed by the heavy seas, may reach maximum instantaneous velocities 4–5 times larger than hourly drift speeds. A 4000 t bergy bit moving at the maximum fluid particle velocity of 4.5 m s1 in typical Grand Banks storm waves could have about a third of the kinetic energy of a 1 Mt iceberg drifting at 0.5 m s1 (B1 knot). While the influence of waves on ice movement decreases for larger bergs, icebergs as massive as 1 Mt may still exhibit significantly higher maximum instantaneous velocities than their hourly drift values. Seabed Damage

While small icebergs pose a serious threat to structures at the sea surface, seabed structures such as well-heads, pipelines, cables, and mooring systems are endangered by large icebergs, which may possess a deep enough draft to collide with the seafloor. Marine navigators have long known that the keels of icebergs drifting south over the relatively shallow banks of Canada’s eastern continental shelf may touch the seabed and become grounded. Modern iceberg scours appear in the form of linear to curvilinear scour marks and as pits, and occur from the Baffin Bay/Davis Strait region to the Grand Banks of Newfoundland. They are present at water depths up to about 200 m. Seabed scouring has also been documented in Antarctica, but to date has generated little interest because of the absence of seabed structures. A single scour may be as wide as 30 m, as deep as 10 m, and longer than 100 km. An iceberg may also produce pitting when its draft is suddenly increased through splitting or rolling. It may then remain anchored to the seafloor, rocking and twisting, and may produce a pit deeper than the maximum scour depth. Usage of Icebergs

Economic Importance Hazard to Shipping

A large iceberg poses little threat to shipping on the whole. It will not normally exceed a speed of 1 m s1 (2 knots) and it can be detected with normal marine radar at a considerable distance, allowing the ship to alter course. Ironically, a greater danger is posed by

In the past there has been considerable interest in the possibility of transporting icebergs, representing as they do an essentially unlimited supply of fresh water, to arid areas such as Saudi Arabia, Western Australia, and South America. The two seemingly insurmountable problems that need to be solved are propulsion and prevention of in-transit breakup in

(c) 2011 Elsevier Inc. All Rights Reserved.

190

ICEBERGS

warm seas. Proposed means of moving a sufficiently large ice mass over such long distances have ranged from conventional towing to use of a nuclear submarine to wind power. None has proven feasible. Destruction of Icebergs

Attempts at destroying icebergs have been numerous and varied. Perhaps the most-studied technique has involved the use of explosives, which have been extensively tested on glacial ice in the form of glaciers and ice islands. Both crater blasting and bench blasting have been attempted. The results of this testing suggest that ice is as difficult to blast as typical hard rock, and that therefore the use of explosives for its destruction is impractical. Other methods tested include spreading carbon black on the berg’s surface to accelerate melting, and introducing various gases into the ice to create holes or cavities that can then be filled with explosives of choice. Attempts have also been made to cut through the ice using various means. There is little evidence that any great success was achieved with any of these methods. The only report of a successful attempt to break up an iceberg involved the use of thermit, a welding compound that reacts at very high temperatures (B30001C). The explanation was that the very high heat produced by the thermit caused massive thermal shock within the mass of ice that ultimately resulted in its disintegration, much as glass can be fragmented by extreme temperature changes.

Conclusions There is a great deal of uncertainty surrounding iceberg properties, behavior, drift, and other aspects relating to individual icebergs as opposed to laboratory samples or intact glaciers. This is mostly because of the high cost of expeditions to the remote areas where icebergs are most numerous, and the inherent dangers of hands-on measurement and sampling.

Equilibrium line On a glacier, the line above which there is a net gain due to snow accumulation and below which there is a net loss due to melt. Firn Permeable, partially consolidated snow with density between 400 kg m  3 and 830 kg m  3. Growler A small fragment of glacial ice extending less than a meter above the sea surface and having a horizontal area of about 20 m2. Keel The underwater portion of an iceberg. Ram Lobe of the underwater portion of an iceberg that extends outward, horizontally, beyond the sail. Sail The above-water portion of an iceberg.

See also Antarctic Circumpolar Current. Arctic Ocean Circulation. Current Systems in the Southern Ocean. Florida Current, Gulf Stream, and Labrador Current. Ice-induced Gouging of the Seafloor. Sea Ice: Overview. Sea Ice. Sonar Systems. Sub Ice-Shelf Circulation and Processes. Weddell Sea Circulation. Wind Driven Circulation.

Further Reading Colbeck SC (ed.) (1980) Dynamics of Snow and Ice Masses. New York: Academic Press. Husseiny AA (ed.) (1978) First International Conference on Iceberg Utilization for Fresh Water Production, Weather Modification, and Other Applications. Iowa State University, Ames, 1977. New York: Pergamon Press. Vaughan D (1993) Chasing the rogue icebergs. New Scientist, 9 January. International Ice Patrol (IIP): http://www.uscg.mil/lantarea /iip/home.html Library of Congress Cold Regions bibliography: http:// lcweb.loc.gov/rr/scitech/coldregions/welcome.html National Ice Center (NIC): http://www.natice.noaa.gov/

Glossary Calving The breaking away of an iceberg from its parent glacier or ice shelf. Also the subsequent loss of ice from the iceberg itself.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE-INDUCED GOUGING OF THE SEAFLOOR W. F. Weeks, Portland, OR, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1264–1270, & 2001, Elsevier Ltd.

Introduction Inuit hunters have long known that both sea ice and icebergs could interact with the underlying sea floor, in that sea floor sediments could occasionally be seen attached to these icy objects. As early as 1855, no less a scientific luminary than Charles Darwin speculated that gouging icebergs could traverse isobaths, concluding that an iceberg could be driven over great inequalities of [a seafloor] surface easier than could a glacier. Only in 1924 did similar inferences of the possibility of sea ice and icebergs affecting the seafloor again begin to reappear in the scientific literature. However, direct observations of the nature of these interactions were lacking until the early 1970s. The reason for the increased interest in this rather exotic phenomena was initially the discovery of the supergiant oil field on the edge of the Beaufort Sea at Prudhoe Bay, Alaska. Because the initial successful well was located on the coast, there was immediate interest in the possibility of developing offshore fields to the north of Alaska and Canada. It was also apparent that if major oil resources occurred to the north of Alaska and Canada, similar resources might be found on the world’s largest continental shelf located to the north of the Russian mainland. Offshore wells are typically tied together by subsea pipelines which take the oil from the individual wells either to a central collection point where tanker pickup is possible or to the coast where the oil can be fed into a pipeline transportation system. If sea ice processes could result in major distrubances of the seafloor, this would clearly become a major consideration in pipeline design. Some time after the Prudhoe Bay find, another large offshore oil discovery was made to the east of Newfoundland. Here the ice-induced gouging problem was caused not by sea ice but by icebergs. In both cases the initial step in treating these perceived problems was to investigate the nature of these ice–seafloor interactions. One needs to know the frequency of gouging events in time and space as well as the widths and depths of the gouges, the water depth range in which this phenomenon occurs, and the effective lifetime of a gouge after its initial formation. Also of importance are the type of subsea soil

and the nature and extent of the soil movements below the gouges, as this information is essential in calculating safe burial depths for subsea structures. The following attempts to summarize our current knowledge of this type of naturally occurring phenomenon. As extensive studies of ice-induced disruptions of the seafloor have been undertaken only during the last 30 years, there is as yet no commonly agreed upon terminology for this phenomenon. In the literature the process has been described as scouring, scoring, plowing and gouging. Here gouging is used because at least in the case of sea ice, it is felt that it more accurately describes the process than do the other terms.

Observational Techniques A variety of techniques have been used to study the gouging phenomenon. Typically, a fathometer is used to resolve the seafloor relief directly beneath the ship with a precision of better than 10 cm (Figure 1) while, at the same time, a sidescan sonar system provides a sonar map of the seafloor on either side of the ship (Figure 2). Total sonar swath widths have been typically 200 to 250 m not including a narrow area directly beneath the ship that was not imaged. The simultaneous use of these two different types of records allows one to both measure the depth and width of the gouge (fathometer) at the point where it is crossed by the ship track and also to observe the general orientation and geometry of the gouge track (sonar). Along the Alaskan coast the ships used have been small, allowing them to operate in shallow water. In some cases divers have been used to examine the gouges and, at a few deeper water sites off the Canadian east coast, manned submersibles have been used to gather direct observations of the gouging process.

Results Arctic Shelves

The most common features gouging the shelves of the Arctic Ocean are the keels of pressure ridges that are made of deformed sea ice. As the ice pack moves under the forces exerted on it by the wind and currents, pressure ridges typically form at floe boundaries as the result of differential movements between the floes. These features can be very large. Pressure

(c) 2011 Elsevier Inc. All Rights Reserved.

191

192

ICE-INDUCED GOUGING OF THE SEAFLOOR

0

30

Depth (m)

35

40

15

30

45 46

1

16

31

2

17

32

47

3

18

33

48

4

19

34

49

5

20

35

50

6

21

36

51

22

37

52

8

23

38

53

9

24

39

54

25

40

55

10

26

41

27

42

57

43 44

45

15

0

100

200

300 Distance (m)

400

500

Figure 1 Fathogram of an ice-gouged seafloor. Water depth is 36 m. Record taken 25 km NE of Cape Halkett, Beaufort Sea, offshore Alaska. The multiple reflections from the upper layer of the seafloor are the result of the presence of a thawed active layer in the subsea permafrost. (From Weeks et al., 1983.)

125 100 75

Distance across track (m)

50 25 0 0 25 50 75 100 125 0

100

200

300

400

500

Distance along ship's track (m) Figure 2 Sonograph of an ice-gouged seafloor. Water depth is 20 m. Record taken 20 km NE of Cape Halkett, Beaufort Sea, offshore Alaska. (From Weeks et al., 1983.)

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE-INDUCED GOUGING OF THE SEAFLOOR

ridge keels with drafts up to 50 m have been observed in the upward-looking sonar records collected by submarines passing beneath the ice. The distribution of keel depths of these deformed features is approximately a negative exponential, in that there are many shallow ridge keels while deep keels are rare. Deformation can be particularly intense in nearshore regions where the moving ice pack contacts an immovable coast with the formation of large shear ridges that can extend for many kilometers. Such ridges are frequently anchored at shoal areas where large accumulations of highly deformed sea ice can build up. Such grounded ice features, which are referred to using the Russian term stamuki, can have freeboards of as much as 10 m and lateral extents of 10–15 km. As the offshore ice field moves against and along the coast, it exerts force on the sides of any such grounded features, causing them to scrape and plough their way along the seafloor. Considering that surficial sediments along many areas of the Beaufort Sea Coast of Canada and Alaska are fine-grained silts, it is hardly surprising that over a period of time such a process can cause extensive gouging of the seafloor sediments. Figure 3 is a photograph showing active gouging. The relatively undeformed first-year sea ice in the foreground is moving away from the viewer and pushing against a piece of grounded multiyear sea ice (indicated by its rounded upper surface formed during the previous summer’s melt period), pushing it along the coast. The interaction between the first-year and the multiyear ice has resulted in a pileup of broken first-year ice, which when the photograph was taken was higher than the upper surface of the multiyear ice. Also evident is the track cut in the first-year ice as it moves past the

Figure 3 Photograph of active ice gouging occurring along the coast of the Beaufort Sea. The grounded multiyear ice floe that is being pushed by the first-year ice has a freeboard of B2 m. The thickness of the first-year ice is B0.3 m. (Photograph by GFN Cox.)

193

multiyear ice. The fact that both the multiyear ice and the pileup of first year ice are interacting with the seafloor is indicated by the presence of bottom sediment in the deformed first-year ice and on the far side of the multiyear ice. The maximum water depth in which contemporary sea ice gouging is believed to occur is roughly 50–60 m, corresponding approximately to the draft of the largest pressure ridges. Although gouges occur in water up to 80 m deep, it is generally believed that these deep-water gouges are relicts that formed during periods when sea level was lower than at present. Not surprisingly the depths of gouges in the seafloor mirrors the keel depth distributions observed in pressure ridge keels in that they are also well approximated by a negative exponential with the character of the falloff as well as the number of gouges varying with water depth. As with ridge keels, shallow gouges are common and deep gouges are rare. As might be expected, there are fewer deep gouges in shallow water as the large ice masses required to produce them have already grounded farther out to sea. For instance, along the Beaufort Coast in water 5 m deep, a 1 m gouge has an exceedance probability of approximately 10 4; that is, 1 gouge in 10 000 will on the average be expected to have a depth equal to or greater than 1 m. In water 30 m deep, a 3.4 m gouge has the same probability of occurrence. Gouges in excess of 3 m deep are not rare and 8 m gouges have been reported in the vicinity of the Mackenzie Delta. As might be expected, the nature of gouging varies with changes in seafloor sediment types. Along the Beaufort coast where there are two distinct soil types, the gouges in stiff, sandy, clayey silts are typically more frequent and slightly deeper than those found in more sandy sediments. Presumably the gouges in the more sandy material are more easily obliterated by wave and current action. It is also reasonable to assume that the slightly deeper gouges in the silts provide a better picture of the original incision depths. An extreme example of the effect of bottom sediment type on gouging can be found between Sakhalin Island and the eastern Russian mainland. Here the seafloor is very sandy and the currents are very strong. As a result, even though winter observations have conclusively shown that seafloor gouging is common, summer observations reveal that the gouges have been completely erased by infilling. Gouge tracks in the Beaufort Sea north of the Alaskan and western Canadian coast generally run roughly parallel to the coast, with some regions having an excess of 200 gouges per kilometer.

(c) 2011 Elsevier Inc. All Rights Reserved.

194

ICE-INDUCED GOUGING OF THE SEAFLOOR

Histograms of the frequency of occurrence of distances between gouges are also well described by a negative exponential, a result suggesting that the spatial occurrence of gouges may be described as a Poisson process. Gouge occurrence is hardly uniform, however, with significantly higher concentrations of gouges occurring on the seaward sides of shoal areas and barrier islands and fewer gouges on their more protected lee sides. This is clearly shown in Figure 4 in that the degree of exposure to the moving pack ice decreases in the sequence Jones Islands–Lonely–Harrison Bay–Lagoons. Figure 5 is a sketch based on divers’ observations showing gouging along the Alaskan coast of the Beaufort Sea. The second features producing gouges on the surface of the outer continental shelf of the Arctic Ocean are ice islands. These features are, in fact, an unusual type of tabular iceberg formed by the gradual breakup of the ice shelves located along the north coast of Ellesmere Island, the northernmost of the Canadian Arctic Islands. Once formed, an ice island can circulate in the Arctic Ocean for many years. For

4

10

3

Number of gouges observed

10

2

10

n = 16 620

instance, the ice island T-3 drifted in the Arctic Ocean between 1952 and 1979, ultimately completing three circuits of the Beaufort Gyre (the large clockwise oceanic circulation centered in the offshore Beaufort Sea) before exiting the Arctic Ocean through Fram Strait between Greenland and Svalbard (Spitzbergen). The lateral dimensions of ice islands can vary considerably from a few tens of meters to over ten kilometers. Thicknesses, although variable, are typically in the 40–50 m range; ice islands possess freeboards in the same range as exhibited by larger sea ice pressure ridges. In the study of fathometer and sonar data on gouge distributions and patterns, no attempt is usually made to separate gouges made by pressure ridges from gouges made by ice islands, as the ice features that made the gouges are commonly no longer present. However, it is reasonable to assume that many of the very wide, uniform gouges are the result of ice islands interacting with the seafloor. It is relatively easy to characterize the state of the gouging existing on the seafloor at any given time. It is another matter to answer the question of how deep one must bury a pipeline, a cable, or some other type of fixed structure beneath the seafloor to reduce the chances of it being impacted by moving ice to some acceptable level. To answer this question one needs to know the rates of occurrence of new gouges; these values are in many locations still poorly known, as they require replicate measurements over a period of many years so that new gouges can be counted and the rate of infilling of existing gouges can be estimated. The best available data on gouging occurrence along the Beaufort Coast has been collected by the Canadian and US Geological Surveys and indicates that gouging occurs in rather large scale regional events related to severe storms that drive the

1

10

2869 842 41

0

10

10

_1

0

1

2

3

Gouge depth (m) Figure 4 Semilogarithmic plot of the number of gouges observed versus gouge depth for four regions along the Alaskan coast of the Beaufort Sea. , Jones Island and east; W, Lonely;, n Harrison Bay; J, Lagoons. (From Weeks et al., 1983.)

Figure 5 Diver’s sketch of active gouging occurring off the coast of the Alaskan Beaufort Sea. Some sense of the scale of this drawing can be gained from the observation that commonly the thickness of the ice blocks in such ridges ranges between 0.3 and 1.0 m. (Drawing by TR Alpha, US Geological Survey.)

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE-INDUCED GOUGING OF THE SEAFLOOR

pack ice inshore – about once every 4–5 years. Because there is no absolute technique for dating the age of existing gouges, there is no current method for determining whether a particular gouge that one observes on the seafloor formed during the last year or sometime during the last 6000 years (after the Alaskan shelf was submerged as the result of rising sea levels at the end of the Pleistocene). It is currently believed that gouges in shallow water formed quite recently. For instance, during the late summer of 1977 when the Beaufort Coast was comparatively ice free, strong wave action during late summer storms obliterated gouges in water less than 13 m deep and caused pronounced infilling of gouges in somewhat deeper water. It is generally estimated that such storms have an average recurrence interval of approximately 25 years. On the other hand, gouges in water deeper than B60 m are believed to be relatively old in that they are presumed to be below the depth of currently active gouging. The lengths of time represented by the gouges observed in water depths between 20 and 50 m are less well known. Unfortunately, this is the water depth range in which the largest gouges are found. Stochastic models of gouging occurrence, which incorporate approximate simulations of subsea sediment transport, suggest that if seafloor currents are sufficiently strong to exceed the threshhold for sediment movement, gouge infilling will occur comparatively rapidly (within a few years). To date, two different procedures have been used to estimate the depth of gouges with specified recurrence intervals. In the first case, the available rates and geometric characteristics of new gouges at a particular site are used. If a multiyear high-quality dataset of new gouge data is available for the region of interest, this would clearly appear to be the favored approach. In the other case, information on the draft distribution of pressure ridge keels and pack ice drift rates are combined to calculate the rates of gouging. The problem with using pack ice drift rates and pressure ridge keel depths is that drift rates in the near shore where grounding is occurring are undoubtedly less than in the offshore where the ice can move relatively unimpeded. In addition, it is doubtful that offshore keel depths observed in water sufficiently deep to allow submarine operations provide accurate estimates of keel depths in the near-shore. One might suggest that offshore engineers should simply bury pipelines and cables at depths below those of known gouges and forget about all these statistical considerations. Although this sounds attractive, it is not a realistic resolution of the problem. In the first place, for a pipeline to avoid deformation, buckling, and failure it must, depending on the

195

nature of the seafloor sediments, be buried at depths up to two times that of the design gouging event. Considering that gouges in excess of 3 m are not rare at many locations along the Beaufort Coast of Alaska and Canada, this means burial depths 45 m. Burial at such depths is extremely costly and time consuming considering the very short operating season in the offshore Arctic. It is also near the edge of existing subsea trenching technology. Another factor to be considered here is that the deeper the pipeline is buried, the more likely it is that its presence will result in the thawing of subsea permafrost that is known to exist in many areas of the Beaufort and Chukchi Shelves. Such thawing could result in settlement in the vicinity of the pipe, which could threaten the integrity of the pipeline. There are engineering remedies to this problem but, as one might expect, they are expensive. Canadian East Coast

The primary problem to the east of Newfoundland and off the coast of Labrador is not sea ice but icebergs. Although the majority of these originate from the Greenland Ice Sheet calving to the west into Baffin Bay, some icebergs do originate from the East Greenland coast and from the Canadian Arctic (see Icebergs). Considering their great size and deep draft, icebergs are formidable adversaries. Iceberg gouges with lengths 460 km are known and many have lengths 420 km. Gouge depths can exceed 10 m and recent gouging is known to occur at depths of at least 230 m along the Baffin and Labrador Shelves. Icebergs also appear to be able to traverse vertical ranges of bathymetry up to at least 45 m. As with the sea-ice-induced gouges of the Arctic shelves, iceberg gouge depths are exponentially distributed, with small gouges being common and deep gouges being rare. Iceberg gouges can be straight or curved. Pits are also common, which occur when the iceberg draft is suddenly increased through splitting and rolling. The iceberg can then remain fixed to the seafloor while it rocks and twists as a result of the wave and current forces that impinge upon it. At such times pits can be produced that are deeper than the maximum gouge depth otherwise associated with the iceberg. Although considerable information is available on the statistics of iceberg gouge occurrences, in studies of iceberg groundings more attention has been paid to the energetics of the ice–sediment interactions than to the statistics. Also, more emphasis has been placed on iceberg ‘management’ in order either to reduce or to remove the threat to a specific offshore operation. For instance, as icebergs are discrete

(c) 2011 Elsevier Inc. All Rights Reserved.

196

ICE-INDUCED GOUGING OF THE SEAFLOOR

objects as compared to pressure ridges, which are giant piles of ice blocks that are frequently poorly cemented together, icebergs can be deflected away from a production site via the use of tugboats in combination with trajectory modeling. Such schemes can also utilize aerial, satellite, and ship reconnaissance techniques to provide operators with an adequate warning of a potential threat. Surface-based over-the-horizon radar technology has also been developed as an iceberg reconnaissance tool, but at the time of writing it has not been used operationally.

quantitative understanding of processes such as ice gouging becomes essential for safe development. Lack of such understanding leads to poor, inefficient designs and the increased possibility of failures. The last thing that either the environmental community or the petroleum industry needs is an offshore oil spill in the high Arctic.

See also Icebergs. Rigs and offshore Structures. Sea Ice: Overview. Sea Ice. Sub Ice-Shelf Circulation and Processes. Sub-sea Permafrost.

Other Locations

Although the discussion here has primarily focused on the Beaufort Coast of the Arctic Ocean and the offshore regions of Newfoundland and Labrador, this is only because the interest in offshore oil and gas reserves at these locations has resulted in the collection of observational information on the local nature of the ice-induced gouging phenomenon. However, regions where ice-induced gouging is presently active are very large and include the complete continental shelf of the Arctic Ocean, the continental shelves of Greenland, of eastern Canada, and of Svalbard, as well as the continental shelf of the Antarctic continent where typical shelf icebergs are known to have drafts of B200 m with lateral dimensions as large as 100 km. In addition, relict iceberg gouges exist and as a result affect seafloor topography in regions where icebergs are no longer common or no longer occur, such as the Norwegian Shelf and even the northern slope of Little Bahama Bank and the Straits of Florida. Finally, although 230 m is a reasonable estimate for the maximum depth of active iceberg gouging, relict gouges presumed to have occurred during the Pleistocene have been discovered at depths from 450 m to at least 850 m on the Yermak Plateau located to the northwest of Svalbard in the Arctic Ocean proper.

Conclusions There is clearly much more that we need to know concerning processes acting on the poorly explored continental shelves of the polar and subpolar regions. The ice-induced gouging phenomenon is clearly one of the more important of these, in that an understanding of this process is essential to both the engineering and the scientific communities. At first glance, icebergs and stamuki zones might appear to be exotic entities that are out of sight somewhere way to the north or south and that can be safely put out of mind. However, as oil production moves to ever more difficult frontier areas such as the offshore Arctic, a

Further Reading Barnes PW, Schell DM, and Reimnitz E (eds.) (1984) The Alaskan Beaufort Sea: Ecosystems and Environments. Orlando: Academic Press. Colony R and Thorndike AS (1984) An estimate of the mean field of arctic sea ice motion. Journal of Geophysical Research 89(C6): 10623--10629. Darwin CR (1855) On the power of icebergs to make rectilinear, uniformly-directed grooves across a submarine undulatory surface. London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science 10: 96--98. Goodwin CR, Finley JC and Howard LM (1985) Ice Scour Bibliography. Environmental Studies Revolving Funds Rept. No. 010, Ottawa. Lewis CFM (1977) The frequency and magnitude of drift ice groundings from ice-scour tracks in the Canadian Beaufort Sea. Proceedings of 4th International Conference on Port Ocean Engineering Under Arctic Conditions. Newfoundland: Memorial University. vol. 1, 567–576. Palmer AC, Konuk I, Comfort G, and Been K (1990) Ice gouging and the safety of marine pipelines. Offshore Technology Conference, Paper 6371, 235--244. Reed JC and Sater JE (eds.) (1974) The Coast and Shelf of the Beaufort Sea. Arlington, VA: Arctic Institute of North America. Reimnitz E, Barnes PW, and Alpha TR (1973) Bottom features and processes related to drifting ice, US Geological Survey Miscellaneous Field Studies, Map MF532. Washington, DC: US Geological Survey. Vogt PR, Crane K, and Sundvor E (1994) Deep Pleistocene iceberg plowmarks on the Yermak Plareau: sidescan and 3.5 kHz evidence for thick calving ice fronts and a possible marine ice sheet in the Arctic Ocean. Geology 22: 403--406. Wadhams P (1988) Sea ice morphology. In: Leppa¨ranta M (ed.) Physics of Ice-Covered Seas, pp. 231--287. Helsinki: Helsinki University Printing House. Weeks WF, Barnes PW, Rearic DM, and Reimnitz E (1983) Statistical aspects of ice gouging on the Alaskan Shelf of the Beaufort Sea. Cold Regions Research and Engineering

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE-INDUCED GOUGING OF THE SEAFLOOR

Laboratory Report. New Hampshire, USA: Hanover. 83–21. Weeks WF, Tucker WB III, and Niedoroda AW (1985) A numerical simulation of ice gouge formation and infilling on the shelf of the Beaufort Sea. Proceedings, International Conference on Port Ocean Engineering

197

Under Arctic Conditions, Narssarssuaq, Greenland 1: 393--407. Woodworth-Lynas CMT, Simms A, and Rendell CM (1985) Iceberg grounding and scouring on the Labrador continental shelf. Cold Regions Science and Technology 10(2): 163--186.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE–OCEAN INTERACTION J. H. Morison, University of Washington, Seattle, WA, USA M. G. McPhee, McPhee Research Company, Naches, WA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1271–1281, & 2001, Elsevier Ltd.

Introduction The character of the sea ice cover greatly affects the upper ocean and vice versa. In many ways icecovered seas provide ideal examples of the planetary boundary layer. The under-ice surface may be uniform over large areas relative to the vertical scale of the boundary layer. The absence of surface waves simplifies the boundary layer processes. However, thermodynamic and mechanical characteristics of ice–ocean interaction complicate the picture in unique ways. We discuss a few of those unique characteristics. We deal first with how momentum is transferred to the water and introduce the structure of the boundary layer. This will lead to a discussion of the processes that determine the fluxes of heat and salt. Finally, we discuss some of the unique characteristics imposed on the upper ocean by the larger-scale features of a sea ice cover.

Drag and Characteristic Regions of the Under-ice Boundary Layer To understand the interaction of the ice and water, it is useful to consider three zones of the boundary layer: the molecular sublayer, surface layer, and outer layer (Figure 1). Under a reasonably smooth and uniform ice boundary, these can be described on the basis of the influence of depth on the terms of the equation for a steady, horizontally homogeneous boundary layer (eqn [1]). ifV ¼

    q qV q qV n þ K  r1 rh p qz qz qz qz

½1

The coordinate system is right-handed with z positive upward and the origin at the ice under-surface. V is the horizontal velocity vector in complex notation (V ¼ u þ iv), r is water density, and p is pressure. An eddy diffusivity representation is used for turbulent

198

shear stress, KðqV=qzÞ ¼ V 0 w0 , where K is the eddy diffusivity. The term nðqV=qzÞ is the viscous shear stress, where n is the kinematic molecular viscosity. The pressure gradient term, r1rhp is equal to r1 ðqp=qx þ iqp=qyÞ. The stress gradient term due to molecular viscosity is of highest inverse order in z. It varies as z2, and therefore dominates the stress balance in the molecular sublayer (Figure 1) where z is vanishingly small. As a result the viscous stress, nðqV=qzÞ, is effectively constant in the molecular sublayer, and the velocity profile is linear. The next layer away from the boundary is the surface layer. Here the relation between stress and velocity depends on the eddy viscosity, which is proportional to the length scale and velocity scale of turbulent eddies. The length scale of the turbulent eddies is proportional to the distance from the boundary, |z|. Therefore, the turbulent stress term varies as z1 and becomes larger than the viscous term beyond z greater than (1/k)(n/u*0 ), typically a fraction of a millimeter. The velocity scale in the surface layer is u*0 , where ru2*0 is equal to t0, the average shear stress at the top of the boundary layer. Thus, K is equal to ku*0 |z|, where Von Ka¨rma¨n’s constant, k, is equal to 0.4. Because the turbulent stress term dominates the equations of motion, the stress is roughly constant with depth in the surface layer. This and the linear z dependence of the eddy coefficient result in the log-layer solution or ‘law of the wall’ (eqn [2]). u 1 1 z ¼ ln z þ C ¼ ln u* k k z0

½2

C ¼  (ln z0)/k is a constant of integration. Under sea ice the surface layer is commonly 1–3 m thick. The surface layer is where the influence of the boundary roughness is imposed on the planetary boundary layer. In the presence of under-ice roughness, the average stress the ice exerts on the ocean, t0, is composed partly of skin friction due to shear and partly of form drag associated with pressure disturbances around pressure ridge keels and other roughness elements. Observations under very rough ice have shown a decrease in turbulent stress toward the surface, presumably because more of the momentum transfer is taken up by pressure forces on the rough surface. The details of this drag partition are not known. Drag partition is complicated further for cases in which stratification exists at depths shallow

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE–OCEAN INTERACTION

〈w ′T ′〉0 = T dT / dz 〈w ′S ′〉0 = S dS/dz

N  NV ≈ 0 Nz Nz

(

Molecular sublayer

U  =  du /dz

z0 = hs /30

N K NV ≈ 0 N z Nz

(

Ice

0

Z0

(

S

(

Ice

T

199

23°

 = u*20

Depth below ice (m) for u *0 = 0.01 m s

s|

v /u *0

|u| = 13.5u*0 45°

u / u*0

_ 0.1

Vice

10 _ 0.2

15

45° Outer layer

20

ifV ≈ N K NV Nz Nz

(

(

hs

_1

Surface layer

Z s| 5

Stress and velocity vectors in plan view

_ 0.3

25

 = zf /u*0 _5

_ 0.4 0

5

10

15

u / u*0 and v /u *0 Figure 1 Illustration of three regions of the planetary boundary layer under sea ice: molecular sublayer, surface layer, and outer layer. The velocity profiles are from the Rossby similarity solution (eqns [8], [9], [10] and [11]) for u*0 ¼ 0.01 m s1, z0 ¼ 0.06 m, Z* ¼ 1. The stress and velocity vector comparisons are from the same solution.

compared to the depth of roughness elements. Then it also becomes possible to transfer momentum by internal wave generation. However, for many purposes t0 is taken as the turbulent stress evaluated at z0. Laboratory studies of turbulent flow over rough surfaces suggest that z0 may be taken equal to hs/30, where hs is the characteristic height of the roughness elements. In rare situations the ice surface may be so smooth that bottom roughness and form drag are not factors in the drag partition. In such a hydrodynamically smooth situation, the turbulence is generated by shear induced instability in the flow. The surface length scale, z0, is determined by the level of turbulent stress and is proportional to the molecular sublayer thickness according to the empirically derived relation z0 ¼ 0.13(n/u*0 ). In the outer layer farthest from the boundary, the Coriolis and pressure gradient terms in eqn [1], which have no explicit z dependence, are comparable to the turbulent stress terms. The presence of the Coriolis term gives rise to a length scale, h, for the outer boundary layer equal to u*0 /f under neutral stratification. This region is far enough from the boundary so that the turbulent length scale becomes

independent of depth and in neutral conditions has been found empirically to be l ¼ xnu*0 /f, where xn is 0.05. For neutral stratification, u* and h are the independent parameters that define the velocity profile over most of the boundary layer. The ratio of the outer length scale to the surface region length scale, z0, is the surface friction Rossby number, R0 ¼ u*0 /(z0f). Solutions for the velocity in the outer layer can be derived for a wide range of conditions if we nondimensionalize the equations with these Rossby similarity parameters, u*0 /f and u*0 . However, the growth and melt of the ice produce buoyancy flux that strongly affects mixing. Melting produces a stabilizing buoyancy flux that inhibits turbulence and contracts the boundary layer. Freezing causes a destabilizing buoyancy flux that enhances turbulence and thickens the boundary layer. We can account for the buoyancy flux effect by adjusting the Rossby parameters dealing with length scale. We define the scale of the outer boundary layer as hm ¼ u*0 Z* =f . If the mixing length of the turbulence in the outer layer is lm ¼ xn u*0 Z2* =f , it interpolates in a reasonable way between known values of lm for neutral stratification (xnu*0 /f ) and stable stratification (RcL) if Z* is given

(c) 2011 Elsevier Inc. All Rights Reserved.

200

ICE–OCEAN INTERACTION

as eqn [3]. Z* ¼

  x u 1 1=2 1þ n * f Rc L

In the outer layer, eqns [5] and [8] are satisfied for nondimensional velocity given by eqn [10]. ½3

ˆ U ¼ idˆ ed z

for zrz0

½10

Rc is the critical Richardson number; the Obukhov length, L, is the ratio of shear and buoyant production of turbulent energy, ru3*0 =kg/r0 w0 S; and  /r0 w0 Sg/r is the turbulent buoyancy flux. With this Rossby similarity normalization of the equations of motion, we can derive analytical expressions for the under-ice boundary layer profile that are applicable to a range of stratification. For large |z|, V will approach the free stream geostrophic velocity, V¯ g ¼ Ug þ iVg ¼ f 1 r1 rh p. Here we will assume this is zero. However, surface stressdriven absolute velocity solutions can be superimposed on any geostrophic current. We also ignore the time variation and viscous terms and define a normalized stress equal to S ¼ ðKqV=qzÞ=u2*0 . The velocity is nondimensionalized by the friction velocity and the boundary layer thickness, U ¼ Vfhm =u2*0 , and depth is nondimensionalized by the boundary layer thickness scale, z ¼ z/hm. With these changes eqn [1] becomes eqn [4].

Thus the velocity is proportional to stress but rotated 451 to the right. As we see in the derivation of the law of the wall [2], the surface layer variation of the eddy viscosity with depth is critical to the strong shear present there. Thus eqn [10] will not give a realistic profile in the surface layer. We define the nondimensional surface layer thickness, zsl, as the depth where the surface layer mixing length, |z|, becomes equal to the outer layer mixing length, lm ¼ xn u*0 Z2* =f . We find zsl is equal to  Z* xn and applying the definition [6] gives K* as K*sl ¼  kz/Z* in the surface layer. If we approximate the stress profile [8] by a Taylor series, we can integrate [5] with K*sl substituted for K* to obtain the velocity profile in the surface layer.

iU ¼ qS=qz

Eqn [11] is analogous to [2] except for the introduction of the dˆðzsl  zÞ term. This is the direct result of accounting for the stress gradient in the surface layer. This term is small compared to the logarithmic gradient. Figure 1 illustrates the stress and velocity vectors at various points in the boundary layer as modeled by eqns [8] through [11]. For neutral conditions the nondimensional boundary layer thickness is typically 0.4 (dimensional thickness is 0.4u* /f). Through the outer layer, the velocity vector is 451 to the right of the stress vector as a consequence of the idˆ peið451Þ multiplier in [10]. As the ice surface is approached through the surface layer, the stress vector rotates 10–201 to the left to reach the surface direction. However, in the surface layer the velocity shear in the direction of the surface stress is great because of the logarithmic profile. Thus, as the surface is approached, the velocity veers to the left twice as much as stress. At the surface the velocity is about 231 to the right of the surface stress. It is commonly useful to relate the stress on underice surface to the relative velocity between ice and water a neutral-stratification drag coefficient, 2 where Cz is the drag coefficient for ru2*0 ¼ rCz VðzÞ depth z. If z is in the log-layer, eqn [2] can be used to derive the relation between ice roughness and the drag coefficient. We find that Cz ¼ k2[ln(z/z0)]2.

½4

In terms of nondimensional variables the constitutive law is given by eqn [5]. S ¼ K* qU=qz

½5

The nondiemensional eddy coefficient is given by eqn [6]. K* ¼ ku*0 lm =fh2m ¼ kxn

½6

Eqn [6] is the Rossby similarity relation that is the key to providing similarity solutions for stable and neutral conditions. It even provides workable results for slightly unstable conditions. Eqns [4] and [5] can be combined in an equation for nondimensionalized stress (eqn [7]). ði=K* ÞS ¼ dSd=z

½7

This has the solution eqns [8]. ˆ

S ¼ ed z

½8

dˆ ¼ ði=K* Þ1=2

½9

Eqn [8] attenuates and rotates (to the right in the Northern Hemisphere) with depth. It duplicates the salient features found in data and sophisticated numerical models.

UðzÞ  Uðzsl Þ ¼

    Z* z ln sl þ dˆ ðzsl  zÞ k z0

for z  z0 ½11

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE–OCEAN INTERACTION

Clearly values of the drag coefficient can vary widely depending on the under-ice roughness. Typical values of z0 range from 1 to 10 cm under pack ice. A commonly referenced value for the Arctic is 6 cm, which produces a drag coefficient at the outer edge of the log layer of 9.4  103 (Figure 1). If the reference depth is outside the log layer, the drag coefficient formulation is poorly posed because of the turning in the boundary layer. For neutral conditions, eqns [10] and [11] can be used to obtain a Rossby similarity drag law that yields the nondimensional surface drift relative to the geostrophic current for unit nondimensional surface stress (eqn [12]). U0 ¼

V0 1 ¼ ð½lnðR0 Þ  A  iBÞ u *0 k

½12

Here

A¼

sffiffiffiffiffiffiffiffi rffiffiffiffiffiffi! k xn þ 1  ln xn  D2:2 2k 2xn sffiffiffiffiffiffiffiffi rffiffiffiffiffiffi k xn D2:3 þ B¼ 2k 2xn

salt may be expressed in kinematic form as eqns [14] and [15]. q_ ¼ /w0 T 0 S0  w0 QL

ðwith units K m s1 Þ

ðw0 þ wi ÞðS0  Sice Þ ¼ /w0 S0 S0

This Rossby similarity drag law for outside the surface layer results in a surface stress that is proportional to V1.8 rather than V2, a result that is supported by observational evidence, and can be significant at high velocities.

Heat and Mass Balance at the Ice–Ocean Interface: Wintertime Convection The energy balance at the ice–ocean interface not only exerts major influence over the ice mass balance but also dictates the seasonal evolution of upper ocean salinity and temperature structure. At low temperature, water density is controlled mainly by salinity. Salt is rejected during freezing, so that buoyancy flux from basal growth (or ablation), combined with turbulent mixing during storms, determines the depth of the well-mixed layer. Vertical motion of the ice–ocean interface depends on isostatic adjustment as the ice melts or freezes. The interface velocity is w0 þ wi where w0 ¼ ðrice =rÞh_ b ; h_ b is the basal growth rate, and wi represents isostatic adjustment to runoff of surface melt and percolation of water through the ice cover. In an infinitesimal control volume following the ice–ocean interface, conservation of heat and

½14

ðwith units psu m s1 Þ

½15 where q_ ¼ Hice =ðrcp Þ is flux (Hice) conducted away from the interface in the ice; r is water density; cp is specific heat of seawater; /w0 T0 S0 is the kinematic turbulent heat flux from the ocean; QL is the latent heat of fusion (adjusted for brine volume) divided by cp; S0 is salinity in the control volume, Sice is ice salinity, and /w0 S0 S0 is turbulent salinity flux. Fluid in the control volume is assumed to be at its freezing temperature, approximated by the freezing line (eqn [16]). T0 ¼ mS0

½13

201

½16

By standard closure, turbulent fluxes are expressed in terms of mean flow properties (eqns [17] and [18]). hw0 T 0 i0 ¼ ch u*0 dT

½17

hw0 S0 i0 ¼ cS u*0 dS

½18

u*0 is the square root of kinematic turbulent stress at the interface (friction velocity); dT ¼ T  T0 and dS ¼ S  S0 are differences between far-field and interface temperature and salinity; and ch and cS are turbulent exchange coefficients termed Stanton numbers. The isostatic basal melt rate, w0 is the key factor in interface thermodynamics, and in combination with wi it determines the salinity flux. A first-order approach to calculating w0 that is often sufficiently accurate (relative to uncertainties in forcing parameters) when melting or freezing is slow, is to assume that S0 ¼ S, the far-field salinity, and that ch is constant. Combining [14], [16], and [17] gives eqn [19]. w0 ¼

ch u*0 ðT þ mSÞ  q_ QL

½19

Salinity flux is determined from [15]. Note the cS is not used, and that this technique fixes (unrealistically) the temperature at the interface to be the mixed layer freezing temperature. A more sophisticated approach is required when melting or freezing is intense. Manipulation of [14]

(c) 2011 Elsevier Inc. All Rights Reserved.

202

ICE–OCEAN INTERACTION

through [18] produces a quadratic equation for w0 (eqn [20]). SL 2 w þ ðST þ SL cS  Sice Þw0 þ ðu*0 cS þ wi ÞST u*0 0 þ u*0 cS S  wi Sice ¼ 0  ST ¼

 q_  T =m ch u*0

½20

and

SL ¼ QL =ðmch Þ ½21

Here ch and cS (turbulent Stanton numbers for heat and salt) are both important and not necessarily the same. Melting or freezing will decrease or increase S0 relative to far-field salinity, with corresponding changes in T0. The Marginal Ice Zone Experiments (MIZEX) in the 1980s showed that existing ice–ocean turbulent transfer models overestimated melt rates by a wide factor. It became clear that the rates of heat and mass transfer were less than momentum transfer (by an order of magnitude or more), and were being controlled by molecular effects in thin sublayers adjacent to the interface. If it is assumed that the extent of the sublayers is proportional to the bottom roughness scale, z0, then dimensional analysis suggests that the Stanton numbers (nondimensional heat and salinity flux) should depend mainly on two other dimensionless groups, the turbulent Reynolds number, Re* ¼ u*0 z0/n, where n is molecular viscosity, and the Prandtl (Schmidt) numbers, n/nT(S), where nT and nS are molecular diffusivities for heat and salt. Laboratory studies of heat and mass transfer over hydraulically rough surfaces suggested approximate expressions for the Stanton numbers of the form shown in eqn [22]. chðSÞ ¼

  /w0 TðSÞ0 S0 n 2=3 pðRe* Þ1=2 u*0 dTðSÞ nTðSÞ

½22

The Stanton number, ch, has been determined in several turbulent heat flux studies since the original MIZEX experiment, under differing ice types with z0 values ranging from less than a millimeter (eastern Weddell Sea) to several centimeters (Greenland Sea MIZ). According to [22], ch should vary by almost a factor of 10. Instead, it is surprisingly constant, ranging from about 0.005 to 0.006, implying that the Reynolds number dependence from laboratory results cannot be extrapolated directly to sea ice. If the Prandtl number dependence of [14] holds, the ratio ch /cS ¼ (nh/nS)2/3 is approximately 30. Under conditions of rapid freezing, the solution of [20] with this ratio leads to significant supercooling of

the water column, because heat extraction far outpaces salt injection in what is called double diffusion. This result has caused some concern. Because the amount of heat represented by this supercooling is substantial, it has been hypothesized that ice may spontaneously form in the supercooled layer and drift upward in the form of frazil ice crystals. This explanation has not been supported by ice core sampling, which shows no evidence of widespread frazil ice formation beyond that at the surface of open water. The physics of the freezing process suggest that the seeming paradox of the supercooled boundary layer may be realistic without spontaneous frazil formation. When a parcel of water starts to solidify into an ice crystal, energy is released in proportion to the volume of the parcel. At large scales this manifests itself as the latent heat of fusion. However, as the parcel solidifies, energy is also required to form the surface of the solid. This surface energy penalty is proportional to the surface area of the parcel and depends on other factors including the physical character of any nucleating material. In any event, if the parcel is very small the ratio of parcel volume to surface area will be so small that the energy released as the volume solidifies is less than the energy needed to create the new solid surface. For this reason, ice crystals cannot form even in supercooled water without a nucleating site of sufficient size and suitable character. In the clean waters of the polar regions, the nearest suitable site may only be at the underside of the ice cover where the new ice can form with no nucleation barrier. Therefore, it is possible to maintain supercooled conditions in the boundary layer without frazil ice formation. Furthermore, recent results suggest that supercooling in the uppermost part of the boundary layer may be intrinsic to the ice formation process. Sea ice is a porous mixture of pure ice and high-salinity liquid water (i.e., brine). The bottom surface of a growing ice floe consists of vertically oriented pure ice platelets separated by vertical layers of concentrated brine. This platelet–brine sandwich (on edge) structure is on the scale of a fraction of a millimeter, and its formation is controlled by molecular diffusion of heat and salt. The low solid solubility of the salt in the ice lattice results in an increase of the salinity of water in the layer above the advancing freezing interface. Because heat diffuses more rapidly than salt at these scales, the cold brine tends to supercool the water below the ice–water interface. With this local supercooling, any disturbance of the ice bottom will tend to grow spontaneously. The conditions of sea ice growth are such that this

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE–OCEAN INTERACTION

instability is always present. Continued growth results in additional rejection of salt, some fraction of which is trapped in the brine layers, and consequently the interfacial region of the ice sheet continues to experience constitutional supercooling. Also, anisotropy in the molecular attachment efficiency intrinsic to the crystal structure of the ice platelets creates an additional supercooling in the interfacial region. The net result is that heat is extracted from the top of the water column at the rate needed to maintain its temperature near but slightly below the equilibrium freezing temperature as salt is added. This and the convective processes in the growing ice may imply that ch/cS ¼ 1 during freezing. The situation with melting may be quite different, since the physical properties of the interface change dramatically. Observations to date suggest that ch remains relatively unchanged with variable ice type and mixed layer temperature elevation above freezing. A value of 5.5  103 is representative. cS is not so well known, since direct measurements of /w0 S0 S are relatively rare. The dependence of the exchange coefficients on Prandtl and Schmidt numbers is not clear, and will only be resolved with more research.

Effects of Horizontal Inhomogeneity: Wintertime Buoyancy Flux Although the under-ice surface may be homogeneous over ice floes hundreds of meters in extent, the key fluxes of heat and salt are characteristically nonuniform. As ice drifts under the action of wind stress, the ice cover is deformed. Some areas are forced together, producing ridging and thick ice, and some areas open in long, thin cracks called leads. In special circumstances the ice may form large, unit-aspectratio openings called polynyas. In winter the openings in the ice expose the sea water directly to cold air without an intervening layer of insulating sea ice. This results in rapid freezing. As the ice forms, it rejects salt and results in unstable stratification of the boundary layer beneath open water or thin ice. These effects are so important that, even though such areas may account for less than 10% of the ice cover, they may account for over half the total ice growth and salt flux to the ocean. Thus the dominant buoyancy flux is not homogeneous but is concentrated in narrow bands or patches. Similarly, in the summer solar radiation is reflected from the ice but is nearly completely absorbed by open water. Fresh water from summertime surface melt tends to drain

203

into leads, making them sources of fresh water flux as well. The effect of wintertime convection in leads is illustrated in Figure 2. It shows two extremes in the upper ocean response. Figure 2A shows what we might expect in the case of a stationary lead. As the surface freezes, salt is rejected and forms more dense water that sinks under the lead. This sets up a circulation with fresh water flowing in from the sides near the surface and dense water flowing away from the lead at the base of the mixed layer. Figure 2B illustrates the case in which the lead is embedded in ice moving at a velocity great enough to produce a well-developed turbulent boundary layer (e.g. 0.2 m s1). If the mixed layer is fully turbulent, the cellular convection pattern may not occur; rather, the salt rejected at the surface may simply mix into the surface boundary layer. The impact of nonhomogeneous surface buoyancy flux on the boundary layer can also be characterized by the equations of motion. The viscous terms in eqn [1] can be neglected at the scales we discuss here, but the possibility of vertical motion associated with large-scale convection requires that we include the vertical component of velocity. For steady state we have eqn [23].   q qV¯ ¯ ¯ ¯ ¯ K  r1 rp VdrV þ f  V ¼ qz q

½23

V¯ is the velocity vector including the mean vertical velocity w; f¯ is the Coriolis parameter times the vertical unit vector. The advective acceleration term, ¯ ¯ and pressure gradient term are necessary to Vdr V, account for the horizontal inhomogeneity that is caused by the salinity flux at the lead surface. The condition that separates the free convection regime of Figure 2A and the forced convection regime of Figure 2B is expressed by the relative magnitude of the pressure gradient, r1rhp, and turbulent stress, q=qzðKqV=qzÞ, terms in [23]. This ratio can be derived with addition of mass conservation and salt conservation equations, and if we assume the vertical equation is hydrostatic, qp=qz ¼ gr ¼ gMS, where M is the sensitivity of density to salinity. If we nondimensionalize the equations by the ice velocity Ui, mixed-layer depth, d, average salt flux at the lead surface, FS, and friction velocity, u*0, the ratio of the pressure gradient term to the turbulent stress term scales as eqn [24].

(c) 2011 Elsevier Inc. All Rights Reserved.

L0 ¼

gMFS d r0 Ui u2*0

½24

204

ICE–OCEAN INTERACTION

Free convection Vice ~ 0

z

Forced convection

Lead y

Lead

LL

Vice u*

x

FS

du (x)

u (z )

Thick sea ice dd (x)

Increased S dml u (z )

Internal boundary layers, depth = d (x )

 (z )

(A)

(B)

Figure 2 Modes of lead convection. (A) The free convection pattern that results when freezing and salt flux are strong, and the relative velocity of the ice is low. Cellular patterns of convective overturning are driven by pressure gradients that arise from the salinity distrurbance due to ice formation. (B) The forced convection regime that exists when ice motion is strong. The salinity flux and change in surface stress in the lead cause a change in the character of the boundary layer that grows deeper downstream. The balance of forces is primarily Coriolis and turbulent diffusion of momentum. (From Morison JH, MCPhee MG, Curtin T and Paulson CA (1992) The oceanography of winter leads. Journal of Geophysical Research 97: 11199–11218.)

autonomous underwater vehicle. Using the vehicle vertical motion as a proxy for vertical water velocity, it is also possible to estimate the salt flux w0 S0 . The lead was moving at 0.04 m s1, and estimates of salt flux put L0 between 4 and 11 (free convection in Figure 3). Salinity increased in the downstream direction across the lead and reached a sharp maximum Water temperature _ air temperature (˚C)

If this lead number is small because the ice is moving rapidly or the salt flux is small, the pressure gradient term is not significant in [23]. In this forced convection case, illustrated in Figure 2B, the boundary layer behaves as in the horizontally homogeneous case except that salt is advected and diffused away from the lead in the turbulent boundary layer. If the lead number is large because the ice is moving slowly or the salt flux is large, the pressure gradient term is significant. In this free convention case the salinity disturbance is not advected away, but builds up under the lead. This creates pressure imbalances that can drive the type of cellular motion shown in Figure 2A. Figure 3 shows conditions for which the lead number is unity for a range of ice thickness. Here the salt flux has been parametrized in terms of the air– sea temperature difference, and stress has been parametrized in terms of Ui. The figure shows the locus of points where L0 is equal to unity. For typical winter and spring conditions, L0 is close to 1, indicating that a mix of free and forced convection is common. Conditions where lead convection features have been observed are also shown in Figure 3. Most of these are in the free convection regime, probably because they are more obvious during quiet conditions. There have been several dedicated efforts to study the effects of wintertime lead convection. The most recent example was the 1992 Lead Experiment (LeadEx) in the Beaufort Sea. Figure 4 illustrates the average salinity profile at 9 m under a nearly stationary lead. The data was gathered with an

30 20 15 10 hi = 5 cm

′71

25 A3

20

A4

′92 Lead 3

′92 Lead 4

h i = 0 cm A

15 Free convection

10

Forced convection

5

L0 = 1 L0 t = 1

0 0

0.02

0.04 0.06 0.08

0.10 0.12

0.14 0.16

_

Ice velocity (m s 1) Figure 3 Air–water temperature difference versus Ui for L0 equal to 1 for various ice thicknesses, hi. Also shown are the temperature difference and ice velocity values for several observations of lead convection features such as underice plumes. Most of these are in the free convection regime: 0 71 denotes the AIDJEX pilot study; A3 denotes the 1974 AIDJEX Lead Experiment – lead 3 (ALEX3); A4 denotes ALEX4; A denotes the 1976 Arctic Mixed Layer Experiment; and 0 92 Lead 4 denotes the 1992 LeadEx lead 4. LeadEx lead 3 (0 92 Lead 3) was close to L0 ¼ 1. (From Morison JH, McPhee MG, Curtin T and Paulson CA (1992) The oceanography of winter leads. Journal of Geophysical Research 97: 11199–11218.)

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE–OCEAN INTERACTION

205

ICE

ICE

LEAD

2

0

Relative current

_

Salt flux down (10 5 kg m 2 s 1)

_3

Salinity fluctuation (10 PSU)

4

_

_

~0.02−0.04 m s

_1

Lead average

5

= 6.0 × 10

_6

0

_ 250

_ 200

_ 150

_ 100

_ 50

0

50

100

150

200

Cross-lead distance (m) Figure 4 Composites of S 0 and w 0 S 0 at 9 m depth measured with an autonomous underwater vehicle during four runs under lead 4 at the 1992 Lead Experiment. The horizontal profile data have been collected in 1 m bins. (From Morison JH, McPhee MG (1998) Lead convection measured with an autonomous underwater vechicle, Journal of Geophysical Research 103: 3257–3281.)

at the downstream edge. The salt flux was highest near the lead edges, but particularly at the downstream edge. With even a slight current, the downstream edge plume is enhanced by several factors. The vorticity in the boundary layer reinforces the horizontal density gradient at the downstream edge and counters the gradient at the upstream edge. The salt excess is greatest at the downstream edge by virtue of the salt that is advected from the upstream lead surface. The downstream edge plume is also enhanced by the vertical motion of water at the surface due to water the horizontal flow being forced downward under the ice edge. Figure 5 shows the salt flux beneath a 1000 m wide lead moving at 0.14 m s1 with L0 equal to about 1 (Figure 3). Here the salt flux is more evenly spread under the lead surface. The salt flux derived from the direct w0 S0 correlation method does show some enhancement at the lead edge. This may be partly due to the influence of pressure gradient forces and the reasons cited for the free convection case described above. The other factor that influences the convective pattern is the lead width. In the case of the 100 m lead in even a weak current, the convection may not be fully developed until the downstream edge is reached. For the 1000 m lead of the second case, the convection under the downstream portion of the

lead was a fully developed unstable boundary layer. The energy-containing eddies filled the mixed layer and their dominant horizontal wavelength was equal to about twice the mixed layer depth.

Effects of Horizontal Inhomogeneity: Summertime Buoyancy Flux The behavior of the boundary layer under summer leads is relatively unknown compared to the winter lead process. Because of the important climate consequences, it is a subject of increasing interest. Summertime leads are thought to exhibit a critical climate-related feature of air–sea–ice interaction, icealbedo feedback. This is because leads are windows that allow solar radiation to enter the ocean. The proportion of radiation that is reflected (albedo) from sea ice and snow is high (0.6–0.9) while that from open water is low (0.1). The fate of the heat that enters summer leads is important. If it penetrates below the draft of the ice, it warms the boundary layer and is available to melt the bottom of the ice over a large area. If most of the heat is trapped in the lead above the draft of the ice, it will be available to melt small pieces of ice and the ice

(c) 2011 Elsevier Inc. All Rights Reserved.

206

ICE–OCEAN INTERACTION

30

LEAD

_6

_2

_1

Salt flux (10 kg m s )

25 20 Salt flux from Ebp-IDM

15 10

Lead average flux _6 7.762 × 10

Downstream average _ flux = 3.362 × 10 6

5 0 w ′S ′ salt flux, 56-m bin composite

_5

−6

average = 9.41× 10 , lead average = 1.19 ×10

−5

_ 10 _ 300

_ 200

_ 100

0

100

200

300

400

Distance upstream from the lead edge (m) Figure 5 Composite average for autonomous underwater vehicle runs 1 to 5 at lead 3 of the 1992 Lead Experiment. The salinity is band-passed at 1 rad m1 and is indicative of the turbulence level and is used to estimate the salt flux by the Ebp  IDM method of Morison and McPee (1998). The salt flux is elevated in the lead and decreases beyond about 72 m downstream of the lead edge. The composite average of w 0 S 0 in 56 m bins for the same runs is also shown in the center panel. The average flux and the decrease downstream are about the same as given by the Ebp  IDM method, but w 0 S 0 suggests elevated fluxes near the lead edge. (From Morison JH, McPhee MG (1998) Lead convection measured with an autonomous underwater vechicle, Journal of Geophysical Research 103: 3257–3281.)

floe edges. In the latter case the area of ice will be reduced and the area of open water increased. This allows even more solar radiation to enter the upper ocean, resulting in a positive feedback. This process may greatly affect the energy balance of an ice-covered sea. The critical unknown is the partition of heating between lateral melt of the floe edges and bottom melt. There are fundamental similarities between the summertime and wintertime lead problems. The equations of motion ([15]–[24]) are virtually identical. Only the sign of the buoyancy flux is opposite. The heat flux is important to summer leads and tends to decrease the density of the surface waters. However, as with winter leads, the buoyancy flux is controlled mainly by salt. As the top surface of the ice melts, much of the water that does not collect in melt ponds on the ice surface instead runs into the leads. If the ambient ice velocity is low, ice melt from the bottom surface will tend to flow upward and collect in the leads as well. Thus leads are the site of a concentrated flux of fresh water accumulated over large areas of ice. If this flux, FS, into the lead is negative enough relative to the momentum flux represented by u*0, the lead number, L0, will be a large negative number and shear production of turbulent energy will not be able to overcome the stabilizing buoyant production. This means turbulent mixing will be weak beneath the lead surface and a layer of fresh water will accumulate near the surface of the lead. The stratification

at the bottom of this fresh water layer may be strong enough to prevent mixing until a storm produces a substantial stress. This will be made even more difficult than in the winter situation because of the effect of stabilizing buoyancy flux on the boundary layer generally. The only way the fresh water will be mixed downward is by forced convection; there is no analogue to the wintertime free convection regime. When there is sufficient stress to mix out a summertime lead, the pattern must resemble that of the forced convection regime in Figure 2A. At the upstream edge of the lead, fresh warm water will be mixed downward in an internal boundary layer that increases in thickness downstream until it reaches the steady-state boundary layer thickness appropriate for that buoyancy flux or the ambient mixed layer depth. The rate of growth should scale with the local value of u*0 (or perhaps u*0 Z* ). At the downstream edge, another boundary layer conforming to the under-ice buoyancy flux and surface stress will begin to grow at a rate roughly scaling with the local u*0. In spite of the generally stabilizing buoyancy flux, this process has the effect of placing colder, more saline water from under the ice on top of fresher and warmer (consequently lighter) water drawn from the lead. Thus, even embedded in the stable summer boundary layer, the horizontal inhomogeneity due to leads may create pockets of instability and more rapid mixing than might be expected on the basis of average conditions.

(c) 2011 Elsevier Inc. All Rights Reserved.

207

ICE–OCEAN INTERACTION

Recent studies of summertime lead convection at the 1997–98 Surface Heat Budget of the Arctic experiment saw the salinity decrease in the upper 1 m of leads to near zero and temperatures increase to more than 01C. Only when ice velocities were driven by the wind to speeds of nearly 0.2 m s1 were these layers broken down and the fresh, warm water mixed into the upper ocean. At these times the heat flux measured at 5 m depth reached values over 100 W m2. The criteria for the onset of mixing are being studied along with the net effect of the growing internal boundary layers. Even with an understanding of the mixing process, it will be a challenge to apply this information to larger-scale models, because the mixing is nonlinearly dependent on the history of calm periods and strong radiation.

Internal Waves and Their Interaction with the Ice Cover One of the first studies of internal waves originated with observations made by Nansen during his 1883 expedition. It did not actually involve interaction with the ice cover, but with his ship the Fram. He found that while cruising areas of the Siberian shelf covered with a thin layer of brackish water, the Fram had great difficulty making any headway. It was hypothesized by V. Bjerknes and proved by Ekman that this ‘dead water’ phenomenon was caused by the drag of the internal wave wake produced by the ship’s hull as it passed through the shallow surface layer. This suggests that internal wave generation by deep keels may cause drag on moving ice. Evidence of internal wave generation by keels has been observed by several authors, but estimates of the amount of drag vary widely. This is due mainly to wide differences in the separation of the stratified pycnocline and the keels. The drag produced by under-ice roughness of amplitude h0 with horizontal wavenumber b moving at velocity Vi (magnitude vi) over a pycnocline with stratification given by Brunt–Vaisala frequency, N, a depth d below the ice–ocean interface, can be expressed as an effective internal wave stress (eqn [25]), where Cwd (eqn [26]) accounts for the drag that would exist if there were no mixed layer between the ice and the pycnocline. Siw ¼ GCwd Vi

½25

above which the waves are evanescent (bc ¼ N/vi). G is an attenuation factor that accounts for the separation of the pycnocline from the ice by the mixed layer of depth d (eqn [27]). 8" 911 #2 < = 2 bDb N G ¼ @sinh2 ðbdÞ cothðbdÞ  2 2 þ 2 2  1 A : ; ui b x ui bx 0

½27 Db is the strength of the buoyancy jump at the base of the mixed layer. For wavenumbers of interest and d much bigger than about 10 m, G becomes small and internal wave drag is negligible. Thus it is not a factor in the central Arctic over most of the year. However, in the summer pack ice, and many times in the marginal ice zone, stratification will extend to or close to the surface. Then internal wave drag can be at least as important as form drag. The ice cover also uniquely affects the ambient internal wave field. In most of the world ocean the internal wave energy level, when normalized for stratification, is remarkably uniform. It has been established by numerous studies that the internal wave energy in the Arctic Ocean is typically several times lower. In part this may be due to the absence of surface gravity waves. The other likely reason is that friction on the underside of the ice damps internal waves. Decomposing the internal wave field into vertical modes, one finds the mode shapes for horizontal velocity are a maximum at the surface. This is perfectly acceptable in the open water situation. However, at the horizontal scales of most internal waves, an ice cover imposes a surface boundary condition of zero horizontal velocity. The effect of this can be estimated by assuming that a time-varying boundary layer is associated with each spectral component of the internal wave field. This is not rigorously correct because all the modes interact in the same nonlinear boundary layer, and are thereby coupled. However, in the presence of a dominant, steady current due to ice motion, the effect on the internal wave modes can be linearized and considered separately. The near-surface internal wave velocity can be approximated as a sum of rotary components (eqn [28]). VðzÞ ¼

M P

Dn ðzÞeion t ¼

n¼0

Cwd ¼ 12b2x h0 ½ðb2c =b2x Þ  11=2

M P

½An ðzÞ þ iBn ðzÞeion t

n¼0

½26

½28

The wavenumber in the direction of the relative ice velocity, Vi, is bx, and bc is the critical wave number

The internal wave motion away from the boundary DNn can be subtracted from the linear

(c) 2011 Elsevier Inc. All Rights Reserved.

208

ICE–OCEAN INTERACTION

time-varying boundary layer equation (eqn [1] with the addition of the time variation acceleration, qV/qt). This yields an equation for each rotary component of velocity in the boundary layer (eqn [29]). iðon þ f ÞðDn  DNn Þ ¼

Dn ¼ 0 Dn ¼ DNn

q qDn K qz qz

½29

at z ¼ z0 at z ¼ d

This oscillating boundary layer equation can be solved for K of the form K ¼ ku*0 zexp (  6| þ f |zu*0 ). When we do this for representative internal wave conditions in the Arctic and compute the energy dissipation, we find the timescale required to dissipate the internal wave energy through underice friction is 32 days. This is a factor of 3 smaller than is typical for open ocean conditions. Assuming a steady state with internal wave forcing and other dissipation mechanisms in place, the under-ice boundary layer will result in a 75% reduction in steady-state internal wave energy. This suggests the effect of the under-ice boundary layer is critical to the unique character of internal waves in ice-covered seas.

Outstanding Issues The outstanding issue of ice–ocean interaction is how the small-scale processes in the ice and at the interface affect the exchange between the ice and water. This is arguably most urgent in the case of heat and salt exchange during ice growth. When we apply laboratory-derived concepts for the diffusion of heat and salt to the ice–ocean interface, we get results that are not supported by observation, such as spontaneous frazil ice formation and large ocean heat flux under thin ice. These results are causing significant errors in largescale models. They stem from a molecular sublayer model of the ice–ocean interface (Figure 1) and the difference between the molecular diffusivities of heat and salt. What seems to be wrong is the molecular sublayer model. Recent results in the microphysics of ice growth reveal that the structure and thermodynamics of the growing ice produce instabilities and convection within the ice and extending into the water. The ice surface is thus not a passive, smooth surface covered with a thin molecular layer. Rather it is field of jets emitting plumes of supercooled, high-salinity water at a very small scale. This type of

unstable convection likely tends to equalize the diffusion of heat and salt relative to the apparently unrealistic parameterizations we are using now. Similarly, we do not really understand how the turbulent stress we might measure in the surface layer is converted to drag on the ice. Certainly a portion of this is through viscous friction in the molecular sublayer. However, in most cases the underside of the ice is not hydrodynamically smooth, which suggests that pressure force acting on the bottom roughness elements are ultimately transferring a large share of the momentum. Understanding this will require perceptual breakthroughs in our view of how turbulence and the mean flow interact with a rough surface buried in a boundary layer. Achieving this understanding is complicated greatly by a lack of contemporaneous measurements of turbulence and under-ice topography at the appropriate scales. This drag partition problem is general and not limited to the under-ice boundary layer. However, the marvelous laboratory that the under-ice boundary layer provides may be the place to solve it.

See also Coupled Sea Ice–Ocean Models. Internal Waves. Sea Ice: Overview. Sub Ice-Shelf Circulation and Processes. Under-Ice Boundary Layer.

Further Reading Johannessen OM, Muench RD, and Overland JE (eds.) (1994) The Polar Oceans and Their Role in Shaping the Global Environment: The Nansen Centennial Volume. Washington, DC: American Geophysical Union. McPhee MG (1994) On the turbulent mixing length in the oceanic boundary layer. Journal of Physical Oceanography 24: 2014--2031. Morison JH, McPhee MG, and Maykutt GA (1987) Boundary layer, upper ocean and ice observations in the Greenland Sea marginal ice zone. Journal of Geophysical Research 92(C7): 6987--7011. Morison JH and McPhee MG (1998) Lead convection measured with an autonomous underwater vehicle. Journal of Geophysical Research 103(C2): 3257--3281. Smith WO (ed.) (1990) Polar Oceanography. San Diego, CA: Academic Press. Wettlaofer JS (1999) Ice surfaces: macroscopic effects of microscopic structure. Philosphical Transactions of the Royal Society of London A 357: 3403--3425.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE SHELF STABILITY C. S. M. Doake, British Antarctic Survey, Cambridge, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1282–1290, & 2001, Elsevier Ltd.

Introduction Ice shelves are floating ice sheets and are found mainly around the Antarctic continent (Figure 1). They can range in size up to 500 000 km2 and in thickness up to 2000 m. Most are fed by ice streams and outlet glaciers, but some are formed by icebergs welded together by sea ice and surface accumulation. They exist in embayments where the shape of the bay and the presence of ice rises, or pinning points, plays an important role in their stability. Ice shelves lose mass by basal melting, which modifies the properties of the underlying water mass and eventually influences the circulation of the global ocean, and by calving. Icebergs, sometimes more than 100 km in length, break off intermittently from the continent and drift to lower latitudes, usually breaking up and melting by the time they reach the Antarctic Convergence. The lowest latitude that ice shelves can exist is determined by the mean annual air temperature. In the Antarctic Peninsula a critical isotherm of about 51C seems to represent the limit of viability. In the last 40 years or so, several ice shelves have disintegrated in response to a measured atmospheric warming trend in the western and northern part of the Antarctic Peninsula. However, this warming trend is not expected to affect the stability of the larger ice shelves further south such as Filchner–Ronne or Ross, in the near future.

What are Ice Shelves? Physical and Geographical Setting

An ice shelf is a floating ice sheet, attached to land where ice is grounded along the coastline. Nourished mainly by glaciers and ice streams flowing off the land, ice shelves are distinct from sea ice, which is formed by freezing of sea water. Most ice shelves are found in Antarctica where the largest can cover areas of 500 000 km2 (e.g. Ross Ice Shelf and Filchner–Ronne Ice Shelf). Thicknesses vary from nearly 2000 m, around for example parts of the grounding line of Ronne Ice Shelf, to about 100 m at the seaward edge known as the ice front.

Typically, ice shelves exist in embayments, constrained by side walls until they diverge too much for the ice to remain in contact. Thus the geometry of the coastline is important for determining both where an ice shelf will exist and the position of the ice front. There are often localized grounding points on sea bed shoals, forming ice rises and ice rumples, both in the interior of the ice shelf and along the ice front. These pinning points provide restraint and cause the ice shelf to be thicker than if it were not pinned. Ice flows from the land to the ice front. Input velocities range from near zero at a shear margin (e.g., with a land boundary), to several hundred meters per year at the grounding line where ice streams and outlet glaciers enter. At the ice front velocities can reach up to several kilometers per year. A characteristic of ice shelves is that the (horizontal) velocity is almost the same at all depths, whereas in glaciers and grounded ice sheets the velocity decreases with depth (Figure 2). In the Antarctic, most ice shelves have net surface accumulation although there may be intensive summer melt which floods the surface. The basal regime is controlled by the subice circulation. Basal melting is often high near both the grounding line and the ice front. Marine ice can accumulate in the intermediate areas, where water at its in situ freezing point upwells and produces frazil ice crystals. Surface temperatures will be near the mean annual air temperature whereas the basal temperature will be at the freezing point of the water. Therefore the coldest ice is normally in the upper layers. Ice shelves normally form where ice flows smoothly off the land as ice streams or outlet glaciers. In some areas, however, ice breaks off at the coastline and reforms as icebergs welded together by frozen sea ice and surface accumulation. The processes of formation and decay are likely to operate under very different conditions. An ice shelf is unlikely to reform under the same climatic conditions which caused it to decay. Complete collapse is a catastrophic process and rebuilding requires a major change in the controlling parameters. However, there is evidence of cyclicity on periods of a few hundred years in some areas. Collapse of the northernmost section of Larsen Ice Shelf within a few days in January 1995 indicates that, after retreat beyond a critical limit, ice shelves can disintegrate rapidly. The breakup history of two northern sections of Larsen Ice Shelf (Larsen A and Larsen B) between 1986 and 1997 has been used to determine a stability criterion for ice shelves.

(c) 2011 Elsevier Inc. All Rights Reserved.

209

210

ICE SHELF STABILITY

Figure 1 Distribution of ice shelves around the coast of Antarctica.

Figure 2 Cartoon of ice shelf in bay with ice streams.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE SHELF STABILITY

Analysis of various ice-shelf configurations reveals characteristic patterns in the strain-rates near the ice front which have been used to describe the stability of the ice shelf.

Why are Ice Shelves Important? Ice shelves are one of the most active parts of the ice sheet system. They interact with the inland ice sheet, glaciers and ice streams which flow into them, and with the sea into which they eventually melt, either directly or as icebergs. A typical Antarctic ice shelf will steadily advance until its ice front undergoes periodic calving, generating icebergs. Calving can occur over a wide range of time and spatial scales. Most ice fronts will experience a quasi-continuous ‘nibbling’ away of the order of tens to hundreds of meters per year, where ice cliffs collapse to form bergy bits and brash ice. The largest icebergs may be more than 100 km in size, but will calve only at intervals of 50 years or more. Thus when charting the size and behavior of an ice shelf it can be difficult to separate and identify stable changes from unstable ones. Mass balance calculations show that calving of icebergs is the largest factor in the attrition of the Antarctic Ice Sheet. Estimates based on both ship and satellite data suggest that iceberg calving is only slightly less than the total annual accumulation. Ice shelf melting at the base is the other principal element in the attrition of the ice sheet, with approximately 80% of all ice shelf melting occurring at distances greater than 100 km from the ice front. Although most of the mass lost in Antarctica is through ice shelves, there is still uncertainty about the role ice shelves play in regulating flow off the land, which is the important component for sea-level changes. Because ice shelves are already floating, they do not affect sea level when they breakup. Ice shelves are sensitive indicators of climate change. Those around the Antarctic Peninsula have shown a pattern of gradual retreat since about 1950, associated with a regional atmospheric warming and increased summer surface melt. The effects of climate change on the mass balance of the Antarctic Ice Sheet and hence global sea level are unclear. There have been no noticeable changes in the inland ice sheet from the collapse of the Antarctic Peninsula ice shelves. This is because most of the margins with the inland ice sheet form a sharp transition zone, making the ice sheet dynamics independent of the state of the ice shelf. This is not the case with a grounding line on an ice stream where there is a smooth transition zone, but generally the importance of the local stresses there diminishes rapidly upstream.

211

An important role of ice shelves in the climate system is that subice-shelf freezing and melting processes influence the formation of Antarctic Bottom Water, which in part helps to drive the oceanic thermohaline circulation (see Sub Ice-Shelf Circulation and Processes).

How have Ice Shelves Changed in the Past? As the polar ice sheets have waxed and waned during ice age cycles, ice shelves have also grown and retreated. During the last 20–25 million years, the Antarctic Ice Sheet has probably been grounded out to the edge of the continental shelf many times. It is unlikely that any substantial ice shelves could have existed then, because of unsuitable coastline geometry and lack of pinning points, although the extent of the sea ice may have been double the present day area. At the last glacial maximum, about 20 000 years ago, grounded ice extended out to the edge of the continental shelf in places, for example in Prydz Bay, but not in others such as the Ross Sea. When the ice sheets began to retreat, some of the large ice shelves seen today probably formed by the thinning and eventual flotation of the formerly grounded ice sheets. Large ice shelves also existed in the Arctic during the Pleistocene. Abundant geologic evidence shows that marine Northern Hemisphere ice sheets disappeared catastrophically during the climatic transition to the current interglacial (warm period). Only a few small ice shelves and tidewater glaciers exist there today, in places like Greenland, Svalbard, Ellesmere Island and Alaska. There have been periodic changes in Antarctic iceshelf grounding lines and ice fronts during the Holocene. The main deglaciation on the western side of the Antarctic Peninsula which started more than 11 000 years ago was initially very rapid across a wide continental shelf. By 6000 years ago, the ice sheet had cleared the inner shelf, whereas in the Ross Sea retreat of ice shelves ceased about the same time. Retreat after the last glacial maximum was followed by a readvance of the grounding line during the climate warming between 7000 and 4000 years ago. Open marine deposition on the continental shelf is restricted to the last 4000 years or so. Short-term cycles (every few hundred years) and longer-term events (approximately 2500 year cycles) have been detected in marine sediment cores that are likely related to global climate fluctuations. More recently, ice-shelf disintegration around the Antarctic Peninsula has been associated with a regional atmospheric warming which has been occurring

(c) 2011 Elsevier Inc. All Rights Reserved.

212

ICE SHELF STABILITY

since about 1950. Ice front retreat of marginal ice shelves elsewhere in Antarctica (e.g., Cook Ice Shelf, West Ice Shelf) is probably also related to atmospheric warming. There is little sign of any significant impact on the grounded ice.

Where are Ice Shelves Disintegrating Now? Two Case Studies The detailed history of the breakup of two ice shelves, Wordie and Larsen, illustrate some of the critical features of ice-shelf decay. Wordie Ice Shelf

Retreat of Wordie Ice Shelf on the west coast of the Antarctic Peninsula has been documented using high resolution visible satellite images taken since 1974 and the position of the ice front in 1966 mapped from aerial photography. First seen in 1936, the ice front has fluctuated in position but with a sustained retreat starting around 1966 (Figure 3). The ice shelf area has decreased from about 2000 km2 in 1966 to about 700 km2 in 1989. However, defining the position of the ice front can be very uncertain, with large blocks calving off to form icebergs being difficult to

classify as being either attached to, or separated from, the ice shelf. Ice front retreat until about 1979 occurred mainly by transverse rifting along the ice front, creating icebergs up to 10 km by 1 km. The western area was the first to be lost, between 1966 and 1974. Results from airborne radio-echo sounding between 1966 and 1970 suggested that the ice shelf could be divided into a crevassed eastern part and a rifted western part where brine could well-up and infiltrate the ice at sea level. By 1974 there were many transverse rifts south of Napier Ice Rise and by 1979 longitudinal rifting was predominant. Many longitudinal rifts had formed upstream of several ice rises by 1979, some following preexisting flowline features. A critical factor in the break-up was the decoupling of Buffer Ice Rise; by 1986 ice was streaming past it apparently unhindered, in contrast to the compressive upstream folding seen in earlier images. Between 1988 and 1989 the central part of the ice shelf, consisting of broken ice in the lee of Mount Balfour, was lost, exposing the coastline and effectively dividing the ice shelf in two. By 1989, longitudinal rifting has penetrated to the grounding line north of Mount Balfour. Further north, a growing shear zone marked the boundary between fast-flowing ice from the north Forster Ice Piedmont and slower moving

Figure 3 Cartoon showing retreat of Wordie ice front between 1966 and 1989.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE SHELF STABILITY

ice from Hariot Glacier. There has been no discernible change in the grounding line position. The major ice rises have played several roles in controlling ice-shelf behavior. When embedded in the ice shelf, they created broken wakes downstream, and zones of compression upstream which helped to stabilize the ice shelf. During ice front retreat, they temporarily pinned the local ice front position and also acted as nucleating points for rifting which quickly stretched upstream, suggesting that a critical fracture criterion had been exceeded. At this stage, an ice rise, instead of protecting the ice shelf against decay, aided its destruction by acting as an indenting wedge. Breakup was probably triggered by a climatic warming which increased ablation and the amount of melt water. Laboratory experiments show that the fracture toughness of ice is reduced at higher temperatures and possibly by the presence of water. Instead of refreezing in the upper layers of firn, free water could percolate down into crevasses and, by increasing the pressure at the bottom, allow them to grow into rifts or possibly to join up with basal crevasses. Processes like these would increase the production rate of blocks above that required for a ‘steady state’ ice front position. The blocks will drift away as icebergs if conditions, such as bay geometry and lack of sea ice, are favorable. Thus, ice front retreat would be, inter alia, a sensitive function of mean annual air temperature. Some ice shelves, such as Brunt, are formed from blocks that break off at the coast line and, unable to float away, are ‘glued’ together by sea ice and snowfall. These heterogeneous ice shelves contrast with those, such as Ronne, where glaciers or ice streams flow unbroken across the grounding line to form a more homogeneous type of ice shelf. Before 1989, Wordie Ice Shelf consisted of a mixture of both types, the main tongues being derived from glacier inputs, while the central portion was formed from blocks breaking off at the grounding line and at the sides of the main tongues. The western rifted area that broke away between 1966 and 1974 was described in early 1967 as ‘snowed-under icebergs’ and it was the heterogeneous central part that broke back to the coastline around Mount Balfour in 1988/89. Increased ablation would not only enhance rifting along lines of weakness but would also loosen the ‘glue’ that held the blocks together. Larsen Ice Shelf

The most northerly ice shelf in the Antarctic, Larsen Ice Shelf, extends in a ribbon down the east coast of the Antarctic Peninsula from James Ross Island to

213

the Ronne Ice Shelf. It consists of several distinct ice shelves, separated by headlands. The ice shelf in Prince Gustav Channel, between James Ross Island and the mainland, separated from the main part of Larsen Ice Shelf in the late 1950s and finally disappeared by 1995. The section between Sobral Peninsula and Robertson Island, known as Larsen A, underwent a catastrophic collapse at the end of January 1995 when it disintegrated into a tongue of small icebergs, bergy bits and brash ice. Previously the ice front had been retreating for a number of years, but in a more controlled fashion by iceberg calving. Before the final breakup, the surface that had once been flat and smooth had become undulating, suggesting that rifting completely through the ice had occurred. Collapse during a period of intense north-westerly winds and high temperatures was probably aided by a lack of sea ice, allowing ocean swell to penetrate and add its power to increasing the disintegration processes. The speed of the collapse and the small size of the fragments of ice (the largest icebergs were less than 1 km in size) implicate fracture as the dominant process in the disintegration (Figure 4). The ice front of the ice shelf known as Larsen B, between Robertson Island and Jason Peninsula, steadily advanced for about 6 km from 1975 until 1992. Small icebergs broke away from the heavily rifted zone south of Robertson Island after July 1992 and a major rift about 25 km in length had opened up by then. This rift formed the calving front when an iceberg covering 1720 km2 and smaller pieces corresponding to a former ice shelf area of 550 km2 broke away between 25 and 30 January 1995, coincident with the disintegration of Larsen A. The ice front has retreated continuously by a few kilometers per year since then (to 1999) and the ice shelf is considered to be under threat of disappearing completely.

Why do Ice Shelves Break-up? Calving and Fracture

Calving is one of the most obvious processes involved in ice shelf breakup. Although widespread, occurring on grounded and floating glaciers, including ice shelves, as well as glaciers ending in freshwater lakes, it is not well understood. The basic physics, tensile propagation of fractures, may be the same in all cases but the predictability of calving may be quite different, depending on the stress field, the basal boundary conditions, the amount of surface water, etc. Fracture of ice has been studied mainly in laboratories and applying these results to the

(c) 2011 Elsevier Inc. All Rights Reserved.

214

ICE SHELF STABILITY

Figure 4 Photo of the Larsen breakup. Picture taken approximately 10 km north of Robertson Island, looking west north-west. In the far background the mountains of the Antarctic Peninsula can be seen.

conditions experienced in naturally occurring ice masses requires extrapolations which may not be valid. Questions such as how is crevasse propagation influenced by the ice shelf geometry and by the physical properties of the ice such as inhomogeneity, crystal anisotropy, temperature and presence of water, need to be answered before realistic models of the calving process can be developed. The engineering concepts of fracture mechanics and fracture toughness promise a way forward. Fracture of ice is a critical process in ice shelf dynamics. Mathematical models of ice shelf behavior based on continuum mechanics, which treat ice as a nonlinear viscous fluid will have to incorporate fracture mechanics to describe iceberg calving and how ice rises can initiate fracture both upstream and downstream. High stresses generated at shear margins around ice rises, or at the ‘corner points’ of the ice shelf where the two ends of the ice front meet land, can exceed a critical stress and initiate crevassing. These crevasses will propagate under a favorable stress regime, and if there is sufficient surface water may rift through the complete thickness. Transverse crevasses parallel to the ice front act as sites for the initiation of rifting and form lines where calving may eventually occur. Calving is a very efficient method of getting rid of ice from an ice shelf – once an iceberg has formed, it can drift away within a few days. Sometimes,

however, an iceberg will ground on a sea bed shoal, perhaps for several years. Once an Antarctic iceberg has drifted north of the Antarctic Convergence, it usually breaks-up very quickly in the warmer waters. Climate Warming

Circumstantial evidence links the retreat of ice shelves around the Antarctic Peninsula to a warming trend in atmospheric temperatures. Observations at meteorological stations in the region show a rise of about 2.51C in 50 years. Ice shelves appear to exist up to a climatic limit, taken to be the mean annual 51C isotherm, which represents the thermal limit of ice-shelf viability. The steady southward migration of this isotherm has coincided with the pattern of iceshelf disintegration. There are too few sea temperature measurements to show if the observed atmospheric warming has resulted in warmer waters, and thus increased basal melting. Although basal melting may have increased, the indications are that it is surface processes that have played the dominant role in causing breakup. The mechanism for connecting climate warming with ice-shelf retreat is not fully understood, but increased summer melting producing substantial amounts of water obviously plays a part in enhancing fracture processes, leading to calving.

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE SHELF STABILITY

The retreat of ice shelves around the Antarctic Peninsula that has occurred in the last half of the twentieth century has raised questions about whether or not this reflects global warming or whether it is only a regional phenomenon. The warming trend seems to be localized to the Antarctic Peninsula region and there is no significant correlation with temperature changes in the rest of Antarctica. Stress Patterns

Numerical models of ice shelves have been used to examine their behavior and stability criteria. The strain-rate field can be specified by the principal values (e_1 and e_2 where e_1 4e_2 ) and the direction of the principal axes. A simple representation is given by the trajectories, which are a set of orthogonal curves whose directions at any point are the directions of the principal axes. Analyses of the strainrate trajectories for Filchner Ronne Ice Shelf and for different ice shelf configurations of Larsen Ice Shelf show characteristic patterns of a ‘compressive arch’ and of isotropic points (Figure 5). The ‘compressive arch’ is seen in the pattern formed by the smallest principal component (e_2 ) of the strain-rate trajectories. Seaward of the arch both principal strain-rate components are extensive, whereas inland of the arch the e_2 component is compressive. The arch extends from the two ends of the ice front across the whole width of the ice shelf. It is a generic feature of ice shelves studied so far and structurally stable to small perturbations. It is probably related to the geometry of the ice shelf bay. A critical arch, consisting entirely of compressive trajectories, appears to correspond to a criterion for stability. If the ice front breaks back through the arch then an irreversible retreat occurs, possibly catastrophically, to another stable configuration. The exact location of the critical arch cannot yet be determined a priori, but it is probably close to the compressive arch delineated by the transition from extension to compression for e_2. Another pattern seen in the strain-rate trajectories is that of isotropic points. They are indicators of generic features in the surface flow field which are stable to small perturbations in the flow. This means that their existence should not be sensitive either to reasonable errors in data used in the model or to simplifications in the model itself. They act as (permanent) markers in a complicated (tensor) strainrate field and are often located close to points where the two principal strain-rates are equal or where the velocity field is stationary (usually either a maximum or a saddle point). Isotropic points are classified by a number of properties, but only two categories can be

215

reliably distinguished observationally, by the way the trajectory of either of the principal strain-rates varies around a path enclosing the isotropic point. In one case the trajectory varies in a prograde sense and the isotropic point is called a ‘monstar’, whereas in the other case the trajectory varies in a retrograde sense and the isotropic point is called a ‘star’ (Figure 5). The two categories of isotropic points can be identified in the model strain-rate trajectories, ‘stars’ occurring near input glaciers and ‘monstars’ near ice fronts if they are in a stable configuration. Icebergs calving off Filchner Ice Shelf in 1986 moved the position of the ‘monstar’ up to the newly formed ice front, whereas on Ronne Ice Shelf the ‘monstar’ is about 50 km inland of the ice front. This suggests that the existence of a monstar can be used as a ‘weak’ indicator of a stable ice front. Calving can remove a monstar even if the subsequent ice front position is not necessarily an unstable one and may readvance.

How Vulnerable are the Large Ice Shelves and how Stable are Grounding Lines? The two largest ice shelves (Ross and Filchner– Ronne) are too far south to be attacked by the atmospheric warming that is predicted for the twentyfirst century. Basal melting will increase if warmer water intrudes onto the continental shelf, but the effects of a warming trend could be counterintuitive. If the warming was sufficient to reduce the rate of sea ice formation in the Weddell Sea, then the production of High Salinity Shelf Water (HSSW) would reduce as well. This would affect the subice shelf circulation, replacing the relatively warm HSSW under the Filchner–Ronne Ice Shelf with a colder Ice Shelf Water which would reduce the melting and thus thicken the ice shelf. Further climate warming would restore warmer waters and eventually thin the ice shelf by increasing the melting rates, possibly to values of around 15 m year 1 as seen under Pine Island Glacier. The likelihood of this is discussed in Sub Ice-Shelf Circulation and Processes. The supplementary question is what effect, if any, would the collapse of the ice shelves have on the ice sheet, especially the West Antarctic Ice Sheet (WAIS). It has been a tenet of the latter part of the twentieth century that the WAIS, known as a marine ice sheet because much of its bed is below sea level, is potentially unstable and may undergo disintegration if the fringing ice shelves were to disappear. However, the theoretical foundations of this belief are shaky and more careful consideration of the relevant

(c) 2011 Elsevier Inc. All Rights Reserved.

216

ICE SHELF STABILITY

66˚S

60˚W

65˚S

Scale 20

Kilometers 0 66˚S

62˚W

Figure 5 Modeled strain-rate trajectories on Larsen Ice Shelf superimposed on a Radarsat image taken on 21 March 1997. The trajectory of the smallest principal strain-rate (e_2 ) is negative (compressing flow, blue) over nearly all the Larsen Ice Shelf in its posticeberg calving configuration (January 1995) but is positive (red) in the region where the iceberg calved. The pattern shows isotropic points (‘stars,’ green) near where glaciers enter the ice shelf and, for the ‘iceberg’ area a ‘monstar’ (yellow cross) near the ice front. ‘Stars’ are seen for all the different ice front configurations, indicating that no fundamental change is occurring near the grounding line due to retreat of the ice front. A ‘monstar’ is only seen for Larsen B before the iceberg calved in 1995. The disappearance of the isotropic point is perhaps a ‘weak’ indication that the iceberg calving event was greater than expected for normal calving from a stable ice shelf and suggests further retreat may be expected.

dynamics suggests that the WAIS is no more vulnerable than any other part of the ice sheet. The problem lies in how to model the transition zone between ice sheet and ice shelf. If the transition is sharp, where the basal traction varies over lengths of the order of the ice thickness, then it can be considered as a passive boundary layer and does not affect the mechanics of the sheet or shelf to first order. Mass conservation must be respected but there is no need to impose further constraints. Requiring the ice sheet to have the same thickness as the ice shelf at the grounding line permits only two stable grounding line positions for a bed geometry which deepens inland. Depending on the depth of the bed

below sea level, the ice sheet will either shrink in size until it disappears (or, if the bed shallows again, the grounding line is fixed on a bed near sea level), or grow to the edge of the continental shelf. The lack of intermediate stable grounding line positions in this kind of model has been used to support the idea of instability of marine ice sheets. However, permitting a jump in ice thickness at the transition zone means the ice sheet system can be in neutral equilibrium, with an infinite number of steady-state profiles. Another kind of transition zone is a smooth one, where the basal traction varies gradually, for example, along an ice stream. The equilibrium dynamics have not been worked out for a full three-dimensional flow,

(c) 2011 Elsevier Inc. All Rights Reserved.

ICE SHELF STABILITY

but it seems that the presence of ice streams destroys the neutral equilibrium and helps to stabilize marine ice sheets. This emphasizes the importance of understanding the dynamics of ice streams and their role in the marine ice sheet system. Attempts to include moving grounding lines in whole ice sheet models suffer from incomplete specification of the problem. Assumptions built into the models predispose the results to be either too stable or too unstable. Thus there are no reliable models that can analyze the glaciological history of the Antarctic Ice Sheet. Predictive models rely on linearizations to provide acceptable accuracy for the near future but become progressively less accurate the longer the timescale.

See also Bottom Water Formation. Icebergs. Ice–ocean interaction. Sub Ice-Shelf Circulation and Processes.

Further Reading Doake CSM and Vaughan DG (1991) Rapid disintegration of Wordie Ice Shelf in response to atmospheric warming. Nature 350: 328--330.

217

Doake CSM, Corr HFJ, Rott H, Skvarca P, and Young N (1998) Breakup and conditions for stability of the northern Larsen Ice Shelf, Antarctica. Nature 391: 778--780. Hindmarsh RCA (1993) Qualitative dynamics of marine ice sheets. In: Peltier WR (ed.) Ice in the Climate System, NATO ASI Series, vol. 12, pp. 67--99. Berlin: Springer-Verlag. Kellogg TB and Kellogg DE (1987) Recent glacial history and rapid ice stream retreat in the Amundsen Sea. Journal of Geophysical Research 92: 8859--8864. Robin G and de Q (1979) Formation, flow and disintegration of ice shelves. Journal of Glaciology 24(90): 259--271. Scambos TA, Hulbe C, Fahnestock M, and Bohlander J (2000) The link between climate warming and break-up of ice shelves in the Antarctic Peninsula. Journal of Glaciology 46(154): 516--530. van der Veen CJ (ed.) (1997) Calving Glaciers: Report of a Workshop 28 February–2 March 1997. BPRC Report No. 15. Columbus, Ohio: Byrd Polar Research Center, The Ohio State University. van der Veen CJ (1999) Fundamentals of Glacier Dynamics. Rotterdam: A.A. Balkema. Vaughan DG and Doake CSM (1996) Recent atmospheric warming and retreat of ice shelves on the Antarctic Peninsula. Nature 379: 328--331.

(c) 2011 Elsevier Inc. All Rights Reserved.

IGNEOUS PROVINCES M. F. Coffins, University of Texas at Austin, Austin, TX, USA O. Eldholm, University of Oslo, Oslo, Norway

dynamic, nonsteady-state circulation within the Earth’s mantle, and suggests a strong potential for LIP emplacements to contribute to, if not instigate, major environmental changes.

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1290–1298, & 2001, Elsevier Ltd.

Introduction Large igneous provinces (LIPs) are massive crustal emplacements of predominantly Fe- and Mg-rich (mafic) rock that form by processes other than normal seafloor spreading. LIP rocks are readily distinguishable from the products of the two other major types of magmatism, midocean ridge and arc, on the Earth’s surface on the basis of petrologic, geochemical, geochronologic, geophysical, and physical volcanological data. LIPs occur on both the continents and oceans, and include continental flood basalts, volcanic passive margins, oceanic plateaus, submarine ridges, seamounts, and ocean basin flood basalts (Figure 1, Table 1). LIPs and their small-scale analogs, hot spots, are commonly attributed to decompression melting of hot, low density mantle material known as mantle plumes. This type of magmatism currently representB10% of the mass and energy flux from the Earth’s deep interior to its crust. The flux may have been higher in the past, but is episodic over geological time, in contrast to the relatively steady-state activity at seafloor spreading centers. Such episodicity reveals

Composition, Physical Volcanology, Crustal Structure, and Mantle Roots LIPs are defined by the characteristics of their dominantly Fe- and Mg-rich (mafic) extrusive rocks; these most typically consist of subhorizontal, subaerial basalt flows. Individual flows can extend for hundreds of kilometers, be 10s to 100s of meters thick, and have volumes as great as 104–105 km3. Sirich rocks also occur as lavas and intrusive rocks, and are mostly associated with the initial and late stages of LIP magmatic activity. Relative to midocean ridge basalts, LIPs include higher MgO lavas, basalts with more diverse major element compositions, rocks with more common fractionated components, both alkalic and tholeiitic differentiates, basalts with predominantly flat light rare earth element patterns, and lavas erupted in both subaerial and submarine settings. As the extrusive component of LIPs is the most accessible for study, nearly all of our knowledge of LIPs is derived from lavas forming the uppermost 10% of LIP crust. The extrusive layer may exceed 10 km in thickness. On the basis of geophysical, predominantly seismic data from LIPs, and from comparisons with normal oceanic crust, LIP crust beneath the extrusive layer is believed to consist of

Figure 1 Phanerozoic global LIP distribution (red), with LIPs labeled (Table 1).

218

(c) 2011 Elsevier Inc. All Rights Reserved.

IGNEOUS PROVINCES

an intrusive layer and a lower crustal body, characterized by compressional wave velocities of 7.0– 7.6 km s1, at the base of the crust (Figure 2). Beneath continental crust this body may be considered as a magmatically underplated layer. Seismic wave velocities suggest an intrusive layer that is most likely gabbroic, and a lower crust that is ultramafic. If the LIP forms on pre-existing continental or oceanic crust or along a divergent plate boundary, dikes and sills are probably common in the middle and upper crust. The maximum crustal thickness, including extrusive, intrusive, and the lower crustal body, of an oceanic LIP is B35 km, determined from seismic and gravity studies of the Ontong Java Plateau (Figure 1, Table 1). Low-velocity zones have been observed recently in the mantle beneath the oceanic Ontong Java Plateau, Oceanic plateau off-axis

x LCB

on-axis

x MC LCB

Ocean basin flood basalt x

LCB Continental flood basalt x Continental crust LCB Volcanic margin Rift Zone Postopening sediment COB x MC LCB

Continental crust

LIP CRUST Oceanic crust

Intrusives

Figure 2 Schematic LIP plate tectonic settings and gross crustal structure. LIP crustal components are: extrusive cover (X), middle crust (MC), and lower crustal body (LCB), continentocean boundary (COB), Normal oceanic crust is gray.

219

as well as under the continental Deccan Traps and Parana´ flood basalts (Figure 1, Table 1). Interpreted as lithospheric roots or keels, the zones can extend to at least 500–600 km into the mantle. In contrast to high-velocity roots beneath most continental areas, and the absence of lithospheric keels in most oceanic areas, the low-velocity zones beneath LIPs apparently reflect residual chemical and perhaps thermal effects of mantle plume activity. High-buoyancy roots extending well into the mantle beneath oceanic LIPs would suggest a significant role in continental growth via accretion of oceanic LIPs to the edges of continents.

Distribution, Tectonic Setting, and Types LIPs occur worldwide, in both continental and oceanic crust in purely intraplate settings, and along present and former plate boundaries (Figure 1, Table 1), although the tectonic setting of formation is unknown for many features. If a LIP forms at a plate boundary, the entire crustal section is LIP crust (Figure 2). Conversely, if one forms in an intraplate setting, the pre-existing crust must be intruded and sandwiched by LIP magmas, albeit to an extent not resolvable by current geological or geophysical techniques. Continental flood basalts, the most intensively studied LIPs due to their exposure, are erupted from fissures on continental crust (Figure 1, Table 1). Most continental flood basalts overlie sedimentary basins that formed via extension, but it is not clear what happened first, the magmatism or the extension. Volcanic passive margins form by excessive magmatism during continental breakup along the trailing, rifted edges of continents. In the deep ocean basins, four types of LIPs are found. Oceanic plateaus, commonly isolated from major continents, are broad, typically flat-topped features generally lying 2000 m or more above the surrounding seafloor. They can form at triple junctions (e.g., Shatsky Rise), midocean ridges (e.g., Iceland), or in intraplate settings (e.g., northern Kerguelen Plateau). Submarine ridges are elongated, steep-sided elevations of the seafloor. Some form along transform plate boundaries, e.g., Ninetyeast Ridge. In the oceanic realm, oceanic plateaus and submarine ridges are the most enigmatic with respect to the tectonic setting in which they are formed. Seamounts, closely related to submarine ridges, are local elevations of the seafloor; they may be discrete, form a linear or random grouping, or be connected along their bases and aligned along a ridge or rise. They commonly form in

(c) 2011 Elsevier Inc. All Rights Reserved.

220

IGNEOUS PROVINCES

Table 1

Large igneous provinces

(c) 2011 Elsevier Inc. All Rights Reserved.

IGNEOUS PROVINCES

(c) 2011 Elsevier Inc. All Rights Reserved.

221

222

IGNEOUS PROVINCES

intraplate regions, e.g., Hawaii. Ocean basin flood basalts, the least studied type of LIP, are extensive submarine flows and sills lying above and postdating normal oceanic crust.

Ages Age control for all LIPs except continental flood basalts is sparse due to their relative inaccessibility, but the 40Ar/39Ar dating technique is having a particularly strong impact on studies of LIP volcanism. Geochronological studies of continental floor basalts (e.g., Siberian, Karoo/Ferrar, Deccan, Columbia River; Figure 1) suggest that most LIPs result from mantle plumes which initially transfer huge volumes (B105–107 km3) of mafic rock into localized regions of the crust over short intervals (B105–106 years), but which subsequently transfer mass at a far lesser rate, albeit over significantly longer intervals (107– 108 years). Transient magmatism during LIP formation is commonly attributed to mantle plume ‘heads’ reaching the crust following transit through all or part of the Earth’s mantle, whereas persistent magmatism is considered to result from steady-state mantle plume ‘tails’ penetrating the lithosphere which is moving relative to the plume (Figure 3). However, not all LIPs have obvious connections to mantle plumes or hot spots, suggesting that more than one source model may be required to explain all LIPs. LIPs are not distributed uniformly in time. During the past 150 million years for example, many LIPs formed between 50 and 150 million years ago, whereas few have formed during the past 50 million

ge

n ea

rid

oc

d

Large Igneous Provinces and Mantle Dynamics The formation of various sizes of LIPs in a variety of tectonic settings on both continental and oceanic lithosphere suggests a variety of thermal anomalies in the mantle that give rise to LIPs as well as strong lithospheric control on their formation. Equivalent mantle plumes beneath continental and oceanic lithosphere should produce more magmatism in the latter scenario, as oceanic lithosphere is thinner, allowing more decompression melting. Similarly, equivalent mantle plumes beneath an intraplate region (e.g., Hawaii) and a divergent plate boundary (e.g., Iceland) (Figure 1, Table 1) will produce more magmatism in the latter setting, again because decompression melting is enhanced. Recent seismic tomographic images of mantle plumes beneath Iceland and Hawaii show significant differences between the two. Only recently, seismic tomography has revealed that slabs of subducting lithosphere can penetrate the entire Earth’s mantle to the D00 layer at the boundary Crust Upper mantle

Pl um int e _ er Ri ac dg tio e n

b

g sla

uctin

uc Su

bd

Core

tin

gs

lab

Subd

Lower mantle D″

Plu

me

Mi

Plume

LIP

years (Figure 4). Such episodicity likely reflects variations in rates of mantle circulation, and this is supported by high rates of seafloor spreading during a portion of the 50–150 million year interval. Thus, although LIPs manifest types of mantle processes distinct from those resulting in seafloor spreading, waxing and waning rates of overall mantle circulation probably affect both sets of processes. A major question that emerges from the global LIP production rate is whether the mantle is circulating less vigorously as the Earth ages.

660 km

2900 km

Figure 3 Model of the Earth’s interior showing plumes, subducting slabs, and two mantle layers that move in complex patterns, but never mix.

(c) 2011 Elsevier Inc. All Rights Reserved.

IGNEOUS PROVINCES

223

km3 year

_1

10

5

0 0

75 Ma

50

25

100

150

125

Figure 4 LIP production, corrected for subduction and averaged over a 15 million year running window, since 150 million years ago (Ma).

between the mantle and core at B2900 km depth (Figure 3). If we assume that the volume of the Earth’s mantle remains roughly constant through geological time, then the mass of crustal material fluxing into the mantle must be balanced by an equivalent mass of material fluxing from the mantle to the crust. Most, if not all of the magmatism associated with the plate tectonic processes of seafloor spreading and subduction is believed, on the basis of geochemistry and seismic tomography, to be derived from the upper mantle (above B660 km depth). It is most reasonable to assume that the lithospheric material that enters the lower mantle is eventually recycled, in some part contributing to plume magmas.

Large Igneous Provinces and the Environment The formation of LIPs has had documented environmental effects both locally and regionally. The global

effects are less well understood, but the formation of some LIPs may have affected the global environment, particularly when conditions were at or near a threshold state. Eruption of enormous volumes of basaltic magma during LIP formation releases volatiles such as CO2, S, Cl, and F (Figure 5). A key factor affecting the magnitude of volatile release is whether eruptions are subaerial or submarine; hydrostatic pressure inhibits vesiculation and degassing of relatively soluble volatile components (H2O, S, Cl, F) during deep-water submarine eruptions, although low solubility components (CO2, noble gases) are mostly degassed even at abyssal depths. Investigations of volcanic passive margins and oceanic plateaus have demonstrated widespread and voluminous subaerial basaltic eruptions. Another important factor in the environmental impact of LIP volcanism is the latitude at which the LIP forms. In most basaltic eruptions, released volatiles remain in the troposphere. However, at high latitudes, the tropopause is relatively low, allowing

Increased planetary albedo

N O 2 5 CIONO2 HCl

hv and OH SO2, HCI, Ash SO2

Stratosphere

Heterogeneous chemistry HNO3 CIO

H2SO4

Warming

Nucleation and particle growth Rainout H2O, HCI, Ash

Troposphere

Removal processes Cirrus modification

x

Infrared

Land Bridges Gateways

Continental crust

Hydrothermal activity

x

Sea level Upwelling Oceanic crust

LCB LCB

Continental Flood Basalt

Oceanic Plateau (off-axis)

Figure 5 Environmental effects of LIP formation. LIP eruptions can perturb the Earth–ocean–atmosphere system significantly. Note that many oceanic plateaus form at least in part subaerially. LCB, lower crustal body; X, extrusive cover.

(c) 2011 Elsevier Inc. All Rights Reserved.

224

IGNEOUS PROVINCES

large mass flux, basaltic fissure eruption plumes to transport SO2 and other volatiles into the stratosphere. Sulfuric acid aerosol particles that form in the stratosphere after such eruptions have a longer residence time and greater global dispersal than if the SO2 remains in the troposphere; therefore they have greater effects on climate and atmospheric chemistry. The large volume of subaerial basaltic volcanism, over relatively brief geological intervals, at highlatitude LIPs would contribute to potential global environmental effects. Highly explosive felsic eruptions, such as those documented from volcanic passive margins, an oceanic plateau (Kerguelen) (Figure 1, Table 1), and continental flood basalt provinces, can also inject both particulate material and volatiles (SO2, CO2) directly into the stratosphere. The total volume of felsic volcanic rocks in LIPs is poorly constrained, but they may account for a small, but not negligible fraction of the volcanic deposits in LIPs. Significant

volumes of explosive felsic volcanism would further contribute to the effects of plume volcanism on the global environment. Between B145 and B50 million years ago, the global oceans were characterized by variations in chemistry, relatively high temperatures, high relative sea level, episodic deposition of black shales, high production of hydrocarbons, mass extinctions of marine organisms, and radiations of marine flora and fauna (Figure 6). Temporal correlations between the intense pulses of igneous activity associated with LIP formation and environmental changes suggest a causal relationship. Perhaps the most dramatic example is the eruption of the Siberian flood basalts (Figure 1, Table 1) B250 million years ago, coinciding with the largest extinction of plants and animals in the geological record. Around 90% of all species became extinct at that time. On Iceland, the 1783–84 eruption of Laki provides the only human record of experience with the type of volcanism that

Figure 6 Temporal correlations among geomagnetic polarity, crustal production rates, LIPs, seawater strontium (Sr), sea level, climate, black shales, and extinctions.

(c) 2011 Elsevier Inc. All Rights Reserved.

IGNEOUS PROVINCES

constructs LIPs. Although Laki produced a basaltic lava flow representing B1% of the volume of a typical (103 km3) LIP flow, the eruption’s environmental impact resulted in the deaths of 75% of Iceland’s livestock and 25% of its population from starvation.

Conclusions Oceanic plateaus, volcanic passive margins, submarine ridges, seamounts, ocean basin flood basalts, and continental flood basalts share geological and geophysical characteristics indicating an origin distinct from igneous rocks formed at midocean ridges and arcs. These characteristics include: (1) broad areal extent (>104 km2) of Fe- and Mg-rich lavas; (2) massive transient basaltic volcanism occurring over 105–106 years; (3) persistent basaltic volcanism from the same source lasting 107–108 years; (4) lower crustal bodies characterized by compressional wave velocities of 7.0–7.6 km s1; (5) some component of more Si-rich volcanic rocks; (6) higher MgO lavas, basalts with more diverse major element compositions, rocks with more common fractionated components, both alkalic and tholeiitic differentiates, and basalts with predominantly flat light rare earth element patterns, all relative to midocean ridge basalts; (7) thick (10s–100s of meters) individual basalt flows; (8) long (r750 km) single basalt flows; and (9) lavas erupted in both subaerial and submarine settings. There is strong evidence that LIPs both manifest a fundamental mode of mantle circulation commonly distinct from that which characterizes plate tectonics, and contribute episodically, at times catastrophically, to global environmental change. Nevertheless, it is important to bear in mind that we have literally only scratched the surface of oceanic, as well as continental LIPs.

See also Deep-Sea Drilling Results. Geophysical Heat Flow. Gravity. Magnetics. Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Propagating Rifts and Microplates. Seamounts and Off-Ridge Volcanism. Seismic Structure

225

Further Reading Carlson RW (1991) Physical and chemical evidence on the cause and source characteristics of flood basalt volcanism. Australian Journal of Earth Sciences 38: 525--544. Campbell IH and Griffiths RW (1990) Implications of mantle plume structure for the evolution of flood basalts. Earth and Planetary Science Letters 99: 79--93. Coffin MF and Eldholm O (1994) Large igneous provinces: crustal structure, dimensions, and external consequences. Reviews of Geophysics 32: 1--36. Cox KG (1980) A model for flood basalt vulcanism. Journal of Petrology 21: 629--650. Davies GF (2000) Dynamic Earth: Plates, Plumes and Mantle Convection. New York: Cambridge University Press. Duncan RA and Richards MA (1991) Hotspots, mantle plumes, flood basalts, and true polar wander. Reviews of Geophysics 29: 31--50. Hinsz K (1981) A hypothesis on terrestrial catastrophes: wedges of very thick oceanward dipping layers beneath passive continental margins – their origin and paleoenvironmental significance. Geologisches Jahrbuch E22: 3--28. Macdougall JD (ed.) (1989) Continental Flood Basalts. Dordrecht: Kluwer Academic Publishers. Mahoney JJ and Coffin MF (eds.) (1997) Large Igneous Provinces: Continental, Oceanic, and Planetary Flood Volcanism. Washington: American Geophysical Union Geophysical Monograph 100. Morgan (1981) Hotspot tracks and the opening of the Atlantic and Indian oceans. In: Emiliani C (ed.) The Oceanic Lithosphere, The Sea vol. 7, pp. 443--487. New York: John Wiley. Richards MA, Duncan RA, and Courtillot VE (1989) Flood basalts and hot-spot tracks: plume heads and tails. Science 246: 103--107. Saunders AD, Tarney J, Kerr AC, and Kent RW (1996) The formation and fate of large oceanic igneous provinces. Lithos 37: 81--95. Sigurdsson H, Houghton BF, McNutt SR, Rymer H and Stix J (eds.) (2000) Encyclopedia of Volcanoes, San Diego Academic Press: p. 1147. Sleep NH (1992) Hotspot volcanism and mantle plumes. Annual Review of Earth and Planetary Science 20: 19--43. White R and McKenzie D (1989) Magmatism at rift zones: the generation of volcanic continental margins and flood basalts. Journal of Geophysical Research 94: 7685--7729. White RS and McKenzie D (1995) Mantle plumes and flood basalts. Journal of Geophysical Research 100: 17543--17585.

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS M. Fieux, Universite´ Pierre et Marie Curie, Paris Cedex, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1298–1308, & 2001, Elsevier Ltd.

Introduction Dynamically the equatorial area is a singular region on the earth because the Coriolis force is small, vanishing exactly at the equator. This results in a current structure that differs from that at other latitudes. Moreover, the equatorial current system of the Indian Ocean is entirely different from the current system found near the equator in the Pacific and Atlantic. This is principally due to its different wind forcing, which is described in the first section below. The systems of strictly equatorial currents at surface and at depth are reviewed in the second section. The third and fourth sections describe the North-East and South-West Monsoon Currents, north of the Equator, and the South Equatorial Countercurrent and the South Equatorial Current, south of the Equator.

component of the winds corresponding to the reversals between the NE and the SW monsoons, particularly in the western region along the Somali coast where the winds are the strongest. Between the monsoons, during the two transition periods, at the Equator, moderate eastward winds blow in spring (April–May) and in fall (October– November), with maxima between 701E and 901E (Figure 2). They could blow into one or several eastward bursts with a large seasonal and interannual variability. During the NE monsoon, the mean zonal component of the wind is weakly westward west of 801E, increasing in strength near the Somali coast. During the SW monsoon, the zonal components of the winds is weakly westward between 601E and 701E; west and east of that region, they are weakly eastward 40oE

60oE

80oE

100oE

120oE

NE monsoon o

20 N

o 20 N

0o

0o ITCZ

40oS

40oS

40oE

60oE

120oE

Maldives Is.

ia ric a So m al Af

Indon esia Java

0o

ra

at

m Su

Sri Lanka Gan Is.

o 20 N

nes

Cape Delgado

20oS

100oE

SW monsoon India Bay of Arabian Bengal Sea

20oN

0o

80oE

care

226

20oS

Mas

The winds over the tropical Indian Ocean are quite different from the winds over the Atlantic and the Pacific tropical oceans, where the NE and SE trade winds blow always in the same direction. Instead, the Indian Ocean (Figure 1), north of 101S, is under the influence of a monsoonal circulation, with complete reversal of the winds twice a year. The winter monsoon (December–March) blows from the NE in the Northern Hemisphere and from the NW south of the Equator toward the intertropical convergence zone (ITCZ) located near 101S. The change in direction at the Equator comes from the change of sign of the Coriolis force. The summer monsoon (June–September) blows from the SW in the Northern Hemisphere in continuity with the SE trade winds of the Southern Hemisphere, particularly in the western part of the equatorial ocean. The winds are stronger during the summer monsoon season. At the equator, during the monsoons, the winds have a preponderant meridional component, southward during the winter monsoon and northward during the summer monsoon, particularly near the western boundary. The result is a strong annual cycle in the meridional

20oS

Mo za Bamb Ma sin ique dag asc ar

Atmospheric Circulation over the Equatorial Indian Ocean

20oS

Mauritius Australia

o 40 S

o 40 S

o 40 E

60oE

80oE

100oE

120oE

Figure 1 Winds over the Indian Ocean during the NE monsoon (A) and during the SW monsoon (B) with locations used in the text.

(c) 2011 Elsevier Inc. All Rights Reserved.

227

INDIAN OCEAN EQUATORIAL CURRENTS

0.4

D

0.3

N O

0.2

S

J J

A

o

50 E

o

60 E

o

o

70 E

o

80 E

90 E

o

100 E

-0.2

M

-0.3

F J o

50 E

o

60 E

o

o

70 E

80 E

0.

o

90 E

o

100 E

-0.1

0

D

(B)

-0.4

D

N

0.1

N

O

O

0.

0

S

0.0

0.0

S

A

A 0.

J

J

1

0.

0

J

-0.1

0.1

M

A

A

M

M

F

F

J

0.0

o

60 E

o

70 E

o

80 E

o

90 E

o

100 E

0.0

0.1

0.1

J o

50 E

0.0

J

M

(C)

0.0

-0.1

M

(A)

0.1

dyn cm

_2

A

(D)

0.1

o

50 E

o

60 E

o

70 E

o

80 E

o

90 E

o

100 E

Figure 2 Longitude time plots of equatorial zonal wind stress averaged from 11S to 11N, determined from the FSU pseudostress climatology for the period 1970–1996: (A) the total wind in dyn cm2; (B) time-averaged mean zonal wind stress; (C) annual zonal wind stress component; (D) semiannual zonal wind stress component. Contour interval is 0.05 dyn cm2, eastward wind stress regions are shaded. (From W Han, JP McCreary, DLT Anderson, AJ Mariano (1999) Dynamics of the eastern surface jets in the equatorial Indian Ocean. Journal of Physical Oceanography 29: 2191–2209, (1999) copyright by the American Meteorological Society.)

with a maximum between 801E and 901E. As a result, the annual mean zonal wind is eastward, the opposite of what is found in the other equatorial oceans, and is maximum around 801E to 851E. This particular wind regime is dominated by an annual and a semiannual cycle with similar amplitudes for the zonal components. The amplitude of the annual component is maximum near the western boundary associated with the stronger monsoon winds off Somalia. Relative annual component maxima are also found near 821E and near the eastern boundary due to the southward monsoon winds extension south of Sri Lanka and along the Indonesian coast. In contrast, the semiannual component of the equatorial zonal winds has a simple structure with a single maximum in the central ocean between 601E and 801E during spring and fall.

Currents at the Equator Currents in the Extreme West, along the Somali Coast

Compared to the other major oceans, there are very few direct current measurements in the Indian Ocean and most of the surface currents information comes

from the ship drifts and recently from satellitetracked drifting buoys. Along the Somali Coast, as seen above, the annual period is the dominant variability in the wind forcing. It is only relatively recently (1984–1986) that long-term direct current measurements were carried out at the equator within 200 km off the Somali coast. Above 150 m they show a seasonally reversing flow in phase with the NE and the SW monsoons, but a striking asymmetry between both monsoon seasons. While during the stronger SW summer monsoon the north-eastward Somali Current decays monotonically in the vertical, during the weaker NE winter monsoon the south-westward surface current is limited to the surface layer and a north-eastward countercurrent exists between about 150 m and 400 m, remnant of the SW monsoon current, followed again by weak south-westward current underneath down to about 1000 m (Figure 3). This sliced structure is confined to the equatorial region corresponding to the equatorial waveguide. Surface Currents at the Equator, East of 521E

Ship-drift climatology indicates that, at the equator, the surface currents reverse direction four times a

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS

0

0

SHIP DRIFTS ±1o

Depth (m)

45oE 50o

500

500

D

July 1500 _100 0

0

50

100

20

Depth (m) Depth (m)

0

50 X

100

0

500

1000

A 40

40

J J

20

60

0

M A

-20

-40

M

0 -60

0 o

F J

20 o

55

60

o

65

o

o

70

75

o

o

o

80

85

90

o

o

95 E

Figure 4 Mean ship drift estimates of the current between 11N and 11S in cm s1. Dots correspond to westward currents and stripes to eastward currents. (From G. Reverdin (1987). The upper equatorial Indian Ocean: The climatological seasonal cycle. Journal of Physical Oceanography 17: 903–927, copyright by the American Meteorological Society.)

150

500

S

0

45 E 50

1500 0

O

80

o

_50

o

95 E

20

J

Jan.

o

90

40

F

1500 _100 (B)

o

20

-40

M

1000

85

N

-20

A

1000

o

80

0

M

500

o

60

J

0

500

75

80

J

150

X

o

70

D

A

1500

_50

o

65

O

1000

S

(A)

o

60

20 20

N

1000

o

55

1000 ADCP 18 April 1985 equator

1500 _100 (C)

_50

Apr. 1500

0

50 _ Speed (cms 1)

100

150

Figure 3 Monthly mean current profiles observed on three moorings in the Somali current at the equator during the NE monsoon (January 1985), the SW monsoon (July 1985), and during the transition period (April 1985); monthly mean of component parallel to coast. The crosses at the surface are historical ship drift data from Cutler and Swallow (1984). The dotted near-surface profile in April is from the ship mounted acoustic Doppler current profiler. (From F Schott (1986). Seasonal variation of cross-equatorial flow in the Somali current. Journal of Geophysical Research 91(C9): 10581– 10584, copyright by the American Geophysical Union.)

year. During the two transition periods, under the influence of the equatorial eastward winds, a strong surface eastward jet (known as the ‘Wyrtki jet’ because Klaus Wyrtki first identified it by looking at different shipdrift atlases in the 1960s) sets up in a

narrow band, trapped in the equatorial wave guide within 2–31 of the equator (Figure 4). The Wyrtki jet exists mostly in the central and eastern regions and disappears in the western part where the currents have a strong meridional component as shown above. Owing to the efficiency with which zonal winds can accelerate zonal currents at the equator where the Coriolis force disappears, the current speed can reach 1 m s1. The ship drift climatology indicates roughly the same strength for the two jets, with the October–November one (1 m s1) slightly stronger than the April–May one (0.90 m s1). As there is a large seasonal and interannual variability in the eastward winds, strong currents could be found in April–May instead of October–November during some years. At the equator, in the middle of the Indian Ocean, the first long-term current measurements (1973–1975), near Gan Island (73110E–0141S), show energetic eastward currents throughout the upper

dyne cm

_2

1.5

150

us

100

1.0 0.5

50

τx

0

_ cm s 1

228

0

_50

_0.5 I.0 J

F M

A M

J

J

A S O N D

J

F M A M

J

J A

S

1973 1974 Figure 5 Comparison of weekly average zonal wind stress (tx) in dyne s2 and surface zonal currents (us) in cm s1 at Gan Island (731100 E–01410 S) in 1973–1975. (Reprinted from RA Knox. On a long series of measurements of Indian Ocean equatorial currents near Addu Atoll, copyright 1976, Deep-Sea Research 23: 211–221, with permission from Elsevier Science.)

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS

229

Buoy 1885 20

15

10

5

0

_5

_ 10 40

50

60

70

80

90

100

110

Figure 6 Trajectory of one drifting buoy (drogued at 20 m) entrained eastward into the equatorial Wyrtki jet in November–December 1981, then in the reversal in December–January, crossing the Arabian Sea, and undergoing the seasonal reversal of the current south of Sri Lanka. (From M. Fieux (1987) Oce´an Indien et Mousson ‘Unesco Anton Bruun Lecture’ 15 pages, unpublished manuscript.)

100 m mixed layer in phase with the zonal eastward component of local winds during the two transition periods between the monsoons (Figures 5, 6). The establishment of this eastward jet could be explained through eastward-propagating Kelvin waves triggered by the eastward-going winds during the transition periods. The stopping of the jet is thought to be the result of the reflection of the eastward-propagating Kelvin waves on the eastern coast into westward-propagating equatorially trapped Rossby waves that progressively impede the jet from east to west. This has been observed with drifting buoys (Figure 7). The eastward currents at the equator are convergent (contrary to the westward currents). Consequently, drifting buoys launched in the jet stay in it and are a good means of observing its variability. In Figure 8 the drifting buoys were entrained into the May jet. They then stopped during the SW monsoon drifting slowly south of the equator and were taken up again eastwards into the November jet. Between the strong eastward flow periods, the equatorial surface currents are westward and much weaker (Figures 4 and 6). The associated change of current direction during each transition period produces semiannual variations in the thermocline depth and sea level. During periods of eastward flow when warm water is carried toward the east, the thermocline deepens off Sumatra and rises off Africa, corresponding to opposite displacements of the sea surface. Strong eastward flow at the equator entrains a surface convergence and as a result a downwelling in the upper layer, contrary to what is found in the

Figure 7 Drifting buoy trajectories between 601E and Sumatra during the period of September–February. Full lines are for 1979– 1980 and dashed lines are for 1980–1981. Dashed arrows indicate different speeds on this time–longitude diagram. During the first year the reversal propagation speed toward the west was around 55 cm s1, and around 40 cm s1 during the second year. (From G Reverdin, M Fieux, J Gonella and J Luyten (1983) Free drifting buoy measurements in the Indian Ocean equatorial jet. In: Nihoul JCJ (ed.) Hydrodynamics of the Equatorial Ocean. Amsterdam: Elsevier Science, 99–120.)

(c) 2011 Elsevier Inc. All Rights Reserved.

230

INDIAN OCEAN EQUATORIAL CURRENTS

Figure 8 (A) Satellite-tracked drifting drogued (at 20 m) buoys trajectories in 1979; dots on the trajectories are spaced at 15-day intervals. (B) Wind vectors estimated from cloud trajectories on satellite images during the equatorial eastward wind period (24 May 1979) and during the fully developed SW monsoon (19 June 1979). (The short tick on the arrow corresponds to 5 knots, the longer one to 10 knots, and the triangle to 50 knots.) (Reprinted with permission from JR Luyten, M Fieux and J Gonella. Equatorial currents in the Western Indian Ocean, Science 209: 600–603, copyright 1980 American Association for the Advancement of Science.)

other oceans. It is only during the weak westward flow that there is a weak upwelling. Long-term direct current measurements carried between Sri Lanka and 01450 S, in 1993–1994, reveal a large seasonal asymmetry that year in the semiannual eastward jet transport, with 35 Sv in November 1993 (surface velocities exceed 1.3 m s1) and only 5 Sv in May 1994. These transports were calculated with a lower boundary sets at 200 m as, sometimes, the eastward currents are not confined to one core but extend into the ray-like structures of the equatorial waves that continue to greater depths as in November 1993 (Figure 9). This large variability is due to a large seasonal and interannual variability of the zonal winds partly related to the Southern Oscillation. For example, during the El Nin˜o of the century in 1997, the equatorial eastward winds in the Indian Ocean completely disappeared in October–November and were replaced by westward winds. Numerical model results show that direct wind forcing is the dominant forcing mechanism of the equatorial surface jets. The semiannual response of the current to the wind is nearly three times as strong as the annual one, despite similar amplitudes in the corresponding wind components. This is due to the simpler zonal structure of the semiannual wind and to the resonance of the basin to the semiannual component. The mixed layer shear flow seems also to

enhance the semiannual response. The jet is strengthened when the mixed layer is reduced, particularly when the precipitation during the northern hemisphere summer and fall thins the mixed layer in the eastern ocean and so could strengthen the November jet in the east. The reflection of equatorial Kelvin waves into westward going Rossby waves at the eastern boundary is also important in weakening and even canceling the directly forced eastward jet about 2 months after the wind onset. The presence of the Maldives islands blocks part of the equatorially trapped waves. The effect on the semiannual waves is to weaken both jets, and the effect on the annual waves tends to weaken the May jet and to strengthen the November jet. Currents at Depth at the Equator

It is only during the NE monsoon, the season when the large-scale wind structure resembles the Pacific and the Atlantic ones, that a similar eastward equatorial undercurrent (EUC) embedded in the thermocline exists. Observations of the equatorial undercurrent are scarce. It was first observed during the International Indian Ocean Expedition (IIOE) in March–April 1963, between 531E and 921E, with speeds up to 0.8 m s1, then at Gan Island (731E) in March 1973 with velocities of up to 1 m s1, but not the following year. In 1975 and 1976, it was found

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS

Sept. 93

Aug. 93

231

Oct. 93

20 10

100 m 0

10

_ 10

0

200 m

0

0

300 m

0o

2oN

4oN

0o

2oN

4oN

0o

2oN

Dec. 93

Nov. 93

4oN

Jan. 94 _2

100 m

_ 25

0

7 0

0

200 m

0

_ 10

_ 10

300 m

0o

2oN

4oN

0o

2oN

4oN

0o

2oN

Mar. 94

Feb. 94

4oN

Apr. 94

_20 0

20

_20 0

100 m

10

0

0

10

30

0

200 m 0

300 m

0o

2oN

4oN

0o

2oN

0o

June 94

May 94 10 0

100 m

4oN

2oN July 94 _10

30 0

10

4oN

0

0

200 m 0

_10

300 m

0o

2oN

4oN

0o

2oN

4oN

0o

2oN

4oN

Aug. 94

10

100 m

0

0

200 m 300 m 0o

2oN

4oN

Figure 9 Monthly mean zonal velocities in cm s1 for the upper 300 m, south of Sri Lanka, along 801300 E, between August 1993 and August 1994. Contour interval is 10 cm s1 and shaded areas indicate eastward currents. Note the unusual reappearance of the EUC in August 1994. (From J Reppin, F Schott and J Fischer (1999). Equatorial currents and transports in the upper central Indian Ocean: Annual cycle and interannual variability, Journal of Geophysical Research 104(C7): 15495–15514, copyright by the American Geophysical Union.)

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS

from January extending into May–June at 551300 E with observed speeds reaching 0.8 m s1 in February and March at the end of the NE monsoon (Figure 10). Direct measurements give a maximum transport of 17 Sv, in March–April 1994, at the longitude of Sri Lanka (801300 E) 0 (Figure 9). It is confined between 2130 N–21300 S with slight meandering and is weak east of 801E. Its core is around 100 m in the upper equatorial thermocline. During the NE monsoon, it flows under a weak westward current until April–May when the eastward Wyrtki jet starts, then the whole upper layer flows eastward. It stops in May–June but surprisingly, in 1994, it reappeared in August. The existence of the EUC is related to equatorial westward winds, which force a westward surface current that builds up a zonal pressure gradient below the mixed layer that maintains the eastward undercurrent. The slope of the sea surface is opposite to the slope of the thermocline. The reappearance of the undercurrent in August 1994 is effectively related to anomalous onset of westward winds in the eastern part of the ocean, during May and June 1994. They force a westward surface current, again building a subsurface zonal pressure gradient, and thus an eastward undercurrent reappeared with some delay in August 1994. In 1976, current profiles show that the vertical velocity structure in the vicinity of the equator is characterized by small vertical scales of order of 50– 100 m in the upper layer increasing with depth throughout the water column. This deep jetlike structure is trapped to within 11 of the Equator and has a timescale of the order of several months. One-year current-meter measurements made at 200 m, 500 m, and 750 m at the Equator, between 481E and 621E, and recently at 801E, present a dominant semiannual reversal of the zonal component at all depths, much deeper than could be explained by direct wind forcing (Figure 11). Furthermore, these reversals do not happen at the same time at different depths. This seasonal cycle, with larger vertical scales, has been shown to penetrate vertically. The measurements suggest a mixture of equatorial Kelvin waves and long equatorial Rossby waves propagating downward from the surface where they are forced by the winds. The zonal velocity shows upward phase propagation, and downward energy propagation away from the surface. The behavior of the zonal currents is characteristic of an eastward-propagating equatorial Kelvin wave and a westward-propagating long equatorial Rossby wave. The ratio of the semiannual energy in the east current component to that in the north component is 40 to 1. This shows how

1975

0

20

80

100

< _60 0

200 Feb 0

Depth (m)

232

_80

> 40

> 60 0

100

< _20 0

200 Mar

0

0

60 100

0

200

Apr oS

3

2oS

1oS

0o

1oN

2oN

Figure 10 Meridional section along 531E of zonal velocities in February, March, and April 1975 showing the equatorial undercurrent. Contours interval is 20 cm s1. Plain contours correspond to eastward currents and dashed contours to westward currents. (From A Leetmaa and H Stommel (1980) Equatorial current observations in the Western Indian Ocean in 1975 and 1976, Journal of Physical Oceanography, 10: 258–269, copyright by the American Meteorological Society.)

much the equatorial ocean is a barrier to the meridional motions.

Currents North of the Equator South of Sri Lanka

North of the Equator, the monsoon forcing drives a general eastward flow, the South-west Monsoon Current (SMC), during the fully developed SW monsoon, and a general westward flow, the Northeast Monsoon Current (NMC), during the NE monsoon, extending south to about 21S in January– February. The exchanges between the Arabian Sea and the Bay of Bengal are restricted to the south of India and Sri Lanka. Drifting-buoy trajectories show

(c) 2011 Elsevier Inc. All Rights Reserved.

Depth (m)

INDIAN OCEAN EQUATORIAL CURRENTS

AFRICA 0

AREA OCCUPIED BY MOORED ARRAY

233

SUMATRA

1000

3000 6000 (A) 43oE 45o

50o

55o

60o

65o

70o

40

75o

80o

85o

90o

95o

97oE

a

_1

East (cm s )

20

0

_20

_40

-15

50

100

15

0

-1

-5

-15

15 15 20

5

-20

-10

-25

-10

500

400

-5

Depth (m)

200

750 (C)

200 300 100 Days (0=1 Apr. 1979)

0

10 5

(B)

10

200

300

Time (days)

Figure 11 (A) Vertical section of the Indian Ocean along the Equator with the location of the current-meter array between 471E and 621E. (B) Time series of east velocity from seven current-meters at 750 m depth starting on April 1 1979. (C) Contour plot of lowfrequency east velocity as a function of depth and time (day 01 April 1979). At each nominal depth (200, 500, 750 m), four records are filtered by a 30-day running mean and averaged together. The contour interval is 5 cm s1. (From JR Luyten and DH Roemmich (1982) Equatorial currents at semiannual period in the Indian Ocean, Journal of Physical Oceanography 12: 406–413, copyright by the American Meteorological Society.)

the seasonal current reversal in that restricted region (Figure 6). Direct observations carried out in 1991–1994 south of Sri Lanka show that the monsoon currents are mostly confined in the upper 100 m. In August 1993, the eastward SMC extended to the equator and retracted north of 41300 N in September. It was replaced in October by the westward NMC, extending south to about 21N, with speeds between 0.3 and 0.8 m s1 (Figure 9). Shipdrifts show that, around the Maldives, the NMC splits into a branch that bends south-westward and a branch that follows the western coast of India, bringing low-salinity Bay of Bengal water into the Arabian Sea. The NMC maximum flow lies north of 41N with a mean transport of 10–12 Sv (Figure 12). In May the current reverses eastwards into the SMC again. The SMC transport reaches about 8–10 Sv with speeds up to 0.75 m s1 in July. The SMC is sometimes separated from the coast of Sri Lanka by a coastal westward countercurrent bringing Bay of Bengal water. The annual mean flow past Sri Lanka was 2–3 Sv westward in 1993–1994. Numerical models show

that these monsoon currents are driven by the largescale tropical wind field. Contrary to the equatorial circulation where the semiannual period prevails, the annual component dominates the upper layer flow north of 41N.

Currents South of the Equator Apart from shipdrifts and satellite-tracked drifting buoys, there are no direct long-term current measurements in the South Equatorial Countercurrent and only two in the South Equatorial Current. The South Equatorial Countercurrent

In contrast to the Pacific and the Atlantic oceans, where a countercurrent is found year-long north of the Equator between the North Equatorial Current and the South Equatorial Current, in the Indian Ocean the countercurrent is found south of the Equator and has a large extent only during a short period of the year.

(c) 2011 Elsevier Inc. All Rights Reserved.

234

INDIAN OCEAN EQUATORIAL CURRENTS

30

20

SMC

Transport (Sv)

10

0

+ _ 10

_ 20

_ 30

NMC

D

J

F

M

A

M

J J 1991

A

S

O

N

D

J

F M 1992

Figure 12 Seasonal variability of the transport in the upper 300 m from moorings south of Sri Lanka, between 31450 N and 51520 N, in 1991–1992. (The solid circles are from ship-mounted acoustic Doppler current profiler measurements in December 1990–January 1991; the solid circle in parentheses is for July 1991, and the cross is from Pegasus profiler measurements in March 1992.) (From F Schott, J Reppin, J Fisher and D Quadfasel, Currents and transports of the monsoon current south of Sri Lanka, Journal of Geophysical Research 99(C12): 25127–25141, copyright by the American Geophysical Union.)

During the winter monsoon, the Somali current flows southward, crosses the Equator, and merges with the northward flowing East Africa Coastal Current (EACC), at about 2–41S, to form the eastward South Equatorial Countercurrent (SECC). It is a region of high eddy activity. The SECC is found between 21S and 81S during January–March. During that season the winds have an eastward component at those latitudes just north of the ITCZ. The speeds vary between 0.5 and 0.8 m s1 in the west, getting weaker in the east. In March 1995, at 801E, observed surface velocities exceed 0.7 m s1 and the current extended to about 1100 m, with an eastward geostrophically deduced transport of about 55 Sv. In the east, part of it continues into the Java Current and part of it recirculates southward into the South Equatorial Current (SEC). It is detected in the meridional slope of the thermocline which slopes downward toward the Equator, with an opposite upward slope of the sea surface (Figure 13). Together with the SEC and the EACC, the SECC forms an elongated cyclonic gyre. In the latitude range about 10–131S and between 201S and 251S, recent measurements of the sea level variability through satellite altimetry show westward propagation of sea level anomalies corresponding to semiannual and annual Rossby wave characteristics. The wind-driven model results show westward propagation of Rossby waves in the shear zone between the SECC and the SEC that are obstructed and partially reflected by the Mascarenes banks

(55–601E). In the west, at the end of the winter season, in March–April, when the eastward winds start on the equator, the outflow from the EACC into the SECC begins to move northward toward the Equator and the eastward flow at that time is mostly equatorial.

The South Equatorial Current

The westward South Equatorial Current forced by the south-east trade winds extends south of the SECC. It represents the northern branch of the South Indian subtropical gyre. It is seen in the meridional downward slope of the thermocline towards the south (Figure 13). It is partly fed, in its northern part, between 101S and 141S, by the low-salinity throughflow jet originating from the western Pacific Ocean through the Indonesian Seas carrying about 4 to 12 Sv with larger extremes depending on the year. The low salinity extends down to about 1200 m. In March 1995, at 801E, measured surface velocities reached 0.7 m s1. The SEC is the limit of the influence of the monsoon system and separates the northern and southern Indian Ocean. It is stronger during July–August when the SE trade winds are stronger. Its indirectly computed mean transport relative to the 1000 m level varies from 39 Sv in July– August to 33 Sv in January–February with large uncertainty. Its mean transport increases from east to west. Its latitudinal range varies between 8–221S in

(c) 2011 Elsevier Inc. All Rights Reserved.

INDIAN OCEAN EQUATORIAL CURRENTS

235

Figure 13 Meridional temperature section along 54–551E in May 1981 with 11C isotherm interval. The downward slope of the thermocline form 91S toward the south corresponds to the SEC, and the downward slope toward the Equator corresponds to the SECC. Close to the Equator, the accentuated slope corresponds to the May Wyrtki jet. (From M Fieux and C Levy (1983) Seasonal observations in the western Indian Ocean. In: Nihoul JCJ (ed.) Hydrodynamics of the Equatorial Ocean. Amsterdam: Elsevier Science, 17–29.)

Buoy 1886 10

5

0 SEC C _5 _ 10

SEC

_ 15 _ 20

40

50

60

70

80

90

100

110

Figure 14 Satellite-tracked drifting buoy trajectory between November 1981 and February 1984. The buoy, drouged at 20 m, was entrained eastward in the November jet then in the SECC and back westward into the SEC. One year later it was again entained in the same elongated gyre. (From M Fieux (1987) Oce´an Indien et Mousson, ‘Unesco Anton Bruun Lecture’, unpublished manuscript.)

July–August and 10–201S in January–February. Its velocities range between 0.3 m s1 and 0.7 m s1. The SEC impinges both on the east coast of Madagascar and on the east African coast, resulting

in several intensified boundary currents along these coasts. East of Madagascar, the SEC splits near 171S into a northward flow, carrying about 27 Sv near 121S, and a southward flow, transporting 20 Sv

(c) 2011 Elsevier Inc. All Rights Reserved.

236

INDIAN OCEAN EQUATORIAL CURRENTS

between the surface and 1100 m near 231S. At 121S energetic boundary current transport variations occur at the 40–55-day period, contributing to about 40% to the total transport variance, while at 231S the 40–55-day period fluctuations contribute only 15% to the total transport variance. The northern branch of the SEC splits again east of the African coast near Cape Delgado (111S) into the southward flowing Mozambique Current (MC) and the northward flowing East African Coastal Current (EACC). Drifting-buoy trajectories describe an elongated cyclonic gyre of the equatorial current system composed of the equatorial eastward jet or the SECC, depending on the season, the off Sumatra SE current, the westward SEC, and the EACC. Some drifting buoys, launched during a transition period at the Equator, carried into the eastward Wyrtki jet and in the SECC, crossed the basin, then were driven southeastward into the Sumatra current, then crossed the basin westward into the SEC, and flowed back to the Equator in the western region exactly one year later when they were again carried into the SECC (Figure 14).

is eastward, associated with a strong convergence at the surface inducing an equatorial downwelling. The upwelling regions are found instead north of the equator, along the Somalia, the Arabian, and the Indian coasts, and are seasonally depending. The equatorial current structure in the Indian ocean is complex and further long-term observations as well as modeling efforts are needed to better understand its seasonal and interannual variability and its role in the large-scale meridional exchanges between the northern and the southern Indian Ocean.

See also Agulhas Current. Coastal Trapped Waves. Current Systems in the Indian Ocean. Elemental Distribution: Overview. El Nin˜o Southern Oscillation (ENSO). Rossby Waves. Somali Current. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Conclusion Owing to stronger winds in the tropical Indian ocean, the currents are stronger than in the Pacific and Atlantic Oceans but seasonally are highly variable. Away from the western boundary the equatorially trapped long waves (Kelvin and Rossby) explain most of the observed seasonal variations of the equatorial currents. In contrast to the Atlantic and Pacific Oceans, equatorial upwelling is weak in the Indian Ocean. During the period of strong current, the surface flow

Cutler AN and Swallow JC (1984) Surface Currents of the Indian Ocean (to 251S, 1001E): Compiled from Historical Data Archived by the Meteorological Office, Brack-nell, UK. Wormley, UK: Institute of Oceanographic Science (unpublished report) 187, 8pp. and 36 charts. Fein JS and Stephens PL (eds.) (1987) Monsoons. Washington DC: NSF and Wiley. Open University, Oceanography Course Team (1993) Ocean Circulation. Oxford: Pergamon Press. Tomczak M and Godfrey S (1994) Regional Oceanography: An Introduction. Oxford: Pergamon Elsevier.

(c) 2011 Elsevier Inc. All Rights Reserved.

INDONESIAN THROUGHFLOW J. Sprintall, University of California San Diego, La Jolla, CA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The Indonesian seas play a unique role in providing the only open pathway that connects two major ocean basins at tropical latitudes. On average, the sea level is higher on the western Pacific side of the Indonesian archipelago compared to the eastern Indian side. This pressure gradient generates a transport of water and their properties from the Pacific toward the Indian Ocean. This flow from the Pacific Ocean into the Indian Ocean is known as the Indonesian Throughflow. The warmer water that enters from the Pacific can be traced throughout the Indonesian seas, and then followed within the surface to intermediate depths as a distinct low-salinity tongue in the Indian Ocean. As such, the Throughflow forms the ‘warm’ water route for the global thermohaline circulation and therefore impacts the regional and global climate system. Because of its proximity to Asia and Australia, the circulation and transport in the Indonesian seas has a large seasonal variation due to the influence of the reversing annual wind patterns associated with the Asian–Australian monsoon system. During the different monsoon seasons, waters of different sources from both the Indian and Pacific Oceans flow into the Indonesian archipelago and cause variability in temperature, salinity, and other properties. Local processes within the Indonesian seas related to the regional monsoon winds such as upwelling and downwelling, along with the tides, air–sea heat fluxes, and voluminous precipitation and associated river runoff, also act to change the Pacific temperature and salinity stratification into the distinctly fresh Indonesian seas profile. These changes in the physical properties of the water are linked to the behavior, migration pattern, and the seasonal distribution of the phytoplankton and pelagic fish species that live within the Indonesian seas. Thus a knowledge and understanding of the pathways and variability of the Indonesian Throughflow and its properties is important for the region’s people, who depend on the sea for their very food and livelihood, and also to help develop management plans to sustain these valuable and limited maritime resources.

Pathways and Water Masses The combination of numerous islands, with narrow straits that connect a series of large, deep ocean basins within the Indonesian seas, provides a winding route for the Indonesian Throughflow (Figure 1). The surface to upper thermocline waters in the Indonesian seas are primarily drawn from the North Pacific subtropical waters and North Pacific Intermediate Water that flows southward in the Mindanao Current, east of the Philippines. These waters mostly take the ‘western’ Throughflow route through the Sulawesi Sea (sometimes known as the Celebes Sea) into the Makassar Strait. Within Makassar Strait, the 650-m-deep Dewakang sill permits only the upper thermocline waters to enter the Flores and Banda Seas, or to directly exit into the Indian Ocean via the shallow (350 m) Lombok Strait. Smaller contributions of North Pacific surface water may also take the ‘eastern’ Throughflow route, through the Maluku Sea and over the deeper (1940 m) sill of Lifamatola Strait. Lower thermocline and deeper water masses – Antarctic Intermediate Water (AAIW) and Circumpolar Deep Water – that originate in the South Pacific in the New Guinea Coastal Undercurrent, also enter the Indonesian seas via the eastern route through the Lifamatola Strait. These deeper water masses are saltier than the upper waters that originate from the North Pacific. Upper waters of South Pacific origin can also flow directly over the Halmahera Sea (blocked below 700-m depth) and into the internal Seram and Banda Seas. Most of the transformation of the Pacific water masses into the distinctive Indonesian water profile of relatively isohaline water from the thermocline to near bottom occurs within the Banda Sea. As the different water masses flow toward the Banda Sea over the various Indonesian sills and straits, the abrupt changes in bottom topography cause intense diapycnal mixing. The region-average vertical diffusivities at the sills are greater than 10  3 m2 s  1, which is an order of magnitude larger than both that typically estimated for the interior ocean, and also for that which occurs within the interior of the Banda Sea itself. At least some of this diapycnal mixing is attributable to the strong tides within the Indonesian seas. At the sills, mixing can also occur via entrainment and shear-induced turbulence, including internal wave generation. It is also likely that hydraulic effects can restrict the flow and affect the circulation pattern in some straits due to the shoaling

(c) 2011 Elsevier Inc. All Rights Reserved.

237

238

INDONESIAN THROUGHFLOW

0 8° N −500

4° N

−1000

−1500

0°

−2000 4° S −4000

Java Sea 8° S

−6000

−8000

12° S

−10 000 100° E

110° E

120° E

130° E

140° E

Figure 1 Map of the Indonesian region showing the bathymetry and pathways of the mean Indonesian Throughflow. Shallow and upper thermocline pathways from the North Pacific are shown by the red arrows, and lower thermocline to intermediate depth pathways are shown by the yellow arrows. Islands, seas, and straits referred to in the text are indicated. The 200-m isobath is shown by the dashed line.

and contraction of sills, although as yet there have been no suitable direct observations or process studies undertaken that may resolve this response. Density-driven (or isopycnal) flow dominates the deeper layers of South Pacific origin that enter through Lifamatola Strait into the Banda Sea, driving deep circulation patterns that keep the internal basins well ventilated. In the surface to upper thermocline layer, the diapycnal mixing combines with the tidal forcing, wind-driven upwelling, air–sea heat flux, and surface precipitation and river runoff to form the low-salinity Indonesian Throughflow Water (ITW: 34–34.5 psu). Below this sits the Indonesian Intermediate Water (IIW), which has a silica maximum and is slightly saltier (36 psu) as a result of diapycnal mixing of the intermediate Pacific water masses with a larger contribution from the saltier water masses of the South Pacific found at depth in the Banda basin. From the Banda Sea, the Indonesian Throughflow exits into the southeast Indian Ocean via the Timor Passage (1250 m eastern sill depth at Leti Strait, and 1890 m at the western end), or through Ombai Strait (sill depth 3250 m) and then through the Savu Sea to

Sumba Strait (900 m) and Savu Strait (1150 m). The Indonesian Throughflow waters are apparent as a freshwater jet in the westward-flowing South Equatorial Current (SEC) across the entire tropical Indian Ocean between 81 and 171 S. In the Indian Ocean, the total Throughflow is recognizable as the lowsalinity ITW in the upper 100–200 m of surface water and the deeper-salinity minimum of the IIW which is found at B600–1200-m depth, and hence above the eastern sill depth of Timor Passage. Although the ITW and IIW are separate cores even directly west of Timor Passage as they exit the Indonesian seas, they gradually increase in salinity and become even more distinct entities as the highsalinity maximum subtropical and equatorial waters of the open Indian Ocean encroach between them. The ITW and IIW lose their low-salinity signature more slowly than the intervening salinity maximum during westward advection across the Indian Ocean because the adjacent water masses on their isopycnals (surface layer and AAIW) are considerably fresher than other Indian Ocean water masses. Nonetheless, the slight increase in salt results in an increase in density and thus the Throughflow waters

(c) 2011 Elsevier Inc. All Rights Reserved.

INDONESIAN THROUGHFLOW

become deeper and erode toward the western Indian Ocean, although the latitude of both the core ITW and IIW remains remarkably constant within the SEC because of the nearly exact zonality of potential vorticity in the Tropics. At the western edge of the Indian Ocean, it is likely that much of the Throughflow enters the Agulhas Current system, although because of mixing with other water masses in this region and beyond, it becomes more difficult to directly trace their origin from the Indonesian seas. Nonetheless, numerous Agulhas eddies in the Atlantic have been identified as containing relatively fresh Indonesian thermocline water, providing observational evidence to support the numerical models that show the Indonesian Throughflow playing an important role in the global circulation.

Mean Throughflow Mass and Heat Flux Estimates The magnitude and variability of the Indonesian Throughflow are still sources of major uncertainty for both the modeling and observational oceanography communities. They are the dominant sources of error in the basin-wide heat and freshwater budgets for the Pacific and Indian Oceans. Though general circulation models are gradually improving, they are unable at present to reproduce the narrow passages and convoluted bottom topography of the internal Indonesian seas in order to adequately resolve the structure and variability of the ITF transport. Thus the models still largely rely on the scantly available observations to provide guidance in their initialization and validation for simulations of the ocean circulation and climate. Historical estimates of the mean mass transport of the Indonesian Throughflow are wide-ranging, from near zero to 30 Sv (1 Sv ¼ 106 m3 s  1). In part, this sizeable difference is due to the lack of direct measurements, but also because of the real variation that can severely alias mean estimates if survey periods are not sufficiently long. Recent mooring measurements of velocity in Makassar Strait in 1986–87 suggest a volume transport of around 8 Sv for the surface to upper thermocline waters that pass through this inflow channel. Long-term direct measurements of the lower thermocline to intermediate water masses have yet to be obtained, although water mass considerations and synoptic survey measurements of the Throughflow at the eastern edge of the Indian Ocean suggest a volume transport of 2–3 Sv. A 3-month current meter record at Lifamatola Strait in early 1985 determined a 1.5-Sv contribution of deep water to the Throughflow. Direct long-term measurements

239

of the Throughflow within the exit passages have also been obtained at different times over the past few decades. The transport through Timor Passage (1988–89) and Ombai Strait (1995) appears to be roughly equally partitioned, at c. 4–6 Sv through each strait, whereas the transport through the shallower exit passage of Lombok Strait (1985) is B2 Sv. Hence the total volume transport associated with the fulldepth Throughflow is probably around 10–14 Sv, although because of the different sampling years of the different direct measurement programs this total volume transport estimate should be treated with caution. As will be discussed below, the volume and properties of the Throughflow are known to change with various phases of the El Nin˜o–Southern Oscillation (ENSO) cycle, and hence this may have impact on the transport estimates from a given passage in a given year. Only with multiyear, simultaneous measurements of the full-depth velocity structure in all the major Throughflow passages can the volume transport be accurately determined. Direct measurements that meet these criteria to determine transport will become available from an array of moorings that were deployed as part of the International Nusantara Stratification and Transport (INSTANT) program. The 11 moorings were deployed in the major inflow and outflow passages in December 2003, turned around in June 2005, and are to be recovered in December 2006. The heat and fresh water carried by the Indonesian Throughflow potentially impact the basin budgets of both the Pacific and Indian Oceans. Earlier estimates of the heat transport ranged from 0.5 to 1.0 PW (petawatt, with 1 PW ¼ 1015 W) as the Throughflow was assumed to be surface-intensified, and reflect the relatively warm temperatures found in its western Pacific source waters. However, the recent observations from the Makassar moorings suggest that the strongest Throughflow velocities (southward in this strait) are found in the thermocline, and hence the cool transport-weighted temperature of B15 1C leads to a mean heat transport of B0.4–0.5 pW (relative to 0 1C). Including an as yet unresolved colder component from the IIW would only further reduce the Throughflow impact on the heat transport of the Southern Hemisphere. Because of the lack of subsurface salinity measurements, the contribution of the freshwater flux from the Indonesian Throughflow to the Indian Ocean and beyond is even less well constrained. However, since the salinity of the Pacific waters entering the Indonesian seas is fresher than the inflow from the Southern Ocean northward into the Pacific, the Throughflow probably represents a net transport of fresh water out of the Pacific and into the Indian Ocean where the

(c) 2011 Elsevier Inc. All Rights Reserved.

240

INDONESIAN THROUGHFLOW

salinity increases before being exported southward in the Agulhas Current.

Variability in the Throughflow Properties and Transport Pacific, Indian, and regional Indonesian wind forcing modulates the Indonesian Throughflow mass and property transport over a wide range of timescales that combine to produce variability of the same order of magnitude as the long-term mean estimates cited above. In this section, we will discuss the cause of the different timescales of variability and their impact on the Throughflow. Because it is the main driving force of the Throughflow we begin with a discussion of the seasonal variability related to the Asian–Australasian monsoon regime. Forcing at other timescales from remote Pacific and Indian Ocean sources generally tends to modulate this fundamental seasonal response. Monsoonal-related Variability: Annual Timescales

The large-scale sea level difference between the easterly trade winds in the western Pacific and the reversing monsoonal winds over the southern Indonesian seas drives the annual change in the Throughflow. During the northwest monsoon from December to February, south of the equator the winds over Indonesia are to the southeast (Figure 2(a)). Through Ekman dynamics, the southeastward winds cause a drop in sea level along the northern Nusa Tenggara islands and warm surface waters accumulate in the Banda Sea, acting to reduce the Throughflow transport into the Indian Ocean. The downwelling south of Nusa Tenggara leads to warmer surface waters here. The convective rainfall associated with the northwest monsoon spans the Indonesian archipelago to produce heavy rainfall and subsequently high river runoff that together combine to freshen the surface waters, particularly in the Java Sea region. The flow from the Banda Sea toward the Indian Ocean is strongest during the southeast monsoon, from June through September, when the winds are to the northwest and more intense (Figure 2(b)). The strong winds during this monsoon result in a lower sea level to the south of Nusa Tenggara, and hence the flow through the exit passages is thought to be enhanced by this local Ekman response. Surface waters are cooler in the Banda Sea, and also south of Nusa Tenggara due to the winddriven upwelling. The convective activities move northward of the equator so conditions in Indonesia are drier and surface salinity higher compared to the

northwest monsoon. The wind-driven regime also leads to changes in the chlorophyll concentrations, which is elevated in regions of upwelling as nutrients from the deep are bought to the surface, and reduced where there is downwelling. Chlorophyll indicates phytoplankton abundance and therefore gives a broad view of the biological activity and associated fishery distribution during the different monsoon seasons. While the surface response to the reversing monsoonal winds is clearly evident in Figure 2, the monsoonal effect on the Throughflow velocity and transport is much more complicated. While all the observations presently available in the exit passages do indeed show the expected maximum Throughflow occurring during the southeast monsoon period, the Makassar Strait mooring data showed the maximum southward flow occurred during the 1996–97 northwest monsoon period. The large differences in peak transport timing between the inflow and outflow straits are most likely due to storage of waters in the Banda Sea on seasonal timescales. It appears that the Banda Sea acts as a reservoir, filling up and deepening the thermocline during the northwest monsoon. During the more intense southeast monsoon, Ekman divergence in the Banda Sea combined with the lower sea level off the south coast of the Nusa Tenggara are more conducive to drawing waters into the Indian Ocean. The maximum Ekman convergence (or gain) of 1.7 Sv causes the thermocline to heave almost 40 m during October–November, although there is matching divergence (or loss) occurring in April–May of each year. Convergences and divergences of this magnitude can have a substantial impact on thermocline stratification, and thus may affect the sea surface temperature by changing the temperature of the water being entrained into the mixed layer. In addition, since the Banda Sea is also the primary site for conversion of the Pacific waters into the distinct Indonesian stratification profile, it is likely that the composition and magnitude of the stored waters could have a significant impact on the Indian Ocean heat, fresh water, and mass budgets. The response of the surface properties to the shifting monsoonal winds has recently been suggested to play another important role in modifying the Throughflow velocity and transport. The buoyant low-salinity water present in the Java Sea and southern Makassar Strait during the northwest monsoon (Figure 2(a)) creates a northward pressure gradient that inhibits the warm surface water in Makassar Strait from flowing southward into the Indian Ocean, even though the predominant winds are southward during this monsoon. During the

(c) 2011 Elsevier Inc. All Rights Reserved.

INDONESIAN THROUGHFLOW

(a)

Northwest monsoon (DJF)

(b) 250 245 240 235 230 225 220 215 210 205 200 195 190 185 180 175 170

4° N

0°

4° S

8° S

12° S 110° E 120° E 130° E 0.150 N m−2

Latitude

4° S

8° S

12° S 105° E

125° E 115° E Longitude

0°

4° S

8° S

12° S 125° E

12° S 110° E

135° E

120° E 130° E 0.150 N m−2

140° E

31.5 31 30.5 30 29.5 29 28.5 28 27.5 27 26.5 26 25.5

4° N

0°

4° S

8° S

12° S 105° E

34.8 34.5 34.2 33.9 33.6 33 32.7 32.4 32.1 31.8 31.5 31.2 30.9 30.6 30.3 30 115° E

8° S

135° E

4° N

105° E

4° S

100° E

31.5 31 30.5 30 29.5 29 28.5 28 27.5 27 26.5 26 25.5

0°

250 245 240 235 230 225 220 215 210 205 200 195 190 185 180 175 170

0°

140° E

4° N

Southern monsoon (JJA)

4° N

Latitude

100° E

241

115° E 125° E Longitude

135° E

4° N

34.8 34.5 34.2 33.9 33.6 33 32.7 32.4 32.1 31.8 31.5 31.2 30.9 30.6 30.3 30

0°

4° S

8° S

12° S 105° E

115° E

125° E

135° E

Figure 2 The average wind stress (N m  2) and sea surface height (cm; top panels), sea surface temperature (1C; middle panels) and sea surface salinity (psu; bottom panels) during the (a) northwest monsoon (Dec.–Feb.) and (b) southeast monsoon (Jun.–Aug.). The 200-m isobath is shown by the dotted lines in the lower panel.

southeast monsoon, the surface waters are more saline (Figure 2(b)), the northward pressure gradient is eliminated, and the northward winds tend to constrain the surface Throughflow. Thus, it appears that during both monsoon seasons, the strongest southward Throughflow occurs at thermocline depths

(100–200 m) in Makassar Strait, resulting in the recently determined cooler ocean heat transport cited in the above section. Although the data are still preliminary, the INSTANT moorings in the shallower Lombok Strait also show a subsurface southward velocity maximum at B50 m during the

(c) 2011 Elsevier Inc. All Rights Reserved.

242

INDONESIAN THROUGHFLOW

southeast monsoon. During the northwest monsoon in Lombok Strait the main southward flow is pushed even deeper to B100 m, as strong northward flows are found at the surface. Whether these northward surface flows are related to remote Indian Ocean or local wind forcing is further discussed in the following section. Variability Related to Indian Ocean Forcing: Intraseasonal and Semiannual Timescales

Remote wind forcing from the Indian Ocean is responsible for variability in the Indonesian Throughflow transport and properties via the generation of coastal Kelvin waves. Anomalous wind bursts in the equatorial Indian Ocean force an eastward equatorial jet associated with an equatorial Kelvin wave that, upon impinging the west coast of Sumatra, results in poleward propagating coastal Kelvin waves. Westerly wind bursts force downwelling Kelvin waves semiannually during the monsoon transitions, nominally in May and November, and intraseasonally (30–90 days) most likely in response to the intraseasonal wind forcing of the Madden–Julian Oscillation. The poleward propagation of the downwelling coastal Kelvin wave raises the sea level along the Indonesian wave guide of Sumatra and Nusa Tenggara, transporting warm and fresh surface water and causing maximum eastward flow of the boundary current that flows along southern Nusa Tenggara, the South Java Current. While there is some discrepancy between theoretical models as to whether the Indian Ocean-forced Kelvin wave energy can penetrate northward through Lombok Strait, the available observational data clearly show the northward flow of warm water that indicates the presence of Kelvin waves in both Lombok Strait and Makassar Strait, particularly during the May transition period. Furthermore, some of the Kelvin wave energy is also readily apparent in observations from Ombai Strait, further east along the coastal waveguide, during the May and November transition periods, suggesting that Indian Ocean water can penetrate through the Savu Sea and beyond. Commensurate with vertical ray-tracing theory, the semiannual energy from remotely forced Kelvin waves extends deeper in the water column with distance along the waveguide, and so the reversals in flow are evident well below the thermocline in all three straits. What proportion of the Kelvin wave energy is ‘lost’ through Lombok Strait versus that remaining in the coastal waveguide, and how this may be modulated by the annual cycle, is the focus of active research. Similarly, whether the northward surface flows observed during the

northwest monsoon in the INSTANT data from Lombok Strait and Ombai Strait, and to some extent in Makassar Strait, are related to Kelvin wave intrusions or to locally generated wind forcing is also the subject of ongoing research. The northward flows during this monsoon phase have a clear intraseasonal variability that may be related to the remote winds associated with the intraseasonal Madden–Julian Oscillation in the eastern equatorial Indian Ocean. However, given the relatively shallow vertical penetration of these reversals, they could also be a locally wind-driven Ekman dynamical response.

Variability Related to Pacific Ocean Forcing: Interannual Timescales

Remote winds in the Pacific Ocean drive large-scale circulation that impacts the Indonesian Throughflow on low-frequency, interannual timescales. The weakening of Pacific trade winds that occurs during El Nin˜o reduces the sea level gradient between the Pacific and Indian Oceans that is thought to be the driving force for the Throughflow. The changes are of the sense that mass and heat transport are smaller and the thermocline is shallower during El Nin˜o, with the converse being true during La Nin˜a time periods. Within the inflow Makassar Strait, the 1996–98 mooring velocity and temperature time series show a strong correlation with the Southern Oscillation Index that tracks the ENSO-related variability in the Pacific. Along the Nusa Tenggara exit passages, the relationship of transport and temperature to ENSO variability is much less certain, primarily because we lack the multiyear, full-depth measurements that are required to establish such a relationship. Surface geostrophic transport variability (0–100-m depth) over the period 1996–99 through the exit passages, inferred from the cross-strait pressure gradient measured by shallow pressure gauges, suggest diminished flow through Lombok and Ombai Straits during the El Nin˜o years 1997–98, with increased flow through Timor Passage. This is consistent with modeling studies, in which there is stronger flow through Timor Passage and weaker flow through Ombai Strait during El Nin˜o events. Interestingly, the surface geostrophic transport through Lombok Strait was strongly northward during the northwest monsoon of the 1997–98 El Nin˜o event, amplified over the expected northward flows in this strait during this monsoon phase. It is possible, as discussed above, that the main southward Throughflow occurred at depth during this time period and hence was not measured by the shallow pressure gauges.

(c) 2011 Elsevier Inc. All Rights Reserved.

INDONESIAN THROUGHFLOW

Interannual variability in sea level is also evident within the Banda Sea and along the Northwest Australian shelf region down to around 201 S. Both models and observations suggest that the remote Pacific tropical winds associated with ENSO drive both equatorial and off-equatorial Rossby waves that are responsible for the observed sea level and associated thermocline response. The Rossby wave signal is radiated through the Banda Sea, as well as entering the coastal waveguide at the western tip of New Guinea, and propagates poleward through Timor Strait and along the Australian shelf as coastally trapped waves. In turn, these trapped waves excite free westward propagating Rossby waves at the Australian coast that can be detected several hundreds of kilometers offshore in the southeast tropical Indian Ocean. Interannual Kelvin waves forced remotely by equatorial interannual zonal winds in the Indian Ocean have also been documented as impacting the thermocline and sea level variability along the Nusa Tenggara coastal waveguide, similar to those observed on intraseasonal and semiannual timescales. Again, both models and observations suggest these Kelvin waves penetrate through Lombok Strait and into the internal Indonesian seas. Thus, as suggested by Wijffels and Meyers, the Indonesian seas comprise the intersection of two oceanic wave guides from remotely generated wind forcing in both the Pacific and Indian Oceans on interannual timescales.

Conclusions The mass and property characteristics of the Indonesian Throughflow vary over the full range of tidal to interannual timescales. Within the internal Indonesian seas, the Pacific temperature and salinity stratification, as well as the local sea surface temperature, are modified by the strong air–sea heat and freshwater fluxes, seasonal wind-induced upwelling, and large tidal forces. These ocean circulation and processes in the Indonesian seas influence not only the local climate but also the global climate through connections with the Pacific and Indian Ocean. For example, ruinous drought conditions in Australia have been linked to cool sea surface temperature anomalies in the

243

Indonesian region. Similarly basin-wide anomalies in sea surface temperature, wind, and precipitation that spanned the Indonesian region during 1997–98 resulted in severe drought in Indonesia and catastrophic floods in eastern Africa. Thus, undoubtedly the oceanic heat and freshwater flux of the Indonesian Throughflow into the Indian Ocean affect the atmospheric– ocean coupling and are strongly linked to the evolution of interannual climate anomalies such as occur through the ENSO and monsoon systems. Understanding the variation in the Indonesian Throughflow is therefore crucial for understanding the coupled air– sea climate system, and the storage of the heat and fresh water that are ultimately redistributed throughout the world oceans by thermohaline circulation.

See also Coastal Trapped Waves. Current Systems in the Indian Ocean. Ekman Transport and Pumping. Heat and Momentum Fluxes at the Sea Surface. Ocean Circulation: Meridional Overturning Circulation. Rossby Waves. Upper Ocean Heat and Freshwater Budgets. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Gordon AL (2001) Interocean exchange. In: Seidler G, Church J, and Gould J (eds.) Ocean Circulation and Climate, ch. 4.7, pp. 303--314. New York: Academic Press. Talley LD and Sprintall J (2005) Deep expression of the Indonesian Throughflow: Indonesian intermediate water in the South Equatorial Current. Journal of Geophysical Research 110: C10009 (doi:10.1029/2004JC002826). The Oceanography Society. (2005). Special Issue: The Indonesian Seas. Oceanography 18(4), 144pp. Wijffels S and Meyers GA (2004) An intersection of oceanic wave guides: Variability in the Indonesian Through flow region. Journal of Physical Oceanography 34: 1232--1253. Wyrtki K (1961) Physical oceanography of the Southeast Asian waters. Scripps Institution of Oceanography NAGA Report 2. San Diego, CA: Scripps Institution of Oceanography.

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE T. D. Dickey, University of California, Santa Barbara, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1313–1323, & 2001, Elsevier Ltd.

Introduction Light is of great importance for the physics, chemistry, and biology of the oceans. In this article, a brief introduction is provided to the two subdisciplines focusing on light in the ocean: ocean optics and biooptics. A few of the problems addressed by these subdisciplines are described. Several of the in situ sensors and systems used for observing the subsurface light field and optical properties are introduced along with general explanations of the operating principles for measuring optical variability in the ocean. Some of the more commonly used ocean platforms and optical systems are also discussed. Finally, some examples of oceanographic optical data sets are illustrated. Solar radiation, which includes visible radiation or light, impinges on the surface of the ocean. On average, a small fraction or percentage is reflected back into the atmosphere (roughly 6% on average) while a high fraction penetrates into the ocean. This fraction (or percentage), defined as the albedo, varies in time and space as a function of several factors including solar elevation, wave state, surface roughness, foam, and whitecaps. Radiative transfer is a branch of oceanography termed ‘ocean optics,’ a term that denotes studies of light and its propagation through the ocean medium. Radiative transfer processes depend on the optical properties of the components lying between the radiant source (e.g., the sun) and the radiation sink (the ocean and its constituents). Another commonly used term is ‘bio-optics,’ which invokes the notion of biological effects on optical properties and light propagation and vice versa. Some of the data used for examples here focus on bio-optical and physical interactions. Solar radiation spans a broad range of the electromagnetic energy spectrum. The visible portion, roughly 400–700 nm, is of primary concern for ocean optics for many practical problems. These include: underwater visibility, photosynthesis and primary production of phytoplankton, upper ocean

244

ecology, red tides, pollution, biogeochemistry including carbon cycling, photochemistry, light-keyed migration of organisms, heating of the oceans, global climate change, and remote sensing of ocean color. It is worth noting that optical oceanographers are also interested in the ultraviolet (UV) range and the infrared. For example, UV radiation can damage phytoplankton and can also lead to modifications in genetic material, and thus evolution, whereas the penetration of the infrared is critical for near surface radiant heating. Solar radiation is clearly a primary driver for ocean physics, chemistry, and biology as well as their complex interactions. Thus, it is not surprising that virtually all important oceanographic problems require interdisciplinary approaches and necessarily atmospheric, physical, chemical, biological, optical, and geological data sets. These data sets should be collected concurrently (concept of synopticity) and span sufficient time and space scales to observe the relevant processes of interest. For local studies involving optics, variability at timescales of one day and one year are especially important. For global problems, variability extends well over ten orders of magnitude in space [O(millimeters) to O(104 kilometers)] and much longer in time for climate problems. Multiplatform observing approaches are essential. Capabilities for obtaining atmospheric and physical oceanographic data are relatively well advanced in contrast to those for chemical, biological, optical, acoustical, and geological data. This is not surprising, because of the greater complexity and nonconservative nature of the chemistry and biology of the oceans. Yet, remarkable advances are being made in these areas as well. In fact, several bio-optical, chemical, geological, and acoustical variables can now be measured on the same time and space scales as physical variables; however many more variables still need to be measured.

Fundamentals of Ocean Optics A few operation definitions need to be introduced before discussing the various optical sensors. Light entering the ocean can be absorbed or scattered. The details of these processes comprise much of the study of ocean optics. First, it is convenient to classify bulk optical properties of the ocean as either inherent or apparent. Inherent optical properties (IOPs) depend

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

only on the medium and are independent of the ambient light field and its geometrical distribution. Inherent optical properties include: spectral absorption coefficient, aðlÞ, spectral scattering coefficient, bðlÞ, and spectral beam attenuation coefficient (or, sometimes denoted as ‘beam c’), cðl). To define these coefficients, consider a beam of monochromatic light impinging perpendicularly on a thin layer of water. The fraction of the incident radiant flux, F0 (energy or quanta per unit time), which is absorbed by the medium, Fa/F0, divided by the thickness of the layer is the absorption coefficient, a, in units of m1. Similarly, the analogous scattering coefficient, b (m1) is the fraction of the incident radiant flux scattered from its original path (primarily forward), Fb/F0, divided by the thickness of the layer. The beam attenuation coefficient is simply c ¼ a þ b. The spectral volume scattering function, bðc; l) represents the scattered intensity of light per unit incident irradiance per unit volume of water at some angle c (with respect to the exiting, nonscattered incident beam) into solid angle element DO. Units for bðc; l) are m1 sr1. This is essentially the differential scattering cross-section per unit volume in the parlance of nuclear physics. The spectral scattering coefficient is obtained by integrating the volume scattering function over all directions (solid angles). The forward scattering coefficient, bf , is obtained by integrating over the forward-looking hemisphere (c ¼ 0 to p/2) and the backward scattering coefficient, bb, is calculated by integrating the backlooking hemisphere (c ¼ p=2 to p). The spectral volume scattering phase function is the ratio of bðc; l) to bðlÞ with units of sr1. Other important IOPs include spectral single-scattering albedo, o0 ðlÞ ¼ bðlÞ=cðl), and fluorescence, which can be considered a special case of scattering. (Note: fluorescence is not strictly an IOP; however it is often used as a proxy for chlorophylla.) The proportion of light which is scattered versus absorbed is characterized by o0 ðl); that is, if scattering prevails then o0 ðl) approaches a value of 1 and if absorption dominates, o0 ðl) approaches 0. IOPs obey simple mathematical operations. For many applications, it is often convenient to partition the total absorption coefficient, at ðl), the total scattering coefficient, bt ðl), and the total beam attenuation coefficient, ct ðlÞ, in terms of contributing constituents such that: at ðlÞ ¼ aw ðlÞ þ aph ðlÞ þ ad ðlÞ þ ag ðlÞ

½1

bt ðlÞ ¼ bw ðlÞ þ bph ðlÞ þ bd ðlÞ þ bg ðlÞ

½2

ct ðlÞ ¼ cw ðlÞ þ cph ðlÞ þ cd ðlÞ þ cg ðlÞ

245

½3

The subscripts indicate contributions by pure sea water (w), phytoplankton (ph), detritus (d), and gelbstoff (g) with units of m1 for all variables. Detritus is the term for particulate organic debris (e.g., fecal material, plant and animal fragments, etc.) and gelbstoff (sometimes called yellow matter or gilvin) is the term for optically active dissolved organic material. Note that eqns [1]–[3] are intended for use where bottom and coastal sediment contributions are minimal. The spectral absorption coefficient of pure sea water is well characterized and is nearly constant in space and time with greater absorption in the red than blue portions of the visible spectrum. The magnitudes of the detrital and gelbstoff absorption spectra tend to decrease monotonically with increasing wavelength and can be modeled. Phytoplankton spectral absorption varies significantly in relation to community composition and environmental changes; however characteristic peaks are typically found near wavelengths of 440 nm and 683 nm. Much bio-optics research focuses on the temporal and spatial variability of aph ðl). Ship-based ocean water samples have often been used for studies of the IOPs, however, it is quite preferable to obtain in situ measurements to insure representative local values as well as to characterize temporal and spatial variability, preferably with simultaneous physical, chemical, and biological measurements.

Instrumentation Beam transmissometers have been the most commonly used instruments for measuring IOPs with a variety of applications ranging from determinations of suspended sediment volume to phytoplankton biomass and productivity to particulate organic carbon (POC). The principle of operation involves the measurement of the proportion of an emitted beam, C, which is lost through both absorption and scattering as it passes to a detector through some pathlength DL. The beam attenuation coefficient, c, is then given as ½lnð1  CÞ=DL; or light intensity, I, is given as IðDLÞ ¼ Ið0Þ expðcDL). Similar expressions apply for absorption coefficient, a, and scattering coefficient, b. Beam transmissometers have often used a red light emitting diode (660 nm) for the collimated light source and pathlengths of 25 or 100 cm (long pathlengths are preferable for clearer waters). The wavelength of 660 nm was selected in order to minimize attenuation by gelbstoff, which attenuates strongly at shorter wavelengths but minimally in the red. One of the measurement

(c) 2011 Elsevier Inc. All Rights Reserved.

246

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

complications for beam c is that most scattering in ocean waters is in the forward direction. The acceptance angle of the detector is thus made as small as possible (o 1–21) to minimize underestimation of beam c. Corrections are done for this effect as well as for instrument temperature and pressure effects. In situ multispectral attenuation–absorption (acmeters) instruments have recently become commercially available. These concurrently measure spectral absorption and attenuation coefficients, which are spectral signatures of both particulate and dissolved material. These devices use dual light sources and interference filters on a single rotating wheel (nine wavelengths from 400 to 715 nm with 10 nm bandwidth). Whereas earlier beam transmissometers were open to sea water, the newer ac-meters use two tubes through which sea water is pumped. The inside of the c-tube is flat black to minimize reflections whereas the a-tube is reflective (shiny) in order to maximize internal reflection to better estimate absorption. The deployment of these meters requires minimal lateral and torsional stressing in mounting (sensitivity of beam alignment) and good flow through the tubes. Calibrations involve temperature, salinity, and pure water measurements. A correction is also needed to account for the fact that not all scattered light is collected (causing a biased overestimate of absorption). Similar, though more capable instruments, have recently been developed to measure a and c at 100 wavelengths from 400 to 726 nm wavelengths with 3.3 nm resolution. A single white light source is used with fiberoptics to provide light to each of the two tubes (for a and b) as well as for a reference path (for correcting changes in lamp output). Each light path has its own spectrometer (using a 256 pixel photodiode array). Biofouling is an issue for all optical instruments. Closed (pumped) systems are less susceptible to fouling as they are primarily in the dark. A variety of special antifouling methods have been attempted with mixed success. These have included copper screens, copper flow tubes, bromide solutions, and other cleaning agents. Spectral scattering coefficient can be computed from ac-meter data by simply performing the difference b ¼ c  a. Spectral a  c meter data and relevant spectral decomposition models can be used to provide in situ estimates of aph ðl) and other components (e.g., gelbstoff) for a broad range of environmental conditions as a function of time at fixed depths using moorings or as a function of depth from ships or profilers deployed from moorings. The volume scattering function is one of the more important and challenging measurements of optical oceanography. Until recently, few instruments existed for this measurement and data collected with an

instrument developed in the 1970s were commonly used for many problems. The challenge arises because dominant scattering typically occurs at small angles (roughly 50% in the forward direction between 0 and 2–61) making it difficult to sense the weak scattering light, which is near the intense illuminating beam. Ideally, measurements should be done at several angles to better resolve the function. However, the backscatter signal, bb, can be estimated with a few measurement angles provided the form of the volume scattering function is relatively well known. Recently developed volume scattering and backscatter instruments take advantage of this function. An important consideration is that backscattering is related to the size distribution, shapes, and composition of the particles sampled (e.g., coastal versus open ocean particles). Thus, this type of information can be obtained in principle. Radiative transfer models need both absorption and volume scattering function information to estimate the propagation of light. Also, an important remote sensing parameter, remote sensing reflectance (discussed below) is related to bb and a. Key wavelengths are selected for the spectral measurements on the basis of the application and water types (e.g., coastal versus open ocean). Particle size distributions can also be measured using laser (Frauenhofer) diffraction instruments. These devices employ charged coupled device (CCD) array photodetectors (linear or circular ring geometries). Modified versions of some of these instruments can also measure particle settling velocities. Particles in the range of 5–500 mm can be measured with resolution dependent on the number of individual detector rings. These devices can also measure beam transmission by sensing the undeviated light. Irradiance is the radiant flux per unit surface area (units of W m2 or quanta (or photons) s1 m2 or mol quanta (or photons) s1 m2. Note that 1 mol of photons is 6.02  1023 (Avogadro’s number) and that one mole of photons is commonly called an einstein. It is convenient to define downwelling irradiance, Ed, as the irradiance of a downwelling light stream impinging on the top face of a horizontal surface (e.g., ideally flat light collector oriented perpendicular to the local gravity vector) and upwelling irradiance, Eu, as the irradiance of an upwelling light stream impinging on the bottom face of a horizontal surface. Irradiance reflectance, R, is Eu/Ed. Irradiance measurements are fundamentally important for quantifying the amount of light available for photosynthesis and for radiative transfer theory and computations. Another useful radiometric variable is radiance, L ¼ Lðy; FÞ, which is defined as the radiant flux at a specified point in a given direction per unit solid

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

angle per unit area perpendicular to the direction of light propagation (units of W (or quanta s1) m2 sr1). Zenith angle, y, is the angle between a vertical line perpendicular to a flat plane and an incident light beam, and azimuthal angle, F, is the angle with respect to a reference line in the plane of the flat plate. By integrating Lðy; FÞcosy over the solid angle ðdo ¼ siny dy dFÞ, of the upper hemisphere (say of a flat plate from y ¼ 0 to p/2 and F from 0 to 2p), one obtains Ed. Similarly, upwelling irradiance, Eu, is calculated by integrating over the bottom hemisphere. Net downward irradiance is the difference Ed  Eu, or the integral of Lðy; FÞcosy over the entire sphere (full solid angle 4p). Scalar irradiance, E0, is defined as the integral of Lðy; FÞ over the entire sphere. It should be noted that all definitions are for a specific wavelength of light. If one integrates scalar irradiance over the visible wavelengths (roughly 400–700 nm; note that the visible range is sometimes defined as 350–700 nm), then the biologically important quantity called photosynthetically available radiation (PAR) is obtained. Apparent optical properties (AOPs) depend on both the IOPs and the angular distribution of solar radiation (i.e., the geometry of the subsurface ambient light field). To a reasonable approximation, the attenuation of spectral downwelling incident solar irradiance, Ed ðl; zÞ, can be described as an exponential function of depth: Ed ðl; zÞ ¼ Ed ðl; 0 Þexp½Kd ðl; zÞz

½4

where z is the vertical coordinate (positive downward), Ed ðl; 0 Þ is the value of Ed just below the air– sea interface, and Kd ðl; zÞ is the spectral diffuse attenuation coefficient of downwelling irradiance. Kd ðl) (here depth dependence notation is suppressed for convenience) is one of the important AOPs. The contributions of the various constituents (e.g., water, phytoplankton, detritus, and gelbstoff) to Kd ðl) are often represented in analogy to the absorption coefficients given in eqn [1][1]. This leads to the term ‘quasi-inherent’ optical property as it is suggestive that inherent optical properties, i.e., like at ðl), are closely related to Kd ðlÞ, which is often described as a quasi-inherent optical property. However, a key point is that Kd ðlÞ is in fact dependent on the ambient light field. Analogous spectral attenuation coefficients are defined for diffuse upwelling irradiance, downwelling radiance, and upwelling radiance, KL ðlÞ. It should be noted that for periods of low sun angle and/or when highly reflective organisms or their products (e.g., coccolithophores and coccoliths) are present, then multiple scattering becomes increasingly more important and simple relations

247

between IOPs and AOPs break down. It is important to mention that spectral radiometric measurements of AOPs have been far more commonly made than measurements of IOPs. The relationships between IOPs and AOPs are central to developing quantitative models of spectral irradiance in the ocean. Radiative transfer theory is used to provide a mathematical formalism to link IOPs and the conditions of the water environment and light forcing to the AOPs of the water column. One motivation of this research is the desire to predict AOPs given environmental forcing conditions and estimates or actual measurements of IOPs. Another related, inverse goal is to be able to estimate or determine the vertical (and ideally horizontal) structure of IOPs and their temporal variability given normalized water-leaving radiance, Lwn ðlÞ, measured remotely from satellite or airplane color imagers. The extrapolation of subsurface values of LðlÞ to the surface has been the subject of considerable research, because a major requirement of ocean color remote sensing is to match in situ determinations of those from satellite- or plane-based spectral radiometers which necessarily must account for the effects of clouds and aerosols. As mentioned above, the important remote sensing parameter, spectral radiance reflectance, Lu ðlÞ=Ed ðl), has been found to be proportional bb/(a þ bb) or bb/a from Monte Carlo calculations for waters characterized by b/a values ranging from 1.0 to 5.0. The spectral radiance reflectance evaluated just above the ocean surface is defined as the remote sensing reflectance Rrs ðl). It should also be noted that remote sensing estimates of pigment (typically chlorophyll a or chlorophyll a þ phaeopigments) concentrations use empirical relations involving ratios such as Lwn(443 nm)/ Lwn(550 nm) or Rrs(490 nm)/Rrs(555 nm) or these ratios with other wavelength combinations. One of the most commonly used instruments for measuring AOPs is a broadband scalar irradiance E0 or PAR sensor, because of the need for determining the availability of light for phytoplankton and its relative simplicity. A spherical light collector made of diffusing plastic receives the light from approximately 4p sr. The light is transmitted via a fiberoptic connector or quartz light conducting rod to a photodetector which records an output voltage. The instrument is calibrated with a standard lamp. Analogous sensors use flat plate cosine or hemispherical collectors. Spectroradiometers are irradiance or scalar irradiance meters which use a variable monochromator placed between a light collector and the photodetector. Light separation can be achieved using sets of interference filters (usually 10 nm in bandwidth) selected for particular purposes (e.g., absorption peaks and hinge points for pigments).

(c) 2011 Elsevier Inc. All Rights Reserved.

248

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

Alternatively, higher spectral resolution (B3 nm from 400 to 800 nm) can be achieved using grating monochromators. Cosine (flat-plate), spherical, or hemispherical collectors are used for several different measurements or calculations of the AOPs described above (e.g., Ed, Lu, Kd, etc.). Irradiance cosine collectors are designed such that their responses to parallel radiant flux should be proportional to the angle between the normal to the collector surface and the direction of the radiant flux. This is necessary as the angle between the incoming radiant flux and the collector is variable, as is the area of the projection. To account for the differences in refractive indices of air and water, an immersion correction factor (typically 1.3–1.4) is applied. It should be noted that radiance sensors are designed to accept light in a small solid angle (typically a viewing angle of a few degrees). Irradiance and radiance instruments are calibrated using well-characterized, standard light sources. Because of the growing concern about UV radiation, a number of instruments are designed to measure into the UV portion of the electromagnetic spectrum as well. A grand challenge of oceanography is to greatly increase the variety and quantity of ocean measurements. Optical and other ocean measurements are expensive. Thus, a major goal is to develop new sensors and systems, which can be efficiently deployed from a host of available ocean platforms including ships, moorings, drifters, floats, and autonomous underwater vehicles (AUVs). A further need is to telemeter the data in near real-time. Shipbased observations are useful for detailed profiling (high vertical resolution), but moorings are better suited for high temporal resolution, long-term measurements. Drifters, floats, and AUVs can provide horizontal coverage unattainable from the other in situ platforms. Ultimately, all of these platforms, along with satellite- and plane-based systems, are needed to fill in the time–space continuum. Several optical systems, which have been used for mooringbased time series studies, are illustrated in Figure 1. These collective instrument systems include most of the sensors described earlier. The sampling for these instruments is typically done every few minutes (in some cases once per hour) for periods of several months. At this point in time, biofouling, rather than power or data storage, is the limiting factor. However, even this aspect is becoming less problematic with several new antibiofouling methods. In addition, more optical sensors are being deployed from other emerging platforms such as AUVs. The optimal utilization of optical data from the various platforms will require the use of advanced data merging methods and data assimilation models for predictions.

Optical Experiment Data Sets Finally, a few examples of data, which have been collected with some of the optical instruments described above (Figure 1), are presented. An interdisciplinary experiment devoted to the understanding of relations between mixing processes and optical variability was conducted south of Cape Cod, Massachusetts on the continental shelf. Several spectral optical instruments collected IOP and AOP data sets using ship-based profiling and towed systems, moorings, and bottom tripods. The period of the experiment covered almost one year (July 1996– June 1997). Some instruments sampled at several minute to hourly intervals whereas others sampled with vertical resolution on the scale of centimeters for a few weeks during two intensive field campaigns. Several interesting observations were enabled. These included: sediment resuspension forced by two passing hurricanes (hurricane Edouard and hurricane Hortense), spring and fall phytoplankton blooms, water mass intrusions, and internal solitary waves. Time series data collected using the BIOP system (e.g. BIOPS and MORS in Figure 1) which was deployed from a mooring and a bottom tripod located near each other in 70 m waters, are shown in Figure 2. Figure 2(A) shows the 37-m time series of total spectral absorption (water absorption has been subtracted) using a nine-wavelength ac-meter. The major feature is related to hurricane Edouard. The spectral absorption contributions due to phytoplankton are shown for two days in the summer in Figure 2(B); the peaks at 440 and 683 nm are caused by phytoplankton. Figure 2(C, D) shows ac-meter time series of spectral (nine wavelengths) scattering and attenuation coefficients. The record is dominated by sediment resuspension caused by mixing and waves created by hurricane Edouard and the more distant hurricane Hortense. The complete mooring data set showed that sediments were lifted more than 30 m above the ocean floor. The second example highlights optical data collected from a ship-based profiling system (similar to the system shown in Figure 1A). The purpose of the measurements was to study the dispersion of contaminants (treated wastewaters) discharged about 4 km offshore (70 m water depth) into Mamala Bay, off the coast of Honolulu, Hawaii in the fall of 1994. Optical and physical measurements were made from shipboard using both profile and towing modes in the vicinity of the outfall plume. Optical instrumentation included a beam transmissometer (660 nm), spectral ac-meters (nine wavelengths for a and c), PAR sensor, chlorophyll fluorometer, a particle size analyzer (laser diffraction method), and a

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

spectral absorption and fluorescence instrument. Physical measurements of temperature, salinity and pressure were also done. Sampling was designed to track both the horizontal and vertical structure of effluent as manifest in the optical signals. The water contributions to a, b, and c were removed for the analyses. The collective optical and physical measurements enabled the partitioning of particle types into categories: particulate versus dissolved components, phytoplankton opposed to detrital components, shallow layer versus deep layer phytoplankton, and old versus newly discharged sewage plume waters. Briefly, a profile (see Figure 3) taken near the end of the outfall

249

diffuser (water depth B 80 m) displayed the following features: (1) sewage plume waters centered near 60 m as characterized by low salinity and very high values of spectral attenuation and absorption (greater in the shorter wavelengths), and (2) a shallow phytoplankton layer near 20 m with modest relative maxima in chlorophyll fluorescence and spectral absorption and attenuation coefficients in the blue. The profile of spectral single-scattering albedo, o0 ðlÞ ¼ bðlÞ=cðlÞ, shows a general trend of increased importance of scattering for greater wavelengths with the most pronounced effect for the sewage plume waters centered at 60 m (Figure 3C).

(A)

(B) Par sensor

PAR sensor Pump

Transmissometer

HiSTAR Pump

9  Wetlabs ac-9

Hydroscat-6 Temperature sensor

Temperature sensor

Pressure sensor

HiROS

BIOPS PAR sensor

(C)

6  Ex, 16  Em Wetlabs safire Pump Temperature sensor (D)

SAFIRE

(E)

7  Ed's

Anti-fouling shuttered spectral radiometers

PAR sensor Data logger

Spectral radiometers Acoustic telemetry module Temperature sensor 7  Lu's

3  Ed's UCSB data logger/ Battery pack 3  Lu's Fluorometer & VSF sensors

MORS

MOSS

Figure 1 Schematic showing a variety of optical systems for mooring applications. (A) System used for measuring inherent optical properties (IOPs) including beam c (660 nm), spectral attenuation and absorption coefficients at nine wavelengths, along with PAR and temperature. (B) System used for measuring inherent optical properties including spectral attenuation and absorption coefficients at 100 wavelengths, spectral backscatter at six wavelengths, along with PAR, temperature, and pressure. (C) System measuring spectral fluorescence with 6 excitation wavelengths and 16 emission wavelengths; also PAR and temperature sensors are included. (D) System for measuring apparent optical properties (AOPs) including spectral downwelling irradiance and upwelling radiance at seven wavelengths and PAR along with temperature. A telemetry module is also included. (E) System for measuring apparent optical properties (AOPs) including spectral downwelling irradiance and upwelling radiance at three wavelengths along with instruments for measuring chlorophyll fluorescence, volume scattering function. An antifouling shutter system is also utilized.

(c) 2011 Elsevier Inc. All Rights Reserved.

250

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

37 m

0.10 E

_

a t_w() (m 1)

_

aph () (m 1)

0.3

o = YD 239 + = YD 204

0.08

0.2

0.1

0.06 0.04 0.02

715

210 Year da y

(A)

0 412

650

0 190 230 (1996)

250

n

( th 488 ng e l ve 412 Wa 532

m)

(B)

650 715 510 Wavelength (nm)

25 68 m

E

20

_

b t_w() (m 1)

H 15

10  (nm) 5

(C)

412 440 488 510

0 30 68 m

532

E

25

555 650 676

H

715

_

c t_w() (m 1)

20 15 10 5 0 235 (D)

240

245

250 Year day (1996)

255

260

265

Figure 2 Illustrations showing data collected from the BIOPS system shown in Figure 1(A) on the continental shelf south of Cape Cod Massachusetts in the summer and fall of 1996. (A) A time series of the total spectral absorption coefficient of light (after subtracting the clear water component) at a depth of 37 m (total water depth is 70 m). (B) The spectral absorption on YD 204 and 239. (C) Time series of spectral scattering coefficient computed by differencing the total spectral attenuation and absorption coefficients (again, the clear water coefficient has been subtracted) at 68 m. The large peaks are attributed to passages of hurricanes Edouard (E) and Hortense (H). (D) as for (C) except for spectral beam attenuation coefficients. (Figures based on Chang and Dickey, 1999.)

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

251

_

a () (m 1)

0.15 0.10 0.05 0 10

20

30

40 50 60 Depth (m)

70

80

(A)

700 750 600 650 550 500 m) 400 450 ngth (n Wavele

_

c () (m 1)

0.8 0.6 0.4 0.2 0 10

20

30

(B)

40 50 60 Depth (m)

70

80

650 550 600 500 m) 400 450 ngth (n Wavele

700 750

1.2 0

1.0 0.8 0.6 0.4 10 (C)

20

30

40 50 Depth (m

60

)

70

80

700 750 600 650 550 500 m) 400 450 ngth (n Wavele

Figure 3 Data collected near an ocean sewage outfall in Mamala Bay off Honolulu, Hawaii in the fall of 1994. Vertical profiles of (A) spectral absorption coefficient, (B) spectral attenuation coefficient, and (C) single scattering albedo. Data were obtained from an acmeter system similar to the one shown in Figure 1(A). Eight wavelengths are displayed and the major peak near 60 m is caused by sewage plume waters with high scattering and absorption coefficents. (Figures based on Petrenko et al., 1997.)

The final example describes an important linkage between in situ and remote sensing observations of ocean color. The Bermuda Testbed Mooring (BTM) is used to test new oceanographic instrumentation (including systems shown in Figure 1), for scientific studies devoted to biogeochemical cycling and climate change, and for groundtruthing (verification using in situ data) and algorithm development for ocean color satellites such as SeaWiFS. The latter aspect is the focus here. Two BTM optical systems utilizing radiometers for measurements of spectral downwelling irradiance and spectral upwelling radiance are illustrated in Figure 1(D) and (E). These systems measure at wavelengths, which are coincident with those of the SeaWiFS ocean color satellite, and sample at hourly intervals. Two or more systems are deployed at different depths and another radiometer system is deployed from the surface buoy to collect incident

spectral downwelling irradiance. Derived quantities include time series of several of the AOP quantities introduced earlier (e.g., spectral diffuse attenuation coefficients, spectral reflectances including remote sensing reflectance, and water-leaving radiance). Time series of spectral upwelled radiance from the BTM radiometer system (Figure 1D) located at 14 m are shown for the period April through July, 1999 in the top set of panels of Figure 4. The variability is primarily caused by the daily cycle of solar insolation, cloud cover, and optical properties associated with phytoplankton and their products. The bottom panels of Figure 4 show a subset of the water-leaving radiance data as determined from the BTM radiometers (extrapolation to surface) and SeaWiFS satellite sensors for six days during the particular sampling period. These two types of data are critical for quantifying both the temporal and horizontal spatial

(c) 2011 Elsevier Inc. All Rights Reserved.

252

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

3 2

412 nm

BTM 14 m Radiance (Lu)

1 0 3

443 nm

_

1

_

Radiance (W cm 2 nm 1 sr 1)

2 0

_

3 2

490 nm

1 0 3

510 nm

2 1 0 0.3

555 nm

0.2

0 March 31

April 22

1.6

April 21

May 14 June 5 Day (1999)

June 27

July 21

_

_

_

_

_

_

L w (W cm 2 nm 1 sr 1) L w (W cm 2 nm 1 sr 1)

0.1

1.6

April 26

1.6

1.2

1.2

1.2

0.8

0.8

0.8

0.4

0.4

0.4

0 400 450 500 550 600 1.6

July 17

1.2

0 0 400 450 500 550 600 400 450 500 550 600 1.6

July 19

1.6

1.2

1.2

0.8

0.8

0.8

0.4

0.4

0.4

0 400 450 500 550 600 Wavelength

May 10

SeaWiFs BTM

July 20

0 0 400 450 500 550 600 400 450 500 550 600 Wavelength Wavelength

Figure 4 BTM radiometer data collected with the system shown in Figure 1(D). The top set of panels shows a time series of spectral upwelling radiance (five wavelengths) from the 14-m spectral radiometer system (April–July, 1999). The bottom set of panels shows spectral water-leaving radiance derived from the BTM radiometers and from the SeaWiFS ocean color satellite for six days of the sampling period.

variability of water clarity, PAR, phytoplankton biomass, and primary productivity.

Toward the Future The last two decades have been marked by the emergence of a host of new optical instruments and systems; their applications are growing at a rapid rate. Looking toward the future, it is anticipated that optical instrumentation will be (1) more capable in spectral resolution, (2) smaller and require less

power, (3) suitable for deployment from a variety of autonomous platforms, (4) designed for ease in telemetering of data for real-time applications, and (5) less costly as demand increases enabling higher volumes of data collection.

See also Absorbance Spectroscopy for Chemical Sensors. Bioluminescence. Estimates of Mixing. IR Radiometers. Optical Particle Characterization. Penetrating Shortwave Radiation.

(c) 2011 Elsevier Inc. All Rights Reserved.

INHERENT OPTICAL PROPERTIES AND IRRADIANCE

Further Reading Agrawal YC and Pottsmith HC (1994) Laser diffraction particle sizing in STRESS. Continental Shelf Research 14: 1101--1121. Bartz R, Zaneveld JRV, and Pak H (1978) A transmissometer for profiling and moored observations in water. Ocean Optics V, SPIE 160: 102--108. Booth CR (1976) The design and evaluation of a measurement system for photsynthetically active quantum scalar irradiance. Limnology and Oceanography 19: 326--335. Chang GC and Dickey TD (1999) Partitioning in situ total spectral absorption by use of moored spectral absorption and attenuation meters. Applied Optics 38: 3876--3887. Chang GC and Dickey TD (2001) Optical and physical variability on time-scales from minutes to the seasonal cycle on the New England continental shelf: July 1996– June 1997. Journal of Geophysical Research 106: 9435--9453. Chang GC, Dickey TD, and Williams AJ 3rd (2001) Sediment resuspension over a continental shelf during Hurricanes Edouard and Hortense. Journal of Geophysical Research 106: 9517--9531. Dana DR, Maffione RA, and Coenen PE (1998) A new in situ instrument for measuring the backward scattering and absorption coefficients simultaneously. Ocean Optics XIV 1: 1--8. Dickey T (1991) Concurrent high resolution physical and bio-optical measurements in the upper ocean and their applications. Reviews in Geophysics 29: 383--413. Dickey T, Granata T, Marra J, et al. (1993) Seasonal variability of bio-optical and physical properties in the Sargasso Sea. Journal of Geophysical Research 98: 865--898. Dickey T, Marra J, Stramska M, et al. (1994) Bio-optical and physical variability in the sub-arctic North Atlantic Ocean during the spring of 1989. Journal of Geophysical Research 99: 22541--22556. Dickey T, Frye D, Jannasch H, et al. (1998) Initial results from the Bermuda Testbed Mooring Program. Deep-Sea Research I 45: 771--794. Dickey T, Marra J, Weller R, et al. (1998) Time-series of bio-optical and physical properties in the Arabian Sea: October 1994–October 1995. Deep-Sea Research II 45: 2001--2025.

253

Dickey TD, Chang GC, Agrawal YC, Williams AJ 3rd, and Hill PS (1998) Sediment resuspension in the wakes of Hurricanes Edouard and Hortense. Geophysical Research Letters 25: 3533--3536. Dickey T, Zedler S, Frye D, et al. (2001) Physical and biogeochemical variability from hours to years at the Bermuda Testbed mooring site: June 1994–March 1998. Deep-Sea Research. Foley D, Dickey T, McPhaden M, et al. (1998) Longwaves and primary productivity variations in the equatorial Pacific at 01, 1401W February 1992–March 1993. Deep-Sea Research II 44: 1801--1826. Gentien P, Lunven M, Lehaitre M, and Duvent JL (1995) In situ depth profiling of particle sizes. Deep-Sea Research 42: 1297--1312. Griffiths G, Knap A, and Dickey T (1999) Autosub experiment near Bermuda. Sea Technology, December. Jerlov NG (1976) Marine Optics. Amsterdam: Elsevier. Kirk JTO (1994) Light and Photosynthesis in Aquatic Ecosystems. Cambridge: Cambridge University Press. Mobley CD (1994) Light and Water: Radiative Transfer in Natural Waters. San Diego: Academic Press. Moore CC, Zaneveld JRV, and Kitchen JC (1992) Preliminary results from in situ spectral absorption meter data. Ocean Optics XI, SPIE 1750: 330--337. O’Reilly JE, Maritorena S, Mitchell BG, et al. (1998) Ocean color chlorophyll algorithms for SeaWiFS. Journal of Geophysical Research 103: 24937--24953. Petrenko AA, Jones BH, Dickey TD, Le Haitre M, and Moore C (1997) Effects of a sewage plume on the biology, optical characteristics and particle size distributions of coastal waters. Journal of Geophysical Research 102: 25 061--25 071. Petrenko AA, Jones BH, and Dickey TD (1998) Shape and near-field dilution of the Sand Island sewage plume: observations compared to model results. Journal of Hydraulic Engineering 124: 565--571. Petzold TJ (1972) Volume scattering functions for selected waters, Scripps Institution of Oceanography Reference 72–78. La Jolla, California: Scripps Institution of Oceanography. Smith RC, Booth CR, and Star JL (1984) Oceanographic bio-optical profiling system. Applied Optics 23: 2791--2797. Spinrad RW, Carder KL, and Perry MJ (1994) Ocean Optics. Oxford: Oxford University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDAL MIXING W. Munk, University of California San Diego, La Jolla, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1323–1327, & 2001, Elsevier Ltd.

Introduction Any meaningful attempt towards understanding how the ocean works has to include an allowance for mixing processes. A great deal is known about the transport of heat and solutes by molecular processes in laboratory-scale experiments. Quite naturally, at the dawn of ocean science, these concepts were borrowed to speculate about the oceans. On that basis it was suggested by Zoeppritz that the temperature structure T(z) at 1000 m depth could be interpreted in terms of the time history T(t) at the surface some 10 million years earlier. It would indeed be nice if past climate could be inferred in this simple manner. Once it was realized that molecular diffusion failed miserably to account for the transports of heat and salt, generations of oceanographers attempted to patch up the situation by replacing the molecular coefficients of conductivity and diffusivity by enormously larger eddy coefficients, but leaving the governing laws (equations) unchanged. By an arbitrary choice of the magnitude of the coefficients it was generally possible to achieve a satisfactory (to the author) agreement between theory and observation, a result aided by the uncertainty of the measurements. But the clear danger signal was there; each experiment, each process required a different set of values. The present generation of oceanographers has come to terms with the need for understanding the mixing processes, just as they had come to terms some years ago with the need for understanding how ocean currents, ocean waves etc. are generated. Wind mixing and tidal mixing are very different processes. And the widespread use of parametrization of mixing processes will not succeed unless there is an underlying understanding of what is being parametrized. Here we have come a long way and have a long way to go.

Stirring and Mixing In a Newtonian fluid the down-gradient flux of a quantity y is given by Fy ¼ k dy=dx

254

½1

It is the very smallness of the molecular diffusivity k which requires large gradients dy/dx in order to attain significant fluxes Fy . The fundamental distinction between stirring and mixing was first made in 1948. Stirring produces the gradients whereas molecular mixing reduces the gradients. For the purpose of this article stirring and mixing are both included in the discussion of the contribution of internal tides to mixing processes. For an ocean in steady-state there needs to be an overall balance between the generation and dissipation of mean-square gradients. Nearly all of ocean dynamics deals with processes that generate gradients. This takes place over a wide variety of scales, all the way up to the scales of ocean basins. Dissipation takes place on the ‘microscale,’ i.e., millimeters to centimeters. This is the scale which includes the dominant contributions to the gradient spectrum. A further increase in the spatial resolution of the measurements does not lead to a significant increase in the measured mean-square gradients.

The Battle for Spatial Resolution It is very difficult to attain a quantitative measure of mixing in the turbulent ocean interior. The problem is the need for very high spatial resolution. An eddy coefficient can be estimated as follows: kmolecular  rmsðdy=dxÞ ¼ keddy  meanðdy=dxÞ

½2

(The subscript ‘molecular’ is introduced here to emphasize the distinction.) The ratio: mean square gradient/square mean gradient (the ‘Cox number’) has been used to estimate the ratio of the eddy coefficient to the molecular coefficient. A typical value away from boundaries is keddy ¼ 105 m2s1, two orders of magnitude in excess of the molecular coefficient. Achieving the required resolution has been a major accomplishment; but there are many problems with the measurements, and even more problems with the interpretation of the measurements along the lines of eqn[2]. It was only with the confirmation by a tracer release experiment that the community has come to accept the value kpelagic ¼ 105 m2s1 for the eddy diffusivity in the interior ocean, away from rough topography. There is of course considerable variability from place to place, but the surprising finding is not how large this variability is but how small it is.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDAL MIXING

Maintaining the Stratification A quite different estimate of eddy diffusivity associated with pelagic mixing can be made from the following considerations. Bottom water is formed in the winter by convective overturning in just a few places: the Greenland Sea, the Labrador Sea and along the Antarctic continent. The formation is estimated at Q ¼ 25 Sverdrups (25  106 m3s1). This would fill the oceans with cold water in a few thousand years. The reason this does not happen is that turbulent diffusion downward from the warm surface balances the upwelling of cold water. With reasonable assumptions this leads to an estimate of eddy diffusivity kstratification ¼ 104 m2 s1, ten times the measured pelagic value. Measurements near topography do indeed give high diffusivities, orders of magnitude above the pelagic value. One simple interpretation is that there are concentrated areas of mixing (just as there are concentrated areas of bottom water formation) from which the water masses (but not the turbulence) are exported into the interior ocean. We can ask the question whether the global stratification can be maintained by vertical mixing in 10% (say) of the ocean volume with an average diffusivity 100 times the pelagic value? The work done against buoyancy by turbulent mixing in a stratified fluid can be written eb ¼ kðg=rð dr=dzÞÞ ¼ kN2 W kg1

½3

where N is the buoyancy frequency. Only a fraction g (called the ‘mixing efficiency’) of the work goes into increasing potential energy (the rest goes into joule heat). A typical value is g ¼ 0.2. The total work per unit area is etotal ¼ eb/g. For the world ocean of area A, the total work done is ð ½4 D ¼ A retotal dz ¼ gg  1kADr W where Dr ¼ 1 kg m3 is taken as the difference between surface and bottom density. Then for A ¼ 3.6  1014 m2 and k ¼ kstratification ¼ 104 m2 s1, one has D ¼ 2 TW (1 terawatt ¼ 1012 W) for the power required to maintain the global stratification in the face of 25 Sverdrups of bottom water formation. To maintain the pelagic turbulence requires only 0.2 TW.

Tidal Dissipation: The Astronomic Evidence It is interesting to compare these numbers with the dissipation of tidal energy. We know this number with remarkable accuracy to be 2.570.1 TW for

255

the principal lunar tide (M2); it is obtained from the measured rate of 3.8270.07 cm y1 at which the Moon is moving away from the Earth. For all solar and lunar tides the dissipation is 3.7 TW, but with lesser certainty. We note that the tidal dissipation is of the same magnitude as the 2 TW required for maintaining the ocean stratification. Is this an accident? The astronomic evidence tells us nothing about how and where the dissipation takes place. Allowing for dissipation in the solid Earth and atmosphere leaves 3.4 TW to be dissipated somewhere somehow in the ocean. Ever since it was estimated in 1919 that the dissipation in the Irish Sea is at 0.060 TW, the traditional sink has been in the turbulent bottom boundary layers of marginal seas, about 60 Irish Seas for the world. And before the astronomic estimates settled down to their present value, the ocean estimates kept rising and falling with the astronomic estimates.

Boundary Layer Dissipation versus Scatter When we speak of tides we usually refer to surface (barotropic) tides which have a nearly uniform current velocity from top to bottom, and a maximum vertical displacement at the surface. However, there is also a class of internal (baroclinic) tides with velocities that vary with depth and with maximum displacements in the interior. A surface (barotropic) tide has essentially no shear in the interior ocean, There is shear near the bottom boundary, but the barotropic tidal velocities are so low in the deep ocean that the dissipation is negligible. In shallow seas the barotropic tidal currents are amplified, and the dissipation (proportional to the current cubed) is greatly amplified. This is the basis on which the global tidal dissipation has been attributed to the marginal seas. Internal tides are part of a larger class of internal waves with frequencies other than tidal frequencies. A possible mechanism of tidal dissipation is the scattering of surface tides into internal tides, with subsequent transfer of energy into the broad spectrum of internal waves, and finally into turbulent dissipation: surface tides-internal tides-internal waves-turbulence. What is required at the second stage is some nonlinear frequency splitting which converts the low-frequency tidal line spectrum into a closely packed high-frequency continuum that resembles the observed internal wave spectrum. The final step is associated with the fact that the internal wave spectrum is at or near instability in the

(c) 2011 Elsevier Inc. All Rights Reserved.

256

INTERNAL TIDAL MIXING

Richardson sense: the means-square shear is roughly 4 N2 (N is the buoyancy or Brunt-Va¨isa¨la¨ frequency). Scattering of surface tides into internal tides can take place along wavy bottoms. A second possibility is scattering along submarine ridges. An acoustic tomography experiment north at Hawaii detected internal waves of tidal frequency radiating northward.

Satellite Altimetry to the Rescue A subsequent analysis of satellite altimetry clearly showed internal tides emanating from the Hawaii submarine ridge. This is shown in Figure 1. The radiated power was estimated at 0.015 TW. So 14 Hawaiian chains will radiate 0.2 TW, enough to power the pelagic mixing associated with kpelagic ¼ 105 m2 s1. The discovery of internal tide signatures by means of satellite altimetry was an altogether unexpected dividend from a technology that

has revolutionized tidal analysis. The global sampling of surface elevation has introduced a new element into a subject that had gone to bed (in the opinion of some) with the work of Victorian mathematicians. With the Laplace tide equation as a guide for the tidal response of a nondissipative ocean, the assimilation of TOPEX/POSEIDON altimetry can lead to estimates of where one needs to introduce dissipation for agreement with the satellite data. The most recent result allocates roughly 1 TW to the open ocean, mostly over rough terrain. The tentative conclusion is that tidal dissipation is a significant factor in open ocean turbulent mixing. Supporting evidence comes from the measurements of tracer dispersion and microstructure in the Brazil Basin. Diffusivities of k ¼ 2  104  4  104m2 s1 at an elevation of 500 m above the abyssal hills of the Mid-Atlantic Ridge, increasing to 10  104 m2 s1

Figure 1 Surface manifestation of internal M2 tides emanating from the Hawaii’an Island Chain. (Reproduced from Ray and Mitchum 1997.) The wiggly curves show the amplitudes along the ascending orbits of TOPEX/POSEIDON, with positive elevations on the north side. The dashed lines are the inferred crests of the mode 1 component of internal tides. Background shading corresponds to bathymetry, with darker areas denoting shallower water. The triangle to the north east shows the position of the tomographic array. Adapted from Ray & Mitchum (1997).

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDAL MIXING

near the bottom have been obtained. Perhaps the most important result is that over a period of a month the diffusivities vary by a factor of two, with the large values occurring at spring tide and the small values at neap tide.

Discussion There is more than enough tidal dissipation to feed the measured pelagic turbulence associated with kpelagic ¼ 105 m2 s1. With regard to the larger value kstratification ¼ 104 m2 s1, the present best estimates would suggest that tidal dissipation could power half the turbulence needed to account for the observed ocean stratification. Figure 2 attempts an allocation of tidal energy flux, but there are many uncertainties, some by factors of two or more. The assumed onedimensional balance between upward advection and

257

downward diffusion as a measure of kstratification is itself somewhat uncertain. However, the present conclusion is that tidal dissipation is a significant, possibly dominant factor driving mixing in the ocean interior. The combination of detailed in situ measurements of turbulent mixing subject to a global lid on available tidal energy has led to giant strides towards a meaningful parametrization of ocean mixing, whether or not tidally induced. The conversion of wind energy to turbulent mixing plays a major role, particularly in the upper oceans. In Figure 2, equal weight has arbitrarily been assigned to tides and winds. We shall have to await the outcome of this competition.

See also Dispersion and Diffusion in the Deep Ocean. Internal Tides. Internal Waves. Tides. Turbulence in the Benthic Boundary Layer.

Further Reading

2

3

Figure 2 Sketch of proposed flux of tidal energy (modified from Munk and Wunsch, 1997). The traditional sink is in the turbulent boundary layer of marginal seas. Scattering into internal tides over ocean ridges (by the equivalent of 14 Hawaii’s) and subsequent degradation into the internal wave continuum feeds the pelagic turbulence at a level consistent with kpelagic ¼ 105 m2 s1. Most of the ocean mixing is associated with a few concentrated areas of surface to internal mode convergence over regions of extreme bottom roughness and with severe wind events. Light lines represent speculation with no observational support.

Cartwright DE (1999) Tides; a Scientific History. Cambridge: Cambridge University Press. Dushaw BD, Cornuelle BD, Worcester PF, Howe BM, and Luther DS (1995) Barotropic and baroclinic tides in the central North Pacific Ocean determined from longrange reciprocal acoustic transmission. Journal of Physical Oceanography 25: 631--647. Eckart C (1948) An analysis of the stirring and mixing processes in incompressible fluids. Journal of Marine Research 7: 265--275. Gregg MC (1988) Diapycnal mixing in the thermocline; a review. Journal of Geophysical Research 92: 5249--5286. Ledwell JR, Watson AJ, and Law CS (1993) Evidence for slow mixing across the pycnocline from an open-ocean tracer release experiment. Nature 364: 701--703. Munk W (1997) Once again: once again-tidal friction. Progress in Oceanography 40: 7--36. Munk W and Wunsch C (1997) The Moon, of course. Oceanography 10: 132--134. Munk W and Wunsch C (1998) Abyssal recipes II: energetics of tidal and wind mixing. Deep-Sea Research I 45: 1977--2010. Ray RD and Mitchum GT (1997) Surface manifestation of internal tides in the deep ocean: observations from altimetry and island gauges. Progress in Oceanography 40: 135--162. Taylor GI (1919) Tidal friction in the Irish Sea. Philosophical Transactions of the Royal Society A 220: 1--93.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDES R. D. Ray, NASA Goddard Space Flight Center, Greenbelt, MD, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1327–1335, & 2001, Elsevier Ltd.

Introduction Oceanic internal tides are internal waves with tidal periodicities. They are ubiquitous throughout the ocean, although generally more pronounced near large bathymetric features such as mid-ocean ridges and continental slopes. The internal vertical displacements associated with these waves can be extraordinarily large. Near some shelf breaks where the surface tides are strong, internal displacements (e.g., of an isothermal surface) can exceed 200 m. Displacements of 10 m in the open ocean are not uncommon. The associated current velocities are usually comparable to or larger than the currents of the surface tide. Internal tides can occasionally generate packets of internal solitons which are detectable in remote sensing imagery. Other common nonlinear features are generation of higher harmonics (e.g., 6 h waves) and wave breaking. Internal tides are known to be an important energy source for mixing of shelf waters. Recent research suggests that they may also be a significant energy source for deep-ocean mixing. Internal tides were first recognized in the early part of the twentieth century, yet as late as the 1950s arguments were still being waged over what causes them. Their wavelengths, generally shorter than 200 km, are poorly matched to the planetary-scale astronomical tidal potential, so the generation mechanism for surface tides appears inapplicable. Various theories invoking hypothetical resonances at inertial latitudes (where tidal and Coriolis frequencies are equal) were put forward, but they are not compelling, not least because the inertial latitude for the dominant tide M2 is in the far polar latitudes (74.51). The now accepted explanation for internal tides is that they are generated by the interaction of the barotropic surface tide with bottom topography. As the tide sweeps stratified water over topographic features, it disrupts normal (equilibrium) isopycnal layers, setting up pressure gradients that induce secondary internal motions at the same frequency as the tide. Since internal tides are a special kind of internal wave, much of our knowledge of internal waves is immediately applicable. For example, an internal

258

tide always displays current shear – i.e., the associated horizontal current velocities change with depth – whereas the surface tide’s horizontal current is independent of depth. And like other internal waves in smoothly varying density stratification, an internal tide displays the seemingly odd property that its group velocity is in the same vertical plane but perpendicular to its phase velocity. The fundamental properties of internal tides, including whether or not they even exist, are controlled by the relative magnitudes of three basic frequencies: the tidal frequency o, the local Brunt-Va¨isa¨la¨ or buoyancy frequency N, and the local Coriolis frequency f. Depending on which of these frequencies is highest and which lowest, internal tides may propagate freely away from their generation point, they may be reflected in some manner, or they may be evanescent. For midlatitude semidiurnal tides, typically f oooN, a regime allowing free propagation. Given that the generation and propagation of internal tides depend strongly on the stratification, it is not surprising that most observations have found internal tides to be highly variable, sometimes with pronounced seasonal variations. In some places they appear only during spring tides (when solar and lunar tides are at maximum). In other places they appear randomly intermittent, evident for several days and then disappearing. Some observations, primarily from the open ocean, have revealed a component that does remain temporally coherent with the astronomical potential (see below), but the dominant characteristic of internal tides in most regions is one of incoherence, both spatially and temporally.

Modes and Beams Two complementary dynamical frameworks are used for analyzing internal tides: decomposition into vertical modes and propagation along characteristics. Generally, the latter description is more useful near generation points, and the modal description more useful elsewhere, but in any particular situation one or the other approach may be advantageous. Both approaches require knowledge of the stratification, usually parameterized by the buoyancy frequency N. This is the frequency with which a vertically displaced fluid element would oscillate because of restoring buoyancy forces. It is given by N ¼ Oðgr1 @r=@zÞ, where g is the acceleration of gravity and r is the average potential density, a

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDES

function of position and depth. The Coriolis frequency f ¼ 2Osin f, where f is the latitude and O is the Earth’s sidereal rotation frequency (7.2921  105s1).

259

Vertical displacement Sea surface

Modes

Mode 1

The governing dynamical equations for internal tides, under linear, hydrostatic, inviscid, Boussinesq and flat-bottom assumptions, and neglecting the horizonal currents, may be solved by separation of variables. The equation for the vertical displacement leads to an eigenvector problem with eigenvalue a2n :

Mode 3

2

q Gn ðzÞ þ a2n N 2 ðzÞGn ðzÞ ¼ 0 qz2

Mode 5

The eigenvectors Gn ðzÞ, ordered so that anþ1 > an , provide a complete, orthogonal basis for the internal vertical displacements. The corresponding equations for horizontal dependence yield expressions for the horizontal wave number k, phase velocity cp ¼ o=k, and group velocity cg ¼ do=dk in terms of an : k2 ¼ a2n o2  f 2





 c2p ¼ o2 = a2n o2  f 2 cg ¼ cp a2n

Seafloor Figure 1 Baroclinic displacement modes 1, 3, and 5, computed for a buoyancy frequency profile from the deep ocean.


1

If N is taken constant, then Gn can be found analytically: Gn ðzÞ ¼ asinðnpz=DÞ for an ocean depth D. If N is taken more representative of deep-ocean conditions, with a peak at the pycnocline, then the oscillations in Gn are shifted upward (Figure 1). Notice that the displacements are small or zero at the top and bottom of the water column, and that each Gn has n  1 crossings of the origin. Horizontal velocity modes are given by dGn =dz, and they have n crossings and hence nonzero shear for all modes. Most observations of internal tides (except those very near the generation point) are adequately described by a superposition of a few low order modes. In the deep ocean typical phase speeds cp are of order 3 m s1 for n ¼ 1. Corresponding wavelengths l ¼ 2p=k are between 100 and 200 km. Higher order modes have speeds and wavelengths given roughly by cp =n and l=n, respectively. On continental shelves both speeds and wavelengths may be an order of magnitude smaller. These values are for semidiurnal tides; wavelengths of diurnal tides are approximately twice as large. From the above expressions for k and cp it is apparent that internal tides cannot freely propagate unless o > f . They are ‘evanescent’ (exponentially

damped)’ polewards of the critical latitudes where o ¼ f . Freely propagating waves for diurnal tides are therefore confined to the region between latitudes 7301. (In fact, unambiguous observations of diurnal tides are fairly rare, but this is partly due to relatively weak barotropic forcing and higher background noise levels.) Beams

A complementary approach to modal analyses stems from the equation for the two-dimensional stream function, which is hyperbolic in spatial coordinates and may therefore be solved by the method of characteristics. The resulting solution consists of narrow beams of intense motion embedded in an otherwise resting ocean. The group velocity, and hence the energy propagation, follow the characteristics, which are along lines of slope:  2 1=2 o  f2 c ¼ tan y ¼ 7 N 2  o2 From a given internal tide generation point, energy thus propagates along beams at the angle y relative to

(c) 2011 Elsevier Inc. All Rights Reserved.

260

INTERNAL TIDES

horizontal, the angle depending (for a given f and N) only on the tidal frequency. Well defined beams comprise a large number of modes, with modal cancellations occurring outside the allowed beam. A numerical example of beam-like propagation from a shelf break is shown in Figure 2. Generation of internal tides is apparently especially efficient when the seafloor slopes at precisely the critical value c. Barotropic flow is then coincident with the motion plane for free internal waves, resulting in near-resonant conditions in which even quite small surface tides can generate internal tides. With nominal values of NB50 cpd, oB2 cpd, f B0:6 cpd, then y is 21. Continental slopes commonly exceed this, so c would be attained near the shelf break, as depicted in Figure 2. When an internal wave is reflected from the ocean bottom or ocean surface, energy propagation is still confined to the angle y, which makes for a curious variation on the usual laws of reflection. If the wave 0 4 80 4 80 4 80 4 80 4

80 4

is incident upon bathymetry that is steeper than y (supercritical case), energy is reflected backwards into deeper water. If the bathymetry is less steep (subcritical case), energy is reflected forward toward shallower water (see Figure 3). Ocean observations of this behavior are not easy to obtain, since mooring instruments must be precisely placed (depending on the ambient N); yet measurements of internal tides in the Bay of Biscay have not only observed the downward energy propagation from the generation point, but also the subsequent reflection from the ocean bottom. In the Bay of Biscay, as in most places, N diminishes with depth, so y grows larger and the beams become steeper as they approach the bottom (Figure 2 and Figure 3 are drawn for constant N.) With such reflection properties, internal waves incident on a subcritical sloping bottom will be focused into the shallows (as in Figure 3A), with energy density correspondingly intensified. The same mechanism tends to trap internal wave energy within steep 80 4 8 0 4 8 0 4 8 0 4 8 0 4 8 0 4

8

(A)

_ 53.8

0

L (km) 40

20

60

80

0

_ 30

_ 60 150 (B)

_ 60

_ 60 _ 90 _150 _180 _ 240

0 30 90

Figure 2 Theoretical internal wave beam from a shelf edge in a constant N ocean. (A) vertical displacements following beam at constant slope tan y; (B) phase contours (in degrees) of the vertical displacements relative to the surface tide. Notice how phase propagation is at right angles to the beam; i.e., the phase velocity is perpendicular to group velocity (and to the direction of energy propagation). (Reproduced with permission from Prinsenberg SJ and Rattray M (1975) Effects of continental slope and variable BruntVa¨isa¨la¨ frequency on the coastal generation of internal tides. Deep Sea Research 22: 251–263.)

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDES

261

downward edge the tide has generated packets of internal solitons. These solitary waves, starting from an initial 7 m depth, have extraordinarily nonlinear isotherm displacements of 25 m. Although usually less dramatic, this phenomenon is not uncommon in high tide regions on a continental shelf. The internal tide, generated at the shelf break, propagates shorewards and becomes progressively more nonlinear until the bore disintegrates into a group of solitons, the leading one usually of largest amplitide.

(A)

Moored Current Meters

(B)

Figure 3 Successive reflections of an internal wave along (A) a subcritical seafloor and (B) a supercritical seafloor. Arrows denote direction of energy propagation.

(supercritical) canyons, where the canyon sides reflect energy ever deeper, focusing it toward the canyon floor. If the floor is subcritical, then energy is further focused toward the canyon head. Intense internal tide currents and large kinetic energy densities have indeed been observed in canyons, and especially near canyon heads. In the presence of internal viscosity or other dissipative mechanisms, internal tidal beams widen. Because group velocities are smaller and decay scales shorter for higher order modes, beams tend to disintegrate rapidly into the few low order modes that are most commonly observed.

Observations Internal tides have been observed with a great multitude of instruments and technique, both in situ and remote. Four distinctly different examples are given here which serve to highlight a number of characteristic features of internal tides. Except for the first example, emphasis is given to deep-sea tides. Vertical Profilers

Vertical profilers, ranging from echo sounders to repeated hydrographic casts to special yo-yo instruments, provide some of the clearest pictures of internal tides. An especially dramatic example from the continental shelf off Oregon is shown in Figure 4. It shows a clear semidiurnal signal in the isotherm displacements, somewhat distorted into a bore-like shape (akin to the nonlinear distortion seen in shoaling surface tides in very shallow water). Along its

Because of their widespread deployments, current meters provide perhaps the most common for observing, or at least detecting, internal tides, especially in the open ocean. Sufficient vertical sampling is required for decoupling the internal modes from the surface tide (and unfortunately sufficient sampling is not common). Figure 5 is an example of marginally adequate vertical sampling; it shows tidal current estimates extracted from moored meters near 1101W on the Pacific equator. Estimates are given for each of 10 months, at 10 depths throughout the water column. The current ellipses are fairly uniform below 1000 m; these depths are dominated by the stable, depth-independent currents of the surface tide. In contrast, large temporal variation, and occasionally much larger amplitudes, are evident in the shallower estimates; in these depths, where the buoyancy frequency (and its change) is maximum, the tidal signal is dominated by the internal tide. Modal analysis reveals that the internal tide is essentially random, isotropic, and without a dominant mode for these 10 months. Such observations are characteristic of in situ observations of internal tides; but in a few locations in the deep ocean, a component of the internal tide has been observed that is not so variable and that maintains phase lock with the astronomical tide. The famous MODE experiment in the western Atlantic found that approximately 50% of the internal tide variance was temporally coherent with the astronomical tide. Such observations imply a nearly constant ocean stratification, at least to the extent that it determines generation and propagation properties. Satellite Altimetry

Recently satellite altimetry has been shown capable of providing a near-global view of the coherent component of internal tides. It does this by detecting the very small surface displacements associated with internal tides. These are given roughly by the tide’s internal displacements scaled by Dr=r, the fractional difference in water density, typically of order 0.2%,

(c) 2011 Elsevier Inc. All Rights Reserved.

262

INTERNAL TIDES

Temperature (˚C) 18

0

˚C

_5 16

Depth (m)

_ 10 14

_ 15

_ 20

12

_ 25 10 _ 30 _ 35

8 269.2

269

269.4 269.6 Time (days)

269.8

Temperature (˚C) 0

18

˚C

_5 16

Depth (m)

_ 10 14

_ 15

_ 20

12

_ 25 10

_ 30

_ 35

8 268.96

268.98 Time (days)

269

269.02

Figure 4 Color contour time series of temperature profiles from the surface to 35 m depth, obtained by repeated 80 s raising and free-falling of the loose-tethered microstructure profiler, deployed offshore Tillamook, Oregon in October 1995. Top: the semidiurnal internal tide displacement (most clearly seen along the yellow 13.81C isotherm) for a 24 h period. Bottom: a zoom view of a 1.7 h period showing the start of the first soliton displacements. The solitons are separated by roughly 10 min. (Reproduced with permission from Stanton TP and Ostrovsky LA (1998) Observations of highly nonlinear internal solitons over the continental shelf. Geophysical Research Letters 25: 1695–1698.)

thus implying surface displacements of a few centimeters for internal displacement of tens of meters. Altimetry detects such small waves as modulations (with wavelengths 100–200 km for internal mode 1)

of the surface tide as estimated along satellite tracks. Because tides can be estimated from altimeter data only by gathering multi-year time series of elevations at a particular site, only the coherent component of

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDES

263

2 cm/s +

+

20 m +

+

+

50 m +

+

+

+

100 m +

+

MAY

JUN

277 m

727 m 927 m

1027 m

2027 m

2127 m 3027 m APR 1980

JUL

AUG

SEP

OCT

NOV

DEC

JAN 1981

Figure 5 M2 tidal current ellipses by month at each of 10 depths obtained from moored current meters near 01, 1101W. Scale bar for velocity is at upper right. Each ellipse indicates how the direction and magnitude of the horizontal current velocity evolves over one tidal cycle. All ellipses are polarized clockwise except those marked with a plus sign. (Reproduced with permission from Weisberg RH, Halpern D, Tang T, Hwang SM (1987) M2 tidal currents in the eastern equatorial Pacific Ocean. Journal of Geophysical Research 92: 3821–3826.)

the internal tide which maintains phase lock with the surface tide is capable of being detected. Figure 6 gives an example of the first detection of such waves, near the Hawaiian Ridge. The waves are roughly 5 cm amplitude near the ridge and decay slowly with distance, but are still detectable 1000 km away. Phase estimates (not shown) reveal clearly that the waves are propagating away from the ridge. Evidently they are created by the barotropic tide striking the ridge (at nearly right angle from the north) and generating an internal tide that propagates both northwards and southwards. The picture reveals three important aspects of deep-ocean internal tides: (1) that in some locations they maintain temporal coherence over several years, thus allowing altimetry to measure them, (2) that they maintain spatial coherence over a wide area, and (3) that they are capable of propagating hundreds to thousands of kilometers before being dissipated. All three aspects contrast sharply with the usual picture of incoherence obtained from in situ observations.

Waves similar to those in Figure 6 have been detected in many regions throughout the global ocean. However, altimetry is incapable of detecting internal tides in a region where they are temporally incoherent. Such is apparently the case, for example, off the northwest European shelf, a region known for some of the largest internal tides in the world, but where the coherent signals in altimeter data are extremely weak. Internal tide studies with satellite altimetry are relatively new, and further work should reveal new facets from a global perspective. Acoustic Methods

A second example of a powerful, but unconventional, technique for studying coherent internal tides is acoustic tomography. Differences in two-way acoustic travel times between reciprocal transceivers are sensitive to barotropic tidal currents within the acoustic path. Similarly, since vertical isotherm displacements perturb the sound speed

(c) 2011 Elsevier Inc. All Rights Reserved.

264

INTERNAL TIDES

10 cm

30°N

25°N

10

0

20°N

180°

170°W

160°W

61

23

2 11

74

36

5 12

87

49

11

15°N

150°W

Figure 6 Mean elevations at the sea surface of internal tides near Hawaii, deduced from altimeter measurements of the Topex/ Poseidon satellite. Positive values (north of the trackline) indicate that the internal tide’s surface elevation is in phase with the barotropic tide’s elevation. Scale bar for elevations at upper left. Background shading corresponds to bathymetry, with darker shading denoting shallower depths and the main axis of the Hawaiian Ridge. Only internal tides that are coherent with the surface tide over the entire measurement period (here 3.5 years) can be detected in this manner.

within the path, the sums (or averages) of the travel times are sensitive to the internal tide. From a sufficiently long time series the mean tidal characteristics along a given path can be determined. The seemingly coarse spatial resolution is actually an advantage, because it suppresses short-scale internal waves and other noise that typically plague current meter measurements. And, in fact, an array of acoustic transceivers can act as a very sensitive directional antenna for spatially coherent internal tides. Such an array in the central Pacific, consisting of acoustic paths roughly 1000 km long and located just north of the area shown in Figure 6, has measured the same coherent internal tide field seen in the altimetry and indicates that the primary source is the Hawaiian Ridge, even at that great distance.

Implications for Energetics and Mixing Internal tides are an important energy source for vertical mixing, especially in coastal waters where they help maintain nutrient fluxes from deep water to euphotic zones on the shelf. A good example is the

Scotian shelf off Nova Scotia where internal tides are responsible for a strip of enhanced concentrations of nutrients and biomass along the shelf break. During each tidal cycle one or two strong (50 m) internal solitons (compare Figure 4) are generated near the shelf edge, moving shoreward but dissipating rapidly, possibly within 10 km. Estimated energy fluxes of 500 W m1 appear more than adequate to maintain observed nutrient supply to the mixed layer. Similar mixing mechanisms have been observed in the Celtic Sea and elsewhere. In the open ocean it seems reasonable that internal tides dissipate by transferring energy into the internal wave continuum or by directly generating pelagic turbulence, but the associated energy fluxes, and even the dominant mechanisms, are unclear. Nonlinearity is a common feature of internal tides (e.g., occurrences of higher harmonics), so ‘diffusion’ into the continuum is conceivable via nonlinear (resonant triad) interactions, but the evidence for this is so far more anecdotal than convincing. Bottom scattering of low mode tides into higher modes may play a role, as well as wave reflections off sloping bottoms, which tend to intensify kinetic energy densities and may lead to shear instabilities and wave breaking.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL TIDES

The traditional view is that both the internal wave continuum and the pelagic turbulence and mixing are maintained by nontidal mechanisms such as wind generation; whether internal tides play a major or minor role in this is not resolved. At a minimum, improved quantitative estimates are needed for the global internal tide energy budget. The internal tide energy budget also has a bearing on a longstanding geophysical problem: finding the energy sink for the global surface tide. If the generation/dissipation rate for internal tides is sufficiently large, then internal tide generation conceivably supplements the traditional sink of botton friction in shallow seas. Dissipation rates for the surface tide are well determined by space geodesy (e.g., lunar laser ranging) at 3.7 TW, with 2.5 TW for the principal tide M2. How much of this is accounted for by conversion into internal tides is not well determined; published estimates range from o100 GW (0.1 TW) to >1 TW. There is fairly wide agreement that generation of internal tides at continental slopes provides a fairly small energy sink. Both models and measurements suggest that typical energy fluxes at shelf breaks are of order 100 W m1, leading to a global total of order 15 GW. This is perhaps an underestimate, because it may not fully account for shelf canyons and other three-dimensional features, but the order of magnitude seems reliable. Internal tide generation by deep-ocean topography, however, may be far more important. Recent research based on global tide models as well as on empirical estimates of tidal dissipation deduced from satellite altimetry suggests that generation of internal tides by deep-sea ridges and seamounts could account for 1 TW of tidal power. Refining such estimates, and understanding the role that internal tides play in

265

generation of the background internal wave continuum, in vertical mixing, and in maintenance of the abyssal stratification, are some of the outstanding issues of current research.

See also Acoustics in Marine Sediments. Internal Tidal Mixing. Internal Waves. Tides.

Further Reading Baines PG (1986) Internal tides, internal waves, and nearinertial motions. In: Mooers C (ed.) Baroclinic Processes on Continental Shelves. Washington: American Geophysical Union. Dushaw BD, Cornuelle BD, Worcester PF, Howe BM, and Luther DS (1995) Barotropic and baroclinic tides in the central North Pacific Ocean determined from longrange reciprocal acoustic transmissions. Journal of Physical Oceanography 25: 631--647. Hendershott MC (1981) Long waves and ocean tides. In: Warren BA and Wunsch C (eds.) Evolution of Physical Oceanography. Cambridge: MIT Press. Huthnance JM (1989) Internal tides and waves near the continental shelf edge. Geophysical and Astrophysical Fluid Dynamics 48: 81--106. Mun WH (1997) Once again: once again – tidal friction. Progress in Oceanography 40: 7--35. Ray RD and Mitchum GT (1997) Surface manifestation of internal tides in the deep ocean: observations from altimetry and island gauges. Progress in Oceanography 40: 135--162. Wunsch C (1975) Internal tides in the ocean. Reviews of Geophysics and Space Physics 13: 167--182.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL WAVES C. Garrett, University of Victoria, VIC, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1335–1342, & 2001, Elsevier Ltd.

Introduction Waves at the sea surface are a matter of common experience. Surface tension is the dominant restoring force for waves with a wavelength less than 17 mm or so; longer waves are more affected by gravity. They have periods up to about 20 s and amplitudes that may be many meters. Given the stable density stratification of the ocean, it is not surprising that there are also ‘internal gravity waves,’ with a water parcel displaced vertically feeding a gravitational restoring force. The wave periods depend on the degree of stratification but may be as short as several minutes and can be long enough that the Coriolis force plays a major role in the dynamics. Vertical displacements are typically of the order of ten meters or so, with horizontal excursions of several hundred meters. The associated horizontal currents are typically several tens of millimeters per second. An interesting difference from the surface wave field is that internal waves always seem to be present, without the intense storms or periods of calm that exist at the surface. The existence of internal waves complicates the mapping of average currents and depths of particular density surfaces. They have also been the objective of intensive military-funded research because of the possibility that wakes of internal waves generated by submarines might be detectable by remote sensing, thus betraying the submarine’s location. More conventional acoustic means of submarine detection are complicated by the deflection of acoustic rays by the rather random variations in sound speed induced by internal waves. In civilian activities, the currents and buoyancy changes associated with internal waves are a matter of concern in offshore oil drilling. Most importantly, perhaps, the current shear of internal waves, including those of tidal frequency, can lead to instability and turbulence, and so the waves are the main agent for vertical mixing in the ocean interior. This mixing plays a major role in determining the strength of ocean circulation, and hence the poleward heat flux and climate. The mixing, along with the associated circulation, also

266

provides nutrient fluxes into the sunlit upper ocean where primary biological production occurs. Understanding internal waves is thus of vital importance, particularly since they occur at too small a scale to be treated explicitly in computer models of the ocean. Their effects must be ‘parametrized,’ or represented by formulas that involve only the quantities that are carried in the model. In this respect, internal waves in the ocean are somewhat akin to clouds in the atmosphere – they play a vital, perhaps even controlling, role in global-scale problems. (The atmosphere also has internal waves, of course, which are known to play a major role in redistributing momentum.) This short article will first describe the waves that can occur at sharp density interfaces in clear analogy to waves at the sea surface. This will be followed by a description of the waves that can propagate through a continuously stratified ocean, and a discussion of their generation, evolution, and relationship to ocean mixing.

Interfacial Waves If the ocean consists of an upper layer of density r  Dr and thickness h1 above a layer of density r and thickness h2, then waves that have a wavelength much greater than both h1 and h2 travel at a speed [g0 h1h2/(h1 þ h2)]1/2 independent of wavelength, where g0 ¼ gDr/r is known as the ‘reduced gravity’. If h2bh1, this becomes (g0 h1)1/2, in clear analogy to the speed (gh)1/2 for surface waves that are long compared with the water depth h. This formula for the speed of interfacial waves also holds for h2bh1 even if the wavelength is not long compared with h2. The theory behind this requires that the amplitude of the waves is much less than the thickness of the layers. Many observed interfacial waves (Figure 1) violate this assumption and also the requirement that their wavelength is long compared with the layer thicknesses. Finite amplitude is associated with a tendency for waves to steepen, much as in the development of a tidal bore at the sea surface. On the other hand, a horizontal scale that is not very long, compared with at least the thickness of the thinner layer, leads to dispersion, the break-up of a disturbance into waves of different wavelengths traveling at different speeds. Interestingly, these effects can cancel, leading to the possibility of ‘internal solitary waves’, waves of finite amplitude that can be spatially localized and travel without change of shape. They can occur singly, or in groups as in Figure 1.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL WAVES

267

(A) 0 10 20

40

5 10 20 25 t

50 60

90

100 _ 1000° E (B)

_ 800

126° 2′ 50 42′ 100

0

Hoeya Head

50

150

200

125° 56′

50 40′

_ 600

0

80

< 20 cm/s 20 to 40 cm/s 40 to 60 cm/s 60 to 80 cm/s 80 to 100 cm/s 100 to 120 cm/s 120 + cm/s

400

20

70

10

Depth (m)

30

0

_ 400

5 km

_ 200

0

200° W

Distance (m)

Figure 1 A group of interfacial solitary waves generated by tidal flow over the sill at Knight Inlet, British Columbia. (A) The banded surface manifestation. (B) An echo-sounder image of the same waves. Current vectors are also shown. The location and ship track are shown in the bottom inset. The upper left inset of the density (st) profile shows a strongly stratified thin upper layer above a much thicker more homogeneous layer, rather than an ideal two-layer situation. (From Farmer D and Armi L (1999) The generation and trapping of solitary waves over topography. Science 283: 188–190; courtesy of D. Farmer.)

If in a group, the crests pointing away from the thinner layer are sharper than the troughs. Even if they occur at a density interface many meters, or tens of meters, below the surface, they are often visible if the upper layer is turbid, so that the crests appear from above as more opaque tubes than the surrounding water. More frequently they are seen because the associated currents cause visible variations in surface roughness (Figure 1A). (Whether the water is rougher above the crests or troughs of the interfacial waves depends on the relative directions of propagation of the surface waves and the interfacial waves, as can readily be established by considering the interaction in a frame of reference moving with the interfacial waves.) The generation of these packets of internal solitary waves, or trains of waves with similar properties, often occurs when tidal flow over a sill, or off the edge of the continental shelf, leads to a leeward

depression in the interface (as in Figure 1). As the tidal current reverses, this depression propagates back over the sill, or onto the continental shelf, and breaks up into large amplitude interfacial waves. (In the situation shown in Figure 1, interfacial waves have actually formed before the current reversal.) The internal solitary waves, or solitons, typically have periods of tens of minutes. This would appear to be too short for the Earth’s rotation to be a factor, but it does seem that the break-up of an internal tide into internal solitons may be inhibited by rotational effects. While it is generally only the shape of these interfacial waves that propagates, with little net water movement, they can be sufficiently large that they do, in fact, carry water along with them. Remarkable behavior also occurs as the waves approach shore: although they have sharp downward crests offshore where the lower layer is thicker,

(c) 2011 Elsevier Inc. All Rights Reserved.

268

INTERNAL WAVES

they must switch to having sharp upward crests as the lower layer becomes thinner than the upper layer! They do this only with considerable loss of energy into smaller-amplitude dispersive waves and turbulence. The waves described above have a vertical motion that is maximum at the interface and tends to zero at top and bottom boundaries. The horizontal currents in the two layers are in opposite directions. A sharp interface is not necessary; large-amplitude internal motions can persist even if the density jump is smeared out in the vertical. In the case of smallamplitude waves, the motion can then be thought of as the first vertical mode, made up of propagating waves that reflect off the sea surface and seafloor. We therefore turn next to a discussion of these building blocks.

Internal Waves The Basic Physics

A particle displaced vertically in a continuously stratified fluid experiences a restoring buoyancy force. If a whole vertical fluid column is displaced, the vertical uniformity of the motion means that there is no change in the hydrostatic vertical pressure gradient and the restoring force on each particle is just gravity g times the density perturbation, which is minus the vertical displacement times the vertical density gradient dr/dz. This leads to a simple harmonic oscillator equation for the motion of the column, with the frequency of oscillation given by the ‘buoyancy frequency’ N, where N2 ¼  (g/r)(dr/dz). This frequency is independent of the horizontal scale of the fluid columns, suggesting that the frequency of motions that are wavelike in the horizontal is N, independently of scale. If the fluid columns are now allowed to oscillate obliquely, at an angle y to the vertical, the vertical restoring force is reduced by a factor cos y, as is the component of this force parallel to the motion. The extra factor cos2 y in the oscillator equation then means that the frequency is reduced to N cos y, again independently of the lateral scale. Regarding these motions as waves, it is clear that the motion is transverse, as required for an incompressible fluid with =  u ¼ 0. In a rotating world the motion is acted upon by the Coriolis force, so that fluid oscillations in inclined sheets now develop a transverse motion, within the sheet but orthogonal to the motion with no rotation. The relationship between the frequency o of the oscillations now involves the Earth’s rotational frequency. Provided that N is sufficiently greater than the

Coriolis frequency f, which is twice the vertical component of rotation, the connection between frequency o and orientation y becomes eqn [1]. o2 ¼ N 2 cos2 y þ f 2 sin2 y

½1

Equation [2] is an alternative expression, in terms of the wavenumber k ¼ (k, l, m).
N 2 k2 þ l2 þ f 2 m2 o ¼ k2 þ l2 þ m2 2

½2

While the frequency can be as high as N when the particle motion is vertical, it cannot be lower than the Coriolis frequency f. In this limit, the particle motion is horizontal in ‘inertial’ circles, expressing the tendency for steady rectilinear motion with respect to a nonrotating reference frame. Any frequency of motion between these limiting frequencies is possible, depending on y, or, equivalently, the ratio of vertical to horizontal wavenumber. Moreover, at any frequency, any wavelength is possible. The group velocity (the velocity with which a wave packet, or energy, propagates) is given by (qo/qk, qo/ ql, qo/qm). For internal waves this is easily shown from (2) to be at right angles to the wavenumber vector k. In other words, energy propagates parallel to the wave crests, rather than at right angles as for surface waves! This remarkable feature can be demonstrated in a laboratory experiment (Figure 2). In terms of vertical propagation, waves with downwards phase propagation have upwards energy flux, and vice versa. Observations

Measurements of the frequency spectra of internal waves can be obtained from analysis of time-series of measurements, at a fixed point, by current meters or by temperature sensors that show changes associated with vertical motion of the temperature-stratified water. Such measurements do show a block of energy at frequencies between f and N, falling off above and below these frequencies and with an energy distribution in between that seems close to o2. For currents there is typically an extra peak (an ‘inertial cusp’) near f, but this is suppressed in temperature data as the near-inertial motions are largely horizontal. Measurements at a single fixed point do not, however, provide information on the wavenumber content, or spatial scales, of the energy at any frequency. For this one needs information from many locations (such as from many current meters on a mooring) or the continuous vertical profile obtainable from an acoustic Doppler current profiler

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL WAVES

m

269

=m *

1

cg

cg

10 _2

10

k

k

cg

k

k

cg

Figure 2 Internal waves of a fixed frequency less than N are generated by vertical oscillations of a wavemaker in a tank of stratified fluid. In the photograph the light and dark radial lines are contours of constant density perturbation and so are lines of constant wave phase. The schematic diagram shows the directions of wave phase propagation (k) and group velocity (cg).

(ADCP). Current meter arrays are, of course, limited by the cost and logistics of deploying large numbers of instruments, and moored ADCPs with good resolution have a range much less than the depth of the ocean. Invaluable high-resolution vertical profiles of horizontal currents over the whole ocean depth have been obtained from dropped or lowered profiling current meters that measure the tiny electric potentials generated by movement of conducting sea water in the Earth’s magnetic field, and also by lowered ADCPs. These techniques do not, however, provide much information on the frequency content of the motions. Further information has also come from horizontal tows of sensors, or arrays of sensors, hence mapping horizontal scales though, again, not providing frequency information. Various syntheses of the information from these types of experiments have shown a tendency for energy to be distributed in vertical wavenumber and

f 

10

_1

10

N

m

2

1

Figure 3 The so-called ‘Garrett–Munk’ spectrum, giving the distribution of energy in a space defined by frequency o and vertical wavenumber m. The spectrum has a peak at the inertial frequency f and falls off like o2 at higher frequencies up to N. In vertical wavenumber the spectrum is fairly flat at small wavenumbers (large scale), then falls off rapidly to high wavenumber. The scales are as multiples of f for o and in terms of an equivalent vertical mode number (number of halfwavelengths in the ocean depth) for m.

frequency somewhat as shown in Figure 3: the frequency dependence is roughly like o2, and at each frequency there is a tendency for there to be more energy at small vertical wavenumbers (large scales), with a roll off to high wavenumbers with a power law like m2 or m5/2, though the exponent here is certainly not well-established or universal. While a given frequency and vertical wavenumber magnitude are associated with a given magnitude of horizontal wavenumber via [2], the direction of the wavenumber is not specified, either up versus down or in the 3601 available horizontally. It is generally assumed that the energy is horizontally ‘isotropic,’ or distributed evenly among all possible directions of propagation. It is also assumed that there is as much energy propagating up as down, though there is evidence for preferential downward transmission for waves with frequency within about 10% of f at midlatitude. The inertial peak, in fact, deserves special consideration, given its dominance. One interesting aspect is that the current vectors spiral with depth, with a connection between the direction of rotation of the spiral and the direction of energy propagation: in the Northern Hemisphere currents rotate clockwise with time, so that a current vector profile that

(c) 2011 Elsevier Inc. All Rights Reserved.

270

INTERNAL WAVES

shows increasing clockwise rotation with increasing depth below the sea surface must have phase propagation upward and hence group, and energy, propagation downward. The model spectrum shown in Figure 3 is only a very rough approximation. Observed spectra typically have considerable additional energy at tidal frequencies, with this energy also distributed over various vertical scales.

or standing internal waves behind topographic features (as occurs in the atmosphere). These may propagate upward into the ocean. The relative importance of these different mechanisms on a global basis has not been established, but the current view is that wind and tides are the dominant sources and are of comparable importance. Evolution

Generation

It seems that the inertial peak may be generated by fast-moving storms that set up currents in the upper ocean with the corresponding Coriolis forces unmatched by pressure gradients. Near-inertial motions result, with the current vectors rotating with a frequency close to the local f. The large horizontal scale of these motions means, however, that they experience different f at different latitudes. The current vectors at different latitudes then rotate at different rates, increasing the latitudinal gradients in current vectors and decreasing the horizontal scale. The resulting convergence and vertical motion means that the waves can no longer be purely inertial; they retain their frequency but propagate equatorward to a region where f is smaller. At the same time they develop an increasing vertical group velocity and propagate downward into the ocean. The evolution of this important near-inertial part of the internal wave spectrum is also affected by wave interactions with lower-frequency eddies. Higher-frequency waves may be generated by storms that move more slowly, by turbulence in the surface mixed layer, by subtle interactions between surface waves, or as part of the decay process of ocean eddies. They may also arise from interactions between preexisting internal waves, as will be discussed shortly. The tides are another important source of energy for internal waves observed throughout the ocean, as already discussed for interfacial waves near the sea surface. The barotropic, depth-independent, tidal currents associated with tidal changes in sea level move density-stratified water over topographic features on the seafloor, setting up internal oscillations much as if the topographic features were oscillating wavemakers in an otherwise still ocean. These ‘internal tides’ are mainly at the tidal frequencies, though there may also be energy at multiples of these. Lower-frequency currents in the deep ocean are generally much weaker than tidal currents, but in areas, such as the Southern Ocean, where they are significant, they may set up quasi-steady ‘lee waves,’

A number of processes can contribute to the filling in of the continuous spectrum typically observed. One seems to be resonant wave–wave interactions: the nonlinear terms, involving u  =, in the governing fluid dynamical equation vanish identically for a single wave, but produce interaction terms if two waves are present. These terms, in the momentum and density equations, may be regarded as forcing terms with frequencies and wavenumbers given by the sum and difference frequencies and wavenumbers. For some pairs the sum (or difference) frequency is exactly what would be expected for a free wave with the sum (or difference) wavenumber, so this wave is now resonantly excited, acquiring energy from the original two waves. Detailed calculations for this theory, and using a different approach when the assumptions of weak interaction break down, do not actually make it clear how a typical spectrum arises, but suggest that, once it is present, there is a cascade of energy to waves with shorter vertical scales. As will be discussed later, these shorter waves are more likely to become unstable, break down into turbulence, and cause mixing. The direction in which energy flows in frequency is less clear, though one interaction mechanism, akin to the excitation of a simple pendulum by oscillation of its point of support with twice the natural frequency of the pendulum, can produce small-scale waves with half the frequency of a large-scale parent wave (provided that this half-frequency is still greater than f). Bottom Reflection and Scattering

Internal waves have a frequency less than the local value of the buoyancy frequency N. This typically decreases with increasing depth below the sea surface, so that some downward-propagating waves must undergo internal reflection at a level where their frequency matches the local N. Waves with a frequency less than N at the seafloor (or just above some well-mixed bottom boundary layer) will be scattered and reflected there. The reflection process is unlike that for, say, sound waves, in that for internal waves to conserve their

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL WAVES

cgr ki

cgi

i

r kr

 Figure 4 Internal wave rays are reflected at the seafloor at an equal angle to the normal to a bottom slope. Subscripts i and r indicate incident and reflected waves.

frequency on reflection, they must, by eqn [1], conserve their angle to the vertical, not their angle to the normal. As a consequence, waves reflected upslope will have a shorter wavelength and narrower ray tube (Figure 4). The latter effect, combined with a reduction in group velocity, causes an increase in wave amplitude, particularly near the ‘critical frequency’ for which the wave rays are parallel to the slope. The waves may be amplified less, or even reduced, for other azimuthal angles of incidence, but it turns out that overall amplification is expected for an isotropic incident spectrum, as has been well documented for internal waves near the steeply sloping sides of Fieberling Guyot in the Pacific Ocean. This analysis certainly applies if the length scale of the slope is large compared with the wavelength. For smaller-scale topography, some energy may be backscattered without as much amplification, but, in general, internal wave interaction with bottom topography tends to redistribute energy toward shorter wavelengths. This may be just as important as wave–wave interactions in shaping the wavenumber part of the internal wave spectrum, though bottom interactions do not affect the frequency distribution. Energetics and Mixing

The general picture that has emerged for internal waves in the ocean is that there is a cascade of energy to smaller scales as a consequence of wave–wave interactions and bottom scattering. This leads to a tendency for shear instability of the horizontal currents, with an expected vertical scale of the order of 1 m for a typical spectrum. It is generally assumed that a fraction of about 15–20% of the energy lost in the breaking leads to an increase in the potential energy of the water column (with the rest of the energy being dissipated and ultimately appearing as a negligible internal heating rate). The associated vertical mixing rate, or ‘eddy diffusivity,’ is of the order

271

of 105 m2 s1, again for typical spectral energy levels in the main thermocline. This agrees rather well with estimates based on measurements of turbulent microstructure, and with even more direct estimates based on observations of the vertical spread of an artificial tracer. The mixing may be considerably more intense throughout the water column in regions, such as the Southern Ocean, where internal wave energy levels seem higher, perhaps as a consequence of seafloor generation of the waves by strong mean currents over rough topography. Stronger mixing is also observed in general within a few hundred meters of the bottom in areas of rough bottom topography, though it is not clear whether this results directly from the increased shear of the reflected and scattered waves, or via stronger wave–wave interactions at increased internal wave energies. The relative importance of wind-generated internal waves and internal tides in these regions is also still unsettled, though a topic of active research. Rather weak mixing in the main thermocline of the ocean, together with much weaker stratification in abyssal areas of strong mixing, means that the overall energy loss from the internal wave field is small enough that it would take many tens of days to drain the observed energy levels. This may tie in with the remarkable feature, mentioned earlier, that observed internal wave energy levels in the ocean seem to be rather uniform in space and time, at least much more so than for surface gravity waves; there is no such thing as an ‘internal calm.’ The interpretation is that the decay time of internal waves is, unlike the situation for surface waves, considerably longer than the interval between generation events. There is still some seasonal modulation of the internal wave energy levels, but less than that in the wind and also in accord with a decay time of tens of days. There are exceptions to this picture, of course, with, for example, much lower internal wave energy levels and mixing in the Arctic Ocean (except near some topographic features), perhaps as a consequence of less wind generation, because of the protective ice cover, as well as rather weak tidal currents. The overall dynamical balance of the internal wave field in the ocean is qualitatively summarized in Figure 5. Internal Waves on the Continental Shelf

The above discussion of internal waves has been focused on the deep-sea situation. There are some similarities on the much shallower continental

(c) 2011 Elsevier Inc. All Rights Reserved.

272

INTERNAL WAVES

Wind

Inertial oscillations

Surface waves

Mixed layer turbulence

Largescale flows

Topography

Continuum

f

Interior mixing

Barotropic tides

M2

Boundary mixing

Figure 5 A summary of internal wave generation, evolution, and eventual dissipation causing mixing in the ocean interior or near boundaries. The dashed lines indicate conjectured energy pathways. (From Mu¨ller and Briscoe (2000)).

shelves, though with a considerable fraction of observed internal wave energy often being associated with the rather large interfacial waves, or their equivalent in a smoothly stratified fluid, discussed at the beginning of this article. It is not clear how much of the rest of the internal wave field on shelves is locally generated and how much propagates there from the deep ocean.

making the shear across the mixed layer base more destablizing during the shallow phase and enhancing the overall mixing. This is an effect that has been excluded from models of the surface layer, and may be a partial reason why these models sometimes need to include ad hoc extra mixing just below the base of the layer.

Conclusions Other Aspects In the atmosphere, internal waves are crucial in establishing the general circulation by transporting momentum from one location to another and then depositing it when they break. One reason for this breaking is that as internal waves propagate vertically into thinner air they must increase their amplitudes in order to conserve their energy flux, and so become more prone to instability. This is not a factor in the oceans, where the density change is very minor. The ratio of mean flow speeds to wave speeds is also less in the ocean than in the atmosphere, making interactions between waves and currents less important in the ocean. None the less, it does seem likely that there are some locations in the ocean where internal wave breaking should drive mean flows. One possible location is the continental slope; internal waves generated as lee waves at one location may propagate shoreward, break on the slope, and drive an along-slope current, much as ocean swell incident at an angle to a beach may drive longshore currents inside the breaker zone. The role of internal waves in other situations may also have been somewhat unrecognized so far. One is their effect on surface mixed layer deepening. The waves alternately shallow and deepen the layer,

Internal waves are both an unavoidable nuisance and a key ingredient of the behavior of the ocean; perhaps they are like the clouds in the atmosphere. The analogy is certainly a good one when one thinks of modeling the large-scale circulation of the two media for applications such as climate prediction. Numerical models fail by several orders of magnitude to have sufficient resolution to treat them explicitly, so their effects must be parameterized. This requires not just an understanding of their role in present conditions, but also a submodel that will predict their characteristics and effects in a changing mean state. A model for internal waves will need to account for the whole awkward mix of generation, propagation, wave–wave interactions, interactions with the mean state, and reflection and scattering from the rough seafloor. We have a partial understanding of many of the pieces but are a long way from putting them all together.

See also Breaking Waves and Near-Surface Turbulence. Internal Tidal Mixing. Internal Tides. Surface Gravity and Capillary Waves. Wave Generation by Wind.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNAL WAVES

Further Reading Garrett C and Munk WH (1979) Internal waves in the ocean. Annual Review of Fluid Mechanics 11: 339--369. Gill AE (1982) Atmosphere–Ocean Dynamics. New York: Academic Press. Kundu PK (1990) Fluid Dynamics. New York: Academic Press.

273

Mu¨ller P and Briscoe M (2000) Diapycnal mixing and internal waves. Oceanography 13: 98--103. Munk WH (1981) Internal waves and small-scale processes. In: Warren BA and Wansch C (eds.) Evolution of Physical Oceanography, pp. 264--291. Princeton: MIT Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNATIONAL ORGANIZATIONS M. R. Reeve, National Science Foundation, Arlington VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1342–1348, & 2001, Elsevier Ltd.

Introduction This article is limited to a description of the most important organizations related to Ocean Sciences, rather than exhaustively cataloging perhaps hundreds of entities which may cross one or more national borders. I have classified the major international organizations into governmental (or more accurately intergovernmental) and nongovernmental organizations, and within those groupings, global and regional organizations. With the proliferation of the world wide web, much information can be readily accessed from each organization’s web page, some of which I have edited and utilized.

Intergovernmental Organizations (Global) The overarching global intergovernmental organization is the United Nations (UN). Within it, the Intergovernmental Oceanographic Commission (IOC) has the main responsibility for the coastal and deep oceans. Organizationally, the IOC is a component commission of the United Nations Educational, Scientific and Cultural Organization (UNESCO). The World Meteorological Organization (WMO), a UN specialized agency, has strong interactions with the ocean interests, by virtue of the coupling of weather and climate with the circulation of the oceans and its other properties. The Fisheries and Agricultural Organization of the UN also has strong ties to the oceans through its work in marine fisheries.

Intergovernmental Oceanographic Commission (IOC) The work of the IOC, founded in 1960, has focused on promoting marine scientific investigations and related ocean services, with a view to learning more about the nature and resources of the oceans. The IOC focuses on four major themes: (1) facilitation of international oceanographic research programs; (2) establishment and coordination of an operational global ocean observing system; (3) education and

274

training programs and technical assistance; and (4) ensuring that ocean data and information are made widely available. The IOC is currently composed of 126 Member States, an Assembly, an Executive Council and a Secretariat. The Secretariat is based in Paris, France. Additionally the IOC has a number of Subsidiary Bodies. Each Member State has one seat in the Assembly, which meets once every two years. The Assembly is the principal organ of the Commission which makes all decisions to accomplish the objectives of the IOC. The Secretariat is the executive arm of the organization. It is headed by an Executive Secretary who is elected by the Assembly and appointed by the Director-General of UNESCO. Countries contribute dues to support the work of the IOC. Scientific/technical subsidiary bodies of IOC include Ocean Science In Relation To Living Resources, Ocean Science In Relation to Non-Living Resources, Ocean Mapping (OM), Marine Pollution Research and Monitoring, Integrated Global Ocean Services System, Global Ocean Observing System, and International Oceanographic Data And Information Exchange. There are various subprograms attached to most of these including the Global Coral Reef Monitoring Network. IOC also has regional subsidiary bodies with responsibilities for carrying out region-specific programs voted by the Assembly. These are the Subcommission for the Caribbean and Adjacent Regions, Regional Committee for the Southern Ocean, Regional Committee for the Western Pacific, Regional Committee for the Cooperative Investigation in the North and Central Western Indian Ocean, Regional Committee for the Central Indian Ocean, Regional Committee for the Central Eastern Atlantic and Regional Committee for the Black Sea. The work of the Secretariat is accomplished by a small permanent staff and a larger number of scientists usually funded by institutions or agencies in their own countries, who have interest in a specific aspect of the activities of the IOC. The funds provided to do this are in addition to country dues. This mechanism to focus staff support on such areas of interest is good, and helps to counterbalance the opposite tendency of the General Assembly, meeting every two years, to direct the General Secretary to undertake new activities. Because such activities must usually be funded out of the existing budget, the tendency inevitably dilutes existing activities, sometimes to a subcritical level.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNATIONAL ORGANIZATIONS

World Meteorological Organization (WMO) The World Climate Program (WCP) is the main activity of WMO related to the oceans. Established in 1979, the WCP includes the World Climate Research Program (WCRP), of which the Global Climate Observing System, encompassing all components of the climate system, atmosphere, biosphere, cryosphere, and oceans, is a component. WMO and its WCRP are treated fully elsewhere and only a brief summary is provided here, relating to oceans. The World Meteorological Convention, by which the World Meteorological Organization was created, was adopted in 1947, and in 1951 was established as a specialized agency of the United Nations. There at least 185 member countries and territories. The World Meteorological Congress, which is the supreme body of WMO, meets every four years. In order to assess available information on the science, impacts and the cross-cutting economic and other issues related to climate change, in particular possible global warming induced by human activities, WMO and the United Nations Environment Program (UNEP) established the Intergovernmental Panel on Climate Change in 1988. In the early years of their development, WCRP incorporated the Tropical Ocean – Global Atmosphere Study initiated by the Scientific Committee on Oceanic Research (SCOR) (see below), and the World Ocean Circulation Experiment (WOCE). The former was successfully completed and the latter is in its later stages of analysis and synthesis, the major field programs including the WOCE Hydrographic Survey, having been completed. Both programs have been very successful, to the point that an offspring is beginning to be implemented – the Climate Variability and Predictability Study. This 15-year program to observe the atmosphere and oceans will incorporate new technologies developed as a result of the predecessor programs, and rapidly increasing sophistication of computer hardware and software to develop new coupled models. Past climates will also be reconstructed. Related to this effort is the Global Ocean Data Assimilation Experiment which is planned for the first half of this decade. This program was a product of the joint IOC/WCRP Ocean Observations Panel for Climate. In 1999 the governing bodies of WMO and IOC agreed to set up a Joint Technical Commission for Oceanography and Marine Meteorology to act as a coordinating mechanism for the full range of WMO and IOC existing and future operational marine program activities, including the coordinating and

275

managing the implementation of an operational Global Ocean Observing System.

The World Bank The Global Environmental Facility (GEF) of the World Bank was launched in 1991 as an experimental facility and evolved ‘to serve the environmental interests of people in all parts of the world’. By 2000, more than 36 nations pledged $2.75 billion in support of GEF’s mission to protect the global environment and promote sustainable development. This has been complemented by $5 billion in co-financing from GEF partners, which include the UN Development Program and the UN Environment Program, as well as host countries. GEF funds projects in four areas: biodiversity, climate change, international waters, and ozone. Up to 1999 the GEF had allocated over $155 million to international waters initiatives. The term ‘international’ refers to fresh as well as ocean waters, and the projects seek to reverse the degradation of bodies of water controlled by a mosaic of regional and international water agreements. A list of currently funded projects is available from the GEF web site. Examples include Black Sea Environmental Management, Gulf of Guinea Large Marine Ecosystem and various coral reef rehabilitation and projection projects. The GEF Secretariat is located within the World Bank in Washington DC, USA.

Intergovernmental Organizations (Regional) Besides regional activities of global intergovernmental organizations referred to above, there are two main regional intergovernmental organizations. The first, the International Council for the Exploration of the Sea (ICES) (for the North Atlantic), and the second is the North Pacific Marine Science Organization. International Council for the Exploration of the Sea (ICES)

Oceanographic investigations form an integral part of the ICES program of multidisciplinary work aimed at understanding the features and dynamics of water masses and their ecological processes. In many instances emphasis is placed on the influence of changes in hydrography (e.g., temperature and salinity) and current flow on the distribution, abundance, and population dynamics of finfish and shellfish stocks. These investigations are also relevant

(c) 2011 Elsevier Inc. All Rights Reserved.

276

INTERNATIONAL ORGANIZATIONS

to marine pollution studies because physical oceanographic conditions affect the distribution and transport of contaminants in the marine environment. ICES promotes the development and calibration of oceanographic equipment and the maintenance of appropriate standards of quality and comparability of oceanographic data. ICES is the oldest intergovernmental organization in the world concerned with marine and fisheries science. Since its establishment in Copenhagen in 1902, ICES has been a leading scientific forum for the exchange of information and ideas on the sea and its living resources, and for the promotion and coordination of marine research by scientists within its member countries. Each year, ICES holds more than 100 meetings of its various working groups, study groups, workshops, and committees. These activities culminate each September when ICES holds its Annual Science Conference, which attracts 500–1000 government, academic, and other participants. Proceedings of these meetings, and other related activities are published by ICES. Membership has increased from the original eight countries in 1902 to the present 19 countries which come from both sides of the Atlantic and include Canada, the USA and all European coastal states except the Mediterranean countries from Italy eastward. Each country has a vote in the governance, through two delegates, all of whom come together annually at the ‘statutory meeting’ held at the time of the science conference. The Council elects a President at three year intervals from its members, as well as a small executive group (the Bureau) to conduct business intercessionally. The ICES Secretariat in Denmark maintains three databanks, the Oceanographic databank, the Fisheries databank, and the Environmental (marine contaminants) databank. Since the 1970s, a major area of ICES work as an intergovernmental marine science organization has been to provide information and advice to member country governments and international regulatory commissions (including the European Commission) for the protection of the marine environment and for fisheries conservation. This advice is peer-reviewed by the Advisory Committee on Fishery Management and the Advisory Committee on the Marine Environment before being passed on. The structure and operation of ICES has continuously evolved, to meet current needs of the adviceseeking member nations and the oceanographic community. In earlier years it focused mostly on the relationship of oceanography to fisheries, but over the past 15 years, the need has arisen increasingly for advice over the broad range of environmental issues

from marine contaminants to effects of fishing activities on the environment, particularly the seafloor ecosystems. The annual meeting, which was traditionally a business occasion where standing committees reviewed the activities of working groups and passed recommendations on to the Council, has evolved into a major north Atlantic science meeting on oceanography and its application to regional societal problems. ICES has tried to capture its past century of achievements through special lectures, a History Symposium and a soon to be published written history volume. Its future is mapped through the development of its first strategic plan, and ongoing organizational evolution. North Pacific Marine Science Organization (PICES)

PICES held its first Annual Meeting in October 1992, in Victoria, British Columbia. From the beginning, the PICES approach has been multidisciplinary, with standing committees concerned with biological oceanography, fishery science, physical oceanography and climate, and marine environmental quality. There has been growing interaction among these specialties, with joint scientific sessions, interdisciplinary symposia, and a broad study of climate change and carrying capacity (the CCCC program) in the region. Most recently, PICES has taken the lead in joining forces with other international organizations to organize an intersessional Conference under the title of El Nin˜o and Beyond (March 2000). Although PICES is an infant compared with its prototype, ICES, it has already become a major focus for international cooperation in marine science in the northern North Pacific. PICES is an intergovernmental scientific organization. Its present members are Canada, People’s Republic of China, Japan, Republic of Korea, Russian Federation, and the USA. The purposes of the organization are: to promote and coordinate marine research in the northern North Pacific and adjacent seas especially northward of 301N; advance scientific knowledge about the ocean environment, global weather and climate change, living resources and their ecosystems, and the impacts of human activities; and promote the collection and rapid exchange of scientific information on these issues. PICES annual meetings, symposia and workshops provide fora at which marine scientists interested in the North Pacific can exchange latest results, data, and ideas and plan joint research. These meetings have been effective in stimulating and accumulating the interest of the scientific community in member countries to coordinate marine science on the basin

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNATIONAL ORGANIZATIONS

scale. PICES has been successful at bringing oceanographers from a number of Pacific Rim countries and disciplines (physical, chemical, and biological) together talking and working with fisheries scientists to understand and eventually forecasts temporal variability of ocean ecosystems, and their key species. It has also encouraged and facilitated collaborative marine scientific research in the North Pacific. International collaboration in this region is essential since the open North Pacific is too large for any one country to study adequately on its own, and oceanic circulation and biological species do not recognize international boundaries. Comparisons of similar species and ocean environments on the eastern and western sides of the Pacific will be much more revealing about the large-scale processes affecting ecosystem and fish dynamics than isolated studies in national waters alone. PICES was established, in large measure, by the tireless efforts of Warren Wooster, of the University of Washington, USA. He had long been active in ICES, including as its President, and had become convinced that there was a great need for a similar organization focused on the North Pacific. Other Regional Commissions

Throughout the world, governments have formed regional organizations to protect the environment and regulate activities. Two European examples of these are the Helsinki Commission, otherwise known as the Baltic Marine Environment Protection Commission, and OSPAR Commission for the Protection of the Marine environment of the north-east Atlantic. Their members comprise the countries bordering these specific marine environments. There are also several such commissions organized to protect and regulate specific fisheries and/or fishery regions (e.g., the North Atlantic Salmon Commission), listings and explanations of which go beyond the scope of this article.

Nongovernmental Organizations (Global) International Council for Science (ICSU)

Formerly known as the International Council of Scientific Unions, ICSU is a nongovernmental organization, founded in 1931 to bring together natural scientists in international scientific endeavor. It comprises 95 multidisciplinary National Scientific Members (scientific research councils or science academies) and 25 international, single-discipline Scientific Unions to provide a wide spectrum of

277

scientific expertise enabling members to address major international, interdisciplinary issues which none could handle alone. ICSU also has 28 Scientific Associates. The Council seeks to break the barriers of specialization by initiating and coordinating major international interdisciplinary programs and by creating interdisciplinary bodies which undertake activities and research programs of interest to several members. It acts as a focus for the exchange of ideas and information and the development of standards. Hundreds of congresses, symposia and other scientific meetings are organized each year around the world, and a wide range of newsletters, handbooks, and journals is published. The principal source of finance for the ICSU is the contributions it receives from its members. Other sources of income are grants and contracts from UN bodies, foundations, and agencies, which are used solely to support the scientific activities of the ICSU Unions and interdisciplinary bodies. ICSU has a three-tier system of governance. They are the General Assembly (the highest organ), the Executive Board, and the Officers. These are assisted by a Secretariat responsible for the day-to-day work of the Council. Interdisciplinary ICSU bodies are created by the General Assembly as the need for these arises in cooperative projects. Two of these, the International Geosphere-Biosphere Program (IGBP) and SCOR, are currently of particular interest to the ocean sciences community. IGBP, planning for the International GeosphereBiosphere Program: A Study of Global Change, was begun in 1986. The IGBP is a research program with the objective to describe and understand the interactive physical, chemical, and biological processes that regulate the total Earth system, the unique environment that it provides for life, the changes that are occurring in this system, and the manner in which they are influenced by human actions. The program is focused on acquiring basic scientific knowledge about the interactive processes of biology and chemistry of the earth as they relate to global change. Priority is placed on those areas in each of the fields involved that deal with key interactions and significant changes on timescales of decades to centuries, that most affect the biosphere, that are most susceptible to human perturbations, and that will most likely lead to a practical, predictive capability. SCOR, the Scientific Committee on Oceanic Research, established in 1957, is the oldest of ICSU’s interdisciplinary bodies. The recognition that the scientific problems of the oceans required a truly interdisciplinary approach was embodied in plans for

(c) 2011 Elsevier Inc. All Rights Reserved.

278

INTERNATIONAL ORGANIZATIONS

the International Geophysical Year. Accordingly, SCOR’s first major effort was to plan a coordinated, international attack on the least-studied ocean basin of all, the Indian Ocean. The International Indian Ocean Experiment of the early 1960s was the result. For the next 30 years, the reputation of SCOR was largely based on the successes of its scientific working groups. These small international groups of not more than ten members are established in response to proposals from national committees for SCOR, other scientific organizations, or previous working groups. In general, they are designed to address fairly narrowly defined topics (often new, ‘hot’ topics in the field) which can benefit from international attention. Although SCOR does not have the resources to fund research directly, many of its scientific groups have organized international meetings and produced important publications in the scientific literature. Others have proposed and planned large international collaborative efforts such as the Joint Global Ocean Flux Study and Global Ocean Ecosystem Dynamics. Scientists from the 39 SCOR member countries participate in its working groups and steering committees for the larger programs. SCOR often works in association with intergovernmental organizations such as IOC and ICES. The work of SCOR falls into two major categories. The first of these is the traditional mechanism of the SCOR working group, addressing topics which range from the ecology of sea ice to the role of wave breaking on upper ocean dynamics and from coastal modeling to the biogeochemistry of iron in sea water and the responses of coral reefs to global change. For longerterm, complex activities, such as the planning and implementation of large-scale programs SCOR establishes scientific committees. Over the past 15 years there has been a growing awareness of the influence of the ocean in large-scale climate patterns and in moderating global change. By the early 1980s the promise of increased computing capabilities and new satellite instruments for remote sensing of the global ocean permitted oceanographers to conceive of large-scale, internationally planned and implemented experiments of the sort never before possible. The first two of these were the World Ocean Circulation Experiment (WOCE) and the Tropical Ocean–Global Atmosphere Study (TOGA, 1985– 1995). Both grew out of SCOR’s former Committee on Climatic Changes and the Ocean which was also cosponsored by the IOC. A few years ago WOCE was incorporated into the World Climate Research program; its field program is now complete and WOCE is now embarking upon the critical phase of analysis, interpretation, modeling and synthesis.

Since the late 1980s SCOR has played a major role in fostering the development of two newer global change programs, both of which now form part of the IGBP effort. These are the Joint Global Ocean Flux Study and Global Ocean Ecosystem Dynamics. SCOR consists of its ‘members’ – the national committees for oceanic research of its 39 member countries, each of which is represented by three individual oceanographers. The biennial general meetings elect an executive committee, which also includes ex officio members from allied disciplinary organizations, namely, the International Association for Physical Sciences of the Ocean, the International Association for Biological Oceanography, and the International Association for Meteorological and Atmospheric Sciences. The Ocean Drilling Program (ODP)

The ODP is an international partnership of scientists and research institutions organized to explore the evolution and structure of the earth. It uses drilling and data from drill holes to improve fundamental understanding of the role of physical, chemical, and biological processes in the geological history, structure, and evolution of the oceanic portion of the earth’s crust. The ODP provides researchers around the world access to a vast repository of geological and environmental information recorded far below the ocean surface in seafloor sediments and rocks. The National Science Foundation (US Federal Government) supports approximately 60% of the total international effort. Other partners (Germany, France, Japan, the UK, the Australia/Canada/Chinese Taipei/Korea consortium, European Science Foundation consortium, and the People’s Republic of China), comprising 20 other countries, provide 40% of the program costs. Currently, specific studies include documenting the history of volcanic plumes in the western Pacific, examining the formation of mineral deposits near west Pacific island arcs, instrumentation of boreholes to study seismicity of the north-west Pacific, and recovery of gas hydrate deposits to examine their formation along the Oregon margin. Support will also be provided for new scientific and operational developments to extend capabilities for deep biosphere investigations for ocean biocomplexity studies. Other Nongovernmental Organizations (Global and Regional)

Within this category there are many scientific unions and societies, as well as other organizations, which are either explicitly ocean-oriented or have ocean

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERNATIONAL ORGANIZATIONS

components. Several, as noted above are related to ICSU. Others include the American Society of Limnology and Oceanography and the American Geophysical Union (both primarily north American, but with a substantial international membership), the European Geophysical Society, the Oceanography and Challenger Societies, and many others.

•

Acknowledgements

•

I wish to gratefully acknowledge the assistance of my colleague Kandace Binkley in the preparation of this article.

•

Memorial

•

I dedicate this article as a small memorial for my friend and colleague Dr. George Grice, whose untimely death occurred in March 2001. A former Associate Director of the Wood’s Hole Oceanographic Institution and Deputy Director of the Northeast Fisheries Science Center in Wood’s Hole, George was involved with international organizations all his professional life, particularly ICES and IOC. He was serving the latter institution at the time of his death as Senior Science Advisor to the Executive Secretary, Dr. Patricio Bernal.

• • •

•

• • • • • •

Links to International Organizations

• • • • • •

American Society of Limnology and Oceanography (ASLO) – http://aslo.org American Geophysical Union (AGU) – http:// www.agu.org Challenger Society for Marine Science – http:// www.soc.soton.ac.uk/OTHERS/CSMS European Geophysical Society (EGS) – http:// www.mpae.gwdg.de/EGS/EGS.html Global Ocean Ecosystem Dynamics (GLOBEC) – http://www1.npm.ac.uk/globec International Association for Meteorology and Atmospheric Sciences (IAMAS) – http:// iamas.org

• • • • • •

279

International Association for the Physical Sciences of the Oceans (IAPSO) – http://www. olympus.net/IAPSO International Council for the Exploration of the Sea (ICES) – http://www.ices.dk International Council for Science (ICSU) – http:// www.icsu.org International Geosphere-Biosphere Program (IGBP) – http://www.igbp.kva.se Intergovernmental Oceanographic Commission (IOC) – http://ioc.unesco.org/iocweb Joint Global Ocean Flux (JGOFS) – http:// ads.smr.uib.no/jgofs/jgofs.htm Baltic Marine Environment Protection Commission (HELCOM) – http://www.helcom.fi/ oldhc.html North Pacific Marine Science Organization (PICES) – http://pices.ios.bc.ca Ocean Drilling Program (ODP) – http://wwwodp.tamu.edu OSPAR Commission for the Protection of the Marine Environment of the North-East Atlantic – http://www.ospar.org Scientific Committee on Antarctic Research (SCAR) – http://www.scar.org Scientific Committee on Oceanic Research (SCOR) – http://www.jhu.edu/Bscor The Oceanography Society (TOS) - http://tos.org Tropical Ocean – Global Atmosphere Coupled Ocean/Atmosphere Response Experiment (TOGA) – http://trmm.gsfc.nasa.gov/trmm_office/field_campaigns/toga_coare/toga_coare.html United Nations (UN) – http://www.un.org United Nations Environment Program (UNEP) – http://www.unep.ch/index.html United Nations Educational, Scientific and Cultural Organization (UNESCO) – http:// www.unesco.org World Bank, Global Environmental Facility (GEF) – http://www.gefweb.org World Meteorological Organization (WMO) – http://www.wmo.ch World Ocean Circulation Experiment (WOCE) – http://www.soc.soton.ac.uk/OTHERS/woceipo/ ipo.html

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERTIDAL FISHES R. N. Gibson, Scottish Association for Marine Science, Argyll, Scotland Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1348–1354, & 2001 Elsevier Ltd.

Introduction and Classification of Intertidal Fishes The intertidal zone is the most temporally and spatially variable of all marine habitats. It ranges from sand and mud flats to rocky reefs and allows the development of a wide variety of plant and animal communities. The members of these communities are subject to the many and frequent changes imposed by wave action and the ebb and flow of the tide. Consequently, animals living permanently in the intertidal zone have evolved a variety of anatomical, physiological and behavioral adaptations that enable them to survive in this challenging habitat. The greater motility of fishes compared with most other intertidal animals allows them greater flexibility in combating these stresses and they adopt one of two basic strategies. The first is to remain in the zone at low tide. This strategy used by the ‘residents’, requires the availability of some form of shelter to alleviate the dangers of exposure to air and to predators. ‘Visitors’ or ‘transients’, that is species not adapted to cope with large changes in environmental conditions, only enter the intertidal zone when it is submerged and leave as the tide ebbs. The extent to which particular species employ either of these strategies varies widely. Many species found in the intertidal zone spend most of their lives there and are integral parts of the intertidal ecosystem. At the other extreme, others simply use the intertidal zone at high tide as an extension of their normal subtidal living space. In between these extremes are species that spend seasons of the year or parts of their life history in the intertidal zone and use it principally as a nursery or spawning ground. The different behavior patterns used by residents and visitors mean that few fishes are accidentally stranded by the outgoing tide.

Habitats, Abundance, and Systematics Fishes can be found in almost all intertidal habitats and in all nonpolar regions. Most shelter is found on rocky shores in the form of weed, pools, crevices and

280

spaces beneath boulders and it is on rocky shores that resident intertidal fishes are usually most numerous. If, however, fishes are capable of constructing their own shelter in the form of burrows in the sediment, as in the tropical mudskippers (Gobiidae), they may be abundant in such habitats. Visiting species may also be extremely numerous on occasions and are particularly common in habitats such as sandy beaches, mudflats and saltmarshes. Few fishes occupy gravel beaches although species like the Pacific herring (Clupea pallasii), capelin (Mallotus villosus, Osmeridae) and some pufferfishes (Tetraodontidae) may spawn on such beaches. Estimating abundance in terms of numbers per unit area can be difficult because of the cryptic nature of the fishes and the patchiness of the habitat. The difficulty is particularly acute on rocky shores where fishes may be highly concentrated in areas such as rock pools but absent elsewhere. Nevertheless, fish densities can be relatively high, particularly at the time of recruitment from the plankton, and on occasions may exceed 10 individuals per m2. Over 700 species of fishes from 110 families have so far been recorded in the intertidal zone worldwide. This figure represents less than 3% of known fish species but is certainly an underestimate because it is based only on species recorded on rocky shores. Species on soft sediment shores are not included and many areas of the world have yet to be studied in detail. The final count is therefore likely to be much higher. Intertidal fish faunas are frequently dominated by members of a few families (Table 1). Generally speaking, more species are found in the tropics than in temperate zones and each area of the world tends to have its own characteristic fauna. The Atlantic coast of South Africa, for example, is characterized by large numbers of clinid species, New Zealand by triplefins, the northeast Pacific by sculpins and pricklebacks (Stichaeidae) and many other areas by blennies, gobies and clingfishes.

Characteristics of Intertidal Fishes as Adaptations to Intertidal Life Resident intertidal fishes are probably descended from subtidal ancestors and have few, if any, characters that are truly unique. They are thus representatives of families that have convergently evolved morphological, behavioral and physiological traits that enable them to survive in shallow turbulent

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERTIDAL FISHES

281

Table 1 Analysis of 47 worldwide collections of rocky shore intertidal fishes to show the 10 families with the largest number of species. Based on Prochazka et al. in Horn et al. (1999). Note that abundance of species does not necessarily imply that families are also numerically abundant

mudskippers of the genus Periophthalmus and blennies of the genera Alticus and Coryphoblennius.

Family

Common name

Number of species

Blenniidae Gobiidae Labridae Clinidae Pomacentridae Tripterygiidae Cottidae Labrisomidae Scorpaenidae Gobiesocidae

Blennies and rockskippers Gobies and mudskippers Wrasses Clinids, kelpfishes, klipfishes Damselfishes Triplefin blennies Sculpins Labrisomids Scorpionfishes Clingfishes

55 54 44 33 30 30 26 26 25 24

All the common families of intertidal fishes possess the characteristic morphological features of fishes adapted for benthic life in turbulent waters. They are cryptically colored, are rarely more than 15 cm long and are negatively buoyant because they lack a swimbladder or possess one that is reduced in volume. Four basic body shapes can be recognized: elongate, dorsoventrally flattened (depressed), smoothly cylindrical (terete), or laterally compressed (Figure 1). In many species the fins are modified to act as attachment devices to prevent dislodgement by turbulence or to assist movement over rough surfaces. In most blennies, the rays of the paired and ventral fins are hooked at their distal ends and may be covered with a thick cuticle to minimize wear. In the gobies and clingfishes, the pelvic fins are fused to form suction cups and allow the fish to attach themselves firmly to the substratum. There seem to be few ‘typical’ sensory adaptations to intertidal life although many species have reduced olfactory and lateral line systems. These sensory systems would be of limited use for species living in turbulent waters or in those that frequently emerge from the water.

habitats. The distribution of many resident intertidal species also extends below low water mark but they mainly differ from their fully subtidal relatives in the degree to which they can withstand exposure to air and are capable of terrestrial locomotion. Nevertheless, a few species can be considered truly intertidal in their distribution because they never occur below low water mark. Examples are the

Morphology

Behavior

(A)

(B)

(C)

(D)

Figure 1 Sketches of four intertidal fish species to demonstrate the basic body shapes. (A) Terete (Gobius paganellus, Gobiidae); (B) dorsoventrally flattened (Lepadogaster lepadogaster, Gobiesocidae); (C) elongate (Pholis gunnellus, Pholidae); (D) laterally compressed (Symphodus melops, Labridae).

Intertidal fishes also show characteristic behavior that enables them to cope with the rigors of intertidal life. Their modified fins and relatively high density allow them to remain on or close to the bottom with the minimum of effort and to resist displacement by surge. Most are also thigmotactic, a behavior that keeps as much of their body touching the substratum as possible and ensures that they come to rest in contact with solid objects when inactive. Their mode of locomotion also reflects this bottom-dwelling lifestyle. Few excursions are made into open water and those species with large pectoral fins use them as much as the tail for forward movement and swim in a series of short hops. Clingfishes and gobies can progress slowly over horizontal and vertical surfaces using their sucker. Elongate forms, which usually have reduced paired fins, creep along the bottom using sinuous movements of their body or alternate lateral flexions of the tail. Strong lateral flips of the tail are also used by some blennies and gobies that can jump between rock pools at low tide and by the semiterrestrial mudskippers in their characteristic ‘skipping’ movements over the surface of the mud. When progressing more slowly mudskippers use the

(c) 2011 Elsevier Inc. All Rights Reserved.

282

INTERTIDAL FISHES

muscular pectoral fins as ‘crutches’ to move over the substratum. Resident species are generally active at a particular state of the tide, although these tidally phased movements can be modulated by the day/night cycle so that some species may only be active, for example, on high tides that occur during the night. Visitors present a more complicated picture because, although their movements are also basically of tidal frequency, they are modulated by a wider range of cycles of lower frequency. Individuals may migrate intertidally on each tide, on every other tide depending on whether they are diurnal or nocturnal, or only on day or night spring tides. They may enter the zone as juveniles in spring or summer, stay there for several months, during which time they are tidally active, and then leave when conditions become unsuitable in winter or as they grow and mature. The distances over which fishes move in the course of their intertidal movements are dependent on several factors. At one end of the scale are relatively gradual but ultimately extensive shifts in position related to the seasons. At the other end are local, short-term, tidally related foraging excursions. Residents tend to be very restricted in their movements and are often territorial, whereas visitors regularly enter and leave the intertidal zone and may cover considerable distances in each tide. Body size also determines the scale of movement. Residents are small and have limited powers of locomotion whereas visitors usually possess good locomotory abilities and can travel greater distances more rapidly. The small size and poor swimming abilities of resident species partially account for the restricted extent of their movements but there is good evidence that some species also possess good homing abilities. Most evidence comes from experiments in which individuals are experimentally displaced short distances from their ‘home’ pools and subsequently reappear in these pools a short time later. Experiments with the goby Bathygobius soporator suggest that this species acquires a knowledge of its surroundings by swimming over them at high tide. It can remember this knowledge for several weeks and use its knowledge to return to its pool of origin. Homing is also known in some species of blennies and sculpins and is presumably based on the use of visual clues in the environment. Displacement experiments with the sculpin Oligocottus maculosus, however, suggest that this species at least may also use olfactory clues to find its way back to its home pool. The energy expended in these movements at the various temporal and spatial scales described suggest that they play an important part in the ecology of

both resident and visiting species alike. Several functions have been proposed for these movement patterns of which the most obvious is feeding. Visitors move into the intertidal zone on the rising tide to take advantage of the food resources that are only accessible at high water and move out again as the tide ebbs. Residents, on the other hand, simply move out of their low-tide refuge, forage while the tide is high, and return to the refuge before low tide. Following the flooding tide into the intertidal zone may have the added benefit of providing protection from larger predators in deeper water. Movements at both short and long time scales may also be in response to changing environmental conditions. Visitors avoid being stranded above low water mark at low tide because of the lack of refuges or because they are not adapted for the low tide conditions that may arise in possible refuges such as rock pools. Longer term seasonal movements into deeper water can be viewed as responses to changes in such physical factors as temperature, salinity and turbulence. Finally, several species whose distribution is basically subtidal move into the intertidal zone to spawn (see Life histories and reproduction below). In order to synchronize their behavior with the constantly changing environment fishes must be able to detect and respond to the cues produced by these changes. The cues that fish actually use in timing and directing their tidally synchronized movements are mostly unknown. Synchronization could be achieved by a direct response to change. The flooding of a tide pool or the changing pressure associated with the rising tide, for example, could be used to signal the start of activity. In addition, behavior may be synchronized with the external environment by reference to an internal timing mechanism. The possession of such a ‘biological clock’ that is phased with, but operates independently of, external conditions is a feature common to all intertidal fishes in which it has been investigated. In the laboratory the presence of the ‘clock’ can be demonstrated by recording the activity of fish in the absence of external cues. Under these conditions fish show a rhythm of swimming activity in which periods of activity alternate with periods of rest (Figure 2). In most cases the period of greatest activity appears at the time of predicted high tide on the shore from which the fish originated. These ‘circatidal’ activity rhythms, so called because in constant conditions the period of the rhythm only approximates the average period of the natural tidal cycle (12.4 h), can persist in the laboratory for several days without reinforcement by external cues. After this time activity becomes random but in the blenny Lipophrys pholis the rhythm can be restarted (entrained) by exposing fish to

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERTIDAL FISHES

283

80

Activity per hour

Cebidichthys violaceus (Monkeyface prickleback)

60 Large fish

Periophthalmus cantonensis (Chinese mudskipper)

40 20

Small fish

Tomicodon humeralis (Sonora clingfish)

0 0

12

24

36 Time (h)

48

60

72

Figure 2 The ‘circatidal’ activity pattern shown by the blenny Lipohrys pholis in constant laboratory conditions. Peaks of activity correspond initially to the predicted times of high tide (dotted lines) but gradually occur later because the ‘biological clock’ has a period greater than that of the natural tidal cycle (12.4 h). In the sea the clock would be continually synchronized by local tidal conditions.

experimental cycles of wave action or hydrostatic pressure or by replacing them in the sea. The probable function of this circatidal ‘clock’ is that it enables a fish to anticipate future changes in tidal state and regulate its activity and physiology accordingly. If activity is prevented in the wild, by storms for example, then the persistent nature of the clock allows fish to resume activity that is appropriate to the tidal state at the next opportunity. Physiology

The ebbing tide exposes the intertidal zone to air and so the location of a fish is critical to its survival. Over the low tide period resident fish face not only the danger of exposure to air but also to marked changes in other physical and chemical conditions. On rocky shores, pools act as low-tide refuges but even here temperature, salinity, pH and oxygen content of the water can change markedly for the few hours that the pool is isolated from the sea. Consequently, many resident species are usually more tolerant of changes in these factors than subtidal species. Exposure to air could result in desiccation but, surprisingly, resident intertidal fishes show no major physiological or anatomical adaptations for resisting desiccation. Instead, desiccation is minimized by behavior patterns that ensure fish hide in pools, or in wet areas under stones and clumps of weed at low tide. Nevertheless, many species can survive out of water in moist conditions for many hours and tolerate water losses of more than 20% of their body weight, equivalent to that of some amphibians (Figure 3). The ability to tolerate water loss is generally correlated with position on the shore; those fish that occupy higher

Pherallodiscus funebris (Fraildisc clingfish)

Rana clamitans (Green frog)

Scaphiopus couchi (Spadefoot toad)

0

30 10 20 40 % wet body weight

50

Figure 3 Tolerance of water loss as a percentage of wet body weight in four intertidal fishes compared with two amphibians. (Reproduced with permission from Horn and Gibson, 1988.)

levels are the most resistant. Prolonged emersion could also present fish with problems of nitrogen excretion and osmoregulation but those species that have been investigated seem to be able to cope with any changes in their internal medium caused by the absence of water surrounding them. A further consequence of emersion is the change in the availability of oxygen. Although air contains a greater percentage of oxygen than water its density is much lower causing the gill filaments to collapse and reducing the area of the primary respiratory surface. Unlike some freshwater fishes, intertidal fishes that leave the water or inhabit regions where the water is likely to become hypoxic have no specialized airbreathing organs but maximize aerial gas exchange in other ways. Some species have reduced secondary gill lamellae, thickened gill epithelia and their gills are stiffened with cartilaginous rods. Such features reduce the likelihood of gill collapse when the fish is out of water. The skin is also used as an efficient respiratory surface because it is in contact with air and is close to surface blood vessels. In order to be effective, however, the skin must be kept moist and fish that are active out of water frequently roll on their sides or return to the sea to wet the skin. It is probable that some species use vascularized linings of the mouth, opercular cavities, and, possibly, the esophagus as respiratory surfaces. The ability to respire in air is currently known for at least 60 species from 12 families and many of these voluntarily leave the water. Fish capable of

(c) 2011 Elsevier Inc. All Rights Reserved.

284

INTERTIDAL FISHES

respiring out of water have been classified into three main types. ‘Skippers’ are commonly seen out of water and actively feed, display and defend territories on land. They are typified by the tropical and subtropical mudskippers (Gobiidae) and rockskippers (Blenniidae). The second much less terrestrially active group, the ‘tidepool emergers’, crawl or jump out of tide pools mainly in response to hypoxia. Species from several families show this behavior but it has been best studied in the sculpins. The third group, the ‘remainers’, comprises many species that sit out the low-tide period emersed beneath rocks and in crevices or weed clumps. Some may be guarding egg masses and they are not active out of water unless disturbed.

Feeding Ecology and Predation Impact Intertidal fishes are no more specialized or generalized in their diets and feeding ecology than subtidal fishes. Most are carnivorous or omnivorous and a few are herbivores. Herbivores appear to be less common in higher latitudes but a satisfactory explanation of this phenomenon is still awaited. In some cases the diet changes with size so that the youngest stages are carnivorous but larger amounts of algae are included in the diet as the fish grow. The small size of most resident fishes also means that their diet is composed of small items such as copepods and amphipods. In common with many other small fishes, they may also feed on parts of larger animals such as the cirri of barnacles or the siphons of bivalve molluscs, a form of browsing that does not destroy the prey. The extent to which intertidal fishes have an impact on the abundance of their prey and on the structure of intertidal communities is not clear. In some areas no impact has been detected whereas in others fish may have a marked effect, particularly on the size and species composition of intertidal algae. Those species that only enter the intertidal zone to feed contribute to the export of energy from this area into deeper water.

Life Histories and Reproduction The majority of resident intertidal fishes rarely live longer than two to three years although some temperate gobies and blennies have a maximum life span of up to 10 years. Maturity is achieved in the first or second year of life and the females of longer-lived species may spawn several times a year for each year thereafter. Representatives of 25 teleost families are known to spawn in the intertidal zone. Of these,

residents and visitors make up about equal proportions. Intertidal spawning has both costs and benefits. For resident species, intertidal spawning reduces the likelihood of dispersal of the offspring from the adult habitat. It also obviates the need for movement to distant spawning grounds; a process that would be energetically costly for small-bodied demersal fishes with limited powers of locomotion and would at the same time expose them to greater risks of predation. These advantages do not apply to subtidal species many of which are good swimmers and whose adult habitat is offshore. For these species, the benefits are considered to be reduced egg predation rates and possibly faster development if the eggs become emersed. In both residents and visitors alike eggs spawned intertidally may be subject to the costs of increased mortality caused by desiccation and temperature stress. In addition, visitors may be vulnerable to avian and terrestrial predators during the spawning process. The rugose topography of rocky shores and the cryptic sites chosen for spawning has led to the development of complex mating behavior in many species, particularly the blennies and gobies. In these species, courtship displays by the male include elements of mate attraction and a demonstration of the location of the chosen spawning site. Observing these displays is difficult in most groups but field observations have been made on several Mediterranean blennies that live in holes in rock walls and on mudskippers that perform their mating behavior out of water on the surface of the mud. All species spawn relatively few (range approximately 102–105) large eggs (B 1 mm diameter) that are laid on or buried in the substratum. Large eggs produce large larvae which may reduce dispersal from shallow water by minimizing the amount of time spent in the planktonic stage. On hard substrata the eggs are laid under stones, in holes and crevices and in or under weed. They may be attached individually to the substratum surface in a single layer (blennies, gobies, clingfishes) or in a clump (sculpins). In the gunnels and pricklebacks the eggs adhere to each other in balls but not to the surface. In soft sediments the eggs may be buried by the female as in the grunions (Atherinidae), or laid in burrows as in the mudskippers and some gobies. Killifishes (Fundulus) lay their eggs in salt-marsh vegetation. In temperate latitudes spawning usually takes place in the spring and early summer, but spawning rhythms of shorter frequency may be superimposed on this annual seasonality. Subtidal species such as the grunions, capelin and pufferfishes that use the intertidal zone as a spawning ground mostly take advantage of spring tides to deposit their eggs in the

(c) 2011 Elsevier Inc. All Rights Reserved.

INTERTIDAL FISHES

90

Mean % of catch at each distance

sediment high on the shore, usually at night. During the reproductive season such fish, therefore, spawn at fortnightly intervals at the times of the new and full moons. The larvae develop over the intervening weeks and hatch when the eggs are next immersed. Buried eggs are left unattended but eggs laid in layers, clumps or balls are always cared for by the parent. The sex of the individual that undertakes this parental care varies between species. In some it is the male, in others the female and in yet others both sexes participate. In some members of the families Embiotocidae, Clinidae and Zoarcidae fertilization is internal and the young are produced live. Parental care by oviparous species takes a variety of forms but all have the function of increasing egg survival rates and removal of the guardian parent greatly increases mortality. Most species guard the eggs against predators but some also clean and fan the eggs to maintain a good supply of oxygen and reduce the possibility of attack by pathogens. Development time of the larvae depends on species and temperature but when fully formed the eggs hatch to release free-swimming planktonic larvae. In only one case, the plainfin midshipman (Porichthys notatus, Batrachoididae) does the male parent also care for the larvae, which in this species remain attached to the substratum near the nest site. The factors stimulating hatching are mostly unknown although it has been suggested that wave shock and temperature change associated with the rising tide may be involved. Until recently it was assumed that the hatched larvae were dispersed randomly by currents and turbulence. It has been shown, however, that at least some species minimize this dispersion by forming schools close to the bottom and are rarely found offshore (Figure 4). On completion of the larval phase the larvae metamorphose into the benthic juvenile phase and settle on the bottom. The clues used by settling larvae to select the appropriate substratum are poorly known but there is some evidence to suggest that individuals can discriminate

285

80 70

Intertidal families Offshore families

60 50 40 30 20 10 0 0

4 20 Distance from shoreline (m)

500

Figure 4 Comparison of the distribution of fish larvae from four intertidal and five offshore families caught in plankton nets at four distances from the shoreline in Vancouver harbour, British Columbia. (Based on data in Marliave JB (1986) Transactions of the American Fisheries Society 115: 149–154.)

between substratum types and settle on their preferred type.

See also Fish Ecophysiology. Fish Larvae. Mangroves. Rocky Shores. Salt Marshes and Mud Flats. Sandy Beaches, Biology of.

Further Reading Gibson RN (1996) Intertidal fishes: life in a fluctuating environment. In: Pitcher TJ (ed.) The Behaviour of Teleost Fishes, 2nd edn, pp. 513--586. London: Chapman Hall. Horn MH and Gibson RN (1988) Intertidal fishes. Scientific American 256: 64--70. Horn MH, Martin KLM, and Chotkowski MA (eds.) (1999) Intertidal Fishes: Life in Two Worlds. San Diego: Academic Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRA-AMERICAS SEA latitude, and l ¼ 50–981W longitude. Figure 1 summarizes the geographical setting. Early oceanographic explorations of the region were by European scientists who chose to name the IAS (the Caribbean Sea in particular) the ‘American Mediterranean’. While superficially this terminology describes the IAS as a similar semi-enclosed sea where evaporation (E) exceeds precipitation (P) plus river runoff (R), E > P þ R, the Mediterranean Sea is markedly different in character from its western Atlantic counterpart. Also, the IAS was broken into smaller components, and little attention was paid to the Caribbean Sea and by Gulf of Mexico oceanographers and vice versa. Conversely, the Straits of Florida and the water currents of the Gulf Stream system are perhaps the most widely studied oceanographic features on Earth. With the coming of significant international cooperation between scientists from throughout all the Americas, the IAS began to be appreciated as a

G. A. Maul, Florida Institute of Technology, Melbourne, FL, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1354–1363, & 2001, Elsevier Ltd.

Introduction The Intra-Americas Sea (IAS) is a semi-enclosed saltwater body of the tropical and subtropical western Atlantic Ocean that comprises the Caribbean Sea, the Gulf of Mexico, the Straits of Florida, the Bahamas, the Guianas, and the adjacent waters. Biogeographically, the IAS includes the estuarine, coastal, shelf, and pelagic waters from the mouth of the Amazon River at the equator off Brazil, to Bermuda and westward to the shores of North, Central, and South America. Geographically, the boundaries may be set approximately as f ¼ 01 to 321N

35˚N USA Charleston BERMUDA

Mis

sis

sip pi

30˚N

Houston

Panama City

R.

NORTH ATLANTIC OCEAN Cape Canaveral

Tampa

Corpus Christi

R

ra

nd

Mia

m

GULF OF MEXICO

S

MA

Old ca Yu tan

CUB

ha

ma

Ch

an

ne

l

nn

ge

sa

W

HISPANIOLA

Mo n

CAYMAN SEA

VIRGIN ISLANDS

ILL ANT SER LES

PUERTO RICO

Ane

JAMAICA

age

in

Veracruz

aP ass

O

ard dw

gad

s Pa

el

IC EX

a Ch

M

YUCATAN PENINSULA

Ba

A

ag e

Tampico

20˚N

HA

Str aits

25˚N

BA

J

aP as s

R

of F

e

a

G

lorid

io

15˚N

HONDURAS

ES

CARIBBEAN SEA

BARBADOS

NICARAGUA

N. ANTILLES

PA N

TRINIDAD

AM

na

CA

ele

RI

gd

10˚N

ST A

Ma

CO

R.

A

VENEZUELA

co

R.

ino

G as

an

ui

Or

Colombia

NORTH PACIFIC OCEAN 5˚N 100˚W

95˚W

90˚W

85˚W

80˚W

75˚W

70˚W

65˚W

60˚W

55˚W

Figure 1 The Intra-Americas Sea, a semi-enclosed water body of the subtropical and tropical western North Atlantic Ocean. Place names in accord with the US Board on Geographic Names.

286

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRA-AMERICAS SEA

unified body of water distinctly different from the early and mid-twentieth century European perspective. An inclusive term was needed to integrate not only the oceanography of the region, but its meteorology and maritime socioeconomic connectivity as well. Thus the term Intra-Americas Sea was developed from the multilingual and multicultural heritage shared by its people.

Regional Overview Pre-Columbian indigenous peoples of the IntraAmericas Sea certainly knew of the oceanic currents and atmospheric winds. Caribes from the south, Tainos in the middle Antilles, Arawaks from the north, Mayas and Olmecs to the west, all moved freely from island to island to continent, presumably with some knowledge of the currents and winds we have named Caribbean Current, Trade Winds, Gulf Stream, Hurricane, and Guianas Current. European explorers and conquistadors relearned this information, not from the IAS’s inhabitants, but from hardship after experiencing what was so well known already. James A. Mitchner’s 1989 novel Caribbean imagines so well what science could have learned directly. As regards the geological setting, the IAS encompasses three tectonic plates: the North American Plate, the Caribbean Plate, and the South American Plate (the Cocos and Nazca Plates mark Pacific tectonic boundaries but are not significantly involved in the air–sea regime discussed herein). About 3 Ma the Caribbean Plate drifting from west to east closed the gap between North and South America, creating Panama and deflecting oceanic flow northward. Central America and the eastern Caribbean margin are volcanically active today. Associated with tectonics are earthquakes and seismic sea waves (tsunami) that are part of the circulation regime, as well as the complex bottom topography that channels water movement and perturbs the atmospheric circulation. Geological forces then not only form the coasts and islands of the IAS; they are a central element in appreciating the flow of air and water. Geophysical fluids obey P P Newton’s laws of motion, specifically F ¼ m  a (where F is force, m is mass, and a is acceleration), the laws of thermodynamics, and continuity of mass. The air and water of the IAS are accordingly connected to all of Earth’s ocean–atmosphere continuum, and have special complexities as they flow around and through the passages, channels, and straits within the region. Thus detailed knowledge of the water depths and land heights, and their attendant frictional characteristics, is essential to appreciating the flow patterns discussed below.

287

Sill depths control much of the oceanic flow patterns in the IAS. It has been reported that the deepest sill is in the Yucatan Channel (2040 m) between Mexico and Cuba, that the average sill-depth of the Antillean Arc is 1200 m, but that the Jungfern-Anegada Passage is 1815 m deep (between the Virgin Islands and the Lesser Antilles), and that the Windward Passage separating Cuba and Hispaniola is 1690 m deep. Within the IAS are many much deeper basins than these controlling sills, leading to the notion that the waters deeper than the sill depths are moved by convection rather than advection. This infers that the IAS has a two-flow regime: an upper layer mostly influenced by wind-driven advection, and a deep layer controlled by overflows. The controlling sill depth of the Straits of Florida is about 800 m, which means that the outflow of the IAS is topographically accelerated. Figure 2 summarizes the IAS water depth information. Lastly, to appreciate the ocean currents of the IntraAmericas Sea, the structure of atmospheric forcing needs to be mentioned. The southern portion of the IAS is under the influence of the Inter-Tropical Convergence Zone, the ITCZ. In the Northern Hemisphere summer, the ITCZ migrates northward and the easterly Trade Winds at the southern boundary of the Bermuda meteorological high-pressure zone dominate the IAS to its northern extent. As boreal winter approaches, the ITCZ migrates southward and the midlatitude frontal passages sweep across the IAS to south of Cuba and sometimes almost to South America. During early autumn, the atmosphere is characterized by a series of tropical cyclones, that from time to time reach intensities known as the West Indian Hurricane. These severe atmospheric disturbances are cyclonic circulation features that bring not only strong winds, storm surge, and the associated damage, but also much needed precipitation and flushing of shallow bays and estuaries. Climatologically, the Ko¨ppen classification system would place the northern IAS in a Cfa (humid subtropical) category; coastal Central America, the Greater Antilles and the Bahamas as Aw (tropical wet and dry) with sections as Af (tropical rain forest) including the Lesser Antilles. The southern and southeastern coastal IAS is classified as BSh and BW (semiarid or steppe, and arid desert, respectively), with Am (tropical monsoon) along the coasts of the Guianas and Brazil. These classifications are perturbed by interannual and decadal climate oscillations, in particular by El Nin˜o–Southern Oscillation (ENSO) events, which tend to cause cold wet winters in the northern IAS (particularly Florida and Cuba) and warm dry autumn conditions in Panama, coastal Colombia and Venezuela, the Guianas, and northern Brazil.

(c) 2011 Elsevier Inc. All Rights Reserved.

288

INTRA-AMERICAS SEA

30˚N

25˚N

20˚N

15˚N

10˚N 0m 200 m 1000 m 2000 m 3000 m

4000 m 5000 m 6000 m 7000 m 8000 m

080˚W

070˚W

060˚W

Figure 2 Bottom topography of the Intra-Americas Sea; water depths in meters.

Surface Throughflow Regime Perhaps the best way to envision IAS surface currents is to take a Lagrangian perspective, and imagine floating on a northbound satellite-tracked buoy passing the equator off Brazil as part of the ‘oceanic conveyor belt’. As the water parcel travels northward, perhaps entraining some Amazon River water, the dominant physics is conservation of potential vorticity, d=dt½B þ f =H ¼ 0, where relative vorticity B ¼ @v=@x  @u=@y, the Coriolis parameter f ¼ 2O sin f with latitude f, and water depth H. For a given H, as the parcel flows northward, f increases and z must decrease, forcing an anticyclonic (clockwise) turning. The region where this occurs is called the North Brazil Current ‘retroflection’ and can be seen in satellite images as a distinct offshore turning of the current. Some of this water continues toward the east in the North Atlantic Equatorial CounterCurrent, but some advects up the South American coast in the Guianas Current to the Lesser Antilles Arc, sometimes as an anticyclonic eddy. Much of this is evident in Figure 3. The northward-flowing water parcel usually passes Barbados and often flows through the passages of

the Lesser Antilles, carrying anticyclonic vorticity and perhaps Amazon riverine particles and biota into the Caribbean Sea. Under the influence of the Northeast Trade Winds, the Caribbean Current moves westward at a leisurely pace of perhaps 0.2 m s1, but with notable meandering and eddying along the path. Sometimes under this same wind regime, water from the Orinoco River is seen to be carried completely across the eastern Caribbean Sea to Puerto Rico and Hispaniola, particularly in late summer. Similarly, in the Panama-Colombia Bight, the wind-stress (t) curl, @ty =@x  @tx =@y, causes an rE150 km radius eddy to spin-up and spin-down annually, the so-called Panama-Colombia Gyre (PCG). In the vicinity of the PCG a major South American river, Colombia’s Magdalena, flows into the ocean where its waters and its flotsam mix with the sea, and are carried to distant shores by ocean currents. The PCG is but one feature of the IAS surface current variability now being simulated in numerical models, and being observed by systems such as satellite altimeters and radiometers. As the meandering, eddying, Caribbean Current approaches the Central America coast, it is forced anticyclonically northward into the Yucatan

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRA-AMERICAS SEA

289

Figure 3 Ocean color composite of the Intra-Americas Sea showing concentration of chlorophyll þ phaeophytins and suspended sediments. Image from observations of the Coastal Zone Color Scanner (NASA) compiled by Frank Muller-Karger for October 1979.

Channel. In the area off Belize and Yucatan Mexico, this IAS current takes on the characteristics of the Gulf Stream: a deep (zE1200 m) western boundary current with swift surface flows of more than n ¼ 1 m s1 and a distinct cyclonic horizontal velocity shear boundary, @ n=@x, along the western edge. Here the stream is known as the Yucatan Current, and it is a northward flow connecting the Caribbean Sea with the eastern Gulf of Mexico. Gallegos (1996) has shown that the Yucatan Current is highly geostrophic, the balance of forces (per unit mass) being f v ¼ 1=r @p=@ n, where r is the density of sea water, and n is the current speed at right angles to the horizontal pressure gradient, @p=@ n . Gallegos also studied the temperature–salinity (T–S) structure of the Yucatan Current and has concluded that it has T–S properties similar to those in the offing of Cape Hatteras. As this branch of the Gulf Stream system flows northward, it forces a vigorous upwelling regime along the eastern Campeche Bank that supports one of the IAS’s greatest fisheries. Once in the Gulf of Mexico, the Yucatan Current is known as the Gulf Loop Current because of its characteristic anticyclonic looping from northward

to eastward to southward to eastward again as it exits the Gulf of Mexico through the Straits of Florida. This clockwise turning of the Gulf Loop Current is part of a cycle of growth (penetrating into the Gulf of Mexico almost to the latitude of the Mississippi River Delta, fE301N), then turning eastward and southward to run along the west Florida escarpment. Near the latitude of Key West (fE241N), the current turns sharply cyclonically and begins to run eastward in the Straits of Florida, where it is now called the Florida Current. Once the Gulf Loop Current has reached its maximum latitudinal extent, a large anticyclonic current ring, 100– 150 km radius, separates from the flow, and the main current reforms farther south near the latitude of the Florida Keys. In its southernmost position, the Gulf Loop Current flows into the Gulf of Mexico, and turns rather sharply in an anticyclonic turn to exit almost directly into the Straits of Florida (Figure 4). Anticyclonic Gulf Loop Current rings have all the T–S and flow features of the Gulf Stream system, just as in the Yucatan Channel and in the Straits of Florida. The ageostrophic dynamic balance is 7n 2 =r þ f n  1=r @p=@ n ¼ 0, where r is the radius of curvature, and where 7n 2 =r is positive in cyclonic

(c) 2011 Elsevier Inc. All Rights Reserved.

290

INTRA-AMERICAS SEA

95˚W

90˚W

85˚W

80˚W

75˚W

70˚W

65˚W

60˚W

55˚W

30˚N

30˚N

25˚N

25˚N

20˚N

20˚N

15˚N

15˚N

10˚N

10˚N

95˚W

90˚W

85˚W

80˚W

75˚W

70˚W

65˚W

60˚W

55˚W

Figure 4 Sea surface height (h) in meters from numerical model calculations as reported in Mooers and Maul (1998). Geostrophic surface currents are calculated from f n ¼ 1=r@p=@ n ¼ g@h=@ n . Anticyclonic eddies are isolated concentric height maxima, the largest of which is shown in the western Gulf of Mexico; cyclonic eddies have the opposite surface height field.

curvature and negative in anticyclonic. Accordingly, flow is super-geostrophic in anticyclonic turns, and sub-geostrophic in cyclonic turns. Identical dynamics describe midlatitude upper tropospheric flows in Earth’s atmosphere, notably in the Jet Stream. A separated Gulf Loop Current ring travels westward into the western Gulf of Mexico, most probably by a self-propulsion mechanism associated with the beta effect, b ¼ @f =@y, of differing Coriolis parameter between the southern and northern ring edges. Using the hydrostatic equation, @p ¼ rg@z, the horizontal pressure gradient term 1=r @p=@ n may be written as g@h=@ n, and for an anticyclonic Gulf Loop Current ring with a diameter of 300 km and @h ¼ 0:75 m, it is calculated using n 2 =r þ f n  g@h=@ n ¼ 0 that eddies self-propagate at speeds of 5–10 cm s1 ðE5  10 km d1). Direct observations from satellite-tracked buoys and from satellite altimeter measurements of sea surface height (h) substantially agree with such calculations. As these Gulf Loop Current rings travel to the west, they begin to spin down, losing their momentum per unit volume, rn, to horizontal friction and eventually mixing in with the ambient Gulf of Mexico Common Water. The lifetime of the current rings is typically 6 months, and they carry with them

the temperature, salinity, and other characteristics of their source region, the Caribbean Sea. Approximately 3  1013 kg y1 of salt is injected into the western Gulf of Mexico by an average Gulf Loop Current ring. Thus the salt balance of the western Gulf of Mexico is decidedly nonMediterranean, E þ d ¼ P þ R, because in the IAS the classical evaporation/precipitation/runoff equation requires an additional term d to account for the infusion of high salinity water from the rings. Typical values for these terms are E2P E35 cm y1, R E75 cm y1, and the volume of fresh water to maintain the salt balance d E40 cm y1. The Gulf Loop Current interacts with the fourth great riverine system of the IAS, the Mississippi River. As the Gulf Loop Current nears the Mississippi Delta, it is observed to entrain or advect the river water to the east. Mississippi River water has been observed by its low surface salinity all along the eastern edge of the Current, into the Straits of Florida, and up the east coast of the USA at least to Georgia (fE321N), the northern boundary of the IAS. Thus the four great rivers of the IAS, the Amazon, the Orinoco, the Magdalena, and the Mississippi, and the many smaller tributaries, are all known to interact with the oceanic flows and to be

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRA-AMERICAS SEA

carried great distances by them. This is an important transport mechanism in the IAS whereby riverine flotsam and jetsam is found on distant shores. This same flow-through regime is also responsible for the considerable impact of tars from maritime commerce and from oil drilling on the highly valuable tourist beaches of the area (cf . Figure 3). Gulf Loop Current ring shedding seems to have a cycle of 10–11 months on average, with some rings being shed in as few as 6 months and others taking 17 months. This is a surprising frequency since it is not at the annual harmonic where many other oceanic features have a spectral peak. Gulf Loop Current ring shedding has been simulated by IAS numerical models, and super-annual periods are often calculated. The cycle of ring formation does not seem to be forced by the R Runmistakable annual cycle of volume transport, nðx; zÞdx dz in the Gulf Stream system, with its maximum in June and minimum in October, nor by variability in the Caribbean Current along 151N, which has a spectral peak at about 75 days. Connectivity between flow variability in the Caribbean Sea and the Gulf of Mexico seems remarkably weak. In the Straits of Florida, the Florida Current turns cyclonically as it passes between Cuba, Florida, and the Bahamas. The lesser passages of the Straits of Florida contribute small amounts to the total volume transport, which by now is at the level of 30 G s1 (30  106 m3 s1) or about 30 Sv. At the latitude of Palm Beach, Florida (fE271N), the meridional RR oceanic heat transport, rCpTnðx; zÞdx dz, is about 1.5 petawatts (1.5  1015W), an amount comparable with the atmosphere at the same latitude. Transport is known to vary on timescales ranging from days to years. Fortnightly volume transport changes are observed to range from 20 Sv to 40 Sv, a value much larger than that of the annual cycle, which is more likely less than 75 Sv. In the narrow confines of the northern Straits of Florida, lateral friction is significantly larger than in the open sea. Here, from extensive in situ studies, the Guldberg-Mohn friction coefficient J in quasi-geostrophic flow fv ¼ g@h=@x þ Jv has been estimated at approximately 4  105 s1, one to two orders of magnitude larger than away from continental boundaries. In addition, the Florida Current axis meanders to the west when transport is high and to the east when it is less. Such detailed transport variations in other regions of the IAS are not as well documented as in the Straits of Florida between Palm Beach and the Bahamas. North of the Straits of Florida the current is called the Gulf Stream, a name it retains to the offing of Nova Scotia. The flow northward to Cape Hatteras,

291

North Carolina (fE351N), seems to be topographically controlled, whereby H in the potential vorticity equation (above) directs onshore and offshore meandering. Near Charleston, South Carolina (fE321N), a notably shallow area decreases H significantly, and the current responds by turning anticyclonically ðBo0Þ into deeper water. Once offshore, the larger value of H causes the flow to turn cyclonically again ðB > 0Þ in a series of meanders as it progresses downstream (cf . Figure 3). Along the outer boundary of the IAS, a surface flow with a great deal of changeability is observed, called the Antilles Current. This intermittent current, carrying on average about 15 Sv, progresses northward up the margin of the Caribbean Sea, along the eastern outer banks of the Bahamas, and eventually joins the Gulf Stream north of Cape Canaveral, Florida (fE291N). Satellite-tracked buoys and subsurface floats both show the latitude of the Bahamas to be an area of great temporal and spatial variability, with many eddies of various sizes, and with inflows to the Straits of Florida through the Old Bahama Channel and the Northwest Providence Channel. The general sense is that of converging surface flows all feeding the Gulf Stream system.

Subsurface Flow Regime The strong surface currents of the Gulf Stream system decrease with depth. Mathematically, this can be explored by applying Leibnitz’ rule to R z the integral form of the hydrostatic equation p ¼ h rgdz . The geostrophic equation (above) can then be expressed as: Z z g@r=@ ndz rf v ¼ rg@h=@ n þ h

where the first term on the right-hand side is the barotropic term, and the integral term on the righthand side is the baroclinic term. Facing downstream @r=@ n is negative, and the surface current (at z ¼ h) decreases with depth until the two terms on the righthand side become equal and opposite. This depth is called the level-of-no-motion and n ¼ 0. In the Yucatan Channel, the level-of-no-motion is approximately 1200 m, but in the northern Straits of Florida (fE271N) northward-flowing currents as much as 0.3 m s1 reach to the seafloor at 800 m. Details of the flows into and out of the IntraAmericas Sea at depth in other passages are less well known than the surface flows. Numerical models and observations suggest a general inflow though the passages of the Lesser and Greater Antilles into the Caribbean Sea, an outflow through the Yucatan Channel into the Gulf of Mexico, and continuing flow into the North Atlantic Ocean north of the

(c) 2011 Elsevier Inc. All Rights Reserved.

292

INTRA-AMERICAS SEA

CURRENT EDGE

VORTEX STREET

YUCATAN PENINSULA

CLOUD COZUMEL ISLAND

Figure 5 LANDSAT negative image of surface-wave glitter patterns showing von Karman vortices downstream of Cozumel Island in the Yucatan Channel. For scaling, Cozumel Island is approximately 50 km long.

Bahamas. Near the sill of the Yucatan Channel, approximately 100 m above the ocean floor, the flow is decidedly from the Gulf of Mexico into the Caribbean Sea. In the Windward Passage, there is also evidence of north-eastward outflow at depth, but it seems not to be as persistent as that in the Yucatan Channel. A major characteristic of the deep waters of the IAS is their nearly isothermal and isohaline profiles below the depth of the major sills. The ocean, being a stratified fluid, tends to inhibit vertical mixing. Thus the sub-sill depth waters are characterized by nearzero vertical density gradients, @r=@zE0, and are neutrally stable. Deep IAS waters have the T–S characteristics of the offshore waters of the juxtaposed North Atlantic Ocean, which seem to spill over the sills from time to time to replenish and ventilate the water interior to the IAS. Thus there must be a surging of sorts to bring into the Caribbean basin in particular, renewing mid-depth North Atlantic Common Water. Along the eastern margin of the IAS at about 2000–3000 m water depth is a southward-flowing mid-level current called the Deep Western Boundary

Current (DWBC). The DWBC has its genesis in the thermohaline circulation of the North Atlantic Ocean, north of the Denmark Strait, and is part of the global conveyor belt. The DWBC flows along the entire outer boundary of the IAS, and has a volume transport of approximately 15 Sv. This extremely important flow is climatically linked to the role of the ocean in Earth’s heat budget and may participate in the complexities of the IAS’s deeper flow patterns.

Other Currents in the Intra-Americas Sea Tidal currents in the IAS are generally weaker than in other semi-enclosed seas. The tides are typically semi-diurnal on the Atlantic Ocean margin of the IAS (M2 and S2 constituents usually), and progressively become diurnal in the Gulf of Mexico where the K1 and O1 constituents dominate. Estuaries such as the Mississippi Delta are of the saltwedge category, mostly because the tidal currents and ranges are small and the river flows very large

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRA-AMERICAS SEA

(average for the Mississippi River is about 103 km3 y1E0.03 Sv). Tidal currents around many IAS islands are similarly weak, with extremes rarely exceeding n ¼ 1 m s1 even in passes through the many bar-built barrier island lagoons. Inertial currents are a ubiquitous feature of the ocean, and are characterized by periods n ¼ 12h =sin f. In the northern IAS, inertial currents often have periods equal to the dominant diurnal tidal currents, such as the K1 or the O1 because sinfE0.5. At these critical latitudes, the inertial currents are not separable from the diurnal currents in the tidal spectrum. Inertial currents, when they occur, are intermittent and have velocities typically below n ¼ 0.2 m s1. Current flows past islands can induce complex patterns in their lee. Numerical models and observations suggest von Karman vortex streets downstream of many island land masses (Figure 5). Such complex currents can cause engineering design complexities, particularly regarding waste disposal and spills. Similarly, with the normally low wave heights so characteristic of the IAS, the longshore and littoral currents are also weak, although in

Mis

90˚W

F lorida Curre n

Loop C

u lf

G

y

E dd

rent (

LC) G

Grande

ur

G LC

Bermuda

60˚W

pi

Rio

certain areas, especially the east coast of Florida, dangerous wave-induced rip currents are very common. Perhaps the most significant physical marine hazard in the Intra-Americas Sea is the combined storm surge and inverted barometer effect associated with hurricanes. Along linear coasts, the water level elevation from a major storm can exceed 7 m, on top of which may be 3 m or greater wind waves. Little is directly known of the currents associated with storm surges, but indirect evidence suggests that they can exceed n ¼ 1–2 m s1, especially in the vicinity of harbor entrances and inlets. Small islands are less at risk from large storm surge-driven currents than are long, low coasts with shallow offshore bottom topography. While storm surge currents are transient features of the IAS with timescales of approximately half a day, they can be costly to infrastructure, and very dangerous to human life. Upwelling is a vertical current of note in the IAS, especially along the long east–west tending north coast of South America. Here the zonal wind stress with mass transtx can force an Ekman upwelling R port per unit width My ¼ rvdz given by My ¼ tx =f.

75˚W Gulf Stream

si p sis v Ri er

293

t N E

W

An ti Cur ll es ren t

S

n ta ca Yu n rr e Cu

20˚N

t

Caribbean Current

-Colo ma

mb

Current

ia

Pana

North Equatorial

yre

G ui an rre a nt

Cu

Ma gd Ri alen ve a

G

r

noco Ori River

5˚N

Figure 6 Summary of surface currents in the Intra-Americas Sea. Maximum IAS sea surface height variability h ¼7 24 cm is centered in the Gulf Loop Current (GLC) at f ¼ 261N, l ¼ 881W; a second maximum in the Caribbean Current of h ¼ 7 12 cm is centered at f ¼ 151N, l ¼ 771W.

(c) 2011 Elsevier Inc. All Rights Reserved.

294

INTRA-AMERICAS SEA

Accordingly, the easterly Trade Winds force a north- g gravity ward mass transport along the coast, and the ocean h sea surface height responds with classical coastal upwelling as seen in n direction vector parallel to pressure the 31C cooler sea surface temperatures along Venegradient zuela. The same physical circumstances cause open- p pressure ocean Ekman pumping in the wake of hurricanes. The r radius of curvature hurricane’s large cyclonic wind stress, t ¼ rair Cd v2air , t time forces mass transport Mx;y in all directions away from u,v,w eastward, northward, upward speed the storm center with attendant lifting of the ther- n velocity vector orthogonal to n mocline and upwelling. Lower sea surface tempera- x, y, z east, north, vertical Cartesian tures are often observed as a cool streak in the wake coordinates air-sea drag coefRcient of these intense air–sea storm systems. Cd specific heat While the danger from seismic sea waves (tsu- Cp evaporation nami) is recognized as another although largely E gigaliters unappreciated natural hazard of the IAS, it is the Gl water depth waves and wave particle motions that create cur- H mass transport rents of such great danger. Caribbean tsunami M precipitation waves have been observed to exceed 9 m in height. P R river runoff Since a tsunami is a progressive shallow-water pffiffiffiffiffiffiffi salinity wave with celerity c ¼ gH, the maximum S Sverdrups currents come at the wave crest and at the wave Sn temperature trough. These currents probably exceed v ¼ T 10 m s1 , and have timescales of several minutes. In W watts that short amount of time however, even more d salinity anomaly from eddies danger exists than with storm surge currents, and f latitude the small islands are equally as vulnerable as are l longitude the continental coasts. r density z vorticity t wind stress Conclusions O Earth’s rate of angular rotation Intra-Americas Sea surface currents (Figure 6) are dominated by a single fact: the IAS is the formation region of the Gulf stream system. Except along the northern coast of the Gulf of Mexico, the volume See also transport of the interior thermohaline component Sphenisciformes. Tides. of IAS currents is minuscule compared with the wind-driven component. While there are important external and peripheral currents associated with the global thermohaline flow such as the Deep Western Further Reading Boundary Current, it is the North Brazil Current– Guianas Current–Caribbean Current–Yucatan Cur- Gallegos A (1996) Descriptive physical oceanography of the Caribbean Sea. In: Maul GA (ed.) Small Islands: rent–Gulf Loop Current–Florida Current–Gulf Marine Science and Sustainable Development. Stream family of atmospheric wind stress-forced Washington: American Geophysical Union. advective movements that characterizes the region Maul GA (ed.) (1993) Climatic Change in the Intra(cf . Figure 4). All these ‘currents’ are in reality one Americas Sea. London: Edward Arnold. current that, coupled with air–sea heat and mois- Mooers CNK and Maul GA (1998) Intra-Americas Sea ture fluxes and winds, integrate into a single conCirculation. In: Robinson AR and Brink KH (eds.) The tinuum that connect the Intra-Americas Sea and its Sea, Vol. 11. New York: John Wiley & Son. peoples. Murphy SJ, Hurlburt HH, and O’Brien JJ (1999) The

List of Symbols c f

wave celerity Coriolis parameter

connectivity of eddy variability in the Caribbean Sea, the Gulf of Mexico, and the Atlantic Ocean. Journal of Geophysical Research 104: 1431--1453. Schmitz WJ Jr (1995) On the interbasin-scale thermohaline circulation. Reviews in Geophysics 33: 151--173.

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRUSIONS Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1363–1367, & 2001, Elsevier Ltd.

Introduction In most frontal regions, where waters of different salinities and temperatures meet laterally, an interleaving of the different waters is observed. These features are commonly referred to as intrusions. Sometimes a single layer of water from one region is advected into the other region, such as the Mediterranean salt tongue in the North Atlantic, by either a mean flow or eddy motion. Multiple layers of the two different water masses are also seen quite often. The driving mechanism for these multiple intrusions is related to horizontal gradients in salinity and temperature and the small-scale (e.g., smaller than the thickness of the intrusions) mixing occurring between the interleaving layers. In this article, only intrusions produced by this latter process are discussed. Both observational and theoretical studies are presented. Frontal regions are locations where waters of different temperature and salinity meet and interact. They are usually characterized by relatively large horizontal gradients in these two properties. Fronts have been found in the coastal ocean, at the shelfbreak and at the boundaries of major currents, such as the Gulf Stream and Antarctic Circumpolar Current. An example of a front (Figure 1) is shown by the azimuthally averaged salinity structure of a Mediterranean eddy. A Mediterranean eddy (Meddy) is a coherent eddy of Mediterranean Sea water found in the eastern North Atlantic Ocean. The front with its larger horizontal gradients in salinity is located at a depth range of 700–1300 m and with a radius of 15–30 km. The temperature field has a similar structure to the salinity structure shown in Figure 1. With the horizontal change in salinity, it would be expected that there would be a horizontal change in the density of the sea water. However, the effect of the horizontal change of temperature on the density nearly completely compensates the density change due to the salinity change across the front. Thus, the density surfaces are nearly horizontal. However, there is a slight upward (downward) tilt of density surfaces in the lower (upper) half of the Meddy. The resulting pressure gradient balances the geostrophic

flow of the eddy. Along-front geostrophic flows are found at most fronts. A closer look at the structure of temperature and salinity in the frontal region shows an interleaving of water with the characteristics of the temperature and salinity on the two sides of the front. The temperature and salinity of the water in these interleaving layers show evidence that mixing of the two water types has also occurred. Figure 2 shows a section of closely spaced (e.g., 1–2 km) vertical profiles of salinity starting from the center of the Meddy (Figure 1) and moving towards the outside edge. In each profile, there are wiggles in the salinity field, typically of 1–2 km vertical scale, which represents the water moving horizontally from the center of the Meddy to the edge or vice versa. These wiggles are referred to as intrusions. Intrusions like these are found in most frontal regions such as those associated with the Gulf Stream and Antarctic Circumpolar Current. The observed interleaving of temperature and salinity is thought to develop as an instability of the thermohaline front. These fluctuations lead to regions of enhanced double-diffusive mixing. Two types of double-diffusive mixing can occur: salt-fingering under the warm, salty layers and diffusiveconvection under layers that are relatively cold and

Salinity (PSU) 0 36.5 36.0

500 35.5 36.0

Depth (m)

D. L. Hebert, University of Rhode Island, RI, USA

1000

35.5

Meddy

35.5

35.5

1500

2000 0

10

20

30

40

50

Radial distance (km)

Figure 1 Azimuthally averaged cross-section of the salinity of a Mediterranean eddy (Meddy) embedded in eastern North Atlantic water. This survey, the second one of this Meddy, was made in June 1985, PSU, practical salinity units.

(c) 2011 Elsevier Inc. All Rights Reserved.

295

296

INTRUSIONS

500

Depth (m)

750

1000

1250

1500 35.5

36

36.5

37

37.5

38

38.5

39

Salinity (PSU) Figure 2 A set of closely spaced vertical profiles of salinity taken from the center of the Meddy towards its edge during June 1985. Profiles have been offset by 0.25 PSU (practical salinity units).

Density flux

Warm and salty

Cold and fresh

Salt-fingering interface Diffusive-convection interface

Frontal region Figure 3 A schematic of the interleaving layers representing the intrusions. The open arrows indicate the cross-frontal motion driven by the depth-varying density flux (solid arrows). In this diagram, salt-fingering is the dominant form of double-diffusion; thus, the warm, salty water rises as it crosses the front.

fresh (Figure 3). Both forms of double-diffusive mixing generate a downward density flux, that is, a release of potential energy. The convergence or divergence of this density flux makes the intrusion either heavier or lighter, respectively. These density changes produce pressure gradients which drive the interleaving motions across the front. If the density

flux of salt-fingering exceeds that of diffusive-convection, waters in the warm, salty layers become less dense and, therefore, rise as they cross the front. The cold, salty layers become more dense and sink as they cross the front. It is believed that this case applies for the intrusions found for the lower half of the Meddy. If diffusive-convection dominates (which is believed

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRUSIONS

to be the case for the intrusions occurring in the upper half of the Meddy), water in the cold, fresh layers should rise across the front and the warm, salty layers sink.

297

500

750

There have been many observations of intrusions in vertical profiles of salinity and temperature taken in frontal regions. Other than demonstrating the presence of intrusions and indicating their vertical scale, it is difficult to make any other conclusions about the dynamics of the intrusions. Closely spaced profiles (e.g., Figure 2) show that the horizontal structure of intrusions is complex. Although it is possible to track an intrusion across several kilometers (and several profiles), the structure of the individual intrusion changes significantly. In addition, some intrusions appear to start and end abruptly. One of the problems of interpreting this type of data is that the frontal region usually has a horizontal velocity field associated with it. Although the water in the intrusions is moving across the front, it is also being advected along the front by the geostrophic current of the front. Therefore, some of the observed crossfrontal variability could be due to differential advection of the intrusions along the front. Most of the observations of intrusions have been single surveys of the front; the evolution and the dynamics of the intrusions could not be determined. Even if multiple surveys are undertaken, temporal changes in the intrusions cannot be separated from possible alongfrontal variations of the intrusions. However, there has been one study where some of the dynamics of the intrusions could be investigated. This was an experiment to determine the evolution of a Mediterranean eddy. The front (Figure 1) between these two water masses can be thought of as a circular front. Thus, the problem of differential along-front advection of the intrusions is removed since the front loops back on itself. It would be expected that the individual cross-frontal transects could be typical for all radial sections and that alongfrontal variations are small. Thus, the cross-frontal transects could be used to determine the intrusion dynamics. This Meddy was surveyed four times over a two-year period as it decayed. The vertical structure of the intrusions evolves as the cross-frontal temperature and salinity gradients change (Figure 4). For the first year of the study, the Meddy had a core region unaffected by intrusive mixing. During this time, the intrusions appeared to have a similar wiggly vertical structure at all locations for both surveys with vertical scale of about

Depth (m)

Observational Studies

1000

1250

Oct. 86 Oct. 85

1500 35.0

35.5

Oct. 84 June 85

36.0 Salinity (PSU)

36.5

Figure 4 Vertical profiles of salinity made through the intrusive region of the Meddy during four surveys: October 1984 (solid line), June 1985 (dashed line), October 1985 (dotted line) and October 1986 (bold line). The profiles have been offset by 0.2 PSU (practical salinity units) from each other.

20 m. The wiggles are rather smooth (i.e., sinusoidal) and have approximately the same vertical scale. By the time of the third survey, the intrusions just reached the center of the Meddy. We can imagine that there was a constant cross-frontal gradient (driving the intrusions) for the first year of observation and that the intrusions had passed their initial (exponential) growth stage. By the time of the third survey, some of the intrusions appeared to have a step-like structure. A year later, the intrusions had a more pronounced step-like structure with a larger vertical scale, about 50 m (Figure 4). This structure is probably representative of decaying intrusions. Double-diffusive processes were still active vertically but the horizontal advection mean gradients were less important; there was not a supply of new water unaffected by the mixing. One major question that remains concerns the cross-frontal fluxes of heat and salt by the intrusions. In order to address this question, it is necessary to measure the very weak velocities in the intrusive layers or observe the large-scale changes in the

(c) 2011 Elsevier Inc. All Rights Reserved.

298

INTRUSIONS

properties of the frontal region. For the first year of study of the Meddy, it had a core region with very little horizontal variability (Figure 1). Using the rate at which the intrusions moved into this central core region, extremely small cross-frontal velocities, u0 , on the order of 1 mm s1 were found. Using the salinity anomalies associated with the intrusions, S0 , and this order of magnitude estimate for the crossfrontal velocity, the average cross-frontal flux of salt, FS ¼  /u0 S0 S, was calculated. Parameterizing this flux in terms of horizontal diffusion, FS ¼ K H

  1 qS r qr

½1

where KH is the horizontal eddy diffusivity and qS/qr is the mean horizontal (radial) salinity gradient across the front; an eddy diffusivity coefficient of 0.4 m2 s1 was found. The dominant mechanism responsible for the decay and eventual demise of the Meddy was thermohaline interleaving, presumably driven by double-diffusive buoyancy fluxes. Over the 2-year observation period, intrusions at the edge of the Meddy core eroded the warm and salty central region from an initial diameter of 60 km until the core was no longer detectable. Using the rate at which the salinity and temperature of the Meddy at a specific radius changed, an eddy diffusivity could be estimated. qS q2 S 1 qS þ ¼ KH qt qr2 r qr

! ½2

Likewise, by integrating eqn [2] from the center to a specified radius, changes in the salt and heat content of the Meddy can be used to estimate an eddy diffusivity. It was found that an eddy diffusivity of 1–5 m2 s1 could be used to parameterize the crossfrontal fluxes of the intrusions. To date, this Meddy study has been the only one where estimates of the fluxes by intrusions could be made. Estimates of the horizontal eddy diffusivity ranged from 0.5 to 5 m2 s1. An attempt was made to understand the dynamics of the intrusions with these surveys but the temporal sampling (six months) was too infrequent to be of use in investigating the evolution of the intrusions.

Theoretical Studies The driving mechanism for the cross-frontal velocity of intrusions (i.e., the horizontal pressure gradients) is due to divergences in the vertical density fluxes, generally assumed to be due to double-diffusive mixing (Figure 3). Most of the theoretical studies to

date have looked at the initial growth of the intrusions using linear stability analysis with parameterizations for the vertical flux by salt-fingers. In these linear stability calculations, the background frontal structure is assumed to have linear gradients, both horizontally and vertically, of salinity and temperature. The horizontal gradients are chosen such that there is no horizontal gradient in density and thus, no along-front velocity. Vertical gradients are chosen such that the background structure is unstable to double-diffusion, usually saltfingering. Double-diffusive mixing is parameterized as a constant eddy diffusivity for salt (or heat) and a constant ratio of the heat to salt flux. The perturbations are assumed to be small, so there are no inversions in temperature and salinity. The linear stability analysis predicts the vertical scale, crossfrontal and along-frontal slopes of the fastest-growing unstable mode given the salinity and temperature gradients. These properties have been compared to observations and have shown general agreement. For typical horizontal gradients of salinity and temperature found in frontal regions, the growth rate of the fastest mode has an e-folding timescale on the order of 10 days. Inclusion of a background velocity shear due to the sloping isopycnals across the front can produce faster-growing intrusions with an e-folding timescale on the order of several days. Linear stability studies predict properties of the initial growth stage, in which fluxes grow exponentially, but say nothing about the finite amplitude ‘steady’ state properties. When the intrusions reach finite amplitude, the fluxes of heat and salt by the interleaving should reach a constant value. Since fronts in the ocean exist much longer than the time for the intrusions to grow, intrusions spend most of their lives in the finite-amplitude state. Therefore, the usefulness of extrapolating intrusion properties and fluxes from linear theory is questionable. For growing intrusions to reach an equilibrium, a three-way balance between salt-finger, diffusiveconvection and (cross-frontal) advective fluxes is necessary. The initial instability may set the vertical scale of the finite-amplitude intrusions, but the crossfrontal fluxes may depend critically on the form of the equilibrium that the growing intrusions eventually reach. A numerical model verified that small amplitude intrusions, predicted by linear stability analysis, evolved into large amplitude, equilibrium, intrusions. When the amplitude of the intrusion becomes large enough that temperature and salinity inversions occur, the growth of the intrusion slows and reaches an equilibrium state. This equilibrium state is characterized by interleaving layers with saltfingering and diffusive-convection occurring at the

(c) 2011 Elsevier Inc. All Rights Reserved.

INTRUSIONS

299

interfaces separating statically unstable ‘convecting’ layers. As expected, the three-way flux balance is achieved. As well as obtaining a balance in the advection and mixing of the salinity and temperature, there must be a momentum (energy) balance. The double-diffusion mixing lowers the potential energy of the system. This potential energy is converted into kinetic energy within the convecting layers. In addition, the convecting layers allow a large flux of momentum from the salt-fingering interface to the double-diffusive interface. The friction between the interleaving layers balances the pressure gradient produced by the density flux divergence.

dimensional structure of intrusions to evaluate the two-dimensional studies done to date.

Summary

Further Reading

The presence of intrusions in frontal regions has led oceanographers to believe that they must be important in the cross-frontal fluxes of heat and salt. However, at present, these fluxes are almost impossible to observe in the ocean. Thus, we must rely on theoretical and numerical studies to address this important question. In order to be useful, these studies must predict properties of intrusions which can be compared to observations. To date, comparisons have been limited to the vertical length scale of intrusions from single vertical profiles of temperature and salinity. Predictions of the slope of the intrusions (relative to density surfaces) in the crossfrontal and along-frontal directions have been compared to the few cross-frontal sections made. With improvements in navigation with global positioning satellites and the advent of undulating towed bodies, rapid three-dimensional high-resolution mapping of intrusions can be undertaken. Future work, both numerical and observational, will use the three-

See also Double-Diffusive Convection. Meddies and SubSurface Eddies. Shelf Sea and Shelf Slope Fronts. Upper Ocean Mean Horizontal Structure. Upper Ocean Mixing Processes. Water Types and Water Masses.

Hebert D, Oakey N, and Ruddick B (1990) Evolution of a Mediterranean salt lens: scalar properties. Journal of Physical Oceanography 20: 1468--1483. May BD and Kelley DE (1997) Effect of baroclinicity on double-diffusive interleaving. Journal of Physical Oceanography 27: 1997--2008. McDougall TJ (1985) Double-diffusive interleaving. Part II: Finite amplitude, steady state interleaving. Journal of Physical Oceanography 15: 1542--1556. Ruddick B (1992) Intrusive mixing in a Mediterranean salt lens – intrusion slopes and dynamical mechanisms. Journal of Physical Oceanography 22: 1274--1285. Ruddick BR and Hebert D (1988) The mixing of Meddy ‘Sharon’. In: Nihoul JCJ and Jamart BM (eds.) Small-Scale Mixing in the Ocean. Elsevier Oceanography Series, vol. 46. Amsterdam: Elsevier. Toole JM and Georgi DT (1981) On the dynamics and effects of double-diffusively driven intrusions. Progress in Oceanography 10: 123--145. Walsh D and Ruddick B (1998) Nonlinear equilibration of thermohaline intrusions, Journal of Physical Oceanography 28: 1043--1070.

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELING OF TRACERS AND NUTRIENTS R. Schlitzer, Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany & 2009 Elsevier Ltd. All rights reserved.

Introduction Seawater contains a large variety of dissolved and particulate constituents, commonly referred to as tracers. Many of these tracers (such as the dissolved nutrients, phosphate, nitrate, and silicate, or the radioactive carbon isotope 14C (radiocarbon)) are natural and are part of the ocean system since geological times. Other tracers, such as the chlorofluorocarbons (CFCs) and various radioactive isotopes from nuclear bomb testing, are of anthropogenic origin, and have been introduced into the ocean only during the last decades. The distributions of anthropogenic tracers in the ocean change markedly on decadal or annual timescales. Naturally occurring tracers, like the nutrients, are believed to be near steady state. However, recent data indicate decadal and inter-annual changes in ocean oxygen and nutrient distributions, likely caused by changes in circulation and biogeochemical processes associated with global climate change.

0.0 5

0.1

0.2

5 2.2 1.5 1 0.25 0.5 5 0.0

1

5

0.1

0.2

0.1

0.05

0.2

2

2.25 1.5 0.25 0.5 0.1 0.1 0.05

1

2.5

2

0.05

0.5

2000

2

0.

0.5 0.2 5 2 0.

1000

1.5

3000 5

0.1

0.2

1

40° N

0.05

6000

EQ 80° W

70° W

60° W

50° W

40° W

0°

Ocean Data View

5000

5

0.2

WOCE A5 Jul./Aug. 1992 40° W

30° W

Ocean Data View

0.2

4000

.05

Depth (m)

1.5

5

2.25 1

2.2

2

2.25

CFC-11 (pmol kg−1) 2.25

0

Some ocean tracers, such as radiocarbon and the CFCs, act as passive tracers. They are chemically inert and are being transported and redistributed in the ocean by currents after entering the ocean from the atmosphere. When dense surface waters sink and spread in the ocean interior, they carry the tracer signal with them and produce tongues of high tracer concentrations along their spreading paths. An example of such tracer tongues and concentration maxima in the ocean interior can be seen in the zonal CFC section in Figure 1. Note the pronounced CFC maxima at about 1500- and 3500-m depth at the western boundary, revealing the two cores of the deep western boundary current (DWBC) carrying tracer laden waters southward at the western side of the basin. The spatial extent of the tongues and the concentration gradients along the tongues can, in principle, be used to determine the pathways and velocities of the currents. Such inferences require advanced mathematical methods, some of which are described below. Other tracers (like the macronutrients phosphate, nitrate, and silicate, or the micronutrients Fe and Zn) are chemically and biologically active and are a prerequisite for biological production in the ocean’s surface layer. Surface nutrient concentrations are very low in most regions of the world ocean due to

1

0.5

0

20° W

Figure 1 Zonal section of chlorofluorocarbon 11 at about 241 N in the Atlantic, clearly showing the traces of the upper and lower branches of the deep western boundary current (DWBC). Chlorofluorocarbons (CFCs) are man-made gases that invade the ocean from the atmosphere in increasing amounts since the 1940s. Before that time the ocean was CFC-free. High CFC concentrations in the ocean interior are clear indications of vigorous subsurface currents.

300

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELING OF TRACERS AND NUTRIENTS

biological utilization. Biologically produced organic particles as well as calcium carbonate and opal shells sink and dissolve in deeper parts of the water column, thereby returning nutrients to the dissolved pool, while utilizing and drawing down dissolved oxygen. Overall, the biogeochemical processes act as a vertical nutrient and carbon pump creating significant vertical gradients and pronounced subsurface

(a)

301

nutrient maxima at the depths where particle remineralization occurs. Subsurface nutrient maxima are often correlated with oxygen minima, clearly indicating the remineralization of organic material. The signatures of these processes in the nutrient and oxygen distributions are large and easily detected (Figure 2) and reveal information about the underlying processes. The strength of the biological

Salinity (pss-78) 0 37

4000

60˚ N

EQ

6000 40° S

20° S

Eq.

20° N

40° N 60˚ S 80° W

Ocean Data View

36

Ocean Data View

Depth (m)

36.5 2000

35.5 35 34.5 34

20° W

(b) 0

2.5

2000 1.5 Ocean Data View

Depth (m)

2

4000

6000 40° S

20° S

Eq.

20° N

40° N

1 0.5 0

60° N

(c) 0

250

2000

200 Ocean Data View

Depth (m)

300

4000

6000 40° S

20° S

Eq.

20° N

40° N

150 100

60° N

Figure 2 Meridional sections of (a) salinity, (b) phosphate, and (c) oxygen along WOCE section A16 in the Atlantic. In addition to ocean circulation, the phosphate and oxygen distributions are also affected by biological nutrient utilization near the surface and by subsurface remineralization of sinking organic material. Organic matter remineralization releases dissolved nutrients while utilizing dissolved oxygen, and can be clearly identified between 200- and 1500-m depth in the tropical and subtropical areas.

(c) 2011 Elsevier Inc. All Rights Reserved.

302

INVERSE MODELING OF TRACERS AND NUTRIENTS

production, for instance, can be inferred from the observed strength of the vertical nutrient gradient, and the depth range of particle remineralization can be determined from the observed vertical position and spatial extent of the nutrient maxima and oxygen minima. Differences in the vertical structure of different nutrients (e.g., phosphate, silicate, carbon, and alkalinity) reveal differences in the remineralization depths of organic material, CaCO3, and opal. Basin-wide observations of a variety of oceanic tracers have been conducted since the 1950s. The first coordinated and global tracer program, GEOSECS, produced tracer data of unprecedented quality and coverage during the 1970s. More recently, the World Ocean Circulation Experiment (WOCE) has provided an even more detailed tracer data set describing the distributions during the 1990s. These data are publicly available in electronic form over the Internet or as colored distribution plots in printed atlases. Availability of original tracer data is essential for the inverse methods described below.

Inverse Model Concepts Deriving quantitative results about the underlying biogeochemical processes from oceanic nutrient and tracer data, and separating the effects of biogeochemistry from the effects of circulation is a challenging task, and requires the use of coupled physical/biogeochemical numerical models. There are a variety of possible approaches, which can broadly be divided into two categories. The so-called ‘forward models’ assume rates of physical and biogeochemical processes to be known a priori, and require ocean currents (or the physical forcing at the ocean surface), as well as biological production and particle remineralization rates to be specified as input. Oceanic tracer concentrations are treated as unknowns, and the tracer fields that would evolve under the assumed flows and biogeochemical parameters are then simulated. Forward models, in general, lead to mathematical systems that are relatively easy to solve. The conceptual disadvantage of forward models is that physical and biogeochemical rate information that is supposed to be determined from the tracer data and, in fact, only becomes available after the data evaluation, is required a priori to enable the model run. ‘Inverse models’, in contrast, follow a seemingly more intuitive approach, treating the measured tracer concentrations and other auxiliary knowledge formally as knowns, whereas the physical and biogeochemical parameters, to be determined on the basis of the tracer data, are treated as unknowns.

As described in more detail below, the inverse approach generally leads to underdetermined mathematical systems that are much harder to solve than the systems encountered in forward models. Error and resolution analysis are two issues of particular importance when solving underdetermined inverse problems. First, the solved-for physical and biogeochemical parameters depend directly on the tracer data, and errors in the data propagate into errors in the solution. Second, owing to the incompleteness of information in underdetermined systems, the unknowns are usually not fully resolved. Instead, only specific linear combinations of unknowns may be well constrained by the data, while individual unknowns or other combinations of unknowns may remain poorly determined. Both, error and resolution analysis are essential for a quality assessment of the solution of underdetermined systems. Two widely used practical inverse approaches are presented in the next two sections (‘Estimating absolute velocities and nutrient fluxes across sections’ and ‘Estimating carbon export fluxes with the adjoint method’). These serve as examples to describe details of the mathematical methods and to list achievable results. The first method, the section inverse approach, infers nutrient, carbon, and tracer fluxes across sets of sections, based on hydrographic, tracer, and nutrient data along these sections. The second example describes an application of the adjoint method for the determination of ocean currents, biological productivity, and downward particle fluxes. This method is specifically adapted for the utilization of many different tracers and can handle problems with heterogeneous and sparse data coverage.

Estimating Absolute Velocities and Nutrient Fluxes across Sections The section inverse approach developed by Wunsch exploits hydrographic data along sets of oceanographic sections, and allows estimation of absolute flow velocities perpendicular to the sections. The method was later extended to include nutrient and oxygen data allowing to estimate nutrient and oxygen fluxes across the sections. The section inverse method is based on the geostrophic principle, which for a given pair of hydrographic stations allows calculating the geostrophic velocity vg(z) perpendicular to the connecting line between the two stations as function of depth z. A reference velocity b has to be added to the geostrophic velocity to obtain the absolute flow velocity

(c) 2011 Elsevier Inc. All Rights Reserved.

vðzÞ ¼ vg ðzÞ þ b

½1

INVERSE MODELING OF TRACERS AND NUTRIENTS

at a given station pair. The reference velocity b is left unknown by the geostrophic calculation. Determining these unknowns was a major challenge in physical oceanography for many decades. There is one unknown reference velocity bi for every station pair in the sections considered. Thus, for global networks consisting of hundreds of sections, as in the case of the WOCE survey, the number of unknown velocities bi may amount to several thousands. Steady-state mass and tracer conservation equations for all subdomains formed by the intersecting hydrographic sections (see Figure 3) are written as constraint equations for the unknown bi. For a given domain the conservation equation of tracer C is expressed as X ðc¯ v¯ gi þ c¯ bi Þ  di  Dxi ¼ 0

½2

i

where the summation is over all station pairs along the boundary of the domain, c¯ v¯ gi is the vertically averaged tracer flux density arising from the

303

geostrophic flow, c¯ is the vertically averaged tracer concentration, di is the mean water depth of pair i, and Dxi is the distance between the two stations of the pair. Mixing terms are ignored for the sake of simplicity. Mass budget equations are obtained by replacing c with density r. It is important to note that the budget equation [2] is linear in the unknown bi, and that the known components of the fluxes involving the geostrophic velocities can be moved to the right-hand side. Formulation of conservation equations for all domains (whole water column and suitably defined layers) and for all tracers considered results in a set of linear equations Ab ¼ G

½3

where b is the vector of m unknown reference velocities (m is the total number of station pairs), A is an n  m matrix containing the coefficients of the n budget equations, and P G is an n vector with the known right-hand sides i c¯ v¯ gi  di  Dxi of the budget

(a)

T1 d1

T2 d2 ΔX1

ΔX2

Stations

(b) T1 T2 d3

ΔX3

Figure 3 Simple examples of (a) a three-station pair defining sections I and II, and (b) the station configuration and bottom depths di in sections I and II, where dashed lines are supposed layer interfaces. Reproduced from Wunsch C (1978) The North Atlantic general circulation west of 501 W determined by inverse methods. Reviews of Geophysics 16: 583–620, with permission from the American Geophysical Union.

(c) 2011 Elsevier Inc. All Rights Reserved.

304

INVERSE MODELING OF TRACERS AND NUTRIENTS

equations. The objective is to solve the system of linear equations for b. Once this is achieved, the absolute flow velocities can be calculated according to [1], and the fluxes of any tracer across a given cross section can be obtained by integrating c  v over the interface. Finding the solution b of [3] is complicated for two reasons: (1) the number of equations n is typically much smaller than the number of unknowns m, leaving [3] underdetermined, and (2) the right-hand side G is based on hydrographic measurements and therefore contains errors. Because of the errors in G, there is no exact solution to [3], in general, and one has to resort to the least-squares solution that minimizes the residuals r ¼ Ab þ G. It should be noted that row and column weighting of [3] is usually applied before calculating solutions. The reader should refer to the literature for a discussion of the weighting process and a description of the various error contributions to be considered for G. Solving [3] relies on the singular value decomposition of the coefficient matrix A: A ¼ U  S  VT

½4

where U and V are orthogonal matrices of dimension n  n and m  m, respectively, and S is a diagonal matrix with the same dimension n  m as A. The superscript T indicates transposition. The diagonal of S contains the nonnegative singular values si (i ¼ 1, y, n) sorted in decreasing size. The number of singular values greater than zero, p, is equal to the rank of A and reveals the actual number of independent equations in [3]. All subsequent singular values sp þ 1, y, sn are zero, and A can be rewritten as A ¼ Up  Sp  VpT

½5

where Up and Vp are trimmed versions of U and V of dimension n  p and m  p, respectively, containing only the first p columns of U and V. Sp is the p  p submatrix of S consisting of the first p rows and columns only. With these quantities, the smallest least-squares solution bˆ of [3], the covariance matrix of the solution covðbˆÞ, and the resolution matrix R are given by T bˆ ¼ Vp  S1 p  Up  G

½6

T covðbˆÞ ¼ Vp  S2 p  Vp

½7

R ¼ Vp  VpT

½8

In these expressions, the superscript ‘  1’ indicates the inverse of the matrix, and superscript ‘  2’

indicates the product of the inverse matrix with itself. The resolution matrix R describes the relationship between the optimal solution bˆ obtained from the underdetermined system [3] with the ‘true’ solution b that one would find if [3] contained sufficient information and was full rank: bˆ ¼ R  b

½9

Every component bˆj of the solution of the underdetermined system [3] can thus be represented as a linear combination of the ‘true’ unknowns P bˆj ¼ k rjk  bk involving coefficients from row j of R. The analysis of the resolution matrix R and the deciphering of the real significance of the calculated bˆj are important steps for a meaningful interpretation of the results of underdetermined systems. Only if the resolution matrix R is diagonal (this is the case if [3] is full rank) will the calculated unknowns represent the ‘true’ unknowns. The section inverse approach described above has been applied by Ganachaud and Wunsch using hydrographic, nutrient, and oxygen data from 20 WOCE sections worldwide. Results for section integrated top-to-bottom net nitrate fluxes and divergences are shown in Figure 4. Net nitrate transports are found to be southward throughout the Atlantic and Indian Oceans and northward in the South Pacific. However, for many sections the uncertainties are of similar magnitude as the flux values themselves, leaving the net transports essentially indistinguishable from zero. Nitrate divergences in many regions of the world ocean indicate nutrient sinks in the upper part of the water column and nutrient sources below, consistent with the concept of biological production in surface waters and subsequent particle remineralization and release of nutrients below.

Estimating Carbon Export Fluxes with the Adjoint Method The section inverse approach described above requires hydrographic, nutrient, and tracer data along all interfaces of the domains for which budget equations are formulated. It is assumed that tracer fields and flows are in steady state, precluding the use of time-dependent tracers, like the CFCs. The budget equations could be generalized to include the time rate of change of tracer inventory inside the domain. However, one would need repeated tracer measurements inside the domain and along all its boundary for all times considered in the model to correctly describe the temporal inventory changes and the

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELING OF TRACERS AND NUTRIENTS

305

60° N NITA 1 × 102 kmol s−1

10

80

20 × 102 kmol s−1

30° N

24.3(67m)

Uncertainty

26.44(162m)

+1 mol yr−1 m−2 −50

−10

−1 mol yr−1 m−2

26.44(183m)

−70 0° −210

−410

−160

−180

270

−60

−240 26.8(378m)

−70

30° S

190

−90

−240

60° S

4340

4620

4090 4300

60° W

0°

60° E

120° E

4700

180° W

120° W

Figure 4 Global dissolved nitrate transports and divergences. The length of each arrow corresponds to the nitrate transports between continents. The open boxes behind each arrow indicate the uncertainty (one standard deviation). Between sections, nitrate divergences are indicated by the solid boxes, either top-to-bottom (single box) or surface/deep (double box). Adapted from Ganachaud A and Wunsch C (2002) Oceanic nutrient and oxygen transport and bounds on export production during the World Ocean Circulation Experiment. Global Biogeochemical Cycles 16 (doi:10.1029/2000GB001333), with permission from the American Geophysical Union.

tracer fluxes across the boundaries. These massive data requirements cannot be met for any known transient tracer, and it seems necessary to adapt the inverse methodology for inclusion of sparse timedependent and steady-state tracers. A hybrid model consisting of forward and inverse steps and utilizing the Lagrange multiplier method of constrained variational optimization for fitting the model to tracer data has been developed for this purpose. The model exploits data for many tracers, including nutrients, radiocarbons, and CFCs. The objective of the model is to find optimal threedimensional (3-D) global ocean flows, biological production rates, and depth-dependent downward particle fluxes that explain the observed tracer, nutrient, and oxygen distributions best. The optimization is done iteratively, varying the flows as well as

the biogeochemical parameters systematically until the agreement between model simulated tracer fields and observations is optimal. The particular model has a rectangular grid (Figure 5), where grid cell boundaries are not required to match lines of available data. The layout of the grid is decoupled from the that of the available data, and individual grid cells (boxes) may be void of any data. Model tracer values are defined at the center of the boxes, whereas flows are defined on the interfaces. Biological production of particulate material occurs in the top model layers representing the euphotic zone. Particle fluxes below the euphotic zone are assumed to decrease with depth following a functional relationship from the literature

(c) 2011 Elsevier Inc. All Rights Reserved.

jP ðzÞ ¼ a  ðz=zEZ Þb

½10

306

INVERSE MODELING OF TRACERS AND NUTRIENTS

Gas exchange, heat, and freshwater fluxes

ZEZ

Export production

wi Depth

uj

vi

Water-column remineralization

wj 3-D circulation

Particle flux

Bottom remineralization Sedimentation

Figure 5 Vertical model grid and definition of model parameters. Reproduced from Schlitzer R, Usbeck R, and Fiscjer H (2004) Inverse modeling of particulate organic carbon fluxes in the South Atlantic. In: Wefer G, Mulitza S, and Ratmeyer V (eds.) The South Atlantic in the Late Quaternary – Reconstruction of Material Budget and Current Systems, pp. 1–19. Berlin: Springer, with permission of Springer Science Business Media.

where a is the particle flux at the base of the euphotic zone, zEZ, and represents the export production. The parameter b determines the shape of particle flux profile and thus controls the depth of remineralization. Large values for b correspond to steep particle flux decreases and thus large remineralization rates just below the euphotic zone, whereas values for b close to zero result in almost constant particle fluxes with depth with little remineralization in the water column and most of the particle export reaching the ocean floor. Export production a and remineralization parameter b vary geographically, and goal of the model runs is to infer optimal values for a and b (in addition to flow velocities v) based on the available nutrient and tracer data. Figure 6 shows a schematic overview of the computational strategy. All quantities listed under model parameters are to be determined by the model. These parameters are combined in the vector of independent parameters p . They have to be initialized to start the calculation (see the literature for initialization strategies), and they will be varied in a systematic way in the course of the calculations. Using the initial independent parameters, the model simulates the distributions of all steady-state (temperature, salinity, oxygen, phosphate, nitrate, total inorganic carbon, alkalinity, and radiocarbon) and transient (CFCs) tracers, as in normal forward models. The simulated tracer concentrations mi are combined in a ˜ All dependent vector of dependent parameters p.

parameters in p˜ can be calculated uniquely from the independent parameters p using the tracer budget equations for the boxes of the model. Once calculated, the simulated tracer distributions can be compared with measurements. Traditionally, this model/data comparison is done subjectively by analyzing the misfit fields and hypothesizing possible causes for the misfits. Model flows or biogeochemical parameters are then modified, hoping that new simulations with the modified parameters lead to more realistic tracer fields. This manual tuning has been used successfully with small box models; however, for problems with many thousands of parameters, such as the one described above, it is impractical and not successful in most cases. The Lagrange method of constrained optimization (termed ‘adjoint method’ in meteorology and oceanography) offers an alternative to manual parameter tuning. This method varies and optimizes the independent parameters p automatically, while satisfying the set of model equations consisting of all tracer budgets for all boxes exactly. The model equations are usually represented in homogeneous form Ej ¼ 0. The adjoint method can be applied to very large problems with hundred thousands of parameters or more. Here, the evaluation of the model performance is done objectively using a suitably defined cost function F. A cost function typically contains terms for all unwanted features of the model, most importantly terms that penalize deviations between model simulated tracer values mi and

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELING OF TRACERS AND NUTRIENTS

307

Initialize

Model parameters

Model fields

Temperature

Nitrate

Salinity

Silicate

Air−sea fluxes

Phosphate

∑CO2

Export production

Oxygen

Alkalinity

Particle fluxes

CFC

Radiocarbon

Circulation Simulation

Mixing coefficients

Adjoint model

Improve model parameters

Compare with observed fields

Optimum? Figure 6 Schematic overview of model calculations performed for every iteration of the optimization process.

observations di7si:   mi  d i 2 ˜ ¼?þ Fðp ; pÞ þ? si 

½11

A large value of F indicates that the simulated tracer concentrations differ much from the observations. Bringing the model close to the data, therefore, is equivalent to minimizing F, subject to satisfying the model equations (tracer budget equations) Ej ¼ 0 exactly. Details on the Lagrange multiplier method used for the constrained minimization of F can be found in textbooks and the literature on data assimilation or constrained data fitting. Overall, this method allows calculating the direction in parameter space rFp of steepest decrease of F (the negative gradient of F with respect to p ), which can then be used by a descent algorithm to arrive at a new, modified vector of independent parameters p . It is guaranteed that new simulations using the modified p will lead to more realistic tracer simulations and a smaller value of the cost function F. This procedure is repeated until no significant further decrease in F can be achieved or a limit on the number of iterations is reached (see Figure 6). The adjoint step in the calculations replaces the subjective evaluation of misfits mentioned above and provides an automatic

‘learning’ step of the model based on the current model/data misfits. The computational cost of the adjoint run is comparable to the cost of the simulation. The number of iterations required to reach the minimum of F depends on the initial p , but is usually large. It is therefore important to use efficient implementations of the simulation and adjoint steps to keep the computation time of a single iteration as short as possible. Figure 7 shows the observed and simulated radiocarbon values in the bottom waters of the world ocean after optimization. Following common practice, 14C concentrations are given in D-notation expressing the per mille 14C concentration difference of a given sample from the 14C standard (wood grown in 1890, decay corrected to 1950). A sample with D14C ¼  100%, for instance, has a 14C concentration 100% (or 10%) lower than that of the standard. In agreement with the data and consistent with the general concept of the global thermohaline circulation, the model yields highest D14C values in the North Atlantic (  70% to  80%) and lowest radiocarbon concentrations in the northeast Pacific (  235%). D14C values in the Southern and Indian Oceans are intermediate and range between  150 and  165%. There are clear signs of tongues with elevated D14C values along the major deep and bottom water spreading paths. This includes the

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved.

30° S

200

60° S

150

180°W

0

20

90°W

10

0

0

0°

5 15

90°E

Ocean data view

Figure 7 Bottom water D14C simulated by the model (color-shaded field) and from data (colored dots). The mean model-data D14C difference in areas not affected by bomb-14C is only þ 1.3%, and the root-mean-square difference amounts to only 5.2%. From Schlitzer R (2006) Assimilation of radiocarbon and chlorofluorocarbon data to constrain deep and bottom water transports in the world ocean. Journal of Physical Oceanography 37: 259–276.

EQ

30° N

60° N

55

200

310

INVERSE MODELING OF TRACERS AND NUTRIENTS

southward flow of North Atlantic Deep Water (NADW) in the North Atlantic and the northward spreading of Antarctic Bottom Water (AABW) in the western South Atlantic as well as in the South Pacific and the Indian Oceans. The quantitative agreement between simulated and measured D14C values is excellent, and the model-data difference is only 1.3% on average. The root-mean-square (rms) difference amounts to only 5.2% and is of the same order of magnitude as the uncertainties of the radiocarbon data. In the deep North Atlantic and parts of the Southern Ocean, the measurements are systematically higher than the model simulations, owing to the contribution of bomb-14C in these waters. Bombradiocarbon is not included in the model, and contaminated data values are not assimilated in the model. The model fluxes of particulate organic carbon (POC) at the base of the euphotic zone (carbon export production) necessary to reproduce the observed oxygen and nutrient fields are shown in Figure 8. The spatial patterns of carbon export resemble the patterns of primary production in published maps showing high fluxes in coastal upwelling areas off West Africa, along the West American coast, in the Arabian Sea and Bay of Bengal, in the northwest Atlantic and north and tropical Pacific, in the area of the Indonesian archipelago, and in the Southern Ocean. Highest values in the productive areas are on the order of 5–10 mol C m  2 yr  1; in the oligotrophic, open-ocean regions they amount to between 0.5 and 2 mol C m  2 yr  1. The export of POC in the model predominantly occurs at mid-latitudes and in the Southern Ocean, and the contribution of the Northern Hemisphere is comparatively small. Globally integrated, the POC export in the model amounts to about 10 Gt C yr  1. Error analysis of the solution vector p is possible but computationally very expensive. An eigenvector/ eigenvalue analysis of the inverse Hessian matrix H  1 (the Hessian is the square matrix of second partial derivatives of F with respect to parameters p ) reveals directions in parameter space that are well determined (eigenvectors associated with large eigenvalues; values of F increase rapidly when moving away from optimal point along the direction of the eigenvector) and directions that are only poorly determined by the model (eigenvectors associated with smallest eigenvalues; values of F increase slowly when moving away from optimal point along the direction of the eigenvector). The ratio of largest and smallest eigenvalues of H  1 is a measure of the anisotropy of F around the optimal solution p . For relatively small problems with a few hundred parameters in p , the Hessian matrix, its inverse, and the associated eigenvalues and

eigenvectors can actually be calculated and a rigorous error analysis of the solution is possible. For large problems with hundred thousands of parameters, this will remain impossible for the foreseeable future. Strategies for obtaining sensitivity information at least for some directions in parameter space or for finding the largest eigenvalues and associated eigenvectors are described in the literature.

Conclusion Recent progress in applications of inverse models to problems in physical and biogeochemical oceanography clearly shows that inverse methodology is well advanced and being used by a growing number of researchers. Mathematical techniques exist that exploit the available data better than before and produce new and important scientific results. These methods successfully cope with problems, such as sparseness of data and incompleteness of information. The advances were possible because of the tremendous technological progress in computer hardware, combined with breakthroughs in the development of efficient and innovative algorithms. These algorithms finally allowed the numerical application of mathematical principles, such as the Lagrange multiplier method, whose theory was established for centuries already. Still, the widespread use of inverse methods would not have been possible, if there had not been at the same time an increased awareness among scientists for the need of integrated, global databases and an increased willingness to contribute to these data sets. Much more data will become available in the future, enabling inverse-type studies that are still impossible today. While ship observations will continue to be important, the new autonomous floats, gliders, or profiling instruments on moorings will provide data in near real time, even from remote and inaccessible regions. New satellite sensors will complement the water column data with high-resolution data from the ocean surface, and new geochemical programs, such as GEOTRACES, will produce highquality data of the ocean’s trace elements and isotopes, including micronutrients such as Fe and Zn, in unprecedented quality and coverage.

See also Inverse Models.

Further Reading Ganachaud A and Wunsch C (2002) Oceanic nutrient and oxygen transport and bounds on export production during the World Ocean Circulation Experiment.

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELING OF TRACERS AND NUTRIENTS

Global Biogeochemical Cycles 16 (doi:10.1029/2000GB 001333). Gill PE, Murray W, and Wright MH (1981) Practical Optimization. London: Academic Press. Hestenes MR (1975) Optimization Theory. New York: Wiley. Kasibhatla P, Heimann M, Hartley D, Mahowald N, Prinn R, and Rayner P (eds.) (1998) AGU Geophysical Monograph Series, Vol. 114: Inverse Methods in Global Biogeochemical Cycles. Washington, DC: American Geophysical Union. Menke W (1984) Geophysical Data Analysis: Discrete Inverse Theory. San Diego, CA: Academic Press. Rintoul S and Wunsch C (1991) Mass, heat, oxygen and nutrient fluxes and budgets in the north Atlantic Ocean. Deep Sea Research 38(supplement): S355--S377. Schlitzer R (2000) Applying the adjoint method for global biogeochemical modelling. In: Kasibhatla P, Heimann M, Rayner P, Mahowald N, Prinn RG, and Hartley DE (eds.) AGU Geophysical Monograph Series, Vol. 114: Inverse Methods in Global Biogeochemical Cycles, pp. 107--124. Washington, DC: American Geophysical Union. Schlitzer R (2002) Carbon export fluxes in the Southern Ocean: Results from inverse modeling and comparison with satellite based estimates. Deep Sea Research II 49: 1623--1644. Schlitzer R (2004) Export production in the Equatorial and North Pacific derived from dissolved oxygen, nutrient and carbon data. Journal of Oceanography 60: 53--62. Schlitzer R (2006) Assimilation of radiocarbon and chlorofluorocarbon data to constrain deep and bottom water transports in the world ocean. Journal of Physical Oceanography 37: 259--276.

311

Schlitzer R, Usbeck R, and Fiscjer H (2004) Inverse modeling of particulate organic carbon fluxes in the South Atlantic. In: Wefer G, Mulitza S, and Ratmeyer V (eds.) The South Atlantic in the Late Quaternary – Reconstruction of Material Budget and Current Systems, pp. 1--19. Berlin: Springer. Tarantola A (1983) Inverse Problem Theory: Methods for Data Fitting and Model Parameter Estimation. New York: Elsevier. Thacker WC and Long RB (1988) Fitting dynamics to data. Journal of Geophysical Research 93: 1227--1240. Wunsch C (1978) The North Atlantic general circulation west of 501 W determined by inverse methods. Reviews of Geophysics 16: 583--620. Wunsch C (1996) The Ocean Circulation Inverse Problem. Cambridge, UK: Cambridge University Press.

Relevant Websites http://cdiac.esd.ornl.gov – Carbon Dioxide Information Analysis Center. http://whpo.ucsd.edu – Clivar and Carbon Hydrographic Data Office. http://www.coriolis.eu.org – Coriolis. http://www.ewoce.org – eWOCE: Electronic Atlas of WOCE Data. http://www.nodc.noaa.gov – National Oceanographic Data Center, NOAA. http://odv.awi.de – Ocean Data View.

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELS C. Wunsch, Massachusetts Institute of Technology, Cambridge, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1368–1374, & 2001, Elsevier Ltd.

Introduction

w

Inverse methods are formal procedures for making inferences about the ocean (or any other physical system) by using observations in combination with dynamical and kinematic models. As such, they are a part of the general methods of statistical inference done in the presence of known or assumed kinematic and dynamical constraints. Their first use in oceanography occurred in 1977 as a method for addressing the famous so-called level-of-no-motion problem in the oceanic general circulation, and the determination of the general circulation remains a major area of application. Subsequently, inverse methods became a central element of ocean acoustic tomography. More recently, they have begun to be applied widely in all areas of oceanography including biogeochemical problems. In mathematical usage, ‘inverse methods’ often describe procedures directed at solving a variety of ill-posed problems, in the absence of observational noise. Although the terminology is much the same, noise-free observations exist only in textbooks, and this literature is useful, but tangential, to the oceanographic problem. ‘Inverse methods’ are often used in conjunction with, and thereby confused with, ‘inverse models’. For historical reasons, and mathematical convention, one denotes many systems as being ‘forward’ or ‘direct’ problems or models. A simple example derives from a supposed theory that produces a rule for a variable, perhaps oceanic temperature, y, as a function of z, in the form y ð z Þ ¼ a 0 þ a1 z þ a 2 z 2 þ a3 z 3 ¼

3 X

ai z 3

½1

i¼0

If the theory tells us the ai, we can calculate y for any value of z. The theory is often labeled a ‘forward’ or ‘direct model’ and, more generally, may be either an analytical or a very complex numerical one. Equation [1] is called a ‘forward solution’. If, on the other hand, y(z) were known, but one or more of the ai were not, one would have an inverse model, also

312

given by equation [1], the label ‘inverse’ being employed only as a matter of convention – because there is a previously studied ‘forward’ version. Another example comes from the classical advection/diffusion/decay equation for a concentration tracer, C, @C @2C  k 2 þ lC ¼ qðzÞ @z @z

½2

where q is a source, w is the vertical advective velocity, k is a diffusion coefficient and l is a decay constant. All are known constants or functions of z. Given suitable boundary conditions, e.g., C(z ¼ 0)C0, C(z-N) ¼ 0, one has a well-understood, well-posed problem, in which eqn [2] and its boundary conditions are labeled as ‘forward.’ But an equally compelling, and commonplace problem is the following: Given C(z) and l, what are w and k? This type of problem occupies much of the mathematical literature on inverse problems. The oceanographers’ version is, however, more likely to be: Given observations, yi ¼ Cðzi Þ þ ni

½3

where ni are noise, at a finite number of discrete positions zi, what are w, k? Many variations of this problem are possible; for example, given noisy or uncertain observations or estimates of C, w, k, l what was the boundary condition C0? In this context, equation [2], along with any other available information, would now be described as the ‘inverse model,’ with the formal knowns and unknowns of the forward problem having in part, been interchanged. Parameter determination from noisy observations goes back at least to Legendre and Gauss. The modern generalization, inverse theory, is often traced to the pioneering work of G. Backus and F. Gilbert in the solid-earth geophysics context. They initiated the study of model systems in which the number of formal unknowns greatly exceeds (perhaps infinitely so), the number of available data by exploiting the existence of an underlying differential or partial differential system. This form of inverse theory is rooted in functional analysis and is highly developed; see Parker (1994), or for oceanographic applications, Bennett (1992), in the Further Reading. In oceanographic practice, even the simplest models are almost always reduced to some numerical, discrete form, either by expansion into a finite set of

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELS

modes or by finite differences or related methods. The polynomial form eqn [1] is already discrete, with four parameters. Many other forms are possible. For example eqn [2], when written in finite differences, becomes Ciþ1  Ci Ciþ1  2Ci þ Ci1  ki þ lCi ¼ qi ; ziþ1  zi ziþ1  2zi þ zi1 1rirM wi

½4

which is a set of M simultaneous linear equations in the finite discrete set wi, ki, l. Alternatively, if w, k are constant, an analytic solution, subject to a fixed surface concentration C(z ¼ 0) ¼ C0 is readily seen to be (

"   #) w 4lk 1=2 ; 1þ 1þ 2 CðzÞ ¼ C0 exp z 2k w

zr0 ½5

yi ¼ Ci þ ni, the problem has become one of finding an estimate, x˜, satisfying, Ex þ n ¼ y

"   #) w 4lk 1=2 1þ 1þ 2 Cðzi Þ ¼ C0 exp zi 2k w

Ciþ1  Ci xiþ1  xi B ziþ1  zi ziþ1  zi ½6

Because eqns [4] and [6] are algebraic equations in w, k, l the problem of determining them has been reduced from that of an infinite-dimensional Hilbert or Banach space to that of an ordinary finite-dimensional vector space. No matter how great the value of M, the corresponding mathematical simplifications render the methods of inverse theory much more transparent in this case. Reduction of the functional analysis methods of Backus and Gilbert to that of finite-dimensional spaces appears to begin with Wiggins, who used the singular value decomposition of Eckart and Young. In practice, almost all real inverse problems are solved on computers; they are thus automatically discrete and of finite dimension and this mathematical representation is the most useful one. Note that eqn [6] is nonlinear in the parameters, showing that the same problem can be rendered linear depending upon exactly how it is formulated.

Example Assuming then, that observations always have a noise component, we can proceed to estimate whatever is known. For the simple power law of eqn [1], let us suppose that the ‘truth’ is yðzÞ ¼ 30  0:0005z2 ;

100r z r0

½7

but the theory says that it could actually be of the general form (1) so that the correct answer would be a0 ¼ 30, a2 ¼ 5  104, ai ¼ 0, otherwise. Defining

½8

Here the matrix vector notation x ¼ [ai] (called the ‘state vector’), n ¼ [ni], y ¼ [yi] is being used, and E is the coefficient matrix. As stated, this is now a problem in polynomial regression theory, and much of inverse theory overlaps that branch of statistics. The tracer problem (eqn [2]) can be reduced to this same form. Suppose that wi, ki, are believed constant, independent of i, and that we have noisy measurements xi [3] of Ci. Then one can readily make the substitution Ci-xi in eqn [4], producing a set of simultaneous equations for w, ki, where the coefficient matrix E has elements depending upon

which can be evaluated at z¼ zi to produce (

313

½9

etc. and y now involves lxi, qi, etc. n is a noise-vector representing the errors introduced by the observational noise, and any misrepresentation of the true full physics by eqn [2]. In practice, the structure of n can be a complicated function of the structure of E, because equations such as (9) render the problem, in a rigorous sense, nonlinear, too. This nonlinearity is often ignored and there are many inverse problems where E is known exactly. Equation [8] can now be regarded as the ‘model’ instead of, or in addition to, eqn [2]. The state vector consists of the two unknowns xT ¼ [w, ki]T (superscript T denotes a transpose; all vectors here are column vectors). The form of [8] obviously generalizes to any number of unknown elements, N, in x. A common method for dealing with equations of this type is to seek a least-squares solution that renders the noise vector n as small as possible: J ¼ nT n ¼ ðy  ExÞT ðy  ExÞ

½10

By taking the derivatives of the ‘objective’ or ‘cost’ function J with respect to the elements of x and setting them to zero, one obtains the ‘normal equations.’ Here, one finds, using this matrix notation, that the so-called normal equations are ET Ex ¼ ET y

½11

 1 x˜ ¼ ET E y

½12

whose solution is

and which permits an estimate of the noise unknowns,  1 ½13 n˜ ¼ y  Ex˜ ¼ y  E ET E y

(c) 2011 Elsevier Inc. All Rights Reserved.

314

INVERSE MODELS

The forward model [1] was used to generate ‘data’ at seven observation points, and these values were then corrupted with noise having a standard deviation of 1. The least-squares solution produces a state vector x˜T ¼ [31, 0.02, 0.002, 6  105, 4  107], and the corresponding estimated y˜ is shown in Figure 1. It passes near the data points but is clearly not ‘correct’ in the sense of reproducing the known true values. Although a very easy-to-use and common parameterdetermining procedure, least-squares in this form is not an inverse method (statisticians call it ‘curve-fitting’). The reasons for seeking a more powerful method are easy to see. At least as written, one can raise a number of questions about the solution: 1. Why should the smallest mean-square noise, nTn, be regarded as the correct solution to choose? 2. What would happen if the inverse of ETE failed to exist? (E corresponding to [7] is a Vandermonde matrix, and known to be very badly conditioned.) 3. Suppose one knew that some of the noise values were likely to be much greater than others: Could that information be used? 4. Suppose one had a reasonable idea of the magnitude of the ai, i.e., of x: Cannot that information be used? 5. Just how reliable is the solution given the noisiness of the data, and the particular structure of the observations? 6. Are some observations more important than others in determining the solution? 7. Could one require k>0 in the advection–diffusion problem? 8. In general, there are M eqns in (8) and N elements of x. When M>N, the system appears to be comfortably overdetermined. But that comfort disappears

when one recognizes that N noise unknowns have also been determined in eqn [13], producing a total of M þ N previously unknown values. Is it still sensible to term the problem ‘overdetermined.’ 9. Problems such as the advection-diffusion one involve a quantity C that is necessarily positive and that often renders the noise statistics highly nonGaussian. How can one understand how that affects the best solution? These and other issues lead one to find other means to make an estimate of x. It is possible to modify the least-squares procedure so that it will produce solutions obtained by other methods; this correspondence has led, in the literature, to serious confusion about what is going on.

Two More Solution Methods There are a number of methods available for estimating the elements of x, n that produce useful answers to some or all of the questions listed above. We will briefly describe two such methods, leaving to the references listed in Further Reading (see Menke, 1989; Tarantola, 1987; Munk et al., 1995; Wunsch, 1996) both the details and discussion of other approaches. Gauss–Markov Method

We postulate some a priori knowledge of the elements of x, n in a common, statistical, form. In particular, we assume that /xS ¼ /nS ¼ 0, where the brackets denote the expected value, and that there is some knowledge of the second moments, S ¼ /xxT S; R ¼ /nnT S.

0

_ 20

_ 20

_ 40

_ 40

z

z

0

_ 60

_ 60

_ 80

_ 80

_ 100 24 (A)

26

28 θ

30

32

_ 100 24

26

(B)

28 θ

30

32

Figure 1 (A) A hypothetical temperature, y, profile with depth, z, drawn as a solid line. ‘Observations’ of that profile at discrete depths contaminated with unit noise variance errors are shown as open circles. Dashed line is the ordinary least-squares fit. (B) Same as in (A), except that the dashed line is the Gauss–Markov solution computed using a priori knowledge of the statistics of the noise and the solution itself. The fit is less structured, and a specific error estimate is available at every point, showing the true line to lie within the estimated uncertainty.

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELS

The so-called Gauss–Markov method (sometimes known as the ‘stochastic inverse’) produces a solution that minimizes the variance about the true value, i.e., it is a statistical method that minimizes not the sum of squares but, individually, all terms of the form /ðx˜ i  xi Þ2 S where xi is the true value, and x˜ i is the estimate made. The solution is  1 x˜ ¼ SET ESET þ R y

½14

One then computes the singular value decomposition E0 ¼ UKVT

All of the available information has been used. A solution for the simple example is shown in Figure 1, where S¼diag ([100, 11010, 1106, 11010, 11010]), R¼{dij}. One obtains, x˜ T ¼½3170:8;1:41010 71105 ;6:1104 75:1104 ;4:5106 79:1106 ;4:9108 76:9108 

½16

that is, the correct answer now lies within two standard errors and, although it is not displayed here, the P matrix provides a full statement of the extent to which the noise elements in x˜ are correlated with each other (which can be of the utmost importance in many problems). Least-squares by Singular Value Decomposition

As noted, least-squares can be modified so that it is more fully capable of producing answers to the questions put above. There are at least two ways to do this. The more interesting one is that based upon the singular value decomposition (SVD) and the socalled Cholesky decomposition of the covariance matrices, S ¼ ST/2S1/2, R ¼ RT/2R1/2. One takes the original eqn [18] and employs the matrices S, R to rotate and stretch E, n, y, x into new vector spaces in which both observations and solution have uncorrelated structures: RT=2 EST=2 ST=2 x þ RT=2 n ¼ RT=2 y

½17

E0 x0 þ n0 ¼ y0

½18

or

where E0 ¼ RT=2 EST=2 ; x0 ¼ ST=2 x; n0 ¼ RT=2 n; y0 ¼ RT=2 y

½19

½20

where K is a diagonal matrix, and U, V are orthogonal matrices. Let there be K nonzero values on the diagonal of K, and let the columns of U, V be denoted ui, vi respectively. Then it can be shown in straightforward fashion that the uTi uj ¼ dij ; vTi vj ¼ dij and that the solution to [18] is

with solution uncertainty (error estimate)  1 P ¼ /ðxx Þðxx ÞT S ¼ SSET ESET þR ES ½15 ˜ ˜

315

x0 ¼

K X k¼1

vk

N X uTk y0 þ ak vk lk k¼Kþ1

½21

where the ak are completely arbitrary. The physical solution is obtained from x˜ ¼ ST=2 x˜ 0 . A major advantage of this solution (and the uncertainly matrix, P, can also be computed for it), is that it explicitly produces the solution in orthonormal structures, vi, in terms of orthonormal structures of the observations ui, and separates these elements from the second sum on the right of equations [21], which defines the so-called null space of the problem. The null space represents the structures possibly present in the true solution, about which the equations [18] carry no information. A null space is present too, in the Gauss–Markov solution, but the corresponding null space vectors are given coefficients ak so as best to reproduce the a priori values of S. The SVD leads naturally to a discussion of what is called ‘resolution’ both of data and of the solution, and provides complete information about the solution. These issues, with examples, are discussed in Wunsch (1996) (see Further Reading). A Hydrographic Example

The first use of inverse methods in oceanography was the hydrographic problem. In this classical problem, oceanographers are able to determine the density field, r(z) at two nearby ‘hydrographic stations’ where temperature and salinity were measured by a ship as functions of depth. By computing the horizontal differences in density as a function of depth, and invoking geostrophic balance, an estimate could be made of the velocity field, up to an unknown constant (with depth) of integration. Although mathematically trivial, the inability to determine the integration constant between pairs of stations plagued oceanography for 100 years, and led to the employment of the ad hoc assumption that the flow field vanished at some specific depth z0. This depth became known as the ‘level-of-no-motion.’ But by employing physical requirements such as mass and salt conservation (or any other constraint involving

(c) 2011 Elsevier Inc. All Rights Reserved.

316

INVERSE MODELS

B (cm s )

35

_1

_1

B (cm s )

the absolute velocity), it is straightforward to reduce the problem to one of deducing the actual flow field at the level-of-no-motion (and which is thus better called the ‘reference-level’). These constraints are readily written in the form of eqn [8], and solved with error estimates, etc. An example of the practical application of this method is shown in Figure 2. In this calculation, constraints were written for mass and salt conservation in a triangular region bounded by the US coastline, and a pair of hydrographic sections in which the Gulf Stream flowed into the region across one section and out again in the other. Direct velocity measurements obtained from the ship permitted additional equations to be written for the unknown flow velocity at the reference level. The resulting estimated absolute flow at the reference level is shown, with error estimates, in the top panels. The lower panels show the total estimated absolute velocity. In general, any available information can be used to estimate the unknowns of the system as long as one has a plausible estimate of the error contained in the resulting constraint. For the hydrographic problem in particular, a number of variations on the constraints have been proposed, under the labels of

_ 15 42 0

140

45

‘beta-spiral,’ ‘Bernoulli-method,’ etc., which we must leave to the references listed in Further Reading. Extensions

Like eqn [6], many inverse problems are nonlinear. Tarantola (1987) provides some general background and specific oceanographic applications may be seen in Mercier (1989), Mercier et al. (1991), and Wunsch (1994). The use of inequality constraints leads to the general subject of mathematical programming, a part of the wide subject of optimization theory (see Arthnari and Dodge, 1981). The Gauss–Markov solution method and the SVD version of least-squares have a ready interpretation, as minimum variance estimates of the true field. If the fields are all normally distributed, the solutions are also maximum-likelihood estimates, a methodology that is readily extended to nonGaussian fields.

Time-dependent Problems As originally formulated by Backus and Gilbert, and as exploited in most of the oceanographic literature

35 _ 15

49

61 60 58 _ 5 0

100

_ 10

80

55

50 49

160 100

_ 10

40 20 10

1000

_5 5

_5

60

5

1000

20

2000

Pressure (db)

Pressure (db)

10 _5

5

10

_ 10

3000

2000 5

3000 _ 15 _5

4000 _ 20

20 km

4000

20 km

Figure 2 Example of the inversion for the reference level velocity in a triangular region bounded on two sides by hydrographic sections crossing the Gulf Stream, and on the third side by the US coastline. Velocity contours are in centimeters. They are the sum of the so-called thermal wind, which involves setting the velocity to zero at a reference depth. The actual flow at that depth (the ‘reference level velocity’) is shown in the top panels as estimated, with uncertainties, from the singular value decomposition solution, with K ¼ 30. (From Joyce et al., 1986.) The ‘columnar’ structure, which is so apparent here, first appeared in inverse solutions, and was greeted with disbelief by those who ‘knew’ that oceanic flows were ‘layered’ in form. Acceptance of this type of structure is now commonplace.

(c) 2011 Elsevier Inc. All Rights Reserved.

INVERSE MODELS

to date, the problems have been essentially static, with time evolution not accounted for. If one has a system that evolves in time, the observations and state vector, x, will also be time evolving. If one simply writes down the relationship between data and a model in which the two are connected by a set of equations, linear or nonlinear, one sees immediately that, mathematically, the problem is identical to that posed by eqn [8], in which the time of the observation or of the model calculation is just a bookkeeping index. The difficulty that arises is purely a practical one: the potentially extremely large growth in the number of equations that must be dealt with over long time spans. With a sufficiently large and fast computer, the most efficient way to solve such inverse problems would be the straightforward application of the same methods developed for static problems: Write out all of the equations explicitly and solve them all at once. In oceanographic practice however, one rapidly outstrips the largest available computers and the static methods become impractical if used naively. Fortunately, time-evolving oceanographic models have a very special algebraic structure, which permits one to solve the corresponding normal equations [11] by methods not requiring storage of everything in the computer at one time. Numerous such special methods exist, and go by names such as sequential estimators (Kalman filter and various smoothers), adjoint equations (Pontryagin principle or method of Lagrange multipliers), Monte Carlo methods, etc. The generic terminology that has come into use is ‘data assimilation’ borrowed from meteorological forecasting terminology. Such methods are highly developed, if often abused or misunderstood, and require a separate discussion. What is important to know, however, is that they are simply algorithmically efficient solutions to the inverse problem as described here.

Common Misconceptions and Difficulties Some chronic misunderstandings and difficulties arise. The most pernicious of these is the attempt to use inverse models, which are physically inconsistent with the known forward model or physics. This blunder corresponds, for example, in the simplest model above, to imposing a linear relationship (model) between y and z, when it is clear or suspected that higher powers of z are likely to be present. Some writers have gone so far as to show that such an inversion does not reproduce a known forward solution, and then declared inverse methods to be failures. Another blunder

317

is to confuse the inability to resolve or determine a parameter of interest, when the data are inadequate for the purpose, with a methodological or model failure. Inverse methods are very powerful tools. Like any powerful tool (a chain saw, for example), when properly used they are useful and even essential; when improperly used they are a grave danger to the user.

See also Data Assimilation in Models. Elemental Distribution: Overview. General Circulation Models. Tomography. Tracer Release Experiments.

Further Reading Arthnari TS and Dodge Y (1981) Mathematical Programming in Statistics. New York: Wiley. Bennett AF (1992) Inverse Methods in Physical Oceanography. Cambridge: Cambridge University Press. Eckart C and Young G (1939) A principal axis transformation for non-Hermitian matrices. Bulletin of the American Mathematical Society 45: 118--121. Joyce TM, Wunsch C, and Pierce SD (1986) Synoptic Gulf Stream velocity profiles through simultaneous inversion of hydrographic and acoustic doppler data. Journal of Geophysical Research 91: 7573--7585. Lanczos C (1961) Linear Differential Operators. Princeton, NJ: Van Nostrand. Liebelt PB (1967) An Introduction to Optimal Estimation. Reading, MA: Addison Wesley. Menke W (1989) Geophysical Data Analysis: Discrete Inverse Theory 2nd edn. New York: Academic Press. Mercier H (1989) A study of the time-averaged circulation in the western North Atlantic by simultaneous nonlinear inversion of hydrographic and current-meter data. DeepSea Research 36: 297--313. Mercier H, Ollitrault M, and Le Traon PY (1991) An inverse model of the North Atlantic general circulation using Lagrangian float data. Journal of Physical Oceanography 23: 689--715. Munk W, Worcester P, and Wunsch C (1995) Ocean Acoustic Tomography. Cambridge: Cambridge University Press. Parker RL (1994) Geophysical Inverse Theory. Princeton, NJ: Princeton University Press. Seber GAF (1977) Linear Regression Analysis. New York: Wiley. Tarantola A (1987) Inverse Problem Theory. Methods for Data Fitting and Model Parameter Estimation. Amsterdam: Elsevier. Van Trees HL (1968) Detection, Estimation and Modulation Theory. Part I. New York: Wiley. Wiggins RA (1972) The general linear inverse problem: implication of surface waves and free oscillations for earth structure. Reviews in Geophysics and Space Physics 10: 251--285.

(c) 2011 Elsevier Inc. All Rights Reserved.

318

INVERSE MODELS

Wunsch C (1977) Determining the general circulation of the oceans: a preliminary discussion. Science 196: 871--875. Wunsch C (1994) Dynamically consistent hydrography and absolute velocity in the eastern North

Atlantic Ocean. Journal of Geophysical Research 99: 14071--14090. Wunsch C (1996) The Ocean Circulation Inverse Problem. Cambridge: Cambridge University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

IR RADIOMETERS C. J. Donlon, Space Applications Institute, Ispra, Italy

principles described are applicable to satellite sensors treated elsewhere in this volume.

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1374–1385, & 2001, Elsevier Ltd.

Introduction Measurements of sea surface temperature (SST) are most important for the investigation of the processes underlying heat and gas exchange across the air–sea interface, the surface energy balance, and the general circulation of both the atmosphere and the oceans. Complementing traditional subsurface contact temperature measurements, there is a wide variety of infrared radiometers, spectroradiometers, and thermal imaging systems that can be used to determine the SST by measuring thermal emissions from the sea surface. However, the SST determined from thermal emission can be significantly different from the subsurface temperature (471 K) because the heat flux passing through the air–sea interface typically results in a strong temperature gradient. Radiometer systems deployed on satellite platforms provide daily global maps of SSST (sea surface temperature) at high spatial resolution (B1 km) whereas those deployed from ships and aircraft provide data at small spatial scales of centimeters to meters. In particular, the development of satellite radiometer systems providing a truly synoptic view of surface ocean thermal features has been pivotal in the description and understanding of the global oceans. This article reviews the infrared properties of water and some of the instruments developed to measure thermal emission from the sea surface. It focuses on in situ radiometers although the general

Infrared Measurement Theory Infrared (IR) radiation is heat energy that is emitted from all objects that have a temperature above 0 K (  273.161C). It includes all wavelengths of the electromagnetic spectrum between 0.75 mm and B100 mm (Figure 1) and has the same optical properties as visible light, being capable of reflection, refraction, and forming interference patterns. The following total quantities, conventional symbols and units provide the theoretical foundation for the measurement of IR radiation and are schematically shown in Figure 2. Spectral quantities can be represented by restricting each to a specific waveband.

• • •

• •

Radiant energy Q, is the total energy radiated from a point source in all directions in units of joules (J). Radiant flux f ¼ dQ/dt is the flux of all energy radiated in all directions from a point source in units of watts (W). Emittance M ¼ df/dA is the radiant flux density from a surface area A in units of W m2. This is an integrated flux (i.e., independent of direction) and will therefore vary with orientation relative to a nonuniform source. Radiant intensity I ¼ df/do is the radiant flux of a point source per solid angle o (steradian, sr) and is a directional flux in units of W sr1. Radiance L ¼ dI/d(A cos y) is the radiant intensity of an extended source per unit solid angle in a given direction y, per unit area of the source projected in the same y. It has units of W sr1 m2.

Figure 1 Schematic diagram of the electromagnetic spectrum showing the location and interval of the infrared waveband.

(c) 2011 Elsevier Inc. All Rights Reserved.

319

320

IR RADIOMETERS

y (surface normal) Radiant flux 

Surface area, A

x

Figure 3 Black-body radiance as a function of wavelength computed using eqns [1] and [2] for different target temperatures. z (A) y Area = A Radiant flux 

Solid angle 2  = A /R

where h ¼ 6.626  1034 J s is Planck’s constant, c ¼ 2.998  108 m s1 is the speed of light and k ¼ 1.381  1023 J K1 is Boltzmann’s constant. The sea surface is considered a Lambertian source (i.e., uniform radiance in all directions) so that the spectral radiance Ll is related to Ml by Ll ¼

R x

z (B)

y (surface normal) d

Surface area, A

Radiant flux 

 Direction,  x Projected source area, dA cos 

z (C)

Figure 2 Schematic definition of (A) Emittance, E; (B) radiant intensity, I; (C) radiance, L.

Planck’s law describes the emittance of a perfectly emitting surface (or black body) at a temperature T in Kelvin. It is the radiant flux (f) per unit bandwidth centered at wavelength l leaving a unit area of surface in any direction in units of W m2 m1. Ml;T ¼

2phc2 l5 ðehc=lkT  1Þ

Ml p

½2

Figure 3 shows Ll computed for several temperatures as a function of wavelength. Considering temperatures of 273–310 K as representative of the global ocean, maximum emission occurs at a wavelength of 9.3–10.7 mm. Atmospheric attenuation is minimal at B3.5 mm, 9.0 mm and 11.0 mm that are the spectral intervals often termed atmospheric ‘windows’. Instruments operating within these intervals are optimal for sea surface measurements – especially in the case of satellite deployment where atmospheric attenuation can be significant. In the 3–5 mm spectral region. Ll is a strong function of temperature (Figure 3) highlighting the possibility to increase in radiometer sensitivity by utilizing this spectral interval. By measuring Ll using eqn [2] and inverting eqn [1], the spectral brightness temperature, B(T,l), rather than the temperature is determined because these equations assume that sea water is a perfect emitter or black body. In practice, the sea surface does not behave as a black body (it is slightly reflective in the infrared) and therefore its spectral and geometric properties need to be considered. The emittance of a perfect emitter at the actual temperature T, wavelength l, and view angle y is given by

½1

(c) 2011 Elsevier Inc. All Rights Reserved.

MðT;l;yÞ ¼

pLðT;l;yÞ eðl;yÞ

½3

IR RADIOMETERS

B11 mm (rE0.0015) and following Kirchoff’s law

where the emissivity, eðl;yÞ , can be calculated using eðl;yÞ ¼

rl þ tl þ el ¼ 1

MðT;l;yÞ measured MðT;l;yÞ blackbody

321

½5

½4

which has a strong dependence on wavelength and viewing geometry. The effective emissivity, e, integrates eðl;yÞ over all wavelengths of interest for radiometer view angle y. Figure 4A shows the calculated normal reflectivity, r, of pure water as a function of wavelength for the spectral region 1–100 mm. Pure water differs only slightly from sea water in this context. Note that r is minimal at a wavelength of

where rl is the spectral reflectivity and tl is the spectral transmissivity. The e-folding penetration depth (i.e., the depth of 63% emission) or optical depth at a typical wavelength of 11 mm is E 10 mm, tl can be neglected and el can be calculated using el ¼ 1  rl

½6

Although the actual optical depth is wavelength dependent, it is clear that IR radiometers determine

 (zenith angle = 40˚)

0.100

0.010

(h+v) h v 0.001

10

100 Wavelength (μm)

(A)

Sea surface emissivity ( = 1 _ )

1.00

0.95

0.90

0.85

0.80 (B)

0˚ 40˚ 50˚ 60˚ 70˚

10

100 Wavelength (μm)

Figure 4 (A) The normal reflectivity, r, of the pure water as a function of wavelength (full line). Also shown are the horizontal polarized (rh) and the vertically polarized (rv) components of r. (B) The spectral emissivity of pure water as a function of viewing zenith angle. h, height above sea surface.

(c) 2011 Elsevier Inc. All Rights Reserved.

322

IR RADIOMETERS

Figure 5 Schematic diagram showing the radiance components measured by an IR radiometer viewing the sea surface. Fov, Field of view.

the temperature of a very thin ‘skin’ layer of the ocean. This temperature is termed the sea surface skin temperature (SSST) and is distinct (although related) to the subsurface SST. Note that this is in contrast to the situation for short-wave solar radiation (having wavelengths of B0.4–0.7 mm) which, for clear water, penetrates to a depth of B100 m. Figure 4B shows the el of pure water computed from eqn [6] as a function of both viewing zenith angle and spectral wavelength. Inspection of Figure 4 reveals that the best radiant temperature measurement will be made when viewing a calm sea surface at an angle of 0–401 from nadir. An IR radiometer is an optical instrument designed to measure Ll entering an instrument aperture (Figure 5). The radiance measured by radiometer, LðT;l;yÞ , having a spectral bandwidth l, viewing the sea surface at a zenith angle y and temperature T is given by: LðT;l;y ¼

ða

    xl ½eðl;yÞ B Tsurf ; l þ 1  eðl;yÞ BðTatm ; lÞ

0

surface because diffuse downwelling sky radiance measured by a radiometer after reflection at the sea surface is polarized.

IR Radiometer Design There are four fundamental components to all IR radiometer instruments described below. Detector and Electronics System to Measure Radiance and Control the Radiometer

A detector system provides an output proportional to the target radiance incident on the detector. There are two main types of detector: thermal detectors that respond to direct heating and quantum detectors that respond to a photon flux. In general, thermal detectors have a response that is weakly dependent on wavelength and can be operated at ambient temperatures whereas rapid response quantum detectors require cooling and are wavelength dependent.

þLpathðh;l;yÞ dl ½7 where xl is the spectral response of the radiometer, BðTsurf ; l and BðTatm ; l are the Planck function for surface temperature Tsurf, and atmospheric temperature Tatm, and Lpathðh;l;yÞ is the radiance emitted by the atmosphere between the radiometer at height h above the sea surface reflected into the radiometer field-of-view (FoV) at the sea surface. Note that the horizontal and vertical polarization components of reflectivity shown in Figure 4 are unequal. It is important to consider the polarization of surface reflectance when making measurements of the sea

Fore-optics System to Filter, Direct and Focus Radiance

All optical components have an impact on radiometer reliability and accuracy. Mirrors should be free of aberration to minimize unwanted stray radiance reaching the detector. Several materials have good reflection characteristics in the IR including, gold, polished aluminum, and cadmium. Care should be exercised when choosing an appropriate mirror substrate and reflection coating to avoid decay in the marine atmosphere. A glass substrate having a ‘hard’ scratch-resistant polished gold surface provides 498% reflectance and good environmental wear.

(c) 2011 Elsevier Inc. All Rights Reserved.

IR RADIOMETERS

323

1.0 Detector transmission Window transmission Total transmission

Normalized transmission

0.8

0.6

0.4

0.2

0.0 6

8

10 12 Waveband (μm)

14

16

Figure 6 Normalized spectral transmission of an IR window and detector shown together with the combined total response.

Spectral filter windows and lenses require spectral properties that, together with the detector characteristics, define the overall spectral characteristics of a radiometer. Figure 6 shows the combined spectral response for a broadband radiometer together with the component spectral response of the window and detector. Environmental System to Protect and Thermally Stabilize the Radiometer

For any optical instrument intended for use in the harsh marine environment, adequate environmental protection is critical. Rain, sea-water spray, and high humidity can destroy a poorly protected instrument rapidly and components such as electrical connections and fore-optics should be resistant to these effects. In situ radiometer windows are particularly important in this context. They should not significantly reduce the incoming signal or render it noisy, and be strong enough to resist mechanical, thermal and chemical degradation. Figure 7 describes several common materials. Germanium (Ge) windows have good transmission characteristics but are very brittle. Sodium chloride (NaCl) is a low-cost, low-absorption material but is of little use in the marine environment because it is water-soluble. Zinc selenide (ZnSe) has high transmission, is nonhygroscopic

Figure 7 Spectral transmission for common IR window materials. (A) Germanium, (B) sodium chloride, (C) zinc selenide.

and resistant to thermal shock but is soft and requires a protective ‘hard’ surface finish coating. Certain window materials achieve better performance when antireflection (AR) coatings are used to minimize reflection from the window. For example, when an AR coating is used on a ZnSe window the transmission increases from B70% to B90%. Other coatings provide windows that polarize the incident radiance signal such as optically thin interference coatings and wire grid diffraction polarizers. Deposition of marine NaCl on all optical components (especially calibration targets) presents an unavoidable problem. Although NaCl itself has good infrared transmission properties (Figure 7), contaminated surfaces may become decoupled from temperature sensors and the noise introduced by the optical system will increase. Finally, adequate thermal control using reflective paint together with substantial instrument mass is required so that instruments are not sensitive to thermal shock. In higher latitudes, it may be necessary to provide an extensive antifreeze capability. Calibration System to Quantify the Radiometer Output

The role of a calibration system is to quantify the instrument output in terms of the measured radiance incident on the detector. Calibration techniques are specific to the particular design of radiometer and vary considerably from simple bias corrections to

(c) 2011 Elsevier Inc. All Rights Reserved.

324

IR RADIOMETERS

systems providing automatic precision two-point blackbody calibrations. Proper calibration accounts for the following primary sources of error:

• • •

the effect of fore-optics; unavoidable drifts in detector gain and bias; long-term degradation of components.

Finally, careful radiometer design and configuration can avoid many measurement errors. Examples include: poor focusing and optical alignment; filters having transmission above or below the stated spectral bandwidth (termed ‘leaks’); inadequate protection against thermal shock, stability, and reliability of the calibration system; poor electronics and the general decay of optomechanical components.

Application of IR Radiometers There are many different radiometer designs and deployment strategies ranging from simple singlechannel hand-held devices to complex spectroradiometers. Instrument accuracy, sensitivity and stability depends on both the deployment scenario and radiometer design. Accordingly, the following sections describe several different examples. Broadband Pyrgeometers

A pyrgeometer is an integrating hemispheric radiometer which, by definition, measures the bandlimited spectral emittance, E, so that the angle dependency in eqn [7] is redundant. They are used to determine the long-wave heat flux at the air–sea interface by measuring the difference between atmospheric and sea surface radiance either using two individual sensors (Figure 8A) or as a single combined sensor (Figure 8B). A thermopile detector is often used which is a collection of thermocouple detectors composed of two dissimilar metallic conductors connected together at two ‘junctions.’ The measurement junction is warmed by incident

radiance relative to a stable reference junction and a mV signal is produced. The response of a thermopile has little wavelength dependence and a hemispheric dome having a filter (B3–50 mm) deposited on its inner surface typically defines the spectral response. Direct compensation for thermal drift using a temperature sensor located at the thermopile reference junction is sometimes used but regular calibration using an independent laboratory blackbody is mandatory. An accuracy of o10 W m2 is possible after significant correction for instrument temperature drift and stray radiance contribution using additional on-board temperature sensors. Narrow Beam Filter Radiometers

Narrow beam filter radiometers are often used to determine the SSST for air–sea interaction studies and for the validation of satellite derived SSST and there are several low-cost instruments that provide suitable accuracy and spectral characteristics. Many of these use a simple thermopile or thermistor detector together with a low-cost broadband-focusing lens. Typically, they have simple self-calibration techniques based on the temperature of the instrument and/or detector. Consequently they have poor resistance to thermal shock and have fore-optics that readily degrade in the marine atmosphere. However, handled with care, these devices are accurate to 70.1 K, albeit with limited sensitivity. Precision narrow beam filter radiometers often use pyroelectric detectors that produce a small electrical current in response to changes in detector temperature forced by incident radiation. They have a fast response at ambient temperatures but require a modulated signal to operate. Modulation is accomplished by using an optical chopper having high reflectivity ‘vanes’ to alternately view a reference radiance source by reflection and a free path to the target radiance. The most common chopper systems are rotary systems driven by a small electric motor phase locked to the

Figure 8 (A) A typical design of a long-wave pyrgeometer. (B) A net-radiation pyrgeometer for determination of the net long-wave flux at the sea surface.

(c) 2011 Elsevier Inc. All Rights Reserved.

IR RADIOMETERS

325

Figure 9 (A) Schematic layout of a rotary chopper; (B) schematic layout of a tuning fork chopper; (a) open; (b) closed.

detector output by an optoelectronic sensor Figure 9A. An alternative design driven by a small oscillating electromagnetic coil called a tuning fork chopper is shown in Figure 9B. As the coil resonates, the reflective vanes of the chopper oscillate alternately opening and closing an aperture ‘gap.’ Dynamic detector bias compensation is inherent when using an optical chopper. The detector alternately measures radiance from the sea surface Lsrc and a reference blackbody, Lbb (sometimes this is the detector itself) reflected by the chopper vanes resulting in two signals S1 ¼ Lbb þ d

½8

S2 ¼ Lsrc þ d

½9

Assuming Lbb remains constant during a short chopping cycle, the bias term d in eqns [8] and [9] is eliminated DS ¼ S1  S2 ¼ Lbb  Lsrc

½10

It is important to recognize the advantages to this technique, which is widely used:

• • •

There is minimal thermal drift of the detector; The detector is dynamically compensated for thermal shock; A precise modulated signal is generated well suited to selective filtering providing excellent noise suppression and signal stability.

However, in order to compensate for instrument gain changes, an additional blackbody sources(s) is required. These are periodically viewed by the detector to provide a mechanism for absolute calibration. Either the black body is moved into the detector FoV or an adjustable mirror reflects radiance from the black body on to the detector.

Calibration cycles should be made at regular intervals so that gain changes can be accurately monitored and calibration sources need to be viewed using the same optical path as that used to view the sea surface. A basic ‘black-body’ calibration strategy uses an external bath of sea water as a high e (40.95) reference as shown in Figure 10. In this scheme, the radiometer periodically views the water bath that is stirred vigorously to prevent the development of a thermal skin temperature deviation. The view geometry for the water bath and the sea surface are assumed to be identical and, by measuring the temperature of the water bath the radiometer can be absolutely calibrated. An advantage of this technique is that eðl;yÞ is not required to determine the SSST. However, in practice, it is difficult to continuously operate a water bath at sea and surface roughness differences between the bath and sea surface are ignored. On reflection at the sea surface, diffuse sky radiance is polarized and, at Brewster’s angle (B501 from nadir at a wavelength of 11 mm), the vertical vpolarization is negligible for a given wavelength (Figure 11). Only the horizontal h-polarization component remains so that if the radiometer filter response is v-polarized (i.e., only passes v-polarized radiance), negligible reflected sky radiance is measured by the radiometer. In practice, because Brewster’s angle is very sensitive to the geometry of a particular deployment (approximately721) this technique is only applicable to deployments from fixed platforms and when the sea surface is relatively calm. Further, the use of a polarizing filter will significantly reduce the signal falling on the detector increasing the signal-to-noise ratio. The use of fabricated black-body cavities (Figure 12A) provides an accurate, versatile and, compact calibration system. Normally, two

(c) 2011 Elsevier Inc. All Rights Reserved.

326

IR RADIOMETERS

Radiometer

Radiometer beam

Water overflow

Inner water bath

Water bath moved out from radiometer FoV which no views the sea surface

Gimball mounting

Overflow drain

Sea-water input

(A)

(B)

Figure 10 Schematic diagram showing the stirred water bath calibration scheme. (A) The radiometer in calibration mode, (B) the radiometer viewing the sea surface after the water bath has been moved out of the field of view (FoV).

black-body cavities are used, one of which follows the ambient temperature of the instrument and a second is heated to a nominal temperature above this. High e (40.99) is attained by a combination of specialized surface finish and black-body geometry. The cavity radiance is determined as a function of the black body temperature that is easily measured. Figure 12B shows a schematic outline of a typical black-body radiometer design using a rotary chopper

and Figure 12C provides a schematic diagram of a typical output signal. Note that for all calibration schemes, larger errors are expected beyond the calibrated temperature range which can be a problem for sky radiance measurements where clear sky temperatures of o200 K are common.

Multichannel Radiometers  at  = 11μm for angles 0 _ 90˚

1.0000

 at  = 11 μm

0.1000

0.0100

0.0010

h v 0.0001

0

20

40 60 Zenith angle (˚)

80

Figure 11 Polarization of sea surface reflection at 11 mm as a function of view angle. Total polarization is shown as a solid line.

The terms in eqn [7] are directly influenced by the height, h, of the radiometer above the sea surface and are different in magnitude for in situ and spacecraft deployments. In the case of a sea surface in situ radiometer deployment, Lpathðh;l;yÞ is typically neglected because h is normally o10 m unless the atmosphere has a heavy water vapor loading (e.g., 490%) or an aircraft deployment is considered. However, for a spacecraft deployment, this is a significant term requiring explicit correction. Conversely, the BðTatm ; lÞ term is critical to the accuracy of an in situ radiometer deployment but of little impact (except perhaps at the edge of clouds) for a satellite instrument deployment because Lpathðh;l;yÞ dominates the signal. A multispectral capability can be used to explicitly account for Lpathðh;l;yÞ in eqn [7] because of unequal atmospheric attenuation for different spectral wavebands. Multichannel radiometers are exclusively used on satellite platforms for this reason. Many in situ multichannel radiometers are designed primarily for the radiant calibration or validation of specific satellite radiometers and the development of satellite radiometer atmospheric correction algorithms. They often have several selectable filters matched to those of the satellite sensor.

(c) 2011 Elsevier Inc. All Rights Reserved.

IR RADIOMETERS

327

Figure 12 (A) A section through a black-body calibration cavity of the re-entrant cone design. (B) A typical black-body calibration radiometer using a rotary optical chopper. (C) A schematic diagram of a typical detector output signal, showing sea, reference, and calibration signals.

It is worth noting that multiangle view radiometers are also capable of providing an explicit correction for atmospheric attenuation. Often operated from satellite and aircraft, these instruments provide a direct measure of atmospheric attenuation by making two views of the same sea surface area at different angles using a geometry that doubles the atmospheric pathlength (Figure 13). The assumption is made that atmospheric and oceanic conditions are stationary in the time between each measurement.

or satellite

Spectroradiometers

A recent development is the use of Fourier transform infrared spectrometers (FTIR) that are capable of accurate (B0.05 K). High spectral resolution

1+dt

Figure 13 Schematic diagram of dual view, atmospheric path radiometer deployment geometry.

(c) 2011 Elsevier Inc. All Rights Reserved.

e=t1

double

328

IR RADIOMETERS

(B0.5 cm1) measurements over a broad spectral range (typically B3–18 mm) as shown in Figure 14. The FTIR provides a unique tool for the development of new IR measurement techniques and investigation of processes at the air–sea interface. For example, the Marine-Atmospheric Emitted Radiance Interferometer (M-AERI) has pioneered a SSST algorithm that uses a narrow spectral region centered at 7.7 mm that is less susceptible to the influence of cloud cover and sky emissions that at 10–12 mm. The FTIR can also be used to measure air temperatures by viewing the atmosphere at B15 mm (a spectral region opaque due to CO2 emission) that are accurate to o0.1 K. Of considerable interest is the ability of an FTIR provide an indirect estimate of rðl;yÞ so that by using eqn [5], eðl;yÞ can be computed. The sky radiance spectrum has particular structures associated with atmospheric emission–absorbance lines (Figure 14A) that are physically uncorrelated with the smooth spectrum of rðl;yÞ (Figure 4). The spectrum of rðl;yÞ can be derived by subtracting a scaled BðTarm ; lÞ

spectrum to minimize the band-limited variance of the BðTsea ; lÞ spectrum (Figure 14B). Finally, direct measurement of the thermal gradient at the air–sea interface to obtain the net heat flux has been demonstrated using an FTIR in the laboratory. The FTIR uses the 3.3–4.1 mm spectral interval that has an effective optical depth (EOD) depending on the wavelength (EOD ¼ 0 mm at 3.3 mm whereas at 3.8 mm EOD ¼ 65 mm) demonstrating the versatility of the FTIR. However, measurement integration times are long and further progress is required before this technique is applicable for normal field operations. Thermal Imagers

Another recent development is the application of IR imagers and thermal cameras for high-resolution process studies such as fine-scale variability of SSST, wave breaking (Figure 15), and understanding air– sea gas and heat transfer. They are also used during

Figure 14 Spectra of emitted sky and sea view radiation measured by the M-AERI FTIR in the tropical Western Pacific Ocean on March 24, 1996. (A) Spectrum of sky radiance and (B) spectrum of corresponding sea radiance Sky measurements were made at 451 and zenith (red) above the horizon and ocean measurements were made at 451 below the horizon. The cold temperatures in the sky spectra show where the atmosphere is relatively transparent. The ‘noise’ in the 5.5–7 mm range is caused by the atmosphere being so opaque that the radiometer does not ‘see’ clearly the instrument internal black-body targets and calibration is void. The spectrum of upwelling radiation (B) consists of emission from the sea surface, reflected sky emission and emission from the atmospheric pathlength between the sea surface and the radiometer.

(c) 2011 Elsevier Inc. All Rights Reserved.

IR RADIOMETERS

329

Figure 15 A thermal image of a breaking wave. Each pixel is B25  25 cm and the wavelength of the camera is 8–12 mm. (Courtesy of D. Woolf.)

air–sea rescue operations providing a nighttime capability for detecting warm objects such as a life raft or survivors. These instruments use a focal plane array (FPA) detector (a matrix of individual detectors, e.g., 256  256) located at the focal point of incoming radiance together with a charge couple device that is used to ‘read’ the FPA. This type of ‘staring array’ system generates a two-dimensional image either by using a mechanical scanning system (in the case of a small FPA array) or as an instantaneous image. Larger FPA arrays are much more power efficient, lighter and smaller than more elaborate mechanical scanning systems. Rapid image acquisition (415 frames s1) is typical of these instruments that are available in a wide variety of spectral configurations and a typical accuracy of B70.1 K. However, considerable problems are encountered when obtaining sky radiance data due to the difficulty of geometrically matching sea and sky radiance data. The major problem with FPA detector technology is nonuniformity between FPA detector elements and drifts in detector gain and bias. Many innovative self-calibration methods which range in quality are used to correct for these problems. For example, a small heated calibration plate assumed to be at an isothermal temperature is periodically viewed by the detector to provide an absolute calibration. However, further development of this technology will

eventually provide extremely versatile instrumentation for the investigation of fine-scale sea surface emission.

Future Direction and Conclusions In the last 10 years, considerable progress has been made in the development and application of IR sensors to study the air–sea interface. The continued development and use of FTIR sensors will provide the capability to accurately investigate the spectral characteristics of the sea surface in order to optimize the spectral intervals used by space sensors to determine SSST. It can be expected that in the near future, new algorithms will emerge for the direct measurement of the air–sea heat flux using multispectral sounding techniques and the accurate in situ determination of sea surface emissivity. Although still in their infancy, the development and use of thermal cameras will provide valuable insight into the fine-resolution two-dimensional spatial and temporal variability of the ocean surface. These data will be useful in developing and understanding the sampling limitations of large footprint satellite sensors and in the refinement of validation protocols. Finally, as satellite radiometers are now providing consistent and accurate observations of the SSST (e.g., ATSR), there is a need for autonomous

(c) 2011 Elsevier Inc. All Rights Reserved.

330

IR RADIOMETERS

operational in situ radiometer systems for ongoing validation of their data. Such intelligent systems that are extremely robust against the harsh realities of the marine environment are currently being developed.

See also Air–Sea Gas Exchange. Heat and Momentum Fluxes at the Sea Surface. Radiative Transfer in the Ocean. Satellite Remote Sensing of Sea Surface Temperatures.

Further Reading Bertie JE and Lan ZD (1996) Infrared intensities of liquids: the intensity of the OH stretching band revisited, and the best current values of the optical constants H2O (1) at 251C between 15,000 and 1 cm1. Applied Spectroscopy 50: 1047--1057. Donlon CJ, Keogh SJ, Baldwin DJ, et al. (1998) Solid state measurements of sea surface skin temperature. Journal of Atmospherical Oceanic Technology 15: 775--787. Donlon CJ and Nightingale TJ (2000) The effect of atmospheric radiance errors in radiometric sea surface

skin temperature measurements. Applied Optics 39: 2392--2397. Jessup AT, Zappa CJ, and Yeh H (1997) Defining and quantifying micro-scale wave breaking with infrared imagery. Journal of Geophysical Research 102: 23145--23153. McKeown W and Asher W (1997) A radiometric method to measure the concentration boundary layer thickness at an air–water interface. Journal of Atmospheric and Oceanic Technology 14: 1494--1501. Shaw JA (1999) Degree of polarisation in spectral radiances from water viewing infrared radiometers. Applied Optics 15: 3157--3165. Smith WL, Knuteson RO, Rivercombe HH, et al. (1996) Observations of the infrared radiative properties of the ocean – implications for the measurement of sea surface temperature via satellite remote sensing. Bulletin of the American Meteorological Society 77: 41--51. Suarez MJ, Emery WJ, and Wick GA (1997) The multichannel infrared sea truth radiometric calibrator (MISTRC). Journal of Atmospheric and Oceanic Technology 14: 243--253. Thomas JP, Knight RJ, Roscoe HK, Turner J, and Symon C (1995) An evaluation of a self-calibrating infrared radiometer for measuring sea surface temperature. Journal of Atmospheric and Oceanic Technology 12: 301--316.

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION K. H. Coale, Moss Landing Marine Laboratories, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1385–1397, & 2001, Elsevier Ltd.

Introduction

Light 0 Surface mixed layer Euphotic zone 50 Depth (m)

The trace element iron has been shown to play a critical role in nutrient utilization and phytoplankton growth and therefore in the uptake of carbon dioxide from the surface waters of the global ocean. Carbon fixation in the surface waters, via phytoplankton growth, shifts the ocean–atmosphere exchange equilibrium for carbon dioxide. As a result, levels of atmospheric carbon dioxide (a greenhouse gas) and iron flux to the oceans have been linked to climate change (glacial to interglacial transitions). These recent findings have led some to suggest that largescale iron fertilization of the world’s oceans might therefore be a feasible strategy for controlling climate. Others speculate that such a strategy could deleteriously alter the ocean ecosystem, and still others have calculated that such a strategy would be ineffective in removing sufficient carbon dioxide to produce a sizable and rapid result. This article focuses on carbon and the major plant nutrients, nitrate, phosphate, and silicate, and describes how our recent discovery of the role of iron in the oceans has increased our understanding of phytoplankton growth, nutrient cycling, and the flux of carbon from the atmosphere to the deep sea.

ratio (Redfield, 1934, 1958) and can be expressed on a molar basis relative to carbon as 106C : 16N : 1P. Significant local variations in this uptake/regeneration relationship can be found and are a function of the phytoplankton community and growth conditions, yet this ratio can serve as a conceptual model for nutrient uptake and export. The vertical distribution of the major nutrients typically shows surface water depletion and increasing concentrations with depth. The schematic profile in Figure 1 reflects the processes of phytoplankton uptake within the euphotic zone and remineralization of sinking planktonic debris via microbial degradation, leading to increased concentrations in the deep sea. Given favorable growth conditions, the nutrients at the surface may be depleted to zero. The rate of phytoplankton production of new biomass, and therefore the rate of carbon uptake, is controlled by the resupply of nutrients to the surface waters, usually via the upwelling of deep waters. Upwelling occurs over the entire ocean basin at the rate of approximately 4 m per year but increases in coastal and

Major Nutrients Phytoplankton growth in the oceans requires many physical, chemical, and biological factors that are distributed inhomogenously in space and time. Because carbon, primarily in the form of the bicarbonate ion, and sulfur, as sulfate, are abundant throughout the water column, the major plant nutrients in the ocean commonly thought to be critical for phytoplankton growth are those that exist at the micromolar level such as nitrate, phosphate, and silicate. These, together with carbon and sulfur, form the major building blocks for biomass in the sea. As fundamental cellular constituents, they are generally thought to be taken up and remineralized in constant ratio to one another. This is known as the Redfield

100

Temperature NO3

–

150

200 Figure 1 A schematic profile indicating the regions of the upper water column where phytoplankton grow. The surface mixed layer is that region that is actively mixed by wind and wave energy, which is typically depleted in major nutrients. Below this mixed layer temperatures decrease and nutrients increase as material sinking from the mixed layer is regenerated by microbial decomposition.

(c) 2011 Elsevier Inc. All Rights Reserved.

331

332

IRON FERTILIZATION

regions of divergent surface water flow, reaching average values of 15 to 30 or greater. Thus, those regions of high nutrient supply or persistent high nutrient concentrations are thought to be most important in terms of carbon removal.

Nitrogen versus Phosphorus Limitation Although both nitrogen and phosphorus are required at nearly constant ratios characteristic of deep water, nitrogen has generally been thought to be the limiting nutrient in sea water rather than phosphorus. This idea has been based on two observations: selective enrichment experiments and surface water distributions. When ammonia and phosphate are added to sea water in grow-out experiments, phytoplankton growth increases with the ammonia addition and not with the phosphate addition, thus indicating that reduced nitrogen and not phosphorus is limiting. Also, when surface water concentration of nitrate and phosphate are plotted together (Figure 2), it appears that there is still residual phosphate after the nitrate has gone to zero.

The notion of nitrogen limitation seems counterintuitive when one considers the abundant supply of dinitrogen (N2) in the atmosphere. Yet this nitrogen gas is kinetically unavailable to most phytoplankton because of the large amount of energy required to break the triple bond that binds the dinitrogen molecule. Only those organisms capable of nitrogen fixation can take advantage of this form of nitrogen and reduce atmospheric N2 to biologically available nitrogen in the form of urea and ammonia. This is, energetically, a very expensive process requiring specialized enzymes (nitrogenase), an anaerobic microenvironment, and large amounts of reducing power in the form of electrons generated by photosynthesis. Although there is currently the suggestion that nitrogen fixation may have been underestimated as an important geochemical process, the major mode of nitrogen assimilation, giving rise to new plant production in surface waters, is thought to be nitrate uptake. The uptake of nitrate and subsequent conversion to reduced nitrogen in cells requires a change of five in the oxidation state and proceeds in a stepwise fashion. The initial reduction takes place via the nitrate/nitrite reductase enzyme present in phytoplankton and requires large amounts of the reduced nicotinamide– adenine dinucleotide phosphate (NADPH) and of adenosine triphosphate (ATP) and thus of harvested light energy from photosystem II. Both the nitrogenase enzyme and the nitrate reductase enzyme require iron as a cofactor and are thus sensitive to iron availability.

Ocean Regions

Figure 2 A plot of the global surface water concentrations of phosphate versus nitrate indicating a general positive intercept for phosphorus when nitrate has gone to zero. This is one of the imperical observations favoring the notion of nitrate limitation over phosphate limitation.

From a nutrient and biotic perspective, the oceans can be generally divided into biogeochemical provinces that reflect differences in the abundance of macronutrients and the standing stocks of phytoplankton. These are the high-nitrate, high-chlorophyll (HNHC); high-nitrate, low-chlorophyll (HNLC); low-nitrate, high-chlorophyll (LNHC); and low-nitrate, low-chlorophyll (LNLC) regimes (Table 1). Only the HNLC and LNLC regimes are relatively stable, because the high phytoplankton

Table 1 The relationship between biomass and nitrate as a function of biogeochemical province and the approximate ocean area represented by these regimes High-chlorophyll

Low-chlorophyll

High-nitrate

Unstable/coastal (5%)

Low-nitrate

Unstable/coastal (5%)

Stable/Subarctic/Antarctic/ equatorial Pacific (20%) Oligotrophic gyres (70%)

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION

333

Figure 3 A schematic representation of the ‘iron theory’ as it functions in offshore HNLC regions and coastal transient LNHC regions. It has been suggested that iron added to the HNLC regions would induce them to function as LNHC regions and promote carbon export.

growth rates in the other two systems will deplete any residual nitrate and sink out of the system. The processes that give rise to these regimes have been the subject of some debate over the last few years and are of fundamental importance relative to carbon export (Figure 3). High-nitrate, Low-chlorophyll Regions

The HNLC regions are thought to represent about 20% of the areal extent of the world’s oceans. These are generally regions characterized by more than 2 mmol l 1 nitrate and less than 0.5 mg l 1 chlorophyll-a, a proxy for plant biomass. The major HNLC regions are shown in Figure 4 and represent the Subarctic Pacific, large regions of the eastern equatorial Pacific and the Southern Ocean. These HNLC regions persist in areas that have high macronutrient concentrations, adequate light, and physical characteristics required for phytoplankton growth but have very low plant biomass. Two explanations have been

given to describe the persistence of this condition. (1) The rates of zooplankton grazing of the phytoplankton community may balance or exceed phytoplankton growth rates in these areas, thus cropping plant biomass to very low levels and recycling reduced nitrogen from the plant community, thereby decreasing the uptake of nitrate. (2) Some other micronutrient (possibly iron) physiologically limits the rate of phytoplankton growth. These are known as top-down and bottom-up control, respectively. Several studies of zooplankton grazing and phytoplankton growth in these HNLC regions, particularly the Subarctic Pacific, confirm the hypothesis that grazers control production in these waters. Recent physiological studies, however, indicate that phytoplankton growth rates in these regions are suboptimal, as is the efficiency with which phytoplankton harvest light energy. These observations indicate that phytoplankton growth may be limited by something other than (or in addition to) grazing. Specifically, these studies implicate the lack of

(c) 2011 Elsevier Inc. All Rights Reserved.

334

IRON FERTILIZATION

Figure 4 Current HNLC regions of the world’s oceans covering an extimated 20% of the ocean surface. These regions include the Subarctic Pacific, equatorial Pacific and Southern Ocean.

sufficient electron transport proteins and the cell’s ability to transfer reducing power from the photocenter. These have been shown to be symptomatic of iron deficiency.

The Role of Iron Iron is a required micronutrient for all living systems. Because of its d-electron configuration, iron readily undergoes redox transitions between Fe(II) and Fe(III) at physiological redox potentials. For this reason, iron is particularly well suited to many enzyme and electron carrier proteins. The genetic sequences coding for many iron-containing electron carriers and enzymes are highly conserved, indicating iron and iron-containing proteins were key features of early biosynthesis. When life evolved, the atmosphere and waters of the planet were reducing and iron was abundant in the form of soluble Fe(II). Readily available and at high concentration, iron was not likely to have been limiting in the primordial biosphere. As photosynthesis evolved, oxygen was produced as a by-product. As the biosphere became more oxidizing, iron precipitated from aquatic systems in vast quantities, leaving phytoplankton and other aquatic life forms in a vastly changed and newly deficient chemical milieu. Evidence of this mass Fe(III) precipitation event is captured in the ancient banded iron formations in many parts of the world. Many primitive aquatic and terrestrial

organisms have subsequently evolved the ability to sequester iron through the elaboration of specific Fe(II)-binding ligands, known as siderophores. Evidence for siderophore production has been found in several marine dinoflagellates and bacteria and some researchers have detected similar compounds in sea water. Today, iron exists in sea water at vanishingly small concentrations. Owing to both inorganic precipitation and biological uptake, typical surface water values are on the order of 20 pmol l 1, perhaps a billion times less than during the prehistoric past. Iron concentrations in the oceans increase with depth, in much the same manner as the major plant nutrients (Figure 5). The discovery that iron concentrations in surface waters is so low and shows a nutrient-like profile led some to speculate that iron availability limits plant growth in the oceans. This notion has been tested in bottle enrichment experiments throughout the major HNLC regions of the world’s oceans. These experiments have demonstrated dramatic phytoplankton growth and nutrient uptake upon the addition of iron relative to control experiments in which no iron was added. Criticism that such small-scale, enclosed experiments may not accurately reflect the response of the HNLC system at the level of the community has led to several large-scale iron fertilization experiments in the equatorial Pacific and Southern Ocean. These have been some of the most dramatic

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION

0.0 0

0.2 50

Fe (nmol kg–1) 0.4 0.6 0.8 1.0 O2 (μmol kg–1) 100 150 200

1.2 250

1.4 300

0.0 0.5

Depth (km)

1.0 O2

1.5

2.5

–

Fe

NO3

100

150

Si

3.0 3.5

0

50

200

250

300

60

70

Si (μmol kg–1)

0

10

developed in four release experiments in the equatorial Pacific (IronEx I and II) and more recently in the Southern Ocean (SOIREE). At this writing, a similar strategy is being employed in the Caruso experiments now underway in the Atlantic sector of the Southern Ocean. All of these strategies were developed to address certain scientific questions and were not designed as preliminary to any geoengineering effort. Form of Iron

2.0

4.0

335

20

30

40

50

NO3– (μmol kg–1)

Figure 5 The vertical distributions of iron, nitrate, silicate, and oxygen in sea water. This figure shows how iron is depleted to picomolar levels in surface waters and has a profile that mimics other plant nutrients.

oceanographic experiments of our times and have led to a profound and new understanding of ocean systems.

Open Ocean Iron Enrichment The question of iron limitation was brought into sharp scientific focus with a series of public lectures, reports by the US National Research Council, papers, special publications, and popular articles between 1988 and 1991. What was resolved was the need to perform an open ocean enrichment experiment in order to definitively test the hypothesis that iron limits phytoplankton growth and nutrient and carbon dioxide uptake in HNLC regions. Such an experiment posed severe logistical challenges and had never been conducted.

Experimental Strategy The mechanics of producing an iron-enriched experimental patch and following it over time was

All experiments to date have involved the injection of an iron sulfate solution into the ship’s wake to achieve rapid dilution and dispersion throughout the mixed layer (Figure 6). The rationale for using ferrous sulfate involved the following considerations: (1) ferrous sulfate is the most likely form of iron to enter the oceans via atmospheric deposition; (2) it is readily soluble (initially); (3) it is available in a relatively pure form so as to reduce the introduction of other potentially bioactive trace metals; and (4) its counterion (sulfate) is ubiquitous in sea water and not likely to produce confounding effects. Although mixing models indicate that Fe(II) carbonate may reach insoluble levels in the ship’s wake, rapid dilution reduces this possibility. New forms of iron are now being considered by those who would seek to reduce the need for subsequent infusions. Such forms could include iron lignosite, which would increase the solubility and residence time of iron in the surface waters. Since this is a chelated form of iron, problems of rapid precipitation are reduced. In addition, iron lignosulfonate is about 15% Fe by weight, making it a space-efficient form of iron to transport. As yet untested is the extent to which such a compound would reduce the need for re-infusion. Although solid forms of iron have been proposed (slow-release iron pellets; finely milled magnetite or iron ores), the ability to trace the enriched area with an inert tracer has required that the form of iron added and the tracer both be in the dissolved form. Inert Tracer

Concurrent with the injection of iron is the injection of the inert chemical tracer sulfur hexafluoride (SF6). By presaturating a tank of sea water with SF6 and employing an expandable displacement bladder, a constant molar injection ratio of Fe : SF6 can be achieved (Figure 6). In this way, both conservative and nonconservative removal of iron can be quantified. Sulfur hexafluoride traces the physical properties of the enriched patch; the relatively rapid shipboard detection of SF6 can be used to track and

(c) 2011 Elsevier Inc. All Rights Reserved.

336

IRON FERTILIZATION

Figure 6 The iron injection system used during the IronEx experiments utilized two polyethylene tanks that could be sequentially filled with sea water and iron sulfate solution while the other was being injected behind the ship’s propellers. A steel tank of sea water saturated with 40 g of sulfur hexafluoride (SF6) was simultaneously mixed with the iron sulfate solution to provide a conservative tracer of mixing.

map the enriched area. The addition of helium-3 to the injected tracer can provide useful information regarding gas transfer. Fluorometry

The biophysical response of the phytoplankton is rapid and readily detectable. Thus shipboard measurement of relative fluorescence (Fv/Fm) using fast repetition rate fluorometry has been shown to be a useful tactical tool and gives nearly instantaneous mapping and tracking feedback.

Remote Sensing

A variety of airborne and satellite-borne active and passive optical packages provide rapid, large-scale mapping and tracking of the enriched area. Although SeaWiffs was not operational during IronEx I and II, AVHRR was able to detect the IronEx II bloom and airborne optical LIDAR was very useful during IronEx I. SOIREE has made very good use of the more recent SeaWiffs images, which have markedly extended the observational period and led to new hypotheses regarding iron cycling in polar systems.

Shipboard Iron Analysis

Experimental Measurements

Because iron is rapidly lost from the system (at least initially), the shipboard determination of iron is necessary to determine the timing and amount of subsequent infusions. Several shipboard methods, using both chemiluminescent and catalytic colorimetric detection have proven useful in this regard.

In addition to the tactical measurements and remote sensing techniques required to track and ascertain the development of the physical dynamics of the enriched patch, a number of measurements have been made to track the biogeochemical development of the experiment. These have typically involved a series of underway measurements made using the ship’s flowing sea water system or towed fish. In addition, discrete measurements are made in the vertical dimension at every station occupied both inside and outside of the fertilized area. These measurements include temperature salinity, fluorescence (a measure of plant biomass), transmissivity (a measure of suspended particles), oxygen, nitrate,

Lagrangian Drifters

A Lagrangian point of reference has proven to be very useful in every experiment to date. Depending upon the advective regime, this is the only practical way to achieve rapid and precise navigation and mapping about the enriched area.

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION

phosphate, silicate, carbon dioxide partial pressure, pH, alkalinity, total carbon dioxide, iron-binding ligands, 234Th : 238U radioisotopic disequilibria (a proxy for particle removal), relative fluorescence (indicator of photosynthetic competence), primary production, phytoplankton and zooplankton enumeration, grazing rates, nitrate uptake, and particulate and dissolved organic carbon and nitrogen. These parameters allow for the general characterization of both the biological and geochemical response to added iron. From the results of the equatorial enrichment experiments (IronEx I and II) and the Southern Ocean Iron Enrichment Experiment (SOIREE), several general features have been identified.

337

unfavorable redox process is only made possible by substantial reducing power (in the form of NADPH) made available through photosynthesis and active nitrate reductase, an iron-requiring enzyme. Without iron, plants cannot take up nitrate efficiently. This provided original evidence implicating iron deficiency as the cause of the HNLC condition. When phytoplankton communities are relieved from iron deficiency, specific rates of nitrate uptake increase. This has been observed in both the equatorial Pacific and the Southern Ocean using isotopic tracers of nitrate uptake and conversion. In addition, the accelerated uptake of nitrate has been observed in both the mesoscale iron enrichment experiments to date, IronEx and SOIREE. Growth Response

Findings to Date Biophysical Response

The experiments to date have focused on the highnitrate, low-chlorophyll (HNLC) areas of the world’s oceans, primarily in the Subarctic, equatorial Pacific and Southern Ocean. In general, when light is abundant many researchers find that HNLC systems are iron-limited. The nature of this limitation is similar between regions but manifests itself at different levels of the trophic structure in some characteristic ways. In general, all members of the HNLC photosynthetic community are physiologically limited by iron availability. This observation is based primarily on the examination of the efficiency of photosystem II, the light-harvesting reaction centers. At ambient levels of iron, light harvesting proceeds at suboptimal rates. This has been attributed to the lack of iron-dependent electron carrier proteins at low iron concentrations. When iron concentrations are increased by subnanomolar amounts, the efficiency of light harvesting rapidly increases to maximum levels. Using fast repetition rate fluorometry and non-heme iron proteins, researchers have described these observations in detail. What is notable about these results is that iron limitation seems to affect the photosynthetic energy conversion efficiency of even the smallest of phytoplankton. This has been a unique finding that stands in contrast to the hypothesis that, because of diffusion, smaller cells are not iron limited but larger cells are. Nitrate Uptake

As discussed above, iron is also required for the reduction (assimilation) of nitrate. In fact, a change of oxidation state of five is required between nitrate and the reduced forms of nitrogen found in amino acids and proteins. Such a large and energetically

When iron is present, phytoplankton growth rates increase dramatically. Experiments over widely differing oceanographic regimes have demonstrated that, when light and temperature are favorable, phytoplankton growth rates in HNLC environments increase to their maximum at dissolved iron concentrations generally below 0.5 nmol l 1. This observation is significant in that it indicates that phytoplankton are adapted to very low levels of iron and they do not grow faster if given iron at more than 0.5 nmol l 1. Given that there is still some disagreement within the scientific community about the validity of some iron measurements, this phytoplankton response provides a natural, environmental, and biogeochemical benchmark against which to compare results. The iron-induced transient imbalance between phytoplankton growth and grazing in the equatorial Pacific during IronEx II resulted in a 30-fold increase in plant biomass (Figure 7). Similarly, a 6-fold increase was observed during the SOIREE experiment in the Southern Ocean. These are perhaps the most dramatic demonstrations of iron limitation of nutrient cycling, and phytoplankton growth to date and has fortified the notion that iron fertilization may be a useful strategy to sequester carbon in the oceans. Heterotrophic Community

As the primary trophic levels increase in biomass, growth in the small microflagellate and heterotrophic bacterial communities increase in kind. It appears that these consumers of recently fixed carbon (both particulate and dissolved) respond to the food source and not necessarily the iron (although some have been found to be iron-limited). Because their division rates are fast, heterotrophic bacteria, ciliates, and flagellates can rapidly divide and

(c) 2011 Elsevier Inc. All Rights Reserved.

338

IRON FERTILIZATION

Figure 7 Chlorophyll concentrations during IronEx II were mapped daily. This figure shows the progression of the phytoplankton bloom that reached over 30 times the background concentrations.

Nutrient Uptake Ratios

3 H

SiO4 / NO3–

respond to increasing food availability to the point where the growth rates of the smaller phytoplankton can be overwhelmed by grazing. Thus there is a much more rapid turnover of fixed carbon and nitrogen in iron replete systems. M. Landry and coworkers have documented this in dilution experiments conducted during IronEx II. These results appear to be consistent with the recent SOIREE experiments as well.

SiO4: NO3 uptake ratio vs. dissolved iron concentration

J H 2 H HJJ J J H J J JJ H JH JH J H H H 1

J H

JJJ H H

J1 Survey I, Exp.1 J2 Survey I, Exp.2 H3 Survey II, Exp.3 H4 Survey II, Exp.4

0

An imbalance in production and consumption, however, can arise at the larger trophic levels. Because the reproduction rates of the larger micro- and mesozooplankton are long with respect to diatom division rates, iron-replete diatoms can escape the pressures of grazing on short timescales (weeks). This is thought to be the reason why, in every iron enrichment experiment, diatoms ultimately dominate in biomass. This result is important for a variety of reasons. It suggests that transient additions of iron would be most effective in producing net carbon uptake and it implicates an important role of silicate in carbon flux. The role of iron in silicate uptake has been studied extensively by Franck and colleagues. The results, together with those of Takeda and coworkers, show that iron alters the uptake ratio of nitrate and silicate at very low levels (Figure 8). This is thought to be brought about by the increase in nitrate uptake rates relative to silica.

0

1

2 Fe (nmol kg – 1)

3

Figure 8 Bottle enrichment experiments show that the silicate : nitrate uptake ratio changes as a function of the iron added. This is thought to be due to the increased rate of iron uptake relative to silicate in these experimental treatments.

Organic Ligands

Consistent with the role of iron as a limiting nutrient in HNLC systems is the notion that organisms may have evolved competitive mechanisms to increase iron solubility and uptake. In terrestrial systems this is accomplished using extracellularly excreted or membrane-bound siderophores. Similar compounds have been shown to exist in sea water where the competition for iron may be as fierce as it is on land. In open ocean systems where it has been measured, iron-binding ligand production increases with the

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION

addition of iron. Whether this is a competitive response to added iron or a function of phytoplankton biomass and grazing is not yet well understood. However, this is an important natural mechanism for reducing the inorganic scavenging of iron from the surface waters and increasing iron availability to phytoplankton. More recent studies have considerably advanced our understanding of these ligands, their distribution and their role in ocean ecosystems. Carbon Flux

It is the imbalance in the community structure that gives rise to the geochemical signal. Whereas iron stimulation of the smaller members of the community may result in chemical signatures such as an increased production of beta-dimethylsulfonioproprionate (DMSP), it is the stimulation of the larger producers that decouples the large cell producers from grazing and results in a net uptake and export of nitrate, carbon dioxide, and silicate. The extent to which this imbalance results in carbon flux, however, has yet to be adequately described. The inability to quantify carbon export has primarily been a problem of experimental scale. Even though mesoscale experiments have, for the first time, given us the ability to address the effect of iron on communities, the products of surface water processes and the effects on the midwater column have been difficult to track. For instance, in the IronEx II experiment, a time-series of the enriched patch was diluted by 40% per day. The dilution was primarily in a lateral

339

(horizontal/isopycnal) dimension. Although some correction for lateral dilution can be made, our ability to quantify carbon export is dependent upon the measurement of a signal in waters below the mixed layer or from an uneroded enriched patch. Current data from the equatorial Pacific showed that the IronEx II experiment advected over six patch diameters per day. This means that at no time during the experiment were the products of increased export reflected in the waters below the enriched area. A transect through the IronEx II patch is shown in Figure 9. This figure indicates the massive production of plant biomass with a concomitant decrease in both nitrate and carbon dioxide. The results from the equatorial Pacific, when corrected for dilution, suggest that about 2500 t of carbon were exported from the mixed layer over a 7day period. These results are preliminary and subject to more rigorous estimates of dilution and export production, but they do agree favorably with estimates based upon both carbon and nitrogen budgets. Similarly, thorium export was observed in this experiment, confirming some particle removal. The results of the SOIREE experiment were similar in many ways but were not as definitive with respect to carbon flux. In this experiment biomass increased 6-fold, nitrate was depleted by 2 mmol l 1 and carbon dioxide by 35–40 microatmospheres (3.5–4.0 Pa). This was a greatly attenuated signal relative to IronEx II. Colder water temperatures likely led to slower rates of production and bloom evolution and there was no observable carbon flux.

Figure 9 A transect through the IronEx II patch. The x-axis shows GMT as the ship steams from east to west through the center of the patch. Simultaneously plotted are the iron-induced production of chlorophyll, the drawdown of carbon dioxide, and the uptake of nitrate in this bloom.

(c) 2011 Elsevier Inc. All Rights Reserved.

340

IRON FERTILIZATION

Figure 10 Simple calculations of the potential for carbon export for the Southern Ocean. These calculations are based on the necessary amount of iron required to efficiently utilize the annual upwelled nitrate and the subsequent incorporation into sinking organic matter. An estimated 1.8  109 t (Gt) of carbon export could be realized in this simple model.

Original estimates of carbon export in the Southern Ocean based on the iron-induced efficient utilization of nitrate suggest that as much as 1.8  109 t of carbon could be removed annually (Figure 10). These estimates of carbon sequestration have been challenged by some modelers yet all models lack important experimental parameters which will be measured in upcoming experiments.

Remaining Questions A multitude of questions remain regarding the role of iron in shaping the nature of the pelagic community. The most pressing question is whether iron enrichment accelerates the downward transport of carbon from the surface waters to the deep sea? More specifically, how does iron affect the cycling of carbon in HNLC, LNLC, and coastal systems? Recent studies indicate that coastal systems may be ironlimited and the iron requirement for nitrogenase activity is quite large, suggesting that iron may limit nitrogen fixation, but there have been limited studies to test the former and none to test the latter. If iron does stimulate carbon uptake, what are the spatial scales over which this fixed carbon may be remineralized? This is crucial to predicting whether fertilization is an effective carbon sequestration mechanism. Given these considerations, the most feasible way to understand and quantify carbon export from an enriched water mass is to increase the scale of the experiment such that both lateral dilution and submixed-layer relative advection are small with respect to the size of the enriched patch. For areas such as

the equatorial Pacific, this would be very large (hundreds of kilometers on a side). For other areas, it could be much smaller. The focus of the IronEx and SOIREE experiments has been from the scientific perspective, but this focus is shifting toward the application of iron enrichment as a carbon sequestration strategy. We have come about rapidly from the perspective of trying to understand how the world works to one of trying to make the world work for us. Several basic questions remain regarding the role of natural or anthropogenic iron fertilization on carbon export. Some of the most pressing questions are: What are the best proxies for carbon export? How can carbon export best be verified? What are the long-term ecological consequences of iron enrichment on surface water community structure, midwater processes, and benthic processes? Even with answers to these, there are others that need to be addressed prior to any serious consideration of iron fertilization as an ocean carbon sequestration option. Simple technology is sufficient to produce a massive bloom. The technology required either for a large-scale enrichment experiment or for purposeful attempts to sequester carbon is readily available. Ships, aircraft (tankers and research platforms), tracer technology, a broad range of new Autonomous Underwater Vehicles (AUVs) and instrument packages, Lagrangian buoy tracking systems, together with aircraft and satellite remote sensing systems and a new suite of chemical sensors/in situ detection technologies are all available, or are being developed. Industrial bulk handling equipment is available for large-scale implementation. The big questions, however, are larger than the technology. With a slow start, the notion of both scientific experimentation through manipulative experiments, as well as the use of iron to purposefully sequester carbon, is gaining momentum. There are now national, international, industrial, and scientific concerns willing to support larger-scale experiments. The materials required for such an experiment are inexpensive and readily available, even as industrial by-products (of paper, mining, and steel processing). Given the concern over climate change and the rapid modernization of large developing countries such as China and India, there is a pressing need to address the increased emission of greenhouse gases. Through the implementation of the Kyoto accords or other international agreements to curb emissions (Rio), financial incentives will reach into the multibillion dollar level annually. Certainly there will soon be an overwhelming fiscal incentive to investigate, if not implement, purposeful open ocean carbon sequestration trials.

(c) 2011 Elsevier Inc. All Rights Reserved.

IRON FERTILIZATION

341

A Societal Challenge

Further Reading

The question is not whether we have the capability of embarking upon such an engineering strategy but whether we have the collective wisdom to responsibly negotiate such a course of action. Posing the question another way: If we do not have the social, political and economic tools or motivation to control our own population and greenhouse gas emissions, what gives us the confidence that we have the wisdom and ability to responsibly manipulate and control large ocean ecosystems without propagating yet another massive environmental calamity? Have we as an international community first tackled the difficult but obvious problem of overpopulation and implemented alternative energy technologies for transportation, industry, and domestic use? Other social questions arise as well. Is it appropriate to use the ocean commons for such a purpose? What individuals, companies, or countries would derive monetary compensation for such an effort and how would this be decided? It is clear that there are major scientific investigations and findings that can only benefit from largescale open ocean enrichment experiments, but certainly a large-scale carbon sequestration effort should not proceed without a clear understanding of both the science and the answers to the questions above.

Abraham ER, Law CS, Boyd PW, et al. (2000) Importance of stirring in the development of an iron-fertilized phytoplankton bloom. Nature 407: 727--730. Barbeau K, Moffett JW, Caron DA, Croot PL, and Erdner DL (1996) Role of protozoan grazing in relieving iron limitation of phytoplankton. Nature 380: 61--64. Behrenfeld MJ, Bale AJ, Kobler ZS, Aiken J, and Falkowski PG (1996) Confirmation of iron limitation of phytoplankton photosynthesis in Equatorial Pacific Ocean. Nature 383: 508--511. Boyd PW, Watson AJ, Law CS, et al. (2000) A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization. Nature 407: 695-702. Cavender-Bares KK, Mann EL, Chishom SW, Ondrusek ME, and Bidigare RR (1999) Differential response of equatorial phytoplankton to iron fertilization. Limnology and Oceanography 44: 237--246. Coale KH, Johnson KS, Fitzwater SE, et al. (1996) A massive phytoplankton bloom induced by an ecosystemscale iron fertilization experiment in the equatorial Pacific Ocean. Nature 383: 495--501. Coale KH, Johnson KS, Fitzwater SE, et al. (1998) IronExI, an in situ iron-enrichment experiment: experimental design, implementation and results. Deep-Sea Research Part II 45: 919--945. Elrod VA, Johnson KS, and Coale KH (1991) Determination of subnanomolar levels of iron (II) and total dissolved iron in seawater by flow injection analysis with chemiluminescence dection. Analytical Chemistry 63: 893--898. Fitzwater SE, Coale KH, Gordon RM, Johnson KS, and Ondrusek ME (1996) Iron deficiency and phytoplankton growth in the equatorial Pacific. Deep-Sea Research Part II 43: 995--1015. Greene RM, Geider RJ, and Falkowski PG (1991) Effect of iron lititation on photosynthesis in a marine diatom. Limnology Oceanogrography 36: 1772--1782. Hoge EF, Wright CW, Swift RN, et al. (1998) Fluorescence signatures of an iron-enriched phytoplankton community in the eastern equatorial Pacific Ocean. Deep-Sea Research Part II 45: 1073--1082. Johnson KS, Coale KH, Elrod VA, and Tinsdale NW (1994) Iron photochemistry in seawater from the Equatorial Pacific. Marine Chemistry 46: 319--334. Kolber ZS, Barber RT, Coale KH, et al. (1994) Iron limitation of phytoplankton photosynthesis in the Equatorial Pacific Ocean. Nature 371: 145--149. Landry MR, Ondrusek ME, Tanner SJ, et al. (2000) Biological response to iron fertilization in the eastern equtorial Pacific (Ironex II). I. Microplankton community abundances and biomass. Marine Ecology Progress Series 201: 27--42. LaRoche J, Boyd PW, McKay RML, and Geider RJ (1996) Flavodoxin as an in situ marker for iron stress in phytoplankton. Nature 382: 802--805. Law CS, Watson AJ, Liddicoat MI, and Stanton T (1998) Sulfer hexafloride as a tracer of biogeochemical and

Glossary ATP AVHRR

Adenosine triphosphate Advanced Very High Resolution Radiometer HNHC High-nitrate high-chlorophyll HNLC High-nitrate low-chlorophyll IronEx Iron Enrichment Experiment LIDAR Light detection and ranging LNHC Low-nitrate high-chlorophyll LNLC Low-nitrate low-chlorophyll NADPH Reduced form of nicotinamide–adenine dinucleotide phosphate SOIREE Southern Ocean Iron Enrichment Experiment

See also Fluorometry for Biological Sensing. Fluorometry for Chemical Sensing. Nitrogen Cycle. Phosphorus Cycle. Platforms: Autonomous Underwater Vehicles. Primary Production Distribution. Primary Production Processes. Redfield Ratio. Satellite Remote Sensing SAR.

(c) 2011 Elsevier Inc. All Rights Reserved.

342

IRON FERTILIZATION

physical processes in an open-ocean iron fertilization experiment. Deep-Sea Research Part II 45: 977--994. Martin JH, Coale KH, Johnson KS, et al. (1994) Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean. Nature 371: 123--129. Nightingale PD, Liss PS, and Schlosser P (2000) Measurements of air–gas transfer during an open ocean algal bloom. Geophysical Research Letters 27: 2117--2121. Obata H, Karatani H, and Nakayama E (1993) Automated determination of iron in seawater by chelating resin concentration and chemiluminescence detection. Analytical Chemistry 65: 1524--1528. Redfield AC (1934) On the proportions of organic derivatives in sea water and their relation to the composition of plankton. James Johnstone Memorial Volume, pp. 177--192. Liverpool: Liverpool University Press. Redfield AC (1958) The biological control of chemical factors in the environment. American Journal of Science 46: 205--221. Rue EL and Bruland KW (1997) The role of organic complexation on ambient iron chemistry in the equatorial Pacific Ocean and the response of a mesoscale iron addition experiment. Limnology and Oceanography 42: 901--910. Smith SV (1984) Phosphorus versus nitrogen limitation in the marine environment. Limnology and Oceanography 29: 1149--1160.

Stanton TP, Law CS, and Watson AJ (1998) Physical evolutation of the IronEx I open ocean tracer patch. Deep-Sea Research Part II 45: 947--975. Takeda S and Obata H (1995) Response of equatorial phytoplankton to subnanomolar Fe enrichment. Marine Chemistry 50: 219--227. Trick CG and Wilhelm SW (1995) Physiological changes in coastal marine cyanobacterium Synechococcus sp. PCC 7002 exposed to low ferric ion levels. Marine Chemistry 50: 207--217. Turner SM, Nightingale PD, Spokes LJ, Liddicoat MI, and Liss PS (1996) Increased dimethyl sulfide concentrations in seawater from in situ iron enrichment. Nature 383: 513--517. Upstill-Goddard RC, Watson AJ, Wood J, and Liddicoat MI (1991) Sulfur hexafloride and helium-3 as sea-water tracers: deployment techniques and continuous underway analysis for sulphur hexafloride. Analytica Chimica Acta 249: 555--562. Van den Berg CMG (1995) Evidence for organic complesation of iron in seawater. Marine Chemistry 50: 139--157. Watson AJ, Liss PS, and Duce R (1991) Design of a smallscale in situ iron fertilization experiment. Limnology and Oceanography 36: 1960--1965.

(c) 2011 Elsevier Inc. All Rights Reserved.

ISLAND WAKES E. D. Barton, University of Wales, Bangor, Menai Bridge, Anglesey, UK

disturbance or wake is related to the value of the Reynolds number

Copyright & 2001 Elsevier Ltd.

Re ¼

This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1397–1403, & 2001, Elsevier Ltd.

Introduction The ‘island mass effect’ has been documented for about half a century. This refers to a biological enrichment around oceanic islands in comparison to surrounding waters. Despite a relatively large body of evidence to support the existence of such an effect in the vicinity of islands, there have been few studies to investigate the underlying physical causes of this phenomenon. From a physical point of view the presence of an island in the background flow will disturb the flow regime to produce perturbations that ultimately must have biological consequences. Two types of island disturbance have been investigated. The first takes place in shallow, stratified shelf seas with significant tidal regimes but no appreciable mean flow, where as the tide moves water back and forth the island acts as a stirring rod to enhance vertical mixing locally and break down the pycnocline. The second occurs in both shallow and deep water, where flow past an island generates eddies downstream and a wake of disturbed flow extends several island diameters away. This arises in the case of larger islands when a clear ambient flow dominates over tidal variability and also around islands small enough that the tidal stream itself can generate a similar effect. The scale of the eddies is typically close to the island diameter and their time scale will be several days for larger islands but only hours in the case of tidal flows. The nature of the wake and eddies may differ between the oceanic case where conditions can be considered quasigeostrophic and the shallow case where they are frictionally dominated by bottom stress.

Theory Nonrotating Case

The simplest case is where flow past isolated oceanic islands is considered to be analogous to that of channel flow past a circular cylinder. The work of Batchelor presented the case of nonrotating flow in a homogeneous fluid. The form of the downstream

Ud v

½1

where U is the free velocity upstream, d is the diameter of the cylinder and v is the molecular viscosity. Although this formula appears simple, there are certain practical difficulties in applying it to even a nonrotating ocean of homogeneous character. Few islands are isolated or cylindrical, the upstream velocity is generally not well known and finally the molecular viscosity must be replaced by the horizontal eddy viscosity in the ocean, which is in general poorly known. Studies of Aldabra, an Indian Ocean atoll, indicate that the same current speed impinging on the island from different directions produces a different wake because of the asymmetrical form of the island. It is frequently the case that the upstream current in the ocean varies on a range of time scales, or may be subject to horizontal shear, both of which complicate the choice of a suitable value for the free stream velocity. Values of horizontal eddy viscosity coefficients in the ocean based on many experimental determinations increase with the length scale of interest, l, in a nonlinear fashion Kh ðm2 s1 Þ ¼ 2:2 104 l1:13 . Typical values appropriate to oceanic islands vary between 102 and 105 m2 s1. The nature of the wake downstream of the obstacle changes as the Reynolds number increases (Figure 1). For low Reynolds numbers (Reo1) the flow pattern downstream is the same as that upstream and there is no perceptible wake. The flow remains attached to the sides of the cylinder and is laminar throughout the flow field. At Reynolds numbers between 1 and 40, the wake remains basically laminar away from the cylinder and two eddies are formed immediately behind the obstacle where they remain attached. At higher Reynolds number the wake becomes increasingly unstable and counter-rotating eddies form a vortex street. The eddies expand as they move away from the obstacle and gradually decay. For Reynolds numbers Re4 80 eddies formed behind the island no longer remain trapped but are shed alternately into the vortex street. The frequency of eddy shedding n is related to another nondimensional number, the Strouhal number

(c) 2011 Elsevier Inc. All Rights Reserved.

St ¼

nd U

½2

343

344

ISLAND WAKES

rotation rate of the earth and f the latitude. This variation of the Coriolis parameter f can be represented in terms of the ‘beta plane’ where b ¼ df/dy and y is distance north of the equator. If the dimensionless parameter

Laminar flow

U Re < 1 Attached eddies

Laminar wake

U

b0 ¼

1 < Re < 40

½5

is small, the beta effect may be ignored and the rotation rate taken to be constant on the scale of the island in question. If on the other hand it is of order unity, differences occur from the case of uniform rotation. In particular flow separation is enhanced for flow towards the west and inhibited for flow towards the east.

Attached eddies Unstable wake with eddies

U

bd2 4U

40 < Re < 80

Effects of Bottom Friction

U

Studies of flow patterns around small islands only a few kilometers in diameter in shallow shelf seas indicate that the Reynolds number as defined in eqn[1] overestimates the value at transition between the different cases of wake formation. It has been found that the island wake parameter

Detaching eddies

Re > 80

Figure 1 Reynold number regimes for flow past a cylinder.

This number approaches an asymptotic value of 0.21 at higher Reynolds numbers. Effect of Rotation

The ocean of course is on the rotating Earth and so laboratory experiments have been carried out in rotating tanks to investigate the effect of this rotation. Two dimensionless quantities of importance here are the Rossby number R0 ¼

U Od

½3

which represents the importance of rotation at rate O and the Ekman number 2v Ek ¼ Od2

½4

which determines the width of the wake. The ratio of Rossby number to Ekman number is proportional to the Reynolds number, generalizing this concept to the case of rotating flow. The Earth’s rotation enhances the shedding of eddies in the same sense (cyclonic) so that in the Northern Hemisphere predominantly anticlockwise eddies are to be expected whereas in the Southern Hemisphere clockwise should be more common. Because of the spherical shape of the Earth’s surface the rate of rotation about the local vertical varies with latitude f ¼ 2o sin f, where o is the

P¼

Uh2 Kz d

½6

where U is the stream velocity, h is the water depth, Kz is the vertical eddy diffusivity and d is the dimension of the island, provided better agreement than the Reynolds number between observed and predicted wake parameters. However, P is actually a correct formulation for the Reynolds number when the effect of lateral and bottom frictional boundary layers is taken into account.

Observations Until recently observations of island flow effects had been limited mainly to remote sensing reports of eddy production. Attempts to observe vortex production and development in situ behind oceanic islands have been largely unsuccessful because the background flow has been too weak or variable or the methods of observation insufficient to determine the flow regime adequately. However, a classic early report in 1972, based upon sparse observations of the drift of fishing gear and rudimentary surface current measurement, showed drift patterns downstream of Johnston Atoll in the Pacific Ocean in good agreement with a vortex street situation. In the case of Aldabra Atoll, situated in the South Equatorial Current of the Indian Ocean, surveys made with acoustic Doppler current profiler

(c) 2011 Elsevier Inc. All Rights Reserved.

ISLAND WAKES

indicated a single cyclonic (anticlockwise) eddy trapped behind the island in a low Reynolds number regime on two occasions. There was no evidence of continuous eddy or wake production downstream. The flow was variable during the experiment. The first trapped eddy was observed during westward background flow around 20 cm s1. Two subsequent rapid surveys, about ten days apart, of the current field showed predominantly northward and westward flows, respectively. In the first case the free stream flow was about 20 cm s1 and impinged on the island’s widest cross-section. No eddy was observed, only an asymmetrical deviation of currents behind the island. The later case found slightly stronger (30 cm s1) free stream velocity from the east impinging the narrow aspect of the island. This time a second weak eddy of diameter similar to the island width was indicated. Another island showing highly variably flow regimes is Barbados. In Spring 1991, the flow seemed topographically steered around the Barbados ridge and there was no clear evidence of eddy production. The following spring, anticyclonic and cyclonic eddies of similar size to the island were found on either flank (Figure 2). Computer simulations of the flow

14.00



345

regime indicated that typical conditions were conducive to shedding of cyclones and anticyclones alternately with a period of around 10 days. It was not possible to demonstrate the degree to which the observations were attributable to this type of vortex generation, however, and it was concluded that much more extensive observations would be needed to do so. One interesting indication of the simulation was that there was a continuous region of downstream reverse flow towards the island that could provide a return path for fish larvae swept away from the island. Considerable evidence of recurrent eddy shedding has been reported recently in the Canary Island archipelago. There, the Canary Current flows southwestward at an average speed of 5 cm s1. Eddies of both signs have been reported (Figure 3) as being frequently spun off from the island of Gran Canaria. Cyclonic eddies of the same diameter as the island (50 km) and rotation period around 3 days have been observed to develop on the southwestern flank of the island and to move southwest at speeds of 5–15 cm s1. Almost as frequently, anticyclonic eddies have been observed to develop on the south east of the island. They have diameters up to twice

7

 

20 30

 

13.75

67 



40



65





13.50



61 

Latitude





13.25 

80 70 60 50 40 30

20





30 L



58 



H





50  60

BARBADOS

40







13.00

 



84





12.75 _ 60.25

_ 60.00

_ 59.75

_ 59.50 Longitude

_ 59.25

_ 59.00

_ 58.75

Figure 2 Sea surface topography in centimeters relative to 250 dbar 25 April–2 May 1991. Arrows denote the direction of surface flow. Note the overall northwestward flow and anticyclonic (H) and cyclonic (L) eddies either side of the island. (Adapted from Bowman et al., 1996.)

(c) 2011 Elsevier Inc. All Rights Reserved.

346

ISLAND WAKES

29N

1607 23 Aug 1999 C

Lee Gran Lee

28N

Canaria

Lee A

Lee C

Lee

Lee

C

27N

A

19W

18W

17W

16W

15W

14W

Figure 3 Sea surface temperature image showing multiple eddies shed from the Canary Islands during August 1999. Cyclones and anticyclones are labeled C and A, respectively.

the size of the cyclonic eddies, similar rotation rates and appear to persist for many months in the region. One such eddy, seeded with drifters in June 1998, was observed to persist for at least seven months (Figure 4). It initially drifted slowly southwestward but later returned northwestward towards the outer islands of the archipelago. Other large anticyclones have been observed downstream of the island pair, Tenerife and La Gomera, trapped close to the islands for at least several weeks before moving away from the island. Smaller cyclones and anticyclones are

29N Tenerife

Fuerteventura Gran Canaria

28N Release point

27N

26N

Jul _ Oct 1998

18W

17W

16W

15W

14W

Figure 4 Path of a drifter released into an anticyclonic eddy shed from Gran Canaria in June 1998. The eddy persisted for 7 months though only the first three months are shown here. (Courtesy of Dr Pablo Sangra, University of Las Palmas de Gran Canaria.)

frequently seen in sea surface temperature images being spun off from the flanks of the other smaller islands. The generation of these eddies may not be entirely a result of the oceanic flow past the islands. The Canaries are high volcanic islands situated in a regime of strong southwestward trade winds. The high island peaks block the flow of the trade winds to form extended lee regions downwind. These are bounded by localized horizontal shear in the wind field and so are locations of strong Ekman divergence and convergence. Ekman transport takes place in the near surface layer and is to the right of the wind in the Northern Hemisphere. At the western boundary of the lee, upwelling of deeper waters must compensate the divergence, while at the eastern boundary, sinking must occur. The upwelling and downwelling elevates or depresses, respectively, the pycnocline from its unperturbed depth. Because the downwind scale of the lee is limited, the elevation or depression of the pycnocline tends to form an eddy of the same sign as expected from the current past the island. The vertical motions expected from the horizontal wind shear on the lee boundaries are of the order of tens of meters per day, and so potentially could contribute significantly to the observed eddy production. Despite the frequent reports of eddies spun off from the islands, it has proven difficult to obtain time series observations of eddy generation which would allow determination of basic eddy properties such as average shedding frequency, size, propagation speed or whether eddies are shed alternately from opposite island flanks. Remote sensing even in the subtropics is often blocked by cloud cover and time series in situ sampling is difficult to maintain.

(c) 2011 Elsevier Inc. All Rights Reserved.

ISLAND WAKES

A situation similar to the Canaries has been observed in the case of the Hawaiian archipelago, situated in the North Equatorial Current and trade winds of the Pacific. Downstream of these mountainous islands, the trade winds with speeds of 10– 20 m s1 are separated from the calmer lee by strong boundaries of high wind shear. Locally, the depth of the surface mixed layer depends on wind speed: in the channels between islands, deep mixed layers are observed; in the lee, stirring by the wind is too weak to distribute solar heating down below the surface layer and intense surface warming results during the day. Sharp surface temperature fronts up to 41C, are often associated with these wind shear lines, as is also observed in the Canaries. The lee of islands in both archipelagos is often visible in sea surface temperature images as a significantly warmer triangular area extending downwind. Ekman transports associated with the wind pattern produce pycnocline perturbations as in the Canaries, resulting in intense anticlockwise eddies under the northern shear lines, and less intense clockwise eddies under southern shear lines. The depth of the mixed layer in the lee of Hawaii can vary from less than 20 m in the counterclockwise eddy to more than 120 m in the clockwise eddy. Figure 5 shows a diagram of the Hawaiian island situation which applies equally well to the Canaries. Though the wind has long been viewed as an important generating mechanism for the Hawaiian eddies, it is still unclear how the variability of the wind field affects oceanic eddy generation. The wind itself has often been observed in satellite images of low level clouds to form wakes of counter-rotating atmospheric eddies behind Hawaii and the Canaries. The eddy-shedding period in this case is much

347

shorter, about 10 h, than for oceanic eddies, which have a periodicity of many days. Presumably it is the wind field averaged over the timescale of the ocean eddies that is important. However, it is quite possible that atmospheric eddy shedding is intermittent; during periods when a trapped eddy regime dominates the wind shear lines are relatively stationary and able to feed energy into oceanic eddy production. Further observations are required to unravel the mechanisms at work in these island situations. The North Equatorial Current impinging on the Hawaiian islands, of course, will also tend to produce eddies. The cumulative effect of the many eddies that are spun off is the formation of a mean large scale re-circulation behind the Hawaiian (cf. Barbados) chain that has been named the Hawaii Lee Counter Current. The longevity of individual eddies is further illustrated by one of their surface layer drifters which remained trapped in an anticyclonic vortex formed near the southern flank of the main island, drifting westward over 2000 km at 11 cm s1. This and other drifter tracks showed the remarkable phenomenon of vortex doubling, the process of merging of two vortices of the same sign. When two identicalpffiffiffivortices merge, the radius of the merged eddy is 2 of the original and its period of rotation is doubled. The drifter appeared to show three such vortex-doubling episodes during its trajectory. In the case of larger shallow sea islands in a tidal regime, the flow reverses before a wake can be properly set up. However, the relative motion between the island and surrounding body of water allows the island to act as a ‘stirring rod’ because of the increased flow speeds on the island flanks which produce vertical mixing within a tidal mixing front some distance off the shore. In a stratified region this

Figure 5 Hawaii schematic showing the mechanism of eddy generation by wind shear. Ekman surface layer transports lead to depression and elevation of the pycnocline in regions of convergence and divergence, respectively. This in turn generates oceanic anticyclones and cyclones. (Courtesy of Professor P. Flament, University of Hawaii.)

(c) 2011 Elsevier Inc. All Rights Reserved.

348

ISLAND WAKES

introduces nutrients from the lower layers into the surface layers so enhancing productivity around the island. In the case of the Scilly Islands mixing between low density surface layer and high density bottom layer waters produced intermediate density enriched water spreading out between the layers causing enhancement of the mid-depth chlorophyll maximum in the area around. Similar results were found around St Kilda off western Scotland.

Conclusions Island wakes produced by eddy shedding have been observed in both deep ocean and shallow shelf sea situations. In some cases of shelf sea islands there is no wake as such, because of weak mean flows, but vigorous tidal currents can produce regions of wellmixed water in the surrounding area. In the oceanic case, though there are many examples of eddy and wake observations, there has yet to be a definitive study demonstrating the phenomenon over a range of flow conditions. The effect of islands on the flow regime does have biological consequences, in that enhanced mixing related to shallow sea islands can increase primary production. It has often been argued that island shed eddies may provide a mechanism for retaining fish larvae in the vicinity of their spawning areas. Although the evidence for this is not yet convincing the existence of mean return circulation has been demonstrated in both model and drifter studies of oceanic islands. It was suggested that energy dissipation caused by island flow disturbance could account for 10% of the wind kinetic energy input to the Pacific. However, it is also considered that general turbulent processes known within the deep ocean may be too weak by more than an order of magnitude to explain global redistribution of energy input. In this case vertical mixing must be greater at the ocean boundaries with land than previously considered and the role of islands and island chains may be greater than is presently perceived. Simulations of the Barbados wake indicated that the flow disturbance was extensive, reaching at least eight island diameters downstream.

Both the Canaries and Hawaii archipelagoes clearly control physical conditions for large distances downstream, modifying water masses through mixing, enhancing productivity and shedding long-lived eddies.

See also Canary and Portugal Currents. Ekman Transport and Pumping. Fish Migration, Horizontal. Mesoscale Eddies. Pacific Ocean Equatorial Currents. Upper Ocean Vertical Structure. Wind Driven Circulation.

Further Reading Ari´stegui J, Tett P, Herna´ndez-Guerra A, et al. (1997) The influence of island-generated eddies on chlorophyll distribution: a study of mesoscale variation around Gran Canaria. Deep-Sea Research 44(1): 71–96 Barkley RA (1972) Johnston Atoll’s wake. Journal of Marine Research 30: 201--216. Batchelor GK (1967) An Introduction to Fluid Dynamics. Cambridge: Cambridge University Press. Bowman MJ, Dietrich DE, and Lin CA (1996) Observations and modelling of mesoscale ocean circulation near a small island. In: Maul G (ed.) Small IslandsM: arine Science and Sustainable Development. Coastal and Estuarine Studies, vol. 51, pp. 18--35. Washington: American Geophysical Union. Chopra KP (1973) Atmospheric and oceanic flow problems introduced by islands. Advances in Geophysics 16: 297--421. Flament P, Kennan SC, Lumpkin C, and Stroup ED (1998) The Atlas of Hawai’i. In: Juvik S and Juvik J (eds.) The Ocean, pp. 82--86. Honolulu: University of Hawai’i Press. Simpson JH and Tett P (1986) Island stirring effects on phytoplankton growth. In: Bowman MJ, Yentsch M, and Peterson WT (eds.) Lecture Notes on Coastal and Estuarine Studies, vol. 17, Tidal Mixing and Plankton Dynamics, pp. 41--76. Berlin, Heidelberg: Springer-Verlag. Wolanski E, Imberger J, and Heron ML (1984) Island wakes in shallow coastal waters. Journal of Geophysical Research 89C: 10533--10569.

(c) 2011 Elsevier Inc. All Rights Reserved.

KRILL E. J. Murphy, British Antarctic Survey, Marine Life Sciences Division, Cambridge, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1405–1413, & 2001, Elsevier Ltd.

Introduction Krill play a major role in the transfer of energy in marine food webs, being important consumers of phytoplankton and other zooplankton, and prey of many higher trophic level predators that are often commercially important. The importance of krill in the diet of marine predators is reflected in their name; ‘krill’ comes from the Norwegian whaler’s description of the larger food of the great whales. Krill form an order within the Crustacea, the Euphausiacea, which comprises over 80 species in 10 genera. Detailed keys are available to identify individual species that have broadly similar body pattern (Figure 1). The euphausiids occur in a wide range of habitats – coastal, oceanic, and deepoceanregions – and their distributions also extend into the ice-covered regions of the Arctic and the Antarctic. Krill are generally more abundant in higher latitudes and can occur in such large numbers near the surface that they discolor the water. The phylogenetic relationships of many of the euphausiids are unknown, but for some of the key oceanic species their evolutionary development appears to have been associated with the formation of the major circulation patterns of the world’s oceans. This link to large-scale ocean circulation patterns is also reflected in the population distributions and life histories of the euphausiids. Many of the oceanic krill species occur over broad regions in which the centers of the populations tend to be associated with restricted features of the ocean circulation. However, the patterns of flow often result in transport of krill out of their main breeding regions to areas where they do not breed successfully. This also appears to be crucial to their role in many food webs, providing energy input into regions remote from their own main areas of production. The observation that krill are often transported into regions where they do not reproduce also highlights the colonization potential of the group should any changes occur in patterns of ocean circulation. There are several features that mark the euphausiids as unusual plankton. A number of species are

relatively large with a long life span compared with other zooplankton. The largest of the krill grow to over 60 mm and can live for more than 5 years. Another key feature is that in a number of the species the individuals form dense aggregations known asswarms. In some of the larger euphausiids these swarms might more appropriately be thought of as schools, similar to those formed by small fish, where members of the aggregation are aligned and show coherent patterns of behavior. In the Antarctic the term ‘krill’ is often used to denote a single species: the Antarctic krill, Euphausia superba Dana (Figure 1A). This is, as its name suggests, the ‘superb’ krill that is large in size, occurs in vast numbers in the Southern Ocean, and is central to the Antarctic food web. It is the food of not only the now greatly depleted populations of whales but also many of the seals, penguins and other sea birds, and of fish and squid. It is the most studied species and much of the available information on euphausiids in general is based on knowledge of the Antarctic krill, so it is important to remember that this is something of an extreme representative of the group. A number of the euphausiids have been exploited in fisheries. As krill are typically a low trophic level species there has been recognition of the potential impact this could have on the higher trophic levels of marine food webs. The pivotal role of krill in marine food webs has meant that, particularly in the Antarctic, an ecosystem approach to the management of krill fisheries is being developed that has relevance to the sustainable management of marine ecosystems globally.

Species Separation and Geographical Distributions Euphausiids are found throughout the oceans of the world, but their distributions highlight marked differences in habitat and life history amongst apparently similar species. There is a continuing debate about the exact number of species of euphausiids and the degree of separation of subgroups. There are indications from evolutionary studies of mitochondrial DNA that vicariant speciation (separation by formation of a natural barrier) has been important in the development of euphausiid species in the Antarctic. The generation of the Antarctic Polar Front about 25–22 Ma probably led to the separation of

(c) 2011 Elsevier Inc. All Rights Reserved.

349

350

KRILL

Figure 1 Two krill species: (A) the Antarctic krill, Euphausia superba and (B) a North Atlantic krill, Meganyctiphanes norvegica.

the ‘Antarctic clade’ (E. superba and E. crystallorophias) from the sister clade of E. vallentini and E. frigida dated at about 20 Ma. Although some euphausiid species occur in coastal and bathypelagic regions (1000–2500 m), most are foundin oceanic epipelagic (0–200 m) and mesopelagic regions (200–1000 m). Although broad distributions have been described for many of the euphausiids, because these animals frequently occur in only relatively low numbers their local distributions are often not well defined. Generally, there is a trend of increasing abundance of krill at higher latitudes. However, there are variations in this pattern, with strong links between the ocean current systems and the regional distribution of krill species. A feature of the euphausiids is that in Southern Ocean and Southern Hemisphere regions many of the key species occur across the full longitudinal range (Figure 2). In the Southern Ocean the ocean circulation is circumpolar, so the same basic pattern of species distribution is found throughout the connected ocean. The key species in the mainly icecovered regions is E. crystallorophias which inhabits the Antarctic continental shelf, although on occasion it has been found transported northward by the major current flows. Further to the north in the seasonally ice-covered areas of the main flow regions of the Antarctic Circumpolar Current are E. superba and E. frigida, with Thysanossa vicina and T. macrura extending northwards to the Antarctic Polar Front. All of these species have heterogeneous distributions in the region. For E. superba there appear to be centers of population in which they can spawn and reproduce successfully, separated by and possibly connected through, regions that are not favorable to breeding but in which krill are found (Figure 3). E. triacantha overlaps the northern limit of E. superba in the south and in the north it overlaps the southern limit of the range of E. vallentini extending north to south of 401S. Further north still are less abundant species suchas E. longirostris and E. lucens that extend north of 401S in areas encompassed by the eastward flows in the southern regions of the main ocean gyres of the Pacific, Atlantic, and Indian Oceans. To the north of this E. similis occurs in all three ocean basin regions, extending from about 50–601S to 301S, but the species is also present further north inthe north-west region of the Indian Ocean to the north of Madagas car. Across the subtropical and tropical regions there is a wide range of species. One that occurs in all the ocean basins is E. tenera, where it has awide distribution but is not very abundant in any region. There are a number of other species found in both the Atlantic and Pacific Oceans. Some species,

(c) 2011 Elsevier Inc. All Rights Reserved.

KRILL

Atlantic only

351

M. norvegica T. longicaudata

E. similis E. longirostris E. lucens

Atlantic & Pacific

T. inermis

E. vallentini

T. raschii

E. tenera

E. triachantha

E. gibba group E. superba T. macrura, T. vicina E. frigida E. crystallorophias Pacific only

T. longipes E. pacifica

80˚ N

60˚ N

40˚ N

0

20˚ S

40˚ S

60˚ S

80˚ S

Antarctic

Arctic Figure 2 The broad geographical distributions of some key euphausiid species.

particularly in the central North Pacific and North Atlantic, are abundant but only found in one of the ocean basins. Key species that show this pattern are E. pacifica and T. longipes/T. inspinata that occur only in the northern North Pacific, while in the North Atlantic Meganyctiphanes norvegica and T. longicaudata are dominant (Figure 2). In the northern North Atlantic and North Pacific there are species that occur in both oceans and through into the Arctic regions in the far north. In particular there are two important species, T. raschii and T. inermis, with distributions extending from about 451N to about 801N, although breeding is largely restricted to areas south of 701N. As well as geographical differences there are also marked differences in vertical distribution and many of the species show some form of vertical migration. For example, E. pacifica occurs mainly above 300 m during the day, moving nearer the surface (o150m) at night, while M. norvegica occurs between 100 and 500 m during the day and vertically migrates to shallower depths at night, and in the south

E. superba occurs mainly above about 250 m and migrates nearer the surface at night.

Growth, Development, Physiology Krill species show a range of development strategies that vary between species, and also with the environmental conditions to which they are exposed. Studies of krill population dynamics and development are made difficult because of problems in determining the age of a number of the species. Traditional techniques involving the analysis of the population age structure are still relied upon and a range of mathematical techniques have been employed to distinguish different cohorts in length– frequency size distributions. These are not always definitive and a range of other techniques has been explored such as using age pigment analyses, multiple-morphometric analyses, analyses of structures in the eye, and laboratory maintenance of live specimens. None of these techniques has so far

(c) 2011 Elsevier Inc. All Rights Reserved.

352

KRILL

120°

90°

°

90°

120

60°

°

30°

150

°

0°

180°

30 °

150°

60

Figure 3 The main regions of occurrence of Antarctic krill E. superba in the Southern Ocean and the pattern of surface circulation fromthe FRAM model (FRAM Group).

provided a good and practical solution to determining the age of krill. However, there is general agreement about the broad characteristics of growthand development of many of the key species. In the Southern Ocean, early studies of E. superba indicated a 2–3 year life cycle based mainly on samplesfrom open-ocean regions. However, further detailed analyses of the size-structure and development of E. superba populations have led to a revision in the life-span up to 45 years with suggestions that in some areas 5–7 year classes can be identified. Laboratory experiments have maintained krill obtained from the sea for 46 years, indicating that a total age of 7–8 years is probably possible in the wild. The development and growth of the krill will depend on the conditions to which they are exposed. E. superba can reach a size of 460 mm with

indications that growth may be very plastic, as in other euphausiid species, varying with the environmental conditions. Thus, krill in more northern and warmer regions may grow more rapidly and develop earlier than krill further south. In these more northern regions, such as around the Island of South Georgia which lies at about 541S, near the Antarctic Polar Front, the krill do not reproduce successfully, with few indications of any viable larvae being found in thearea. The E. superba population in these regions is probably maintained by advection inputs from further south in the Southern Scotia Sea, Weddell Searegion, and from around the Antarctic Peninsula. Such a plastic range of development is illustrated clearly in the northern species T. inermis and T. raschii. These species have a maximum age of

(c) 2011 Elsevier Inc. All Rights Reserved.

KRILL

1 year at the south of their distribution (about 451N), where as further north they survive to over 2 years old, spawning in each year, although females may not mature until their second year. Continuing northward, the maximum age increases and maturation is delayed further with spawning delayed to year 3, and a maximum age of 3 years. In the high Arctic waters the krill still mature but do not spawn and it is the water circulation bringing krill from further south that maintains the species in these areas. Across this range T. inermis grows to over 20 mm, but the rates involved vary depending on the conditions, with slower growth and development occurring further north in their range. M. norvegica is one of the most abundant North Atlantic krill and individuals can reach a maximum size of over 45 mm in some regions. This species shows less age variation across its range than the more northern Thysanossa species, but the variation is still significant. In the south of its range individuals live up to about 1 year and spawn only once, whereas further north they reach over 2 years of age, spawning more than once. Like T. inermis, this species does not spawn in the extreme northern part of its range, so advection in the current systems is again important in maintaining the distribution. In the Pacific E. pacifica also shows this plastic character of changing maximum age with environmental variation. At the southern limit of its range individuals have a very short life span of only 6–8 months, whereas further north the maximum age is extended to about 15–21 months. In the most northern parts of the range the krill survive to over 2 years old and probably spawn twice. The maximum size across the range is about 20–22 mm, but growth is slower in the regions further north. As well as these general changes in development and life span, there are also sex-related differences. For example, in the northern Thysanossa species the males mature at just over 1 year old while the females mature mainly at over 2 years of age. In E. pacifica both sexes mature and spawn at 1 year old, but females may continue to survive and spawn at over 2 years old. In E. superba the situation can be different, with females spawning and maturing earlier at 2 years old, while males may not mature until over 3 years old. The euphausiids have the potential for rapid growth and development under suitable conditions, moulting as they increase in size. So, for example, E. superba has an energy input of perhaps 20% of body carbon per day or greater, sustained by a high and effective rate of filtration. This level of energy input can result in growth rates of 40.1 mm d 1,

353

particularly for the younger age groups. Krill have the capacity for a large and sustained reproductive output under good conditions,with continuous or multiple spawning occurring through the season in some species. A key question for many of the species is how they survive during winter when food appears scarce, particularly in the extremely seasonal environments of the polar oceans. Studies of polar species show that the krill utilize stored lipids as a major energy source, but the dynamics of storage and utilization vary greatly between species. The lipids are accumulated primarily for winter survival or reproduction, but they may also provide a small degree of buoyancy that may help reduce the costs of swimming. In the Southern Ocean, the diet of E. superba varies with age. Phytoplankton sources are important for the early stages while older groups utilize more animal-based food sources or detritus. Lipids are utilized in winter, but a strong seasonal bloom of production is necessary for reproduction. For E. superba the suggestion is that winter survival is dependent not only on reduced metabolic rate, a potential reduction in size, and use of lipids, but also on the use of alternative food sources. Antarctic krill have been observed to get smaller during poor feeding conditions in the laboratory, but it is unclear how much this occurs in the ocean. Larval E. superba are dependent on sea ice as a habitat and the observation of krill grazing algae associated with the ice indicates that they can utilize this as an alternative food source. It remains unclear how important sea ice algae are for maintaining adult E. superba and this is likely to be a variable contribution to the diet depending on opportunity of access to the right feeding conditions. E. superba are also known to graze other components of the plankton, including copepods, so that a range of possible feeding strategies is likely to be open to them, depending on opportunity. E. crystallorophias occupies the area further south in the Antarctic where the spawning appears to occur before the main bloom, suggesting that lipid stores are used for survival and for reproduction. T. macrura has a similar distribution to E. superba but spawns earlier so it again is dependent on lipid stores for reproduction, but may also utilize other food sources to get through the winter. In northern regions T. inermis converts phytoplankton rapidly into lipids to cope with the seasonal environment, but also utilizes other available organic material such as detritus. M. norvegica also builds up high lipid stores but is more carnivorous, utilizing lipid-rich copepods.

(c) 2011 Elsevier Inc. All Rights Reserved.

354

KRILL

Overall, there appears to be a pattern from high to lower latitudes in the strategies of feeding and energy storage. Truly polar species such as E. crystallorophias and T. inermis rely totally on the seasonal phytoplankton bloom and lipid stores, whereas species such as E. superba, T. raschii and T. macrura survive winter utilizing alternative food sources and require the bloom to reproduce. Further away from the polar regions, species such as T. longicaudata and M. norvegica are more carnivorous, utilizing copepods as their main food source.

Spatial Distribution At large scales there are heterogeneities inthe distributions of euphausiids that extend for tens or hundreds of kilometers. Within these broad aggregations there are also more dense regions where krill form patches, swarms, or schools (Figure 4) forming a distribution generated by a very dynamic system, with aggregation and dispersal over a wide range of scales. These patches can be very dense and compact and it has been suggested that it is likely that on occasion all species of euphausiids aggregate to some extent. This ability to form such dense aggregations is certainly found in a number of species, particularly E. superba, but also

M. norvegica, Nyctyphanes australis, E. pacifica and E. lucens. The generation of such patchy distributions is the result of interactions between biological and physical processes over a range of scales. Over small scales and in very dense aggregations, behavior probably dominates. Swimming speeds can be high – B 20 cm s 1 in short bursts in Antarctic krill – so individuals have a marked ability to undertake directed movement at least over relatively short spatial scales. The formation of these smaller aggregations, along withdiurnal vertical migration, is considered to be mainly a predator avoidance effect. However, it will also lead to changes in the dynamics of the interaction of krill with their food and may generate complex outcomes in the dynamics of planktonic systems. At larger scales physical processes probably dominate so that aggregations are dependent on physical concentration mechanisms in areas of shelf-breaks, around islands, in ice edge regions or associated with eddies. This larger-scale aggregation may be a precondition for behavioral effects to dominate at smaller scales. Krill within aggregations appear to share more similar characteristics in size and maturity compared with those in other aggregations in the same area. The densities of krill within these aggregations can be well in excess of 10 000 m 3;

0.5 km

100 km

Figure 4 A hydroacoustic trace of an aggregation of Antarctic krill, E. superba.

(c) 2011 Elsevier Inc. All Rights Reserved.

KRILL

over 50 000 m 3 has been estimated for E. superba and E. pacifica and over 500 000 m 3 recorded for N. australis, E. lucens and M. norvegica. Some of the aggregations are extremely large and can account for a considerable proportion of the total biomass in an area. So for example, in one survey of Antarctic krill 410% of the regional biomass was recorded in just one aggregation that extended about 1 km horizontally. This large aggregation was observed in the vicinity of a large number of whales, suggesting that some of the very large, dense aggregations may be the result of intense predator–prey interaction and emphasizes the dynamic nature of the spatial distribution of krill. This makes the design of krill distribution surveys using nets or hydroacoustic techniques challenging and the survey data require careful interpretation and analysis. A number of the species undertake diurnal vertical migration, rising to nearer the surface and dispersing at night. This has been shown clearly in M. norvegica, whereas in E. superba vertical migration appears to be highly variable and may depend on local physical conditions, surface predator affects, and predation effects from below, particularly in areas of the shelf. In addition to the importance of predation effects the behavioral tracking of particular isolumes has been suggested as a mechanism involved in diurnal vertical migration and some species appear to show an endogenous rhythm. Seasonal changes in the pattern of aggregation and vertical migration have also been noted in some species. In one area it has been observed that during spring aggregations of E. superba are of the order of 0.7–2 km in length, whereas they are smaller and moredense in summer, and larger and less dense in autumn and winter. In the same study many of the swarms occurred in the upper 70 m during the summer, while in winter many were below 100 m deep. Other studies have found no such vertical change in depth distribution during the year, although diurnal vertical migration did change, being marked only during the spring and autumn.

Role in the Food Web Krill as Consumers

Krill show a range of feeding strategies from complete herbivory to total carnivory, with a full range of capabilities and flexible feeding strategies in between. In the Antarctic they can have a major impact on the large diatoms that form the major components of the intense blooms associated with the summer retreat of the sea ice. Antarctic krill are also known to consume copepods and negative

355

correlations have been shown between the krill occurrence and the density of copepods and the phytoplankton concentration. This indicates that euphausiids are important in the plankton dynamics of these regions. They generate large fecal pellets which have high rates of sinking, suggesting that they can be important in the export of carbon from the surface layers. Rapid grazing of diatom blooms in some areas can therefore lead to a rapid flux of material to deeper ocean regions. The highly aggregated nature of the distributions is also likely to be important in determining the plankton dynamics, not just in terms of producing an interactive mosaic of production and consumption, but also in terms of the nutrient regime. Large krill swarms will generate high concentrations of ammonia that may favor the production of particular size groups of phytoplankton, leading to complex interactions in the plankton. Their role as consumers continues to be studied, but across the order they clearly have the capacity to feed on a wide range of food sources including diatoms, coccolithophores, dinoflagellates, chaetognaths, copepods, and other crustaceans; cannibalism has also been shown in some species. On the basis of observed variations in feeding strategies, it has been suggested that most species of euphausiids can adapt their feeding to utilize what is available, modifying their feeding strategies depending on the food they encounter. Krill as Prey

Krill are prey of many higher trophic level predators and as such play a key role throughout the oceans by transferring energy up the food chain. The baleen whales are the most well known predator, eating krill throughout their range, and despite the massive depletion of their populations due to harvesting they are still important krill predators. So for example, dense aggregations of M. norvegica and T. raschii in the Gulf of St Lawrence are associated with high abundances of fish (capelin) and a range of whale species including minkes, fin, blues, humpbacks, sperm, and beluga. Seals are also major predators of euphausiids in many areas. In Arctic waters, for example, harp seals consume M. norvegica and T. inermis, while in the Southern Ocean, crabeater seals consume E. crystallorophias. Further north, around some of the sub Antarctic islands such as South Georgia, the previously exploited fur seal populations that are now very large consume a considerable quantity of E. superba. Euphausiids also comprise a key component of the diet of a wide range of fish species, many of which

(c) 2011 Elsevier Inc. All Rights Reserved.

356

KRILL

are, or were, exploited. In the North Atlantic these include herring, cod, haddock, whiting, and mackerel. M. norvegica is probably the key species consumed, but others such as T. raschii, T. inermis and T. longicaudata are also important. In the North Pacific and adjacent regions E. pacifica is eaten by most commercial fish species, including Pacific cod, walleye pollack, chub mackerel, and sand lance. Other krill taken include T. raschii and T. inermis. The importance of euphausiids in the diet of many commercially exploited fish species is seen throughout the world. So for example around Australia N. australis is eaten by bluefin tuna and striped tuna, while in the Antarctic E. superba is eaten by the Mackerel icefish. Seabirds are also important predators of euphausiids throughout the world. In the North Atlantic a wide range of bird species consume M. norvegica and T. inermis including gulls, puffins, kittiwakes, and fulmars. The importance of sea-birds as predators of krill is highlighted in the Southern Ocean where E. superba is a key item in the diet and consumed in vast numbers by penguins (gentoos, macaronis, and Adelies) and by flying sea-birds including albatrosses such as grey-headed and black-browed albatross. This broad range view of euphausiids as prey emphasizes the important role that krill play in transferring energy to higher trophic levels in marine food webs worldwide. One of the key reasons for the importance of euphausiid species in food webs is the heterogeneity of krill distribution on a range of spatial and temporal scales. Different predators exploit the aggregation pattern with different foraging strategies, so exploiting different scales of pattern in the prey field. The pattern generated by the biological or biological–physical interactions will thus determine which predators can exploit the prey and hence the structure of the food web.

Krill Fisheries There are extensive fisheries for E. superba in the Southern Ocean, while in the Pacific off Japan there are important fisheries for E. pacifica. There is a more limited E. pacifica fishing off western Canada and there have also been intermittent fisheries for other species. The Southern Ocean fishery for E. superba is the largest and started at beginning of the 1970s, peaking in the early 1980s at 0.5 million tonnes. The fishery has since declined with changes in its economic basis. Catches over recent years have been o100 000 tonnes. The E. superba fishery in the Scotia Sea region is linked to seasonal seaice changes. During winter the fishery operates in the north

around South Georgia, it moves further south in the spring with the ice, to the area near the South Orkney Islands, and then during summer the fishery exploits krill around the Antarctic Peninsula. The management regime for E. superba takes account of krill recruitment variability, growth, and mortality to examine effects of various harvesting levels. Decision rules are included tomaintain stocks at a level that takes into account the dependent predators. At the current time, catch levels are much lower than the allowable catch (o10%) and future expansion of the fishery depends on the development of new products utilizing krill. An ecosystem approach is being developed for managing Southern Ocean fisheries and extensive predatormonitoring programs are operating. The challenge here is to develop management decision rules that consider not just the target species, the krill, but also incorporate ecological information from a number of levels in the food web,taking into account dependent species as well as environmental links. This ecosystem rather than species-based approach is one that will be increasingly relevant elsewhere.

Krill Variability There is considerable evidence of the importance of variation in the physical environment and circulation systems of the oceans in determining the distribution and abundance of krill. For example, links have been noted between variations in the oceanographic regimes associated with El Nin˜o events and the recruitment of E. pacifica in the North Pacific, while water temperature variations have been linked to 2–3 year variations in T. inermis populations. Biological processes associated with the environmental variation are also important in generating the variation observed in euphausiid populations. For example, there are marked interannual variations in the abundance of E. superba in the Southern Ocean, where recruitment strength has been linked to variations in the extent and concentration of sea ice. The current view is that increased sea ice cover and extent lead to favorable conditions for spawning and larval survival. The sea ice is thought to provide better over winter conditions for the krill. Salps compete with krill for phytoplankton – in poor sea ice years salp numbers are increased and krill recruitment is reduced. Further north in their range, E. superba abundance is dependent on the transport of krill in the ocean currents as well as fluctuations in the strength of particular cohorts. Given the importance of euphausiids in marine food webs throughout the world’s oceans, they are

(c) 2011 Elsevier Inc. All Rights Reserved.

KRILL

potentially important indicator species for detecting and understanding climate change effects. Changes in ocean circulation or environmental regimes will be reflected in changes in growth, development, recruitment success, and distribution. These effects may be most notable at the extremes of their distribution where any change in the pattern of variation will result in major changes in food web structure. Given their significance as prey to many commercially exploited species, this may also have a major impact on harvesting activities. A greater understanding of the large-scale biology of the euphausiids and the factors generating the observed variability is crucial. Obtaining good long-term and large-scale biological and physical data will be fundamental to this process.

See also Antarctic Circumpolar Current. Baleen Whales. Copepods. Marine Mammals: Sperm Whales and Beaked Whales.Phalaropes. Plankton. Sea Ice: Overview. Seals.

Further Reading

357

Antarctic marine ecosystem: practical implementation of the Convention on the Conservation of the Antarctic Marine Living Resources (CCAMLR). ICES Journal of Marine Science 57: 778--791. Everson I (ed.) (2000) Krill: Biology, Ecology and Fisheries. Oxford: Blackwell Science. Everson I (2000) Introducing krill. In: Everson I (ed.) Krill: Biology, Ecology and Fisheries. Oxford: Blackwell Science. Falk-Petersen S, Hagen W, Kattner G, Clarke A, and Sargent J (2000) Lipids, trophic relationship, and biodiversity in Arctic and Antarctic krill. Canadian Journal of Fisheries and Aquatic Sciences 57: 178--191. Mauchline JR (1980) The biology of the Euphausids. Advances in Marine Biology 18: 371--677. Mauchline JR and Fisher LR (1969) The biology of the Euphausids. Advances in Marine Biology 7: 1--454. Miller D and Hampton I (1989) Biology and Ecology of the Antarctic Krill. BIOMASS Scientific Series, 9. Cambridge: SCAR & SCOR. Murphy EJ, Watkins JL, Reid K, et al. (1998) Interannual variability of the South Georgia marine ecosystem: physical and biological sources of variation. Fisheries Oceanography 7: 381--390. Siegel V and Nichol S (2000) Population parameters. In: Everson I (ed.) Krill: Biology, Ecology and Fisheries. Oxford: Blackwell Science.

Constable AJ, de la Mare W, Agnew DJ, Everson I, and Miller D (2000) Managing fisheries to conserve the

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS B. Qiu, University of Hawaii at Manoa, Hawaii, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1413–1425, & 2001, Elsevier Ltd.

Introduction The Kuroshio and Oyashio Currents are the western boundary currents in the wind-driven, subtropical and subarctic circulations of the North Pacific Ocean. Translated from Japanese, Kuroshio literally means black (‘kuro’) stream (‘shio’) owing to the blackish – ultramarine to cobalt blue – color of its water. The ‘blackness’ of the Kuroshio Current stems from the fact that the downwelling-dominant subtropical North Pacific Ocean is low in biological productivity and is devoid of detritus and other organic material in the surface water. The subarctic North Pacific Ocean, on the other hand, is dominated by upwelling. The upwelled, nutrient-rich water feeds the Oyashio from the north and leads to its nomenclature, parent (‘oya’) stream (‘shio’). The existence of a western boundary current to compensate for the interior Sverdrup flow is well understood from modern wind-driven ocean circulation theories. Individual western boundary currents, however, can differ greatly in their mean flow and variability characteristics due to different bottom topography, coastline geometry, and surface wind patterns that are involved. For example, the bimodal oscillation of the Kuroshio path south of Japan is a unique phenomenon detected in no other western boundary current of the world oceans. Similarly, interaction with the semi-enclosed and often ice-covered marginal seas and excessive precipitation over evaporation in the subarctic North Pacific Ocean make the Oyashio Current considerably different from its counterpart in the subarctic North Atlantic Ocean, the Labrador Current. Because the Kuroshio and Oyashio Current sexert a great influence on the fisheries, hydrography, and meteorology of countries surrounding the western North Pacific Ocean, they have been the focus of a great amount of observation and research in the past. This article will provide a brief review of the dynamic aspects of the observed Kuroshio and Oyashio Currents: their origins, their mean flow patterns, and their variability on seasonal-to-interannual timescales. The article consists of two sections, the first

358

focusing on the Kuroshio Current and the second on the Oyashio Current. Due to the vast geographical areas passed by the Kuroshio Current (Figure 1), the first section is divided into three subsections: the region upstream of the Tokara Strait, the region south of Japan, and the Kuroshio Extension region east of the Izu Ridge. As will become clear, the Kuroshio Current exhibits distinct characteristics in each of these geographical locations owing to the differing governing physics.

The Kuroshio Current Region Upstream of the Tokara Strait

The Kuroshio Current originates east of the Philippine coast where the westward flowing North Equatorial Current (NEC) bifurcates into the northward-flowing Kuroshio Current and the southward-flowing Mindanao Current. At the sea surface, the NEC bifurcates nominally at 121N–131N, although this bifurcation latitude can change interannually from 111N to 14.51N. The NEC’s bifurcation tends to migrate to the north during El Nin˜o years and to the south during La Nin˜a years. Below the sea surface, the NEC’s bifurcation tends to shift northward with increasing depth. This tendency is due to the fact that the southern limb of the wind-driven subtropical gyre in the North Pacific shifts to the north with increasing depth. Branching northward from the NEC, the Kuroshio Current east of the Philippine coast has a mean geostrophic volume transport, referenced to 1250 dbar, of 25 Sv (1 Sverdrup ¼ 106 m3 s1). Seasonally, the Kuroshio transport at this upstream location has a maximum (B30 Sv) in spring and a minimum (B19 Sv) in fall. Similar seasonal cycles are also found in the Kuroshio’s transports in the East China Sea and across the Tokara Strait. As the Kuroshio Current flows northward passing the Philippine coast, it encounters the Luzon Strait that connects the South China Sea with the open Pacific Ocean (Figure 2). The Luzon Strait has a width of 350 km and is 2500 m deep at its deepest point. In winter, part of the Kuroshio water has been observed to intrude into the Luzon Strait and form a loop current in the northern South China Sea (see the dashed line in Figure 2). The loop current can reach as far west as 1171E, where it is blocked by the presence of the shallow shelf break off the south-east coast of China. The formation of the loop current is probably due to the north-east monsoon, prevailing

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS

60˚ N

359

Bering Sea

E Ka ast m ch

at

ka

C. .

Sea of Okhotsk 50˚ N

ka n Alas

ka

re

Gy

C.

s Ala

Subarctic Current

shio

40˚ N

Mixed water region

Oya

Ts Cu ush rre ima nt

Western Subarctic Gyre

m ea Str

Kuroshio extension 30˚ N

20˚ N

Kuroshio Cu

nt

Recirculation

rre

North Pacific Ocean

Subtropical countercurrent

North Equatorial Current 10˚ N

120˚E

Mindanao Current 130˚E

140˚E

150˚E

160˚E

170˚E

180˚

170˚W

160˚W

150˚W

Figure 1 Schematic current patterns associated with the subtropical and subarctic gyres in the western North Pacific Ocean.

from November to March, which deflects the surface Kuroshio water into the northern South China Sea. During the summer months from May to September when the south-west monsoon prevails, the Kuroshio Current passes the Luzon Strait without intrusion. In the latitudinal band east of Taiwan (221N– 251N), the northward-flowing Kuroshio Current has been observed to be highly variable in recent years. Repeat hydrographic and moored current meter measurements between Taiwan and the southernmost Ryukyu island of Iriomote show that the variability of the Kuroshio path and transport here are dominated by fluctuations with a period of 100 days. These observed fluctuations are caused by impinging energetic cyclonic and anticyclonic eddies migrating from the east. The Subtropical Counter current (STCC) is found in the latitudinal band of 221N– 251N in the western North Pacific. The STCC, a shallow eastward-flowing current, is highly unstable due to its velocity shear with the underlying, westward-flowing NEC. The unstable waves generated by the instability of the STCC-NEC system tend to move westward while growing in amplitude. The

cyclonic and anticyclonic eddies that impinge upon the Kuroshio east of Taiwan are results of these large-amplitude unstable waves. Indeed, satellite measurements of the sea level (Figure 3) show that the Kuroshio east of Taiwan has higher eddy variability than either its upstream counterpart along the Philippine coast or its downstream continuation in the East China Sea. The Kuroshio Current enters the East China Sea through the passage between Taiwan and Iriomote Island. In the East China Sea, the Kuroshio path follows closely along the steep continental slope. Across the PN-line in the East China Sea (see Figure 2 for its location), repeat hydrographic surveys have been conducted on a quarterly basis by the Japan Meteorological Agency since the mid-1950s. Based on the measurements from 1955 to 1998, the volume transport of the Kuroshio across this section has a mean of 24.6 Sv and a seasonal cycle with 24.7 Sv in winter, 25.4 Sv in spring, 25.2 Sv in summer, and 22.8 Sv in fall, respectively. In addition to this seasonal signal, large transport changes on longer timescales are also detected across this section

(c) 2011 Elsevier Inc. All Rights Reserved.

360

KUROSHIO AND OYASHIO CURRENTS

34°N 0

Ky us hu

10 East China Sea

32°N

PN-line

30°N China

Tokara St.

la nd s

28°N

Taiw an

Ry uk y

u

Is

26°N

24°N

Iriomote Is.

100

22°N

20°N

18°N

0 20 1000

Luzon St.

South China Sea

116°E

118°E

Luzon

120°E

122°E

124°E

126°E

128°E

130°E

132°E

Figure 2 Schematic representation of the mean Kuroshio path (solid thick line) along the North Pacific western boundary. The thick dashed line south of Taiwan denotes the wintertime branching of the Kuroshio water into the Luzon Strait in the form of a loop current. PN-line denotes the repeat hydrographic section across which long-term Kuroshio volume transport is monitored (see Figure 4). Selective isobaths of 100 m, 200 m, and 1000 m are depicted.

(see Figure 4). One signal that stands out in the timeseries of Figure 4 is the one with the decadal timescale. Specifically, the Kuroshio transport prior to 1975 was low on average (22.5 Sv), whereas the mean transport value increased to 27.0 Sv after 1975. This decadal signal in the Kuroshio’s volume transport is associated with the decadal Sverdrup transport change in the subtropical North Pacific Ocean. Although the main body of the Kuroshio Current in the East China Sea is relatively stable due to the topographic constraint, large-amplitude meanders are frequently observed along the density front of the Kuroshio Current. The density front marks the shoreward edge of the Kuroshio Current and is located nominally along the 200 m isobath in the East China Sea. The frontal meanders commonly originate along the upstream Kuroshio front north east of Taiwan and they evolve rapidly while propagating downstreamward. The frontal meanders have typical wavelengths of 200–350 km, wave periods of 10–20 days, and downstreamward phase speeds of 10–25 cm s1. When reaching the Tokara Strait, the fully developed

frontal meanders can shift the path of the Kuroshio Current in the strait by as much as 100 km. Around 1281E–1291E and 301N, the Kuroshio Current detaches from the continental slope and veers to the east toward the Tokara Strait. Notice that this area is also where part of the Kuroshio water is observed to intermittently penetrate northward onto the continental shelf to feed the Tsushima Current. The frontal meanders of the Kuroshio described above are important for the mixing and water mass exchanges between the cold, fresh continental shelf water and the warm, saline Kuroshio water along the shelf break of the East China Sea. It is this mixture of the water that forms the origin of the Tsushima Current. The volume transport of the Tsushima Current is estimated at 2 Sv.

Region South of Japan

The Kuroshio Current enters the deep Shikoku Basin through the Tokara Strait. Combined surface current

(c) 2011 Elsevier Inc. All Rights Reserved.

361

KUROSHIO AND OYASHIO CURRENTS

T/ P rms height variability (m) 60˚N

50˚N

0.08 40˚N

0.14

0.16 0 0.26 .18 0.320.30 0.20

0.24

0.06

0.08

30˚N

0.10 0.12 0.14 20˚N

0.10

0.06

0.06

0.08

0.08

0.08 0.10

10˚N

0.08

8

0.0

0 0.1 0˚

140˚E

120˚E

160˚E

0.04

160˚W

180˚

0.08

0.12

0.16

0.20

140˚W

0.24

0.28

120˚W

0.32

0.36

100˚W

80˚W

0.40

Figure 3 Map of the root-mean-square (rms) sea surface height variability in the North Pacific Ocean, based on the TOPEX/ POSEIDON satellite altimetric measurements from October 1992 to December 1997. Maximum rms values of 40.4 m are found in the upstream Kuroshio Extension region south east of Japan. Sea surface height variability is also high in the latitudinal band east of Taiwan. (Adapted with permission from Qiu B (1999) Seasonal eddy field modulation of the North Pacific Subtropical Countercurrent: TOPEX/Poseidon observations and theory. Journal of Physical Oceanography 29: 2471–2486.)

35

Transport (Sv)

30

25

20

15 1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

Year Figure 4 Time-series of the geostrophic volume transport of the Kuroshio across the PN-line in the East China Sea (see Figure 2 for its location). Reference level is at 700 dbar. Quarterly available transport values have been low-pass filtered by the 1-year running mean averaging. (Data courtesy of Dr M. Kawabe of the University of Tokyo.)

(c) 2011 Elsevier Inc. All Rights Reserved.

362

KUROSHIO AND OYASHIO CURRENTS

and hydrographic observations show that the Kuroshio’s volume transport through the Tokara Strait is about 30 Sv. Inference of transport from the sea level measurements suggests that the Kuroshio’s transport across the Tokara Strait is maximum in spring/summer and minimum in fall, a seasonal cycle similar to that found in the upstream Kuroshio Current. Further downstream, offshore of Shikoku, the volume transport of the Kuroshio has a mean value of 55 Sv. This transport increase of the Kuroshio in the deep Shikoku Basin is in part due to the presence of an anticyclonic recirculation gyre south of the Kuroshio. Subtracting the contribution from this recirculation reduces the mean eastward transport to 42 Sv. Notice that this ‘net’ eastward transport of the Kuroshio is still larger than its inflow transport through the Tokara Strait. This increased transport, B12 Sv, is probably supplied by the north-eastward-flowing current that has been occasionally observed along the eastern flank of the Ryukyu Islands. Near 1391E, the Kuroshio Current encounters the Izu Ridge. Due to the shallow northern section of the ridge, the Kuroshio Current exiting the Shikoku Basin is restricted to passing the Izu Ridge at either around 341N where there is a deep passage, or south of 331N where the ridge height drops. On interannual timescales, the Kuroshio Current south of Japan is known for its bimodal path fluctuations. The ‘straight path’, shown schematically by path A in Figure 5, denotes when the Kuroshio flows

closely along the Japan coast. The ‘large-meander path’, shown by path B in Figure 5, signifies when the Kuroshio takes a detouring offshore path. In addition to these two stable paths, the Kuroshio may take a third, relatively stable path that loops southward over the Izu Ridge. This path, depicted as path C in Figure 5, is commonly observed during transitions from a meandering state to a straight-path state. As the meander path of the Kuroshio can migrate spatially, a useful way of indexing the Kuroshio path is to use the mean distance of the Kuroshio axis from the Japan coast from 1321E to 1401E. South of Japan, the Kuroshio axis is well represented by the 161C isotherm. Based on this representation and seasonal water temperature measurements, Figure 6 shows the time-series of the Kuroshio path index from 1955 to 1998. A low index in Figure 6 denotes a straight path, and a high index denotes an offshore meandering path of the Kuroshio. From 1955 to 1998, the Kuroshio large meanders occurred in 1959–62, 1975–79, 1982–88, and 1990. Clearly, the large-meanders occurrence is aperiodic. Once formed, the meander state can persist over a period ranging from a year to a decade. In contrast, transitions between the meander and straight-path states are rapid, often completed over a period of several months. It is worth noting that development of the large meanders is often preceded by the appearance of a small meander south of Kyushu, which migrates eastward and becomes stationary after reaching 1361E.

36°N Honshu

0

100

34°N

A

ku

iko

Kyushu

Sh

C

4000

2000

32°N B

Izu R

30°N

idge

00

10

Tokara Strait

400

28°N 130°E

132°E

134°E

136°E

138°E

140°E

142°E

Figure 5 Schematic stable paths of the Kuroshio Current south of Japan. (Adapted with permission from Kawabe M (1985) Sea level variations at the Izu Islands and typical stable paths of the Kuroshio. Journal of the Oceanography Society of Japan 41: 307–326.) Selective isobaths of 1000 m, 2000 m, 4000 m, 6000 m, and 8000 m are included.

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS

363

2.5

Distance (°lat)

2.0

1.5

1.0

0.5 1955

1960

1965

1970

1975

1980

1985

1990

1995

Year Figure 6 Time-series of the Kuroshio path index from 1955 to 1998, where the Kuroshio path index is defined as the offshore distance of the Kuroshio axis (inferred from the 161C isotherm at the 200 m depth) averaged from 1321 to 1401E. Solid dots denote the seasonal index values and the solid line indicates the annual average. (Adapted with permission from Qiu B and Miao W (2000) Kuroshio path variations south of Japan: Bimodality as a self-sustained internal oscillation. Journal of Physical Oceanography 30: 2124–2137.)

Several mechanisms have been proposed to explain the bimodal path variability of the Kuroshio south of Japan. Most studies have examined the relationship between the Kuroshio’s path pattern and the changes in magnitude of the Kuroshio’s upstream transport. Earlier studies of the Kuroshio path bimodality interpreted the meandering path as stationary Rossby lee wave generated by the protruding coastline of Kyushu. With this interpretation, the Kuroshio takes a meander path when the upstream transport is small and a straight path when it is large. By taking into account the realistic inclination of the Japan coast from due east, more recent studies have provided the following explanation. When the upstream transport is small, the straight path is stable as a result of the planetary vorticity acquired by the north-eastwardflowing Kuroshio being balanced by the eddy dissipation along the coast. When the upstream transport is large, positive vorticity is excessively generated along the Japan coast, inducing the meander path to develop downstream. In the intermediate transport range, the Kuroshio is in a multiple equilibrium state in which the meandering and straight paths coexist. Transitions between the two paths in this case are determined by changes in the upstream transport (e.g. the transition from a straight path to a meander path requires an increase in upstream transport). A comparison between the Kuroshio path variation (Figure 6) and the Kuroshio’s transport in the upstream East China Sea (Figure 4) shows that the 1959–62 large-meander event does correspond to a large upstream transport. However, this correspondence becomes less obvious after 1975, as there were times when the upstream transport was large, but no large meander was present. Assuming

that the upstream Kuroshio transport after 1975 is in the multiple equilibrium regime, the correspondence between the path transition and the temporal change in the upstream transport (e.g. the required transport increase for the transition from a straight path to a meander path) is also inconclusive from the timeseries presented in Figures 4 and 6. Given the low frequency and irregular nature of the Kuroshio path changes, future studies based on longer transport measurements are needed to further clarify the physics underlying the Kuroshio path bimodality. Downstream Extension Region

After separating from the Japan coast at 1401E and 351N, the Kuroshio enters the open basin of the North Pacific Ocean where it is renamed the Kuroshio Extension. Free from the constraint of coastal boundaries, the Kuroshio Extension has been observed to be an eastward-flowing inertial jet accompanied by large-amplitude meanders and energetic pinched-off eddies. Figure 7 shows the mean temperature map at 300 m depth, in which the axis of the Kuroshio Extension is well represented by the 121C isotherm. An interesting feature of the Kuroshio Extension east of Japan is the existence of two quasi-stationary meanders with their ridges located at 1441E and 1501E, respectively. The presence of these meanders along the mean path of the Kuroshio Extension has been interpreted as standing Rossby lee waves generated by the presence of the Izu Ridge. A competing theory also exists that regards the quasi-stationary meanders as being steered by the eddy-driven abyssal mean flows resulting from instability of the Kuroshio Extension jet.

(c) 2011 Elsevier Inc. All Rights Reserved.

364

KUROSHIO AND OYASHIO CURRENTS

2 4

3

5 6

4 40˚

7

Sea of Japan

8

Latitude

5

9 10

10 Japan

11 16

12

16 15 30˚N 130˚E

15

15 140˚

150˚

160˚ Longitude

14 170˚

13 180˚

Figure 7 Mean temperature map (1C) at the 300 m depth from 1976 to 1980. (Adapted with permission from Mizuno K and White WB (1983) Annual and interannual variability in the Kuroshio Current System. Journal of Physical Oceanography 13: 1847–1867.)

Near 1591E, the Kuroshio Extension encounters the Shatsky Rise where it often bifurcates. The main body of the Kuroshio Extension continues eastward, and a secondary branch tends to extend northeastward to 401N, where it joins the eastwardmoving Subarctic Current. After overriding the Emperor Seamounts along 1701E, the mean path of the Kuroshio Extension becomes broadened and instantaneous flow patterns often show a multiplejet structure associated with the eastward-flowing Kuroshio Extension. East of the dateline, the distinction between the Kuroshio Extension and the Subarctic Current is no longer clear, and together they form the broad, eastward-moving North Pacific Current. As demonstrated in Figure 3, the Kuroshio Extension region has the highest level of eddy variability in the North Pacific Ocean. From the viewpoint of wind-driven ocean circulation, this high eddy variability is to be expected. Being a return flow compensating for the wind-driven subtropical interior circulation, the Kuroshio originates at a southern latitude where the ambient potential vorticity (PV) is relatively low. For the Kuroshio to smoothly rejoin the Sverdrup interior flow at the higher latitude, the low PV acquired by the Kuroshio in the south has to be removed by either dissipative or nonlinear forces along its western boundary path. For the narrow and swift Kuroshio Current, the dissipative force is insufficient to remove the low PV anomalies. The consequence of the Kuroshio’s inability to effectively diffuse the PV anomalies along its path results in the accumulation of low PV water in its extension region, which generates an anticyclonic recirculation gyre and provides an energy source for flow instability. Due

to the presence of the recirculation gyre (Figure 8), the eastward volume transport of the Kuroshio Extension can reach as high as 130 Sv south east of Japan. This is more than twice the maximum Sverdrup transport of about 50 Sv in the subtropical North Pacific. The inflated eastward transport is due to the presence of the recirculating flow to the south of the Kuroshio Extension. Although weak in surface velocity, Figure 8 shows that the recirculating flow has a strong barotropic (i.e. depthindependent) component. As a consequence, the volume transport of the recirculation gyre in this case is as large as 80 Sv. In addition to the high meso-scale eddy variability, the Kuroshio Extension also exhibits large-scale changes on interannual timescales. Figure 9A and B compares the sea surface height field in the Kuroshio Extension region in November 1992 with that in November 1995. In 1992, the Kuroshio Extension had a coherent zonal-jet structure extending beyond the dateline. The zonal mean axis position of the Kuroshio Extension from 1411E to 1801E in this case was located north of 351N. In contrast, the jet-like structure in 1995 was no longer obvious near 1601E and the zonal mean axis position shifted to 341N. Note that the changes in the zonal mean axis position of the Kuroshio Extension have interannual timescales (Figure 9C) and are associated with the changes in the strength of the southern recirculation gyre. As the recirculation gyre intensifies (as in 1992), it elongates zonally, increasing the zonal mean eastward transport of the Kuroshio Extension and shifting its mean position northward. When the recirculation gyre weakens (as in 1995), it decreases the eastward transport of the Kuroshio Extension and shifts its zonal mean position southward. At

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS

365

V: (32 CW) 145 _141° E (WHP P10) 6 _ 9 Nov. 1993 0

_ 30 0

90

80

_ 20

5

0

60

70 _ 0_ _ 10 1 5 0 _ 10

5

_ 5

_ 10

2000 _5

0

5 0

10

5

5

5

50

5

_ 10

5 0

_ 10

_5

0

_ 5

0

_ 5

3000

5

_ 5

_ 10

5

0

0

_

Depth (m)

30

15 20

15

_5

5

0

10

0

0

0 _5

5

1000

0

10

50 0

0

4000

0

0

_5

0

10 5 0 5

5

0

0

0

5

5

_5

_ 10

_ 5

_ 5

5 0

_5

0

0

5000

6000 32.5°N

33.0°N

33.5°N

34.0°N Latitude

34.5°N

35.0°N

35.5°N

Figure 8 North-eastward velocity profile from lowered acoustic Doppler current meter profiler (ADCP) measurements along the WOCE P10 line south east of Japan in November 1993 (see the dashed line in Figure 7 for its location). Units are cm s1 and southwestward flow is shaded. (Figure courtesy of Drs E. Firing and P. Hacker of the University of Hawaii.)

present, the cause of the low-frequency changes of the recirculation gyre is unclear.

The Oyashio Current Due to the southward protrusion of the Aleutian Islands, the wind-driven subarctic circulation in the North Pacific Ocean can be largely divided into two cyclonic subgyres: the Alaska Gyre to the east of the dateline and the Western Subarctic Gyre to the west (Figure 1). To the north, these two subgyres are connected by the Alaskan Stream, which flows southwestward along the Aleutian Islands as the western boundary current of the Alaska Gyre. Near the dateline, the baroclinic volume transport of the Alaskan Stream in the upper 3000 m layer is estimated at about15–20 Sv. As the Alaskan Stream flows further westward, the deep passages between 1681E and 1721E along the western Aleutian Islands allow part of the Alaskan Stream to enter the Bering Sea. In the deep part of the Bering Sea, the intruding Alaskan Stream circulates anticlockwise and forms

the Bering Sea Gyre. The western limb of the Bering Sea Gyre becomes the East Kamchatka Current, which flows south-westward along the east coast of the Kamchatka Peninsula. The remaining part of the Alaskan Stream continues westward along the southern side of the Aleutian Islands and upon reaching the Kamchatka Peninsula, it joins the East Kamchatka Current as the latter exits the Bering Sea. As the East Kamchatka Current continues southwestward and passes along the northern Kuril Islands, some of its water permeates into the Sea of Okhotsk. Inside the deep Kuril Basin in the Sea of Okhotsk, the intruding East Kamchatka Current water circulates in a cyclonic gyre. Much of this intruding water moves out of the Sea of Okhotsk through the Bussol Strait (46.51N, 151.51E), where it joins the rest of the south-westward-flowing East Kamchatka Current. The East Kamchatka Current is renamed the Oyashio Current south of the Bussol Strait. Because of the intrusion in the Sea of Okhotsk, the water properties of the Oyashio Current are different from those in the upstream East

(c) 2011 Elsevier Inc. All Rights Reserved.

366

KUROSHIO AND OYASHIO CURRENTS

40°N

2.1 2.1 2.2

2.1

35°N

2.3 2.9

2.7

3.0

2.

8

2.8

2.8

2.9

2.8

(A)

2.6 2.7

3.1

2.8

30°N

2.2

140°E

150°E

160°E

170°E

180°

2.

0

40°N 2.2

2.1

2.0

1.9

2.0

2.3

35°N 2.8 2.6

2.9

2.7

(B)

140°E

2.4

2.8 2.7

7

2.

30°N

2.5 2.6

150°E

160°E

170°E

180°

36°N

35°N

34°N

33°N 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 (C) Year Figure 9 Sea surface height maps on (A) 20 November 1992 and (B) 15 November 1995 from the TOPEX/POSEIDON altimeter measurements. (C) Time-series of the mean axis position of the Kuroshio Extension from 1411E to 1801. (Adapted with permission from Qiu B (2000) Interannual variability of the Kuroshio Extension system and its impact on the wintertime SST field. Journal of Physical Oceanography 30: 1486–1502.)

Kamchatka Current. For example, the mesothermal water present in the East Kamchatka Current (i.e. the subsurface maximum temperature water appearing in the halocline at a depth of 150–200 m) is no longer observable in the Oyashio. While high dissolved oxygen content is confined to above the halocline in the upstream East Kamchatka Current, elevated dissolved oxygen values can be found throughout the upper 700 m depth of the Oyashio water. The baroclinic volume transport of the Oyashio Current along the southern Kuril Islands and off Hokkaido has been estimated at 5–10 Sv from the geostrophic calculation with a reference level of nomotion at 1000 or 1500 m. Combining moored

current meter and CTD (conductivity-temperaturedepth) measurements, more recent observations along the continental slope south east of Hokkaido show that the Oyashio Current has a well-defined annual cycle: the flow tends to be strong, reaching from surface to bottom, in winter/spring, and it is weaker and confined to the layer shallower than 2000 m in summer and fall. The total (baroclinic þ barotropic) volume transport reaches 20–30 Sv in winter and spring, whereas it is only 3–4 Sv in summer and fall. This annual signal in the Oyashio’s total transport is in agreement with the nnual signal in the Sverdrup transport of the wind-driven North Pacific subArctic gyre.

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS

120˚E

140˚E

367

160˚E

1 2

12

10

7

5

17

3

16

23

20

5 10

15

11 11

16 11

15

19 18

22

19

21 23 22

2

4

10

4

16 13

20

20

19

20

19

20

18

21

20

2

16 14

40˚N 10

5 10 10

20

27

19

19

20

23

20˚N Figure 10 Water temperature map at the 100 m depth in September 1989 compiled by the Japan Meteorological Agency. Contour interval is 11C.

After flowing south-westward along the coast of Hokkaido, the Oyashio Current splits into two paths. One path veers offshoreward and contributes to the east-north-eastward-flowing SubArctic Current. This path can be recognized in Figure 10 by the eastward-veering isotherms along 421N south east of Hokkaido. Because the Oyashio Current brings water of subarctic origin southward, the SubArctic Current is accompanied by a distinct temperaturesalinity front between cold, fresher water to the north and warm, saltier water of subtropical origin to the south. This water mass front, referred to as the

Oyashio Front or the Subarctic Front, has indicative temperature and salinity values of 51C and 33.8 PSU at the 100 m depth. Across 1651E, combined moored current meter and CTD measurements show that the SubArctic Current around 411N has a volume transport of 22 Sv in the upper 1000 m layer. The second path of the Oyashio Current continues southward along the east coast of Honshu and is commonly known as the first Oyashio intrusion. As shown in Figure 10, an addition to this primary intrusion along the coast of Honshu, the southerly Oyashio intrusion is also frequently observed further

41°N

40°N

39°N

38°N

37°N

1960

1965

1970

1975

1980

1985

1990

Year Figure 11 Time-series of the annually averaged southernmost latitude of the first Oyashio intrusion east of Honshu. The dashed line shows the mean latitude (38.71N) over the period from 1964 to 1991. (Data courtesy of Dr K. Hanawa of Tohoku University.)

(c) 2011 Elsevier Inc. All Rights Reserved.

368

KUROSHIO AND OYASHIO CURRENTS

60°N

0.7

40°N

0.6 0.4

_ 0.2

0.2

_ 0.2

20°N

0°

0.2

20°S 120°E

140°E

(A)

160°E

180°

160°W

140°W

120°W

100°W

1.00

Temperature (°C)

0.75 0.50 0.25 0 _ 0.25 _ 0.50 _ 0.75 _ 1.00 1950

1955

1960

1965

1970

(B)

1975

1980

1985

1990

Year

Figure 12 (A) Spatial pattern of the second empirical orthogonal function (EOF) mode of the wintertime sea surface temperature anomalies (1950–1992) in the Pacific Ocean. This mode explains 11% of the variance over the domain. (B) Time-series of the wintertime sea surface temperature anomalies averaged in the Kuroshio-Oyashio outflow region (321N–461N,1361E–1761W). (Adapted with permission from Deser C and Blackmon ML (1995) On the relationship between tropical and North Pacific sea surface variations. Journal of Climate 8: 1677–1680.)

offshore along 1471E. This offshore branch is commonly known as the second Oyashio intrusion. The annual mean first Oyashio intrusion east of Honshu reaches on average the latitude 38.71N, although in some years it can penetrate as far south as 371N (see Figure 11). In addition to the year-to-year fluctuations, Figure 11 shows that there is a trend for the Oyashio Current to penetrate farther southward after the mid-1970s. Both this long-term trend and the interannual changes in the Oyashio’s intrusions seem to be related to the changes in the intensity of the Aleutian low atmospheric pressure system and the southward shift in the position of the mid-

latitude westerlies. It is worth noting that the anomalous southward intrusion of the Oyashio Current not only influences the hydrographic conditions east of Honshu, it also affects the environmental conditions in the fishing ground and the regional climate (e.g. an anomalous southward intrusion tends to decrease the air temperature over eastern Japan).

Concluding Remarks Because the Kuroshio and Oyashio Currents transport large amounts of water and heat efficiently

(c) 2011 Elsevier Inc. All Rights Reserved.

KUROSHIO AND OYASHIO CURRENTS

in the meridional direction, there has been heightened interest in recent years in understanding the dynamic roles played by the time-varying Kuroshio and Oyashio Currents in influencing the climate through sea surface temperature (SST) anomalies. Indeed, outside the eastern equatorial Pacific Ocean, the largest SST variability on the interannual-todecadal time-scale in the North Pacific Ocean resides in the Kuroshio Extension and the Oyashio outflow regions (Figure 12). Large-scale changes in the Kuroshio and Oyashio current systems can affect the SST anomaly field through warm/cold water advection, upwelling through the base of the mixed layer, and changes in the current paths and the level of the meso-scale eddy variability. At present, the relative roles played by these various physical processes are not clear. This article summarizes many observed aspects of the Kuroshio and Oyashio Current systems, although due to the constraints of space, important subjects such as the water mass transformation processes in regions surrounding the Kuroshio and Oyashio and the impact of the Kuroshio and Oyashio variability upon the oceanographic conditions in coastal and marginal sea areas have not been addressed. It is worth emphasizing that our knowledge of the Kuroshio and Oyashio Currents has increased significantly due to the recent World Ocean Circulation Experiment (WOCE) program (observational phase: 1990–1997). Fortunately, many of the observational programs initiated under the WOCE program are being continued. With results from these new observations, we can expect an improved description of the Kuroshio and Oyashio Current systems in the near future, especially of the variability with timescales longer than those described in this article.

369

Further Reading Dodimead AJ, Favorite JF, and Hirano T (1963) Review of oceanography of the subarctic Pacific region. Bulletin of International North Pacific Fisheries Commission 13: 1--195. Kawabe M (1995) Variations of current path, velocity, and volume transport of the Kuroshio in relation with the large meander. Journal of Physical Oceanography 25: 3103--3117. Kawai H (1972) Hydrography of the Kuroshio Extension. In: Stommel H and Yoshida K (eds.) Kuroshio – Its Physical Aspects, pp. 235--354. Tokyo: University of Tokyo Press. Mizuno K and White WB (1983) Annual and interannual variability in the Kuroshio Current system. Journal of Physical Oceanography 13: 1847--1867. Nitani H (1972) Beginning of the Kuroshio. In: Stommel H and Yoshida K (eds.) Kuroshio – Its Physical Aspects, pp. 129--163. Tokyo: University of Tokyo Press. Pickard GL and Emery WJ (eds.) (1990) Descriptive Physical Oceanography: An Introduction, 5th edn. Oxford: Pergamon Press. Shoji D (1972) Time variation of the Kuroshio south of Japan. In: Stommel H and Yoshida K (eds.) Kuroshio – Its Physical Aspects, pp. 217--234. Tokyo: University of Tokyo Press. Taft BA (1972) Characteristics of the flow of the Kuroshio south of Japan. In: Stommel H and Yoshida K (eds.) Kuroshio – Its Physical Aspects, pp. 165--216. Tokyo: University of Tokyo Press. Tomczak M and Godfrey JS (1994) Regional Oceanography: An Introduction. Oxford: Pergamon Press.

See also Abyssal Currents. Okhotsk Sea Circulation. Pacific Ocean Equatorial Currents. Wind Driven Circulation.

(c) 2011 Elsevier Inc. All Rights Reserved.

LABORATORY STUDIES OF TURBULENT MIXING J. A. Whitehead, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction As described elsewhere in this encyclopedia, turbulence and diffusion act both vertically in mixing up the density field of the ocean and laterally in mixing adjacent density and tracer fields. Vertical mixing of density in a stratified fluid dissipates turbulent kinetic energy by raising the potential energy of the density field, and the ratio of this dissipation to viscous dissipation of turbulence is a fundamental quantity needed to understand the energy balance of the ocean. Estimates of these rates by theory are not yet readily available in a form useful for the ocean. Numerical approaches and computational fluid dynamics are under development, but they are not capable of investigating a wide range of parameter space in present form, nor can results be recovered that are verified by experimental benchmarks. Direct ocean measurements rely on theoretical assumptions about the form of turbulence, which must ultimately be verified in the laboratory. Therefore, laboratory measurements continue to be essential to the quest of determining the rates of turbulent dissipation within flows of stratified fluid. Both the production of turbulence through instability and its dissipation are markedly different from the case of instability and dissipation within a homogeneous fluid. Ignoring internal wave radiation from turbulent regions, dissipation is partitioned between viscous dissipation and the work that increases potential energy. This is shown below by the two integrals for energy of a simple system with no internal body forces and closed bottom and top boundaries in a field of gravity. In this example, the flow must start with an initial value of kinetic energy, and the decrease in kinetic energy of a volume of incompressible fluid is equal to the rate of buoyancy work plus viscous dissipation: D E 1 dhv˜  v˜ i ¼ ghrwi  v ðrv˜ Þ2 2 dt The change in potential energy is the rate of buoyancy work plus buoyancy flux multiplied by elevation: Z h dhrzi z½rw dz ¼ ghrwi þ g g dt 0

where angle brackets are averages over all three spatial dimensions, and the square brackets are over the two lateral dimensions. Here, v˜ is the threedimensional velocity vector, n is viscous diffusivity, the variables z, w are the vertical direction and velocity (the direction of gravity g), and density is r. Clearly, if the following three conditions are met, then the only two terms that can dissipate the kinetic energy are viscous dissipation and buoyancy (heat) flux: first, the potential energy is not changing in time; second, dense (considering it to be cold) fluid enters the bottom with lighter (warm) fluid leaving the top (it would require a downward heat flow into the volume to allow this); and third, the volume is given an initial value of kinetic energy that is allowed to run down. It is the purpose of laboratory experiments to allow measurements of density and velocity fields and to obtain the partition between viscous dissipation and buoyancy flux. No instrument exists for precisely measuring every term within any of the above brackets, so simplified approaches have been necessitated.

Experiments Two types of experiments generate the turbulence, either a shear-flow instability is set up or eddies are directly generated. In addition, there are two groups of density distribution, one with sharp interfaces and the other having continuous stratification. Since both of the dissipation terms shown above are negative, there are no cases with both of the equations in their steady form. Experiments incorporate either transient setups that run down with time, or utilize flowing tanks (mostly with salt-stratified water but a few wind tunnels with thermal stratification are used too). In the latter case, the flows are transient following a fluid parcel. The techniques to produce eddies are numerous. Figure 1 shows some of the laboratory configurations. Some generate turbulence behind grids in tunnels (Figure 1(a)) and others are driven by buoyancy (Figure 1(b)). In some flumes and wind tunnels, narrowing the sides enhances the shear in a test region. Other experiments have a moving grid or rod stirrer (Figure 1(c)), and still others have a moving lid (Figure 1(d)). Not sketched are special studies of pumped jets directed toward an interface, and experiments with double-diffusion driven flows, both being directed toward explicit mechanisms of mixing. Experiments have been motivated by numerous phenomena in addition to oceanographic ones; some

(c) 2011 Elsevier Inc. All Rights Reserved.

371

372

LABORATORY STUDIES OF TURBULENT MIXING

(a)



z

(b)



(c)



z

z



(d)

z

Figure 1 Configurations of various laboratory experiments to measure stratified mixing.

examples are fire prevention, ventilation, snow avalanches, ecosystem studies, solar ponds, mixing of industrial chemicals, and turbidity currents. Flows have been produced in pumped tunnels, as sketched in Figure 1(a), with the stratification produced by either temperature (using heat-transfer devices) or salinity (using multiple sources). A variation is a closed-circuit tunnel with a special device for turbulence-free propulsion of the water. In some cases the vertical salinity distribution is initially set and the density profile runs down with time. Another variant has layers pumped in at multiple elevations with the same amount withdrawn at each corresponding level downstream and returned to cisterns. Also, there are currents driven by buoyant flow in tanks with sloping tops and bottoms, as shown in Figure 1(b), or with exchange flows in passages between reservoirs containing waters of differing salinity. Then, there are experiments in closed containers with oscillating grids or moving rods as sketched in Figure 1(c). In some cases the experiments are in annular chambers and some are rotating on a turntable (Figure 1(d)). Salt-stratified experiments are the most numerous. They possess a vertical salinity distribution that evolves with time, although

there are also thermal experiments using air or water motivated by engineering applications. The dynamics are all characterized by velocity scale of the turbulence u (which may be the same size as velocity difference in a sheared laminar flow whose instability generates the turbulence) and density difference Dr. The force of gravity makes the density difference equivalent to a buoyancy difference g0 ¼ gDr/r0. Geometrically, there is the separation distance d between regions of different velocity and density. Finally, there are two additional fluid properties, the viscosity n and density diffusivity D. Velocity, reduced gravity, length, viscosity, and diffusivity can be reduced to many combinations of three dimensionless numbers. For stratified mixing, they are picked sequentially to represent the important balances in the flow in rank order. The primary dimensionless number is the bulk Richardson number Ri ¼ g0 d=u2. It is a measure of the ratio of buoyancy to fluid inertia. The second number is the Reynolds number Re ¼ u d=v, which is the ratio of inertial to viscous force. The third is the Schmidt number Sc ¼ n/D. It is the ratio of viscosity to density diffusivity. Experiments generally have the objective to determine a buoyancy work rate (frequently called buoyancy flux) as a function of these three numbers. By the early 1990s it was clear that the buoyancy flux obeyed a range of power-law relations with Ri, and that these are sensitive to details of actual experiments in most cases. The relative roles of Re and Sc are less well documented. Virtually any source of turbulence can be used to mix stratified fluid, and one of the challenges of experiments is to separate the influence of the spatial distribution of the turbulent source from the actual processes within the fluid. Turbulence that is shed from a vertical rod that moves laterally and sheds a turbulent wake in fresh water above a salt layer (Figure 2(a)) produces striations that are strongest near the interface and weaker at higher and lower levels. This produces a divergence in buoyancy flux so the interface gets progressively thicker with time. However, continuous stratification (Figure 2(b)) can spontaneously break down to internal layers. Therefore, in both cases, the flux varies locally. To complicate matters, the variation of the stratification is usually about the same size as the scale of the turbulence, so in almost all experiments it is not obvious that statistical turbulence theory applies. Probably the simplest configuration has salt water under fresh water with grid stirring confined to one layer, for example, the top layer. This easily attains high Reynolds numbers using a horizontal grid moving up and down with vigorous oscillatory motion. To determine u at the level of the interface,

(c) 2011 Elsevier Inc. All Rights Reserved.

LABORATORY STUDIES OF TURBULENT MIXING

(a)

373

(b) 10−1 E ~Ri −1/2

ue /u∗

10−2

10−3 E ~Ri −1 10−4

10−5 Figure 2 Shadowgraphs from a parallel beam of light falling onto a screen. They show the effect of turbulence produced by many excursions of a moving rod (a) for a layered fluid as in Figure 1(c) right (salt water under fresh, the rod has recently reversed direction near the tank wall); and (b) for a stratified fluid that breaks down into layers as in Figure 1(c) left (stratified salt water, the rod is moving toward the left into a placid, previously mixed fluid).

grid velocity should be multiplied by a suitable constant to account for spatial variation of the turbulence between grid and interface. As time progresses, for Ri41 the turbulence in the upper layer causes the interface to remain sharp and it mixes salt water up into the top layer. This ‘entrains’ salt water into the top layer, which increases both the volume and salinity of the top layer and decreases the volume of the bottom one but leaves its salinity unchanged. The interface moves downward with entrainment velocity ue, and the speed quantifies the mixing rate. As density difference between the layers decreases, Ri decreases. The entrainment velocity increases steeply with decreasing Richardson number as shown by solid circles in Figure 3. For Rio1, the interface deflection is as large as d and the subsequent mixing rate is rapid and soon the two layers mix completely. Many other experiments have produced entrainment velocity measurements; a collection of some is shown in the lower cluster of Figure 3. In some of them, the mixing is supplied by shear instability driven by a rotating screen in an annulus with stratified fluid (Figure 1(d)). A mixed layer with a sharp density jump at the bottom penetrates into the stratified fluid. Others involve gravity currents (Figure 1(b)), with buoyant outflows and with counter-flows. All of these configurations successfully give useful data for large stratification (Rio1). The scatter in the points shown in Figure 3 is typical and is due to the statistical nature of the data rather

E ~Ri −3/2 10−1

100

101

102

Ri Figure 3 Entrainment velocity as a function of Richardson number for assorted experiments. Solid circles: Grid experiments where the grid velocity is to define u * . Open circles: Buoyant outflows. Triangles: Density currents. Crosses: Counterflows. Data and slopes taken from Fernando HJS (1991) Turbulent mixing in stratified fluids. Annual Review of Fluid Mechanics 23: 455–493, figure 15, with permission from author, and Turner JS (1973) Buoyant convection from isolated sources. Buoyancy Effects in Fluids, pp. 165–206, Figure 9.3. Cambridge, UK: Cambridge University Press, with permission from author.

than instrumental error. In such experiments, the horizontally averaged density typically breaks up into patches and layers so that local regions have different local values of Richardson number. In spite of such scatter, the results of such experiments are overwhelmingly consistent with each other with respect to the general trends of the data shown in Figure 3. Primarily, the Richardson number is the most important variable governing mixing rate if it is of order 1 or more. The Reynolds number does exhibit some role especially if less than c. 500. Experiments to date range up to almost Reo105 and generally speaking the mixing is sensitive to Reynolds number for the entire range. It is thought that mixing will finally become insensitive for very large Re but data up to such a possible limit are not yet available. Three power laws are sketched as straight lines in this figure. To the left the open circle data have a smaller slope. To the right, the data are inversely proportional to a higher power of Richardson number. The constants of proportionality are functions of the actual configuration and the manner of defining velocity and density. As an example, the

(c) 2011 Elsevier Inc. All Rights Reserved.

374

LABORATORY STUDIES OF TURBULENT MIXING

relation found by Kato and Phillips for shear-driven experiments is

10−1

ue ¼ 2:5 Ri1 u This can be rearranged to gDrue d ¼ 2:5r0 u3 , which states that the rate of change of potential energy is proportional to the rate of delivery of kinetic energy for mixing. The value of power-law dependence has been extensively studied and discussed. There are some circumstances in which there is no dependence because stratification effects are smaller than viscous, diffusive, or turbulent effects. In other cases there are over 40 proposed relations with Ri, but there is still no clear consensus about the range of validity for each of these in Ri, Re, and Sc space. This lack of agreement seems to arise for a number of reasons. First, no tight cluster about one line is found because of the scatter mentioned above. Second, it has always been found that the results tend to be specific for each experimental configuration. Third, the experiments are in water or air, so only a few values of Sc are investigated. One proposed relation between the three dimensionless numbers has the dimensionless entrainment at successively increasing Ri proportional to Ri  3/2, ScRi  1, and Sc  1/2 Re  1/4. It shows that there is an increasingly important role for molecular and turbulence effects as stratification is increased. However, in listing 30 such relations, Fernando found that the  1.5 power law generally tended to be for higher values of Ri rather than lower values. A gravity current down a slope is of particular interest to oceanography because of its relevance to deep overflows in polar regions and salt plumes in aridpregions. In such studies the Froude number Fr ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u = g0 d cos yðERi1=2 Þ (y is angle of the slope) is often used to measure the intensity of mixing compared to stratification. Results for a wide range of gravity-driven currents are shown in Figure 4. As in Figure 3 it is clear that large stratification suppresses entrainment velocity, and that ueEu with smaller stratification. In addition, it is clear that mixing is enhanced at larger Reynolds numbers.

Continuous Stratification The layers of mixed fluid in mechanical stirring experiments are separated by very sharp interfaces, so for large values of Ri the layers remain well defined. Therefore, the results are relatively precise and easy to interpret. Of course, experiments with continuous stratification are more similar to the ocean. Continuously stratified fluid exposed to turbulence tends

10−2 ue u∗ 10−3

10−4 100

101

Fr = u∗ /(g ′dcos())1/2 Figure 4 Entrainment velocity in laboratory gravity currents down a slope compared to estimates of ocean overflows. The solid triangle is for Lake Ogawara, the solid square is for Mediterranean outflow into the Atlantic, the small star is for the Denmark Strait overflow, and the solid diamond is for the Faroe Bank channel overflow. For rotating density currents, the open squares, open triangles, and large stars are experiments with Reo100 and the open diamonds are with ReZ100. The shaded area and open circles are found for large Reynolds number nonrotating density currents. Supplied by C. Cenedese.

to break up into layers for large Ri. Therefore, N varies locally and hence the dimensionless number varies locally. As a result, the mixing becomes concentrated in local regions. In addition, as time progresses the layers evolve and slowly change flux. The dynamics that determine the size of the layers remain controversial. In some cases, the layer depth scales with the scale u/N, and in other cases the scales are linked to vortical modes shed from the stirrers, or even to the stirrer size itself. In continuously stratified experiments, the overall Richardson number can be defined as Ric ¼ (Nd/u )2 or, if shear du/dz is imposed, as Ris ¼ (N/(du/dz))2. The invention of the bathythermograph led to the discovery of extensive layering within the ocean. Although this layering can be explained as a consequence of localized wave breaking, its universal character suggests that there is a more fundamental cause. Laboratory experiments with stirring near the sidewall of a stratified fluid, and later experiments with continuously stratified fluid stirred with a rod, all exhibited the spontaneous growth of layers for Ricc1. Figure 5 shows this growth in salt-stratified water with grid-generated turbulence (with d set to the mesh size) for Ric ¼ 10.7. In some elevations the local stratification (as measured by N) increases, and

(c) 2011 Elsevier Inc. All Rights Reserved.

0

−0.05

−0.05

−0.1

−0.1

−0.15

−0.15

z (m)

0

−0.2

−0.25

−0.3

−0.3

−0.35

−0.35 1020 1030 1040 1050 1060 1070 1080 1090 1100 1110  (kg m−3)

Figure 5 Density profiles at successive times showing the spontaneous emergence of layers in a continuously stratified fluid for large Ric. Adapted from Rehmann CR and Koseff JR (2004) Mean potential energy change in stratified grid turbulence. Dynamics of Atmospheres and Ocean 37: 271–294, with permission from Elsevier.

at other elevations it decreases toward zero. This has a profound influence on propagation of both internal and acoustic waves through the water. As time progresses beyond the stage shown in the figure, the layers begin to interact with each other and some will be eliminated, so finally two layers remain with only one interface in between. Ultimately, the density difference between these two layers decreases to zero, and the fluid becomes fully mixed. In contrast, such an experiment with values of Ric approximately 1 or lower does not produce layers (Figure 6). Instead, a mixed layer forms at the bottom and top of the fluid. Simultaneously, the interior stratification gradually decreases. The result is a fluid with values of N decreasing everywhere. Figures 5 and 6 were produced with data from experiments with the configurations shown in Figure 1(c). Accurate resolution of the vertical density field by a conductivity microprobe (developed for stratified turbulent flume measurements) allows precise measurements of the change of potential energy with time. This change is quantified by a flux Richardson number Rfc, defined as rate of change in potential energy divided by power (rate of energy) exerted by the stirrer (which is estimated for the grid using known drag laws). The resulting data are shown in Figure 7. Starting from Ric ¼ 0, experiments with increasing values have increasing values of Rfc, which level off at a value RfcC0.067 at RicC1. Almost all layered and continuously stratified experiments can be interpreted as having a flux Richardson number that reaches its maximum value

375

−0.2

−0.25

1004

1006

1008

1010

1012

1014

 (kg m−3) Figure 6 The evolution of a density field with stirring source throughout the entire fluid and with Rir1. Adapted from Rehmann CR and Koseff JR (2004) Mean potential energy change in stratified grid turbulence. Dynamics of Atmospheres and Ocean 37: 271–294, with permission from Elsevier.

10−1

Rfc

z (m)

LABORATORY STUDIES OF TURBULENT MIXING

10−2

10−3

10−3

10−2

10−1 Ric

100

101

Figure 7 Mixing efficiency in grid-stirring experiments with continuous stratification and small Ri. Adapted from Rehmann CR and Koseff JR (2004) Mean potential energy change in stratified grid turbulence. Dynamics of Atmospheres and Ocean 37: 271–294, with permission from Elsevier.

of c. 0.1 at a Richardson number of order 1, with values decreasing toward 0 for larger and smaller values. The exact value of the maximum has been widely discussed, with some estimates approaching a maximum value of 0.2 and others only reaching a maximum value of 0.05 or so. Naturally, the exact definition depends on the choice of a length and velocity scale, which is not only a matter of choice, but also subjected to the details of each apparatus and analysis technique. In addition, since most experiments are either transient or possessing a variation in space, the measurement location

(c) 2011 Elsevier Inc. All Rights Reserved.

376

LABORATORY STUDIES OF TURBULENT MIXING

Rfc

10−1

Summary

10−2

10−3 103

104 Re

Figure 8 Mixing efficiency as a function of Reynolds number in mixing grid experiments of a uniformly stratified fluid. Stratification was from salinity in some experiments, temperature in others, and from both in one case. Adapted from Rehmann CR and Koseff JR (2004) Mean potential energy change in stratified grid turbulence. Dynamics of Atmospheres and Ocean 37: 271–294, with permission from Elsevier.

In summary, laboratory experiments to measure the rate of mixing of stratified fluids by turbulence have been conducted using a wide range of devices. Both qualitative and quantitative data provide information to oceanographers and numerical modelers. To estimate the mixing rates and their consequences in many ocean regions from the top mixed layer to the abyss, information about flux Richardson number from controlled laboratory measurements is vital. The buoyancy flux ratio rises from 0 at small Richardson number to values of about 0.1 at Richardson number equal to 1, and then falls off for greater values. At Richardson number greater than 1, a wide range of power laws spanning the range from  0.5 to  1.5 are found to fit data for different experiments. The layering in this range causes flux clustering and may contribute to the relatively large scatter in the data and to the wide range of proposed power laws.

See also and time can influence a value. Thus, a maximum of RfC0.1 with a range of roughly 750% has remained unchanged over the last 20 years. However, so far most of the experiments have been conducted with Reynolds numbers of a few thousand or less, and with salt rather than temperature. Therefore the coverage in Re and Sc space is still limited. Finally there is evidence that a large Reynolds number might produce smaller maximum values (Figure 8), but comprehensive results are not yet available. A few studies have been conducted with two ingredients such as salt and temperature contributing density. If the value of Rf for each component differs, the phenomenon called ‘differential mixing’ is said to exist. At present, laboratory experiments do provide a small amount of evidence for differential mixing. Differential mixing is expected to be most important in oceanic regions where both temperature and salinity variations influence density, for instance, in polar regions. Therefore, for climate studies it would be important to incorporate differential mixing accurately in numerical models, especially since the salinity field is known to influence deep wintertime convection.

Differential Diffusion. Energetics of Ocean Mixing. Estimates of Mixing. Vortical Modes.

Further Reading Breidenthal RE (1992) Entrainment at thin stratified interfaces: The effects of Schmidt, Richardson and Reynolds numbers. Physics of Fluids A 10: 2141--2144. Cenedese C, Whitehead JA, Ascarelli TA, and Ohiwa M (2004) A dense current flowing down a sloping bottom in a rotating fluid. Journal of Physical Oceanography 34: 188--203. Fernando HJS (1991) Turbulent mixing in stratified fluids. Annual Review of Fluid Mechanics 23: 455--493. Park YG, Whitehead JA, and Gnanadesikan A (1994) Turbulent mixing in stratified fluids: Layer formation and energetics. Journal of Fluid Mechanics 279: 279--312. Rehmann CR and Koseff JR (2004) Mean potential energy change in stratified grid turbulence. Dynamics of Atmospheres and Ocean 37: 271--294. Turner JS (1973) Buoyant convection from isolated sources. In: Buoyancy Effects in Fluids, pp. 165--206. Cambridge, UK: Cambridge University Press. Turner JS (1986) Turbulent entrainment: The development of the entrainment assumption and its application to geophysical flows. Journal of Fluid Mechanics 170: 431--471.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS R. S. K. Barnes, University of Cambridge, Cambridge, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1427–1438, & 2001, Elsevier Ltd.

Introduction A ‘lagoon’ is any shallow body of water that is semiisolated from a larger one by some form of natural linear barrier. In ocean science, it is used to denote two rather different types of environment: ‘coastal lagoons’ (the main subject of this article) and lagoons impounded by coral reefs. Coastal lagoons are a product of rising sea levels and hence are geologically transient. Whilst they exist, they are highly productive environments that support abundant crustaceans and mollusks, and their fish and bird predators. Harvests from lagoonal aquaculture of up to 2500 kg ha1y1 of penaeid shrimp, of 400 tonnes ha1y1 of bivalve molluscs, and in several cases of 4200 kg ha1 y1 of fish are taken.

What is a Lagoon? Coastal lagoons are bodies of water that are partially isolated from an adjacent sea by a sedimentary barrier, but which nevertheless receive an influx of water from that sea. Some 13% of the world’s coastline is faced by sedimentary barriers, with only Canada, the western coast of South America, the China Sea coast from Korea to south-east Asia, and the Scandinavian peninsula lacking them and therefore also being without significant lagoons (Table 1, Figure 1). The lagoons behind such barrier coastlines range in size from small ponds of o1 ha through to large bays exceeding 10 000 km2. The median size has been suggested to be about 8000 ha. Although several do bear the word ‘lagoon’ in their name, many do not. Usage of the same titles as for fresh water habitats is widespread (e.g. E´tang de Vaccare`s, France; SwanPool, UK; Oyster Pond, USA; Lake Menzalah, Egypt; Benacre Broad, UK; Ozero Sasyk, Ukraine; Kiziltashskiy Liman, Russia, etc.), as is those for coastal marine regions (Peel-Harvey Estuary, Australia; Great South Bay, USA; Ringkbing Fjord, Denmark; Zaliv Chayvo, Russia; Pamlico Sound, USA; Charlotte Harbor, USA; and even GniloyeMore, Ukraine; Mer des Bibans, Tunisia; Mar Menor, Spain, etc.), whilst lagoons liable to hypersalinity are

often termed Sebkhas in the Arabic-speaking world (e.g. Sebkha el Melah, Tunisia). Conversely, a few systems with ‘lagoon’ in their name fall out with the definition; the Knysna Lagoon in South Africa, for example, is an estuarine mouth dilated behind rocky headlands between which only a narrow channel occurs. Coastal lagoons are most characteristic of regions with a tidal range of o2 m, since large tidal ranges (those 44 m) generate powerful water movements usually capable of breaching if not destroying incipient sedimentary barriers. Furthermore, the usual meaning of ‘lagoon’ requires the permanent presence of at least some water and large tidal ranges are likely to result in ebb of water from open systems during periods of low tide in the adjacent sea. Thus in Europe, for example, lagoons are abundant only around the shores of the microtidal Baltic, Mediterranean, and Black Seas. However, they are also present along some macrotidal coasts – such as the Atlantic north of about 471N (in the east) and 401N (in the west) (Figure 2) – where off shore deposits of pebbles or cobbles (‘shingle’) are to be found as a result of past glacial action. Here, shingle can replace the more characteristic sand of microtidal seas as the barrier material, as indeed it partially does in the lagoon-rich microtidal East Siberian, Chukchi, and Beaufort Seas in the Arctic, because it is less easily redistributed by tidal water movements. Nevertheless, sandy sedimentary barriers have developed (and persisted) in a few relatively macrotidal areas (e.g. in southern Iceland and Portugal), although the regions impounded to landwards retain water only during high tide for the reasons outlined above. Those environments that are true lagoons only during high tide are often referred to as ‘tidal-flat lagoons’. They are therefore the relatively rare macrotidal coast equivalent of the typical lagoons of microtidal seas. In many cases, the salinity of a lagoonal water mass is exactly the same as that of the adjacent sea, although where fresh water discharges into a lagoon its water may be brackish and a (usually relatively stable) salinity gradient can occur between rivermouth and lagoonal entrance channel. In regions where evaporation exceeds precipitation for all or part of a year, lagoons are often hypersaline. The second environment to which the word lagoon is applied is associated with coral reefs; coral here replaces the unconsolidated sedimentary barrier of the coastal lagoon. Circular atoll reefs enclose the ‘lagoon’ within their perimeter, whilst barrier reefs

(c) 2011 Elsevier Inc. All Rights Reserved.

377

378

LAGOONS

are separated from the mainland by an equivalent although less isolated body of water. Indeed barrier reef lagoons are virtually sheltered stretches of coastal sea. Atoll lagoons, however, are distinctive in being floored by coral sand which supports submerged beds of seagrasses and fringing mangrove swamps justas do the coastal lagoons of similar latitudes. The similarity between the two types of lagoon is nevertheless purely physiographic since the atoll lagoon fauna is that typical of coral reefs in general and not related in any way to those of coastal lagoons. Coral lagoons are covered in greater detail in the article Coral Reefs.

Table 1 The contribution of different continents to the world total barrier/lagoonal coastline of 3200 kma Continent

Percent of coastline barrier/lagoonal

Percent of world’s lagoonal resource

N. America Asia Africa S. America Europe Australia

17.6 13.8 17.9 12.2 5.3 11.4

33.6 22.2 18.7 10.3 8.4 6.8

a (Reproduced with permission from estimates by Cromwell JE (1971) Barrier Coast Distribution: A World-wide Survey, p. 50. Abstracts Volume, 2nd National Coastal Shallow Water Research Conference, Baton Rouge, Louisiana.)

The Formation of Lagoons Lagoons have existence only by virtue of the barriers that enclose them. They are therefore characteristic only of periods of rising sea level, when wave action can move sediment on and along shore, and – for a limited period of time – of constant sea level. At times of marine regression, they either drain or their basins may fill with fresh water. Many of the enclosing barriers have today been greatly augmented by wind-blown sand; for example, most of the larger systems along the South African coast (Wilderness, St Lucia, Kosi, etc.), and such enclosed lagoons have a particularly lake-like appearance (Figure 3). The precise physiographic nature of a lagoon then depends on the relationship of the barrier to the adjacent coastline. Starting at a point at which the barriers are some distance offshore, we can erect a sequence of situations in which the barriers are moved ever shorewards. This starting point can be exemplified by Pamlico and Albemarle Sounds in North Carolina, USA (Figure 4). There a chain of long narrow barrier islands located some 30 km off the coast enclosesan area of shallow bay of around 5000 km2. Further landwards movement of the barriers, often to such an extent that some of the larger barriers become attached to the mainland at one end to produce spits, leads to the situation currently characterizing most lagoonal coastlines and to the typical

Lagoonal coasts Figure 1 Major barrier/lagoonal coastlines.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS

379

´ Etang de Kergalan

´ Etang de Trunvel

1 km Figure 2 Shallow lagoons along the coast of the Baie d’Audierne (Brittany, France) formed behind a shingle barrier beach. Sea water enters these systems only via overtopping of the barrier; lagoonal water leaves by percolation through the barrier.

coastal lagoon. Sea water then enters and leaves these lagoons through the channels between the islands or around the end of the spit. More rarely, the lagoons are enclosed by multiple tombolos connecting an offshore island to the mainland, or even tombolos joining a number of islands (as, for example, in Te Whanga, New Zealand). With the exception of tombolo lagoons, the long axis of typical lagoons – and many are extremely elongate – lies parallel to the coastline. Typical coastal lagoons are especially characteristic of microtidal seas. They range in size up to the 10 000 km2 Lagoa dos Patos, Brazil. The abundant lagoons that occur within river deltas form a special case of this category. Several deltas (e.g. those of the Danube and Nile) have developed within – and have now largely or completely obliterated – former lagoons enclosed by spits and intervening barrier island chains. The filling of lagoons with sediment when they receive river discharge is an inevitable process and many surviving examples are but small remnants of their past extents (Figure 5) – islands of water in a sea of marshland.

Some 5000 years BP, the Lake St Lucia lagoon in South Africa, for example, had a length in excess of 110 km and a surface area of nearly 1200 km2; infilling with sediment has reduced the body of free water to a length of 40 km and an area of 310 km2 (see Figure 3). Allowing for the shallowing that has also taken place, this is equivalent to an average input of 623 000 m3 of sediment per year. Infill, coupled with barrier transgression via wash-over fans, is probably the ultimate fate of virtually all of the world’s lagoons. The next stage in the hypothetical sequence sees the barrier very close to the mainland, if not plastered onto it. This has several consequences, of which the first is impingement on emerging river systems. ‘Estuarine lagoons’, which as their name implies merge in to estuaries, have formed where barriers have partially blocked existing drowned river valleys. For this reason they usually have their long axis perpendicular to the coastline. Even if blockage of the river is complete, lagoonal status may remain if sea water can enter by overtopping of the barrier during high water of spring tides.

(c) 2011 Elsevier Inc. All Rights Reserved.

380

LAGOONS

Enkovugeni (8_35) Makhawulani (4_22) Mpungwini (1_ 13) 10 km

Nhlange (1_ 6)

Amanzimnyama (fresh)

Dunes

Marsh

Figure 3 The Kosi Lakes (KwaZulu/Natal, South Africa), with approximate salinity ranges (in practical salinity units) in parentheses. They are enclosed within Pleistocene sand dune fields, and form a classic case of ‘segmentation’ of an original linear lagoon into aseries of rounded basins. (Salinity data reproduced with permission from Begg G (1978) The Estuaries of Natal. Natal Town and RegionalPlanning Report.).

Nevertheless, many former estuarine lagoons are now completely fresh water habitats with seawater entry being prevented by the barrier. In many tropical and subtropical regions, closure of the estuarine lagoons is seasonal. During the wet season, river flow is sufficient to maintain breaches through the barriers. In the dry season, however, flow is reduced and drift can seal the barrier. If fresh water still flows, the whole isolated basin then becomes fresh for several months. The mouths of many natural lagoonal barriers are periodically breached by man for a variety of reasons ranging from ensuring the entry of juveniles of commercially important mollusks, crustaceans, and fish to temporarily lowering water levels to avoid damage to property. Longshore movement of barrier sediment may also divert the mouths of discharging estuaries several kilometers along the coast, as for example the 30 km deflection of that of the Senegal River by the Langue de Barbarie in West Africa, and the creation of the

Indian River Lagoon in Florida, USA. Not infrequently a new mouth is broken through the barrier, naturally or more usually as a result of human intervention, and the diverted stretch can become a dead-end backwater lagoon fed by backflow. Other consequences of on shore barrier migration also involve human intervention. Many small scale lagoons have (unwittingly) been created by land reclamation schemes. Regions to landwards of longshore ridges that were once, for example, lowlying salt marsh but which are now reclaimed usually receive an influx of water from out of the barrier water-table and this collects in depressions that were once part of the creeks draining the marshes. Equivalently, gravel pits or borrow pits in coastal shingle masses, from which building materials have been extracted, similarly receive an influx of water from out of the shingle. In both cases, the shingle water-table is derived from sea water soaking into the barrier during high tide in the adjacent sea

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS

381

Albemarle Sound

Pamlico Sound

Cape Hatteras

Cape Hookout

100 km Figure 4 Pamlico and Albemarle Sounds (North Carolina, USA).

together with such rainfall as has also soaked in. Some of the natural limans of the Black Seacoast also receive their salt input via equivalent seepage. Finally in this category, barriers may come onshore in such a fashion as to straddle the mouth of a preexisting bay. ‘Bahira lagoons’ (from the Arabic for ‘littlesea’) are pre-existing partially land-locked coastal embayments, drowned by the post glacial rise in sea level, that have later had their mouths almost completely blocked by the development of sedimentary barriers. Also included here are systems in which the sea has broken through a pre-existing sedimentary barrier to flood part of the hinterland, but inwhich the entrance/exit channel remains narrow. Bahira lagoons clearly merge into semi-isolated marine bays. Although lagoons may come and go during geological time, some of the larger lagoon systems are not solely features of the present interglacial period. Some, including the Lagoa dos Patos, the Gippsland Lakes of Victoria, Australia, and the Lake St Lucia lagoon, may incorporate elements of barriers and basins formed during previous marine transgressions.

The history of the Gippsland lakes over the last 70 000 years has been reconstructed in some detail (Figure 6). Fossil assemblages of foraminiferans suggest that lagoons may have been a feature of the Gulf Coast of the USA and Mexico at intervals ever since the Jurassic.

Lagoonal Environments It can be argued that the main force structuring lagoons is the extent to which they are connected to the adjacent sea, and they have been divided into three or four types on this basis(Figure 7) which largely reflect points along the hypothetical evolutionary sequence described above. ‘Leaky lagoons’ function virtually as sheltered marine bays (see also Figure 4). As a result of large tidal ranges in the adjacent sea or the existence of many and/or wide connecting channels, interactions with the ocean are dominant and the lagoons have tidally fluctuating water levels, short flushing times and a salinity equal to that ofthe local sea water.

(c) 2011 Elsevier Inc. All Rights Reserved.

382

LAGOONS

Former lagoon

Present delta

50 km

Figure 5 Infill of a former lagoon (established around 6000 years ago) on the Romanian/Ukrainian coast of the Black Sea with sediment carried by the River Danube. The current Danube delta occupies virtually all the former lagoon.

‘Restricted lagoons’ have asmall number of narrow entrance channels through elements of a barrier is land chain or enclosing spits and are therefore more isolated from the ocean(see also Figure 7). Characteristics include: longer flushing times, a vertically well-mixed water column, both wind and tidal water movements as forcing agents, and brackish to oceanic salinity. ‘Choked lagoons’ are connected to the ocean by a single, often long and narrow, entrance channel that serves as a filter largely eliminating tidal currents and fluctuations in water level(see also Figure 4). The 1000 km2 Chilka Lake lagoon, for example, communicates with the Bay of Bengal via a channel 8 km long and 130 m across at its widest. Choked lagoons are typical of coasts with high wave energy and significant longshore drift. Characteristics include: very long flushing times, intermittent stratification of the water column by thermoclines and/or haloclines, wind action as thed ominant forcing agent, purely freshwater zones near inflowing rivers, and otherwise brackish, and in some climatic zones periodic hypersaline, waters. Wet season rainfall in the 44 km2 Laguna Unare, Venezuela (at the end of the dry season), for example, can change the salinity from 60–92 to 18–25, at the same time increasing lagoonal depth by 1 m, area by 20 km2 and volume by 76  106m3.

The ‘closed lagoons’ shown in Figure 2 form the limiting condition of effective isolation from direct inputs from the ocean except via occasional overtopping or after percolation through the barrier system. Most are not only ‘former lagoons’, but also current coastal freshwater lakes; for example, the much studied Slapton Ley in England. Characteristically restricted and choked lagoons are very shallow (usually about 1 m and almost always o5 m deep), floored by soft sediments, fringed by reedbeds(Phragmites and/or Scirpus), mangroves (especially Rhizophora) or salt marsh vegetation, and support dense beds of submerged macrophytes such as seagrasses (e.g. Zostera), pondweeds (Potamogeton) or Ruppia, together with green algae such as Chaetomorpha. The action of the dense beds of submerged vegetation may be to raise levels of pH and to contribute considerable quantities of organic matter. In stratified choked lagoons, decomposition of this organic load below the thermocline or halocline can then lead to anoxic conditions, not withstanding surface waters super saturated with oxygen (Figure 8).

Lagoonal Biotas and Ecology Insofar as is known, although the species may be different, the ecology of coastal lagoons shows the

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS

383

Lag

oon

Prior

barrie

r

(A)

Inner barrier (B)

(C)

(D)

Outer barrier 25 km

(E) Pleistocene land surface

Alluvium and swamps

Outlet Barriers

Figure 6 Evolution of the Gippsland Lagoon system (Victoria, Australia). (Reproduced with permission from Barnes 1980; after Bird ECF (1966) The evolution of sandy barrier formations on the East Gippsland coast. Proceedings of the Royal Society of Victoria 79:75–88.) The prior barrier (A) can be dated to about 70 000 years BP; the inner barrier (B) was formed during the late Pleistocene; the situation in (C) represents a low sea level phase of the late glacial, and (D) and (E) further evolution during the current marine transgression.

(c) 2011 Elsevier Inc. All Rights Reserved.

384

LAGOONS

latitude lagoons of comparable nature; and whilst shallow phytoplankton-dominated lagoons are more productive than deeper ones (depth 42 m), the converse may be true of macrophyte-dominated systems. An order of magnitude calculation (assuming that the total area of the world’s lagoons is around 320 000 km2 and that average lagoonal productivity is 300 g Cm2 y1) yields a total annual lagoonal production of 1011 kg fixed C. The contribution of lagoons to total oceanic carbon fixation is therefore minor, although they are as productive per unit area as are estuaries, and are only less productive than some regions of upwelling, coral reefs, and kelp forests. What happens to this primary production? Data are still scarce. There has been much argument as to whether estuaries are sources or sinks for fixed

10

0

km

same general pattern as seen in estuaries and other regions of coastal soft sediment, although the relative importance of submerged macrophytes may be greater in lagoons (Figure 9). The action of predators as important forces structuring the communities in lagoons as in similar areas, for example, is reflected by the young stages of many species occurring within the dense beds of submerged macrophytes, as the hunting success of predatory species is lower there. The primary productivities of lagoons appear to vary with their general nature, their latitude, and their depth. Choked lagoons tend to be the most productive, with recorded maxima approaching 2000 g C m2 y1, and with productivities (allowing for the effect of latitude) some 50% more than inrestricted lagoons. Temperate zone lagoons achieve only about 50–70% of the productivity of low

15 km

Choked

Restricted

Leaky

25 km Figure 7 ‘Choked’ (Lagoa dos Patos, Brazil), ‘restricted’ (Laguna di Venezia, Italy) and ‘leaky’ (Laguna Jiquilisco, El Salvador) lagoons.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS

385

0

1 Depth (m)

Halocline

Oxycline

2

3 0

5

10

0

15

Salinity

100

50 Oxygen (% saturation)

Figure 8 A halocline and consequent oxycline in the choked Swanpool Lagoon, England, after a prolonged period of high freshwater input (salinity inpractical salinity units). (Reproduced with permission from data in Dorey AE et al. (1973) An ecological study of the Swanpool, Falmouth II. Hydrography and its relation to animal distributions. Estuarine Coastla and Marine Science 1: 153–176.)

Pelican, osprey, fish-eagle, etc.

Mullet

Flamingo

Shore-birds, heron, etc.

Large fish

Waterfowl

Crabs

Amphipod crustaceans, gastropod mollusks, etc.

Prawns

Polychaete worms

Small fish

Bivalve mollusks

Algal productivity and detrital materials Macrophyte material Figure 9 Simplified lagoonal food web.

carbon. In so far as information is available it appears that lagoons are more nearly in balance, as is appropriate for their more isolated nature, although all those so far studied do function as slight sinks.

Even though relatively isolated, however, no lagoon is self-contained, not least in that many elements in the faunas migrate between lagoons and other habitats.

(c) 2011 Elsevier Inc. All Rights Reserved.

386

LAGOONS

Coastal lagoons are heavily used by wetland and shore birds as feeding areas, most noticeably by grebes, pelicans, ibis, egrets, herons, spoonbills, avocets, stilts, storks, flamingoes, cormorants, kingfishers, and fish eagles. Such birds prey on the abundant lagoonal invertebrate and fish faunas. These comprise mixtures of three different elements: both (a) essentially freshwater and (b) essentially marine and/or estuarine species capable of with standing adegree of brackishness, as well as (c) specialist lagoonal or ‘paralic’ species that are not in fact restricted to lagoons but also (or closely related species) occur in habitats like the Eurasian inland seas (e.g. the Caspian and Aral), as well as in (the usually man-made) tideless brackish ponds and drainage ditches that are abundant in reclaimed coastal regions. The main feature of the lagoonal species category is that they are species of marine ancestry that seem to be restricted to shallow, relatively tideless maritime habitats; over their ranges as a whole they are also clearly capable of inhabiting a wide range of salinity, including full-strength and concentrated sea water, but not fresh water. Their relatedness to marine species is evidenced by the fact that they often form species-pairs with marine or estuarine species. The freshwater component of lagoonal faunas is often greater than would be expected in comparable salinities in estuaries, with the common occurrence of adult water beetles and hemipteran bugs besides the omnipresent dipteran larvae. The reasons for this are not understood, but may relate to decreased water movements. Because oceanic and freshwater inputs into alagoon occur at different points around its perimeter, these three elements of the fauna tend to be broadly zoned in relation to these inputs. Indeed, an influential French school, the ‘Group d’Etude du Domaine Paralique’, identifies six characteristic biotic zones in lagoons based not upon salinity but upon ‘a complex and abstract value which cannot be measured in the present state of knowledge’ which the group terms ‘confinement’. This is a function, at least, of the extent to which the water mass at any given point is isolated from oceanic influence. Beginning with the Mediterranean Sea coast, disciples of this school have now published maps of the location of the ‘six degrees of confinement’ in many lagoons throughout the world. Lagoons also form the nursery grounds and adult feeding areas for a large number of commercially important fish and crustaceans that migrate between this habitat and the sea, most of which spawn outside the lagoons and only later move in actively or in some cases probably passively. In respect of the fish, temperate regions are amongst others used byeels

(Anguillidae), sea bass (Moronidae), drum (Sciaenidae), sea bream (Sparidae), grey mullet (Mugilidae), flounder (Pleuronectidae), and various clupeids, and in warmer waters these are joined by many others, including cichlids, milkfish (Chanidae), silver gars (Belonidae), grouper (Serranidae), puffer fish (Tetraodontidae), grunts(Pomadasyidae), rabbit fish (Siganidae), various flatfish (Soleidae, Cynoglossidae), and rays (Dasyatidae). Therefore, throughout the world lagoons support (often artisanal) fisheries, as well as being the location of bivalve shellfish culture (mussels, oysters, and clams). For obvious reasons, most data on secondary productivity have been collected in relation to these fisheries. The fish yield from lagoons is greater than from all other similar aquatic systems, averaging (n ¼ 107) 113 kg ha1 y1 and with a median value of 51 kg (Table 2) and with 13% exceeding 200 kg ha1 y1. Yields vary with the intensity of human involvement in the catching and stocking processes. In the Mediterranean Sea, the region for which most data are available, the yield without intensive intervention is some 82 kg ha1y1; with the installation of permanent fish traps this increases to 185 kg; and with the addition of artificial stocking with juvenile fish it increases to 377 kg. For a 10 year period, the fishermen’s cooperative based on the Stagno di Santa Giusta in Sardinia harvested nearly 700 kg ha1 y1. West African lagoons with ‘fish parks’ – areas that attract fish because of the provision of artificial refuges – are considerably more productive, with average fish yields of 775 kg ha1 y1. Amongst the invertebrates, penaeid shrimp, mudcrabs (Scylla), oysters, mussels, and the arcid ‘cockle’ Anadara are the major harvested organisms. Yields of Penaeus may attain 2500 kg ha1 y1in lagoonal systems of aquaculture, whilst those of the oyster Crassostrea and the mussel Perna can both be 400 tonnesha1 y1. In the Indian subcontinent wholecommunities can be economically entirely dependent

Table 2 Mean fisheries yield from coastal lagoons in relation to those from other aquatic habitatsa Habitat

Fish yield (kg ha1y1)

Percent of sites exceeding yield of 200 kg ha  1y  1

Coastal lagoons Continental shelf Coral reef Fresh waters

113 59 49 34

13 5 0 2

a

(Data reproduced with permission from Chauvet, 1988.)

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGOONS

387

Table 3 Discharges in the early 1990s into two lagoons on the southern (Polish/Russian/ Lithuanian) coast of the Baltic Sea(inputs in tonnes yr1)a Parameter

Curonian Lagoon

Vistula Lagoon

Surface area Mean depth Maximum depth Volume Annual water exchange BOD input Inorganic N Inorganic P Chlorinated substances Cu Zn Pb

1584 km2 3.8 m 5.8 m 6.0 km3 27.3 km3 150 000 20 000 5000 1 30 50 600

838 km2 2.6 m 5.2 m 2.3 km3 ? 47 000 25 000 2100 ‘High concentrations’ ? ? ?

a (Reproduced with permission from data in Coastal Lagoons and Wetlands in the Baltic. WWF Baltic Bulletin 1994, no 1.)

on these shrimp, crab, and fish harvests; for example, the 60 000 people living around Chilka Lake, India. However, many of the world’s lagoons are threatened habitats, suffering deterioration as a result of pollution and destruction, through reclamation and from the natural losses resulting from succession to freshwater and swamp habitats and from landwards barrier migration. The discharge of materials into lagoons and its consequences are essentially similar to those in any other semi-enclosed coastal embayment, but there is one specifically lagoonal feature. As they are often used for aquaculture and yields are generally related to primary productivity, lagoonal waters are frequently deliberately enriched to boost catches. Such enrichment varies from domestic organic wastes from the surrounding communities to commercial processed fish foods. Probably in the majority of such cases, however, the result of this nutrient injection has been eutrophication, loss of macrophytes, deoxygenation, and in several areas a change in the primary producers in the direction of a bacteria-dominated plankton and, across wide areas, benthos as well. In Mediterranean France, this all too frequent state of affairs is known as ‘malai¨gue’. Culture of mussels in the Thau Lagoon in France produces an input to the benthos of some 45 000 tonnes (dry weight) of pseudofecal material. Not surprisingly, at times of minimum throughput of water, malai¨gues can cause mass mortality of the cultured animals and degradation of the whole habitat. Thus in Europe, malai¨gues in the south, pollution in the Baltic lagoons(Table 3), and reclamation ofthose on the Atlantic seaboard have rendered the habitat especially threatened even at a continental level. For this reason they are now a ‘priority habitat’ under the European Union’s

Habitats Directive. Intensiv elagoonal aquaculture also injects not only nutrients, but antibiotics, hormones, vitamins, and a variety of other compounds, and the wider effects of these are giving cause for concern.

See also Crustacean Fisheries. Fish: Demersal Fish (Life Histories, Behavior, Adaptations). Eels. Eutrophication. Geomorphology. Macrobenthos. Mangroves. Molluskan Fisheries. Pelecaniformes. Phytobenthos. Primary Production Distribution. Salt Marshes and Mud Flats.

Further Reading Ayala Castan˜ares and Phleger F.B. (eds.) (1969) Lagunas Costeras. Un Simposio. Universidad Nacional Auto´noma de Me´xico. Barnes RSK (1980) Coastal Lagoons. Cambridge: Cambridge University Press. Bird ECF (1984) Coasts, 3rd edn. London: Blackwell. Chauvet C (1988) Manuel sur l’Ame´nagement des Peˆches dans les Lagunes Coˆtie`res: La Bordigue Me´diterrane´enne. FAO Document Technique sur les Pches No. 290. FAO: Cooper JAG (1994) Lagoons and microtidal coasts. In: Carter RWG and Woodroffe CD (eds.) Coastal Evolution, pp. 219--265. Cambridge: Cambridge University Press. Emery KO (1969) A Coastal Pond Studied by Oceanographic Methods. Elsevier. Gue´lorget O, Frisoni GF, and Perthuisot J-P (1983) La zonation biologique des milieux lagunaires: de´finition d’une e´chelle de confinement dans le domaine paralique

(c) 2011 Elsevier Inc. All Rights Reserved.

388

LAGOONS

mediterrane´en. Journal de Recherche Oce´anographique, Paris 8>: 15--36. Kapetsky JM and Lasserre G (eds.) (1984) Management of Coastal Lagoon Fisheries, vol. 2 vols.. Rome: FAO. Kjerfve B (ed.) (1994) Coastal Lagoon Processes. Elsevier. Lasserre P and Postma H (eds) (1982) Les Lagunes Ctie`res. Oceanologica Acta (Volume Spe´cial.) Rosecchi E and Charpentier B (1995) Aquaculture in Lagoon and Marine Environments. Tour du Valat.

Sorensen J, Gable F, and Bandarin F (1993) The Management of Coastal Lagoons and Enclosed Bays. American Society of Civil Engineers. Special issue (1992) Vie et Milieu 42(2): 59--251. Ya´n˜ez-Arancibia A (ed.) (1985) Fish Community Ecology in Estuaries and Coastal Lagoons. Universidad Nacional Auto´noma de Me´xico.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGRANGIAN BIOLOGICAL MODELS D. B. Olson, C. Paris, and R. Cowen, University of Miami, Miami, FL, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1438–1443, & 2001, Elsevier Ltd.

Introduction The Swiss mathematician Leonhard Euler (1707– 1783) derived the formulations for describing fluid motion by either measuring the properties of the fluid at a fixed point overtime or alternatively following the trajectory of a parcel of fluid as it is carried with the flow. The first of these is known as the Eulerian description of the flow, while the method following a material parcel or particle is known as the Lagrangian description after the French mathematician Joseph Lagrange (1736–1813). Most of the theory used to model ocean currents is posed in an Eulerian frame because of the difficulties in solving the momentum equations in the complicated matrices that arise in the Lagrangian form of the equations. However, it is often useful to use the Lagrangian frame of reference when considering the manner in which mixing occurs in turbulent flows such as those found in the oceans. It is also common to measure these flows by using drifters or floats that trace out oceanic currents. As shown below, the Lagrangian description is also conducive to handling models of marine populations in many cases. This is especially true when the models include quantities that structure the population, such as age, genetics, or physiological traits that depend upon the history of individual organisms that are carried in or swim through oceanic flows. Organisms that drift freely with the currents are termed planktonic, while those that can swim effectively are termed nektonic fauna. Here Lagrangian methods for considering populations of both plankton and nekton are given. Much of the detailed formalism can be found in Okubo (1980) (see Further Reading). The present discussion highlights the application of these methods to marine population dynamics.

Comparing the Eularian and Lagrangian Formulations To understand the difference between the Lagrangian and Eulerian formulations, consider the population dynamic equations for marine organisms. If the ith

population is made up of Ni individuals, one can write an equation for each individual. This will include each organism’s position, Xm ðtÞ, as a function of time t. The total, or Lagrangian, derivative of Xm ðtÞ with respect to time, dXm/dt, gives the individual’s velocity, Vm(t). This can be separated into the influence of the advection of the organism by ocean currents, U (Xm, t), and a swimming contribution, Us(Xm, B,t), where B (Xm, t) is a vector of behavioral clues. These clues involve both physical and biotic components of the environment. The acceleration of the individual organism is then derived by carrying out another differentiation in time (eqn [1]). dVm @U @Us ¼ þ UdrVm þ Us drVm ¼ FðBÞ þ dt @t @t

½1

r is the spatial gradient operator and F is the gravitational and viscous forces imposed on the organism as well as behavioral responses, i.e., swimming. Notice that the Lagrangian derivative on the left leads to a set of Eulerian terms that involve spatial gradients in the fluid velocity and the behavioral clues on the right side of the equation. This equation fully expresses the motion of an individual. To the equation describing the organism’s motion, a set of state relations must be added expressing changes in its physiological state, its age or stage, and the probabilities of its death and reproduction to explain population dynamics. Such a model considering the conditions of each individual explicitly in a population is called an individual-based model (IBM). Individual-based models provide a method for understanding behavior and small groups of organisms as discussed below. For large populations, however, the number of equations involved becomes impossible to handle. It is therefore common to introduce the concept of organism density, ni ¼ Ni/A, where Ni is the number of individuals and A is a given a real measure. The density of the ith taxon is then measured in numbers per square kilometer of ocean surface area. This leads to a continuous spatial field equation. It is typical to consider the mean field of population density and perturbations about it so ni ¼ /ni S þ n0i . Here the first term is the mean population density and the second the perturbations (or variance) about the mean; the mean is over the population. The same separation can be done for the velocity components such that U ¼ /US þ U0 and

(c) 2011 Elsevier Inc. All Rights Reserved.

389

390

LAGRANGIAN BIOLOGICAL MODELS

Day 10 22°N

20°N

20°N

18°N

18°N

Longitude

Longitude

Day 1 22°N

16°N

16°N

14°N

14°N

12°N

12°N

10°N

75°W

70°W 65°W Latitude

10°N

60°W

75°W

60°W

70°W 65°W Latitude Day 30

Day 20 22°N

22°N

20°N

20°N

18°N

18°N

8

Longitude

Longitude

6

16°N

16°N

14°N

14°N

12°N

12°N

4

2

10°N (A)

75°W

70°W 65°W Latitude

60°W

10°N

75°W

65°W 70°W Latitude

60°W

0 log10 ni

22°N

20°N

Latitude

18°N

16°N

14°N

12°N

10°N 75°W (B)

70°W

65°W Longitude

(c) 2011 Elsevier Inc. All Rights Reserved.

60°W

55°W

LAGRANGIAN BIOLOGICAL MODELS

Us ¼ /Us S þ Us 0 . Equation [1] above involves products of velocity components with each other such that the contributions to the mean motion of the population and therefore its average spread will involve both the mean velocities and correlations between velocity perturbations. In the field equations these correlation’s between fluctuations in the turbulent fluid velocity or swimming behavior will lead to turbulent and behavioral diffusion. It is typical to introduce diffusivities, K for the turbulence and Ks for the behavioral related dispersion. There is also a correlation between the variations in the environmental factors, including the distance to members of the same species that come into effect given B ¼ /BS þ B0 . The resulting field equation for the expected mean density of an organism is then given by eqn [2]. dni @ni ¼ þ Vm drni ¼ FðBÞ þ r½ðk þ ks Þrni  dt @t

½2

The k are inside the spatial gradient operators to denote that they are functions of space. Taking the ks term all the way inside the r operators allows density or schooling effects on population density through behavioral preferences for nearest neighbor distance. The results of the field model versus the Lagrangian model following individuals can lead to impressive differences (Figure 1). The diffusion model in Figure 1A using the equations above lead to a finite probability of finding organisms everywhere in a domain immediately. The Lagrangian treatment limits an organism’s spread to the fastest velocities present, so that it takes a finite time for spread. It is important to note, however, that there must also be population losses that are more abrupt than simple exponential decrease of the population across habitat boundaries. This occurs in most population parametrizations such as a logistic growth with linear mortality to achieve realistic population distributions. In the real ocean, while it takes much longer than in the analytical diffusion model, the Lagrangian motions will still lead to a finite possibility of finding organisms everywhere in the domain at large timescales, unless mortality is properly treated.

391

Simple Models in the Lagrangian Frame The most important issue in modeling marine populations is providing an accurate depiction of the physics, the biology, and the intricate biological/ physical interactions that occur. These are represented in the equations above by mean quantities acting on means or perturbations and then by correlations between both physical and biological perturbations. One use of the Lagrangian description of motion is to simulate these interactions along a fluid trajectory in the case of the plankton (Figures 2 and 3). In this case a simple meandering current and its impact on populations is envisioned. The calculation involved is a simple integration that in Figure 2 reveals the basic response without biological non linearities. In Figure 3 the impact of a primary production response as seen in Figure 2 on a density-dependent population conceals a set of more interesting patterns, including extinction. The situation in Figures 2 and 3 involves dynamics that allow exact calculations, i.e, in this case simple integrations of the functions without any use of numerics. This sort of analysis is recommended for testing morecomplicated cases where numerical methods become a major issue. Essentially these simple applications use the Lagrangian frame of viewing advection as a means of allowing simple calculations of population dynamics. The Lagrangian frame becomes indispensable when structured populations are considered.

Simulations of Populations with Demographic Structure The Lagrangian description of the path that biological entities follow through the ocean environment becomes the only feasible method for treating population dynamics where the detailed history of the populations’ interaction with the physical environment and other populations are important. Early works in this area include the plankton models of Wolf and Woods following the details of mixed layer and thermocline development in the North Atlantic over many seasons and the work on zooplanktonin the coastal environment by Hoffman et al. The problem becomes that the Lagrangian

Figure 1 (Left) (A) Simulation of larval reef fish drift from Barbados using the mean flow into the eastern Caribbean and a k of 5000 m2 s1 at 1, 10, 20, and 30 days after spawning. A typical larval mortality rate m ¼ 0.2 (or B18% per day) is applied to larval abundance (Ni). (B) Lagrangian simulation of the same case with trajectories computed from an oceanic general circulation model at day 30 (Cowen et al., 2000). The survivors are indicated by red dots after applying the same m ¼ 0.2. Note that none are on suitable island habitat after 30 days. For the diffusive case (A) there is a finite probability of finding larvae well beyond the range of any of the simulated trajectories. In this case the mortality is truncated by the 30-day duration of planktonic behavior in the fish’s assumed development, i.e., after this time there is assumed recruitment to an island or death.

(c) 2011 Elsevier Inc. All Rights Reserved.

392

LAGRANGIAN BIOLOGICAL MODELS

1.4

Meander phase

K

1.3 1.2 1.1 1.0 0

1 2  (Lagrangian period)

3

Figure 2 Response of phytoplankton to a simple meandering current. The meander is assumed to set up a simple sinusoidal series of upwelling (high) and down welling (low) at meander crests and troughs, respectively, that provide nutrients upon the upwelling phase. The model assumes logistic phytoplankton response to a sinusoidal carrying capacity (K) and is solved analytically by simple integration of the Lagrangian equation in time. The forcing function involves linear response to a cosine shown in the figure.

frame (i.e., following individual trajectories of individuals or individual subpopulations) is the only way to track the biological dynamics where the past history plays a major role in determining the present dynamics. These historical parameters may involve the past history of nutrient or forage availability, the temperature and salinity encountered over the development of the organisms, or the past history of selection on the genetic structure of populations when reproduction occurs. As an example of a structured population simulation in a turbulent ocean gyre, a simulation of a population of physiologically and genetically structured pelagic copepods is described. The population

29 y 19 y 120 9y 100

Biomass

Phytoplankton biomass

1.5

80



60 40

K

2.5

20 0 0

80°

120° 160° 200° 240° 280° 320° 360°

Northern 

K1

Southern

1.1530

1.1570

1.5

× 10

1.0

K2 K3

0.5

0

Growth rate

Phytoplankton biomass

40°

Southern

(A)

2.0

20

1.1560

10

1.1550

Flow

0

1 2  (Lagrangian period)

0

3

Figure 3 A calculation of the zooplankton response to the phytoplankton distribution in a meander like that treated in Figure 2. Here the integration in time along trajectories is slightly more complicated but still analytical. The zooplankton response is parametrized by a Hollings type 2 curve such that the time dependence of zooplankton (Z) is governed by the equation below.

dZ ZP ¼r  dZ dt K0 þ P Here r is a growth rate, P is the sinusoidal phytoplankton field, K0 is a half-saturation term, and d is the death rate for Z. The pattern of the carrying capacity, K is shown at the top of figure. Four different Ks are shown with different magnitudes K1–K3 and fourth that goes extinct. See Olson and Hood (1994) and the literature cited there for further discussion of meander impacts on marine ecosystems.

0

(B)

40°

Southern

80°

1.1540 120° 160° 200° 240° 280° 320° 360°

Northern 

Southern

Figure 4 (A) Biomass of copepods per m2 of surface area as a function of distance (y) around the gyre, at 10-year intervals. Gyre circulation time is 3 years. The carrying capacity at the northern and southern ends of the gyre are indicated by dashed lines. Note that mixing can cause a region to exceed the carrying capacity. The variation in populations is largest at the highest carrying capacity. (B) The mean growth rate (biomass per day) in different segments around the gyre at the three times. The scale at 9 years is 10 times that for the later times (scale at right). Growth rate is going slowly to a constant or fixed state as expected from population genetics grounds. The distribution of growth rates matches the cosine nature of the carrying capacity distribution, but is fully advective in the sense that these patterns advect with the mean flow around the gyre. The direction of this drift is indicated by an arrow.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAGRANGIAN BIOLOGICAL MODELS

is based on the properties of Nanocalanus minor, a copepod found in the subtropical gyre in the North Atlantic. The model is designed to consider the dynamics behind the mitochondrial DNA patterns found in this population in the Gulf Stream. The Gulf Stream and its recirculation gyre are treated as a circular flow with superimposed turbulence. The copepods are simulated as subpopulations, each carried on Lagrangian particles advected in this flow. The populations are subject to a carrying capacity, K ¼ K0 þ K0 sinðy=2Þ, that is high in the northern Gulf Stream and low in the oligotrophic southern portions of the gyre. The population has variable growth rates that are controlled genetically. The growth potential is determined by the statistics of the local breeding subpopulations and by selection induced by competition at a given location and time for food. Selection is local since there is not an optimal growth rate in the sense that it pays to have a high reproductive potential in the northern gyre under low population densities. The offspring of such a population are inevitably at a disadvantage, however, when advected into the southern oligotrophic reaches of the gyre. Since the resulting genetic and physiological attributes at a location depend upon the past history of all of the subpopulations contributing to the interaction at a given time, this sort of simulation becomes computationally impossible in a Eulerian frame. The population distributions in Figure 4, done in a Lagrangian simulation, takes only an hour on a laptop computer. The details of a suite of such simulations are currently being compared to population density and gene sequences.

Conclusions The use of Lagrangian particle-following simulations in modeling population dynamics allows several advantages over Eularian fixed-grid calculations. For simple models the advantage is that the population equations can be simply integrated in time. As new techniques for tracking fluid parcels and therefore planktonic trajectories or individual large pelagic fish or whales become more available, models using real trajectories will become possible. The other advantage that direct Lagrangian simulation of turbulent dispersal of organisms has is that it overcomes the problems that advection/diffusion schemes have with population densities at large distances from their source. Finally, the largest promise in Lagrangian

393

simulations is their use in models that explicitly treat the demographic traits of populations. With the everincreasing realism in physical models of the marine environment and Lagrangian population models, new insights into marine population dynamics are possible.

See also Plankton Viruses. Population Dynamics Models.

Further Reading Bucklin A, LaJeunesse TC, Curry E, Wallinga J, and Garrison K (1996) Molecular genetic diversity of the copepod, Nannocalanus minor: genetic evidence of species and population structure in the N. Atlantic Ocean. Journal of Marine Research 54: 285--310. Carlotti F (1996) A realistic physical-biological model for Calanus finmarchicus in the North Atlantic. A conceptual approach. Ophelia 44: 47--58. Carlotti F and Wolf KU (1998) A Lagrangian ensemble model of Calanus finmarchicus coupled with a 1-D ecosystem model. Fishers and Oceanography 7(7): 191--204. Cowen RK, Lwiza KMM, Sponaugle S, Paris CB, and Olson DB (2000) Connectivity of marine populations: Open or closed? Science 287: 857--859. Flierl G, Gru¨nbaum D, Levin S, and Olson DB (1999) From individuals to aggregation: the interplay between behavior and physics. Journal of Theoretical Biology 196: 397--454. Hoffmann EE, Halstrom KS, Moisan JR, Haidvogel DB, and Mackas DL (1991) Use of simulated drifter tracks to investigate general transport patterns and residence times in the coastal transition zone. Journal of Geophysical Research 96: 15041--15052. Metz JAJ and Diekmann O (eds.) (1986) The Dynamics of Physiologically Structured Populations, 68. Berlin: Springer-Verlag. Okubo A (1980) Diffusion and Ecological Problems: Mathematical Models, 10. Berlin: Springer-Verlag. Olson DB and Hood RR (1994) Modelling pelagic biogeography. Progress in Oceanography 34: 161--205. Wolf KU and Woods JD (1998) Lagrangian simulation of primary production in the physical environment: The deep chlorophyll maximum and nutricline. In: Rothschild BJ (ed.) Toward a Theory on Biological– Physical Interactions in the World Ocean, pp. 51--70. Dordrecht: Kluwer Academic.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAND–SEA GLOBAL TRANSFERS F. T. Mackenzie and L. M. Ver, University of Hawaii, Honolulu, HI, USA

and groundwater flows, thus moving toward the sea in a greater degree than before.

& 2009 Elsevier Ltd. All rights reserved.

The Changing Picture of Land–Sea Global Transfers Introduction The interface between the land and the sea is an important boundary connecting processes operating on land with those in the ocean. It is a site of rapid population growth, industrial and agricultural practices, and urban development. Large river drainage basins connect the vast interiors of continents with the coastal zone through river and groundwater discharges. The atmosphere is a medium of transport of substances from the land to the sea surface and from that surface back to the land. During the past several centuries, the activities of humankind have significantly modified the exchange of materials between the land and sea on a global scale – humans have become, along with natural processes, agents of environmental change. For example, because of the combustion of the fossil fuels coal, oil, and gas and changes in land-use practices including deforestation, shifting cultivation, and urbanization, the direction of net atmospheric transport of carbon (C) and sulfur (S) gases between the land and sea has been reversed. The global ocean prior to these human activities was a source of the gas carbon dioxide (CO2) to the atmosphere and hence to the continents. The ocean now absorbs more CO2 than it releases. In preindustrial time, the flux of reduced sulfur gases to the atmosphere and their subsequent oxidation and deposition on the continental surface exceeded the transport of oxidized sulfur to the ocean via the atmosphere. The situation is now reversed because of the emissions of sulfur to the atmosphere from the burning of fossil fuels and biomass on land. In addition, river and groundwater fluxes of the bioessential elements carbon, nitrogen (N), and phosphorus (P), and certain trace elements have increased because of human activities on land. For example, the increased global riverine (and atmospheric) transport of lead (Pb) corresponds to its increased industrial use. Also, recent changes in the concentration of lead in coastal sediments appear to be directly related to changes in the use of leaded gasoline in internal combustion engines. Synthetic substances manufactured by modern society, such as pesticides and pharmaceutical products, are now appearing in river

394

Although the exchange through the atmosphere of certain trace metals and gases between the land and the sea surface is important, rivers are the main purveyors of materials to the ocean. The total water discharge of the major rivers of the world to the ocean is 36 000 km3 yr 1. At any time, the world’s rivers contain about 0.000 1% of the total water volume of 1459  106 km3 near the surface of the Earth and have a total dissolved and suspended solid concentration of c. 110 and 540 ppm l 1, respectively. The residence time of water in the world’s rivers calculated with respect to total net precipitation on land is only 18 days. Thus the water in the world’s rivers is replaced every 18 days by precipitation. The global annual direct discharge to the ocean of groundwater is about 10% of the surface flow, with a recent estimate of 2400 km3 yr 1. The dissolved constituent content of groundwater is poorly known, but one recent estimate of the dissolved salt groundwater flux is 1300  106 t yr 1. The chemical composition of average river water is shown in Table 1. Note that the major anion in river water is bicarbonate, HCO3 ; the major cation is calcium, Ca2þ , and that even the major constituent concentrations of river water on a global scale are influenced by human activities. The dissolved load of the major constituents of the world’s rivers is derived from the following sources: about 7% from beds of salt and disseminated salt in sedimentary rocks, 10% from gypsum beds and sulfate salts disseminated in rocks, 38% from carbonates, and 45% from the weathering of silicate minerals. Two-thirds of the HCO3  in river waters are derived from atmospheric CO2 via the respiration and decomposition of organic matter and subsequent conversion to HCO3 through the chemical weathering of silicate (B30% of total) or carbonate (B70% of total) minerals. The other third of the river HCO3  comes directly from carbonate minerals undergoing weathering. It is estimated that only about 20% of the world’s drainage basins have pristine water quality. The organic productivity of coastal aquatic environments has been heavily impacted by changes in the

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved. 4.6 4.3 1.4 1.4 4.9 4.9 6.7 5.2 3.8 3.8 3.7 3.4 0.3 8%

6.3 6.3

21.2 20.1

31.7 24.2

15.2 15.0

14.7 13.4 1.3 9%

2.2 2.2

17.8 16.6

5.7 5.3

Mg2 þ

7.2 5.2 2.0 28%

7.6 7.0

16.5 3.2

8.4 6.5

3.3 3.3

8.7 6.6

4.4 3.8

Na þ

1.4 1.3 0.1 7%

1.1 1.1

1.8 1.1

1.5 1.5

1.0 1.0

1.7 1.6

1.4 1.4

Kþ

8.3 5.8 2.5 30%

6.8 5.9

20.0 4.7

9.2 7.0

4.1 4.1

10.0 7.6

4.1 3.4

Cl 

11.5 5.3 6.2 54%

7.7 6.5

35.5 15.1

18.0 14.9

3.8 3.5

13.3 9.7

4.2 3.2

SO42

53.0 52.0 1.0 2%

65.6 65.1

86.0 80.1

72.3 71.4

24.4 24.4

67.1 66.2

36.9 26.7

HCO 3

10.4 10.4 0.0 0%

16.3 16.3

6.8 6.8

7.2 7.2

10.3 10.3

11.0 11.0

12.0 12.0

SO2

110.1 99.6 10.5 10%

125.3 120.3

212.8 140.3

142.6 133.5

54.6 54.3

134.6 123.5

60.5 57.8

TDSb

12.57 9.89 2.67 21%

TOCb

0.574 0.386 0.187 33%

TDNb

0.053 0.027 0.027 50%

TDPb

37.40 37.40

2.40

2.56

5.53

11.04

12.47

3.41

Water discharge (103 km3 yr 1)

0.46 0.46

0.42

0.38

0.41

0.54

0.28

Runoff ratioc

b

Actual concentrations include pollution. Natural concentrations are corrected for pollution. TDS, total dissolved solids; TOC, total organic carbon; TDN, total dissolved nitrogen; TDP, total dissolved phosphorus. c Runoff ratio ¼ average runoff per unit area/average rainfall. Revised after Meybeck M (1979) Concentrations des eaux fluviales en elements majeurs et apports en solution aux oceans. Revue de Geologie Dynamique et de Geographie Physique 21: 215–246; Meybeck M (1982) Carbon, nitrogen, and phosphorus transport by world rivers. American Journal of Science 282: 401–450; Meybeck M (1983) C, N, P and S in rivers: From sources to global inputs. In: Wollast R, Mackenzie FT, and Chou L (eds.) Interactions of C, N, P and S Biogeochemical Cycles and Global Change, pp. 163–193. Berlin: Springer.

a

Africa Actual Natural Asia Actual Natural S. America Actual Natural N. America Actual Natural Europe Actual Natural Oceania Actual Natural World average Actual Natural (unpolluted) Pollution World % pollutive

Ca2 þ

River water concentrationa (mg l 1)

Chemical composition of average river water

By continent

Table 1

396

LAND–SEA GLOBAL TRANSFERS

dissolved and particulate river fluxes of three of the major bioessential elements found in organic matter, C, N, and P (the other three are S, hydrogen (H), and oxygen (O)). Although these elements are considered minor constituents of river water, their fluxes may have doubled over their pristine values on a global scale because of human activities. Excessive riverborne nutrients and the cultural eutrophication of freshwater and coastal marine ecosystems go hand in hand. In turn, these fluxes have become sensitive indicators of the broader global change issues of population growth and land-use change (including water resources engineering works) in the coastal zone and upland drainage basins, climatic change, and sea level rise. In contrast to the situation for the major elements, delivery of some trace elements from land to the oceans via the atmosphere can rival riverine inputs. The strength of the atmospheric sources strongly depends on geography and meteorology. Hence the North Atlantic, western North Pacific, and Indian Oceans, and their inland seas, are subjected to large atmospheric inputs because of their proximity to both deserts and industrial sources. Crustal dust is the primary terrestrial source of these atmospheric inputs to the ocean. Because of the low solubility of dust in both atmospheric precipitation and seawater and the overwhelming inputs from river sources, dissolved sources of the elements are generally less important. However, because the oceans contain only trace amounts of iron (Fe), aluminum (Al), and manganese (Mn) (concentrations are in the ppb level), even the small amount of dissolution in seawater (B10% of the element in the solid phase) results in eolian dust being the primary source for the dissolved transport of these elements toremote areas of the ocean. Atmospheric transport of the major nutrients N, silicon (Si), and Fe to the ocean has been hypothesized to affect and perhaps limit primary productivity in certain regions of the ocean at certain times. Modern processes of fossil fuel combustion and biomass burning have significantly modified the atmospheric transport from land to the ocean of trace metals like Pb, copper (Cu), and zinc (Zn), C in elemental and organic forms, and nutrient N. As an example of global land–sea transfers involving gases and the effect of human activities on the exchange, consider the behavior of CO2 gas. Prior to human influence on the system, there was a net flux of CO2 out of the ocean owing to organic metabolism (net heterotrophy). This flux was mainly supported by the decay of organic matter produced by phytoplankton in the oceans and part of that transported by rivers to the oceans. An example

overall reaction is: C106 H263 O110 N16 S2 P þ 141O2 ) 106CO2 þ 6HNO3 þ 2H2 SO4 þ H3 PO4 þ 120H2 O

½1

Carbon dioxide was also released to the atmosphere due to the precipitation of carbonate minerals in the oceans. The reaction is: Ca2þ þ 2HCO3 ) CaCO3 þ CO2 þH2 O

½2

The CO2 in both reactions initially entered the dissolved inorganic carbon (DIC) pool of seawater and was subsequently released to the atmosphere at an annual rate of about 0.2  109 t of carbon as CO2 gas. It should be recognized that this is a small number compared with the 200  109 t of carbon that exchanges between the ocean and atmosphere each year because of primary production of organic matter and its subsequent respiration. Despite the maintenance of the net heterotrophic status of the ocean and the continued release of CO2 to the ocean–atmosphere owing to the formation of calcium carbonate in the ocean, the modern ocean and the atmosphere have become net sinks of anthropogenic CO2 from the burning of fossil fuels and the practice of deforestation. Over the past 200 years, as CO2 has accumulated in the atmosphere, the gradient of CO2 concentration across the atmosphere–ocean interface has changed, favoring uptake of anthropogenic CO2 into the ocean. The average oceanic carbon uptake for the decade of the 1990s was c. 2  109 t annually. The waters of the ocean have accumulated about 130  109 t of anthropogenic CO2 over the past 300 years.

The Coastal Zone and Land–Sea Exchange Fluxes The global coastal zone environment is an important depositional and recycling site for terrigenous and oceanic biogenic materials. The past three centuries have been the time of well-documented human activities that have become an important geological factor affecting the continental and oceanic surface environment. In particular, historical increases in the global population in the areas of the major river drainage basins and close to oceanic coastlines have been responsible for increasing changes in land-use practices and discharges of various substances into oceanic coastal waters. As a consequence, the global C cycle and the cycles of N and P that closely interact with the carbon cycle have been greatly affected. Several major perturbations of the past three

(c) 2011 Elsevier Inc. All Rights Reserved.

LAND–SEA GLOBAL TRANSFERS

centuries of the industrial age have affected the processes of transport from land, deposition of terrigenous materials, and in situ production of organic matter in coastal zone environments. In addition, potential future changes in oceanic circulation may have significant effects on the biogeochemistry and CO2 exchange in the coastal zone. The coastal zone is that environment of continental shelves, including bays, lagoons, estuaries, and near-shore banks that occupy 7% of the surface area of the ocean (36  1012 m2) or c. 9% of the volume of the surface mixed layer of the ocean (3  1015 m3). The continental shelves average 75 km in width, with a bottom slope of 1.7 m km 1. They are generally viewed as divisible into the interior or proximal shelf, and the exterior or distal shelf. The mean depth of the global continental shelf is usually taken as the depth of the break between the continental shelf and slope at c. 200 m, although this depth varies considerably throughout the world’s oceans. In the Atlantic, the median depth of the shelf-slope break is at 120 m, with a range from 80 to 180 m. The depths of the continental shelf are near 200 m in the European section of the Atlantic, but they are close to 100 m on the African and North American coasts. Coastal zone environments that have high sedimentation rates, as great as 30–60 cm ky–1 in active depositional areas, act as traps and filters of natural and human-generated materials transported from continents to the oceans via river and groundwater flows and through the atmosphere. At present a large fraction (B80%) of the land-derived organic and inorganic materials that are transported to the oceans is trapped on the proximal continental shelves. The coastal zone also accounts for 30–50% of total carbonate and 80% of organic carbon accumulation in the ocean. Coastal zone environments are also regions of higher biological production relative to that of average oceanic surface waters, making them an important factor in the global carbon cycle. The higher primary production is variably attributable to the nutrient inflows from land as well as from coastal upwelling of deeper ocean waters. Fluvial and atmospheric transport links the coastal zone to the land; gas exchange and deposition are its links with the atmosphere; net advective transport of water, dissolved solids and particles, and coastal upwelling connect it with the open ocean. In addition, coastal marine sediments are repositories for much of the material delivered to the coastal zone. In the last several centuries, human activities on land have become a geologically important agent affecting the land–sea exchange of materials. In particular, river and groundwater flows and atmospheric

397

transport of materials to the coastal zone have been substantially altered. Bioessential Elements

Continuous increase in the global population and its industrial and agricultural activities have created four major perturbations on the coupled system of the biogeochemical cycles of the bioessential elements C, N, P, and S. These changes have led to major alterations in the exchanges of these elements between the land and sea. The perturbations are: (1) emissions of C, N, and S to the atmosphere from fossil-fuel burning and subsequent partitioning of anthropogenic C and deposition of N and S; (2) changes in land-use practices that affect the recycling of C, N, P, and S on land, their uptake by plants and release from decaying organic matter, and the rates of land surface denudation; (3) additions of N and P in chemical fertilizers to cultivated land area; and (4) releases of organic wastes containing highly reactive C, N, and P that ultimately enter the coastal zone. A fifth major perturbation is a climatic one: (5) the rise in mean global temperature of the lower atmosphere of about 1 1C in the past 300 years, with a projected increase of about 1.4–5.8 1C relative to 1990 by the year 2100. Figure 1 shows how the fluxes associated with these activities have changed during the past three centuries with projections to the year 2040. Partially as a result of these activities on land, the fluxes of materials to the coastal zone have changed historically. Figure 2 shows the historical and projected future changes in the river fluxes of dissolved inorganic and organic carbon (DIC, DOC), nitrogen (DIN, DON), and phosphorus (DIP, DOP), and fluxes associated with the atmospheric deposition and denitrification of N, and accumulation of C in organic matter in coastal marine sediments. It can be seen in Figure 2 that the riverine fluxes of C, N, and P all increase in the dissolved inorganic and organic phases from about 1850 projected to 2040. For example, for carbon, the total flux (organic þ inorganic) increases by about 35% during this period. These increased fluxes are mainly due to changes in land-use practices, including deforestation, conversion of forest to grassland, pastureland, and urban centers, and regrowth of forests, and application of fertilizers to croplands and the subsequent leaching of N and P into aquatic systems. Inputs of nutrient N and P to the coastal zone which support new primary production are from the land by riverine and groundwater flows, from the open ocean by coastal upwelling and onwelling, and to a lesser extent by atmospheric deposition of nitrogen. New primary production depends on the

(c) 2011 Elsevier Inc. All Rights Reserved.

398

LAND–SEA GLOBAL TRANSFERS

(d) 150

10

Fossil fuel

Carbon

100

5

50 Nitrogen Land use

0

0

Phosphorus

Sewage input (Mt yr −1)

C emissions (Gt yr −1)

(a)

(e)

100 Sulfur

50 0

Nitrogen

N, S emissions (Mt yr −1)

150

1.5 Δ temperature (°C)

(b)

1.0 0.5 0

Fertilizer input (Mt yr −1)

(c) 1700

150

1800

1900 Year

2000

Nitrogen

100 50 0 1700

Phosphorus

1800

1900 Year

2000

Figure 1 Major perturbations on the Earth system over the past 300 years and projections for the future: (a) emissions of CO2 and (b) gaseous N and S from fossil-fuel burning and land-use activities; (c) application of inorganic N and P in chemical fertilizers to cultivated land; (d) loading of highly reactive C, N, and P into rivers and the coastal ocean from municipal sewage and wastewater disposal; and (e) rise in mean global temperature of the lower atmosphere relative to 1700. Revised after Ver LM, Mackenzie FT, and Lerman A (1999) Biogeochemical responses of the carbon cycle to natural and human perturbations: Past, present, and future. American Journal of Science 299: 762–801.

availability of nutrients from these external inputs, without consideration of internal recycling of nutrients. Thus any changes in the supply of nutrients to the coastal zone owing to changes in the magnitude of these source fluxes are likely to affect the cycling pathways and balances of the nutrient elements. In particular, input of nutrients from the open ocean by coastal upwelling is quantitatively greater than the combined inputs from land and the atmosphere. This makes it likely that there could be significant effects on coastal primary production because of changes in ocean circulation. For example, because of global warming, the oceans could become more strongly stratified owing to freshening of polar oceanic waters and warming of the ocean in the tropical zone. This could lead to a reduction in the intensity of the oceanic thermohaline circulation (oceanic circulation owing to differences in density of water masses, also

popularly known as the ‘conveyor belt’) and hence the rate at which nutrient-rich waters upwell into coastal environments. Another potential consequence of the reduction in the rate of nutrient inputs to the coastal zone by upwelling is the change in the CO2 balance of coastal waters: reduction in the input of DIC to the coastal zone from the deeper ocean means less dissolved CO2, HCO3 , and CO32 coming from that source. With increasing accumulation of anthropogenic CO2 in the atmosphere, the increased dissolution of atmospheric CO2 in coastal water is favored. The combined result of a decrease in the upwelling flux of DIC and an enhancement in the transfer of atmospheric CO2 across the air–sea interface of coastal waters is a lower saturation state for coastal waters with respect to the carbonate minerals calcite, aragonite, and a variety of magnesian calcites. The

(c) 2011 Elsevier Inc. All Rights Reserved.

LAND–SEA GLOBAL TRANSFERS

(a) 50

Carbon

Riverine total organic flux 40

Riverine dissolved inorganic flux

30

Organic carbon Accumulation in sediments 20

10 (b) 6

Nitrogen Riverine total organic flux

5 4

Denitrification

3 2

399

lower saturation state in turn leads to the likelihood of lower rates of inorganic and biological precipitation of carbonate and hence deposition and accumulation of sedimentary carbonate. In addition, the present-day burial rate of organic carbon in the ocean may be about double that of the late Holocene flux, supported by increased fluxes of organic carbon to the ocean via rivers and groundwater flows and increased in situ new primary production supported by increased inputs of inorganic N and P from land and of N deposited from the atmosphere. The organic carbon flux into sediments may constitute a sink of anthropogenic CO2 and a minor negative feedback on accumulation of CO2 in the atmosphere. The increased flux of land-derived organic carbon delivered to the ocean by rivers may accumulate there or be respired, with subsequent emission of CO2 back to the atmosphere. This release flux of CO2 may be great enough to offset the increased burial flux of organic carbon to the seafloor due to enhanced fertilization of the ocean by nutrients derived from human activities. The magnitude of the CO2 exchange is a poorly constrained flux today. One area for which there is a substantial lack of knowledge is the Asian Pacific region. This is an area of several large seas, a region of important river inputs to the ocean of N, P, organic carbon, and sediments from land, and a region of important CO2 exchange between the ocean and the atmosphere.

Atmospheric deposition flux

Anticipated Response to Global Warming 1 Riverine dissolved inorganic flux

0 (c)

Phosphorus 0.6

0.4 Riverine total organic flux 0.2 Riverine dissolved inorganic flux 0 1850

1950

1900

2000

Year

Figure 2 Past, present, and predicted fluxes of carbon, nitrogen, and phosphorus into or out of the global coastal margin, in 1012 mol yr  1.

From 1850 to modern times, the direction of the net flux of CO2 between coastal zone waters and the atmosphere due to organic metabolism and calcium carbonate accumulation in coastal marine sediments was from the coastal surface ocean to the atmosphere (negative flux, Figure 3).This flux in 1850 was on the order of 0.2  109 t yr 1. In a condition not disturbed by changes in the stratification and thermohaline circulation of the ocean brought about by a global warming of the Earth, the direction of this flux is projected to remain negative (flux out of the coastal ocean to the atmosphere) until early in the twenty-first century. The increasing partial pressure of CO2 in the atmosphere because of emissions from anthropogenic sources leads to a reversal in the gradient of CO2 across the air–sea interface of coastal zone waters and, hence, invasion of CO2 into the coastal waters. From that time on the coastal ocean will begin to operate as a net sink (positive flux) of atmospheric CO2 (Figure 3). The role of the open ocean as a sink for anthropogenic CO2 is slightly reduced while that of the coastal oceans

(c) 2011 Elsevier Inc. All Rights Reserved.

400

LAND–SEA GLOBAL TRANSFERS

5.0 No change 34% reduction 50% reduction 100% reduction

160

6.0

4.0

5.5

Calcite

120

2.0

40

1.0

(+) flux from atmosphere 0

0

to coastal waters

Carbonate saturation state, Ω

80

(Mol C m−2yr−1)

(1012 mol C yr−1)

3.0

5.0

4.5 Aragonite 4.0

3.5 _1.0

1850

1900

1950 Year

2000

15 mol.% magnesian calcite

3.0 Figure 3 The net flux of CO2 between coastal zone waters and the atmosphere due to organic metabolism and calcium carbonate accumulation in coastal marine sediments, under three scenarios of changing thermohaline circulation rate compared to a business-as-usual scenario, in units of 1012 mol C yr 1.

increases. The net result is the maintenance of the role of the global oceans as a net sink for anthropogenic CO2. The saturation index (O) for calcite or aragonite (both CaCO3) is the ratio of the ion activity product IAP in coastal waters to the respective equilibrium constant K at the in situ temperature. For aqueous species, IAP ¼ aCa2þ  aCO32 (where a is the activity; note that for 15 mol.% magnesian calcite, the IAP also includes the activity of the magnesium cation, aMg2þ ). Most coastal waters and open-ocean surface waters currently are supersaturated with respect to aragonite, calcite, and magnesian calcite containing 15 mol.% Mg, that is, Ocalcite, Oaragonite, and O15%magnesian calcite are 41. Because of global warming and the increasing land-to-atmosphere-toseawater transport of CO2 (due to the continuing combustion of fossil fuels and burning of biomass), the concentration of the aqueous CO2 species in seawater increases and the pH of the water decreases slightly. This results in a decrease in the concentration of the carbonate ion, CO322 , resulting in a decrease in the degree of supersaturation of coastal zone waters. Figure 4 shows how the degree of saturation might change into the next century because of rising atmospheric CO2 concentrations. The overall reduction in the saturation state of coastal

1980

2000

2020

2040

Year Figure 4 Changes in saturation state with respect to carbonate minerals of surface waters of the coastal ocean projected from 1999 to 2040. Calculations are for a temperature of 25 1C.

zone waters with respect to aragonite from 1997 projected to 2040 is about 16%, from 3.89 to 3.26. Modern carbonate sediments deposited in shoalwater (‘shallow-water’) marine environments (including shelves, banks, lagoons, and coral reef tracts) are predominantly biogenic in origin derived from the skeletons and tests of benthic and pelagic organisms, such as corals, foraminifera, echinoids, mollusks, algae, and sponges. One exception to this statement is some aragonitic muds that may, at least in part, result from the abiotic precipitation of aragonite from seawater in the form of whitings. Another exception is the sand-sized, carbonate oo¨ids composed of either aragonite with a laminated internal structure or magnesian calcite with a radial internal structure. In addition, early diagenetic carbonate cements found in shoal-water marine sediments and in reefs are principally aragonite or magnesian calcite. Thus carbonate production and accumulation in shoal-water environments are dominated by a range of metastable carbonate minerals associated with skeletogenesis and abiotic processes, including calcite, aragonite, and a variety of magnesian calcite compositions.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAND–SEA GLOBAL TRANSFERS

With little doubt, as has been documented in a number of observational and experimental studies, a reduction in the saturation state of ocean waters will lead to a reduction in the rate of precipitation of both inorganic and skeletal calcium carbonate. Conversely, increases in the degree of supersaturation and temperature will increase the precipitation rates of calcite and aragonite from seawater. During global warming, rising sea surface temperatures and declining carbonate saturation states due to the absorption of anthropogenic CO2 by surface ocean waters result in opposing effects. However, experimental evidence suggests that within the range of temperature change predicted for the next century due to global warming, the effect of changes in saturation state will be the predominant factor affecting precipitation rate. Thus decreases in precipitation rates should lead to a decrease in the production and accumulation of shallow-water carbonate sediments and perhaps changes in the types and distribution of calcifying biotic species found in shallow-water environments.

Anticipated Response to Heightened Human Perturbation: The Asian Scenario In the preceding sections it was shown that the fluxes of C, N, and P from land to ocean have increased because of human activities (refer to Table 1 for data comparing the actual and natural concentrations of C, N, P, and other elements in average river water). During the industrial era, these fluxes mainly had their origin in the present industrialized and developed countries. This is changing as the industrializing and developing countries move into the twenty-first century. A case in point is the countries of Asia. Asia is a continent of potentially increasing contributions to the loading of the environment owing to a combination of such factors as its increasing population, increasing industrialization dependent on fossil fuels, concentration of its population along the major river drainage basins and in coastal urban centers, and expansion of land-use practices. It is anticipated that Asia will experience similar, possibly even greater, loss of storage of C and nutrient N and P on land and increased storage in coastal marine sediments per unit area than was shown by the developed countries during their period of industrialization. The relatively rapid growth of Asia’s population along the oceanic coastal zone indicates that higher inputs of both dissolved and particulate organic nutrients may be expected to enter coastal waters.

401

A similar trend of increasing population concentration in agricultural areas inland, within the drainage basins of the main rivers flowing into the ocean, is also expected to result in increased dissolved and particulate organic nutrient loads that may eventually reach the ocean. Inputs from inland regions to the ocean would be relatively more important if no entrapment or depositional storage occurred en route, such as in the dammed sections of rivers or in alluvial plains. In the case of many of China’s rivers, the decline in sediment discharge from large rivers such as the Yangtze and the Yellow Rivers is expected to continue due to the increased construction of dams. The average decadal sediment discharge from the Yellow River, for example, has decreased by 50% from the 1950s to the 1980s. If the evidence proposed for the continental United States applies to Asia, the damming of major rivers would not effectively reduce the suspended material flow to the ocean because of the changes in the erosional patterns on land that accompany river damming and more intensive land-use practices. These flows on land and into coastal ocean waters are contributing factors to the relative importance of autotrophic and heterotrophic processes, competition between the two, and the consequences for carbon exchange between the atmosphere and land, and the atmosphere and ocean water. The change from the practices of land fertilization by manure to the more recent usage of chemical fertilizers in Asia suggests a shift away from solid organic nutrients and therefore a reduced flow of materials that might promote heterotrophy in coastal environments. Sulfur is an excellent example of how parts of Asia can play an important role in changing land–sea transfers of materials. Prior to extensive human interference in the global cycle of sulfur, biogenically produced sulfur was emitted from the sea surface mainly in the form of the reduced gas dimethyl sulfide (DMS). DMS was the major global natural source of sulfur for the atmosphere, excluding sulfur in sea salt and soil dust. Some of this gas traveled far from its source of origin. During transport the reduced gas was oxidized to micrometer-size sulfate aerosol particles and rained out of the atmosphere onto the sea and continental surface. The global sulfur cycle has been dramatically perturbed by the industrial and biomass burning activities of human society. The flux of gaseous sulfur dioxide to the atmosphere from the combustion of fossil fuels in some regions of the world and its conversion to sulfate aerosol greatly exceeds natural fluxes of sulfur gases from the land surface. It is estimated that this flux for the year 1990 was equivalent to 73  106 t yr1, nearly 4 times the natural DMS flux from the

(c) 2011 Elsevier Inc. All Rights Reserved.

402

LAND–SEA GLOBAL TRANSFERS

ocean. This has led to a net transport of sulfur from the land to the ocean via the atmosphere, completely reversing the flow direction in preindustrial times. In addition, the sulfate aerosol content of the atmosphere derived from human activities has increased. Sulfate aerosols affect global climate directly as

particles that scatter incoming solar radiation and indirectly as cloud condensation nuclei (CCNs), which lead to an increased number of cloud droplets and an increase in the solar reflectance of clouds. Both effects cause the cooling of the planetary surface. As can be seen in Figure 5 the eastern Asian

(a)

2500 1000 500 250

(b)

100 50 10

Figure 5 Comparison of the magnitude of atmospheric sulfur deposition for the years 1990 (a) and 2050 (b). Note the large increases in both spatial extent and intensity of sulfur deposition in both hemispheres and the increase in importance of Asia, Africa, and South America as sites of sulfur deposition between 1990 and 2050. The values on the diagrams are in units of kg S m 2 yr 1. Revised after Mackenzie FT (1998) Our Changing Planet : An Introduction to Earth System Science and Global Environmental Change. Upper Saddle River, NJ: Prentice Hall; Rodhe H, Langner J, Gallardo L, and Kjellstro¨m E (1995) Global transport of acidifying pollutants. Water, Air and Soil Pollution 85: 37–50.

(c) 2011 Elsevier Inc. All Rights Reserved.

LAND–SEA GLOBAL TRANSFERS

region is an important regional source of sulfate aerosol because of the combustion of fossil fuels, particularly coal. This source is predicted to grow in strength during the early- to mid-twenty-first century (Figure 5).

Conclusion Land–sea exchange processes and fluxes of the bioessential elements are critical to life. In several cases documented above, these exchanges have been substantially modified by human activities. These modifications have led to a number of environmental issues including global warming, acid deposition, excess atmospheric nitrogen deposition, and production of photochemical smog. All these issues have consequences for the biosphere – some well known, others not so well known. It is likely that the developing world, with increasing population pressure and industrial development and with no major changes in agricultural technology and energy consumption rates, will become a more important source of airborne gases and aerosols and materials for river and groundwater systems in the future. This will lead to further modification of land–sea global transfers. The region of southern and eastern Asia is particularly well poised to influence significantly these global transfers.

See also Aeolian Inputs. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Coastal Topography, Human Impact on. Nitrogen Cycle. Ocean Carbon System, Modeling of. Ocean Circulation: Meridional Overturning Circulation. Past Climate from Corals. Phosphorus Cycle.

403

Further Reading Berner EA and Berner RA (1996) Global Environment: Water, Air and Geochemical Cycles. Upper Saddle River, NJ: Prentice Hall. Galloway JN and Melillo JM (eds.) (1998) Asian Change in the Context of Global Change. Cambridge, MA: Cambridge University Press. Mackenzie FT (1998) Our Changing Planet: An Introduction to Earth System Science and Global Environmental Change. Upper Saddle River, NJ: Prentice Hall. Mackenzie FT and Lerman A (2006) Carbon in the Geobiosphere – Earth’s Outer Shell. Dordrecht: Springer. Meybeck M (1979) Concentrations des eaux fluviales en elements majeurs et apports en solution aux oceans. Revue de Geologie Dynamique et de Geographie Physique 21: 215--246. Meybeck M (1982) Carbon, nitrogen, and phosphorus transport by world rivers. American Journal of Science 282: 401--450. Meybeck M (1983) C, N, P and S in rivers: From sources to global inputs. In: Wollast R, Mackenzie FT, and Chou L (eds.) Interactions of C, N, P and S Biogeochemical Cycles and Global Change, pp. 163--193. Berlin: Springer. Rodhe H, Langner J, Gallardo L, and Kjellstro¨m E (1995) Global transport of acidifying pollutants. Water, Air and Soil Pollution 85: 37--50. Schlesinger WH (1997) Biogeochemistry: An Analysis of Global Change. San Diego, CA: Academic Press. Smith SV and Mackenzie FT (1987) The ocean as a net heterotrophic system: Implications from the carbon biogeochemical cycle. Global Biogeochemical Cycles 1: 187--198. Ver LM, Mackenzie FT, and Lerman A (1999) Biogeochemical responses of the carbon cycle to natural and human perturbations: Past, present, and future. American Journal of Science 299: 762--801. Vitousek PM, Aber JD, and Howarth RW (1997) Human alteration of the global nitrogen cycle: Sources and consequences. Ecological Applications 7(3): 737--750. Wollast R and Mackenzie FT (1989) Global biogeochemical cycles and climate. In: Berger A, Schneider S, and Duplessy JC (eds.) Climate and Geo-Sciences, pp. 453--473. Dordrecht: Kluwer Academic Publishers.

(c) 2011 Elsevier Inc. All Rights Reserved.

LANGMUIR CIRCULATION AND INSTABILITY S. Leibovich, Cornell University, Ithaca, NY, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The surface of a wind-driven sea often is marked by streaks roughly aligned with the wind direction. These streaks, or windrows, are visible manifestations of coherent subsurface motions extending throughout the bulk of the ocean surface mixed layer, extending from the surface down to the seasonal thermocline. These may be regarded as the large scales of the turbulence in the mixed layer. Windrows and their subsurface origins were first systematically studied and described by Irving Langmuir in 1938, and the phenomenon since has become known as Langmuir circulation. The existence of a simple deterministic description making these large scales theoretically accessible distinguishes this problem from coherent structures in other turbulent flows. The theory traces these patterns to a convective instability mechanically driven by the wind waves and currents. Recent advances in instrumentation and computational data analysis have led to field observations of Langmuir circulation of unprecedented detail. Although the body of observational data obtained since Langmuir’s own work is mainly qualitative, ocean experiments now can yield quantitative measurements of velocity fields in the near-surface region. New measurement methods are capable of producing data comprehensive enough to characterize the phenomenon, and its effect on the stirring and maintenance of the mixed layer, although the labor and difficulties involved and the shear complexity of the processes occurring in the surface layer leave much work to be done before this can be said to be accomplished. Nevertheless, the combination of new experimental techniques and a simple and testable theoretical mechanism has stimulated rapid progress in the exploration of the stirring of the ocean surface mixed layer.

mechanical processes through the action of the wind, as Langmuir originally indicated. At the surface, rolls act to sweep surface water from regions of surface divergence overlying upwelling water into convergence zones overlying downwelling water. Floating material is collected into lines of surface convergence visible as windrows. In confined bodies of water, such as lakes and ponds, windrows are very nearly parallel to the wind, and can have a nearly uniform spacing as shown in Figure 2. In the open ocean, evidence indicates windrows tend to be oriented at small angles to the wind (typically to the right in the Northern Hemisphere), spacing is more variable, and individual windrows can be traced only for a modest multiple of the mean spacing. A windrow may either terminate, perhaps due to local absence of surface tracers, coalesce with an adjacent windrow, or split into two daughter windrows. Thus in the ocean, the general surface appearance is of a network of lines, occasionally interecting, yet roughly aligned with the wind. Windrows are visible in nature only when both Langmuir circulation and surface tracers are present. In the ocean, bubbles from breaking waves are the most readily available tracers, and Langmuir

z

Wind 35.9 –100

35.5 Shelf 31.8S stns. 59:69

–200

35

Salinity –300

Depth

0

114.4

114.6

114.8

115

115.2

115.4

34.5

115.6

>21

20

–100 15 –200 –300

Shelf 10

Temperature 114.4

114.6

114.8

115 Longitude

115.2

115.4

115.6 0.5

–100 0 Shelf

–200

N–S velocity

–0.5

–300 114.4

114.6

114.8

115

115.2

115.4

115.6

Figure 6 Ship measurements of salinity (practical salinity units) and temperature (1C) with a CTD, and north–south current with an acoustic Doppler current profiler (ADCP) west of Perth (321 S) in Dec. 1994.

Sea surface temperature

20°

35.9 35.8

18° Salinity

16° XBT:

62

63

64

65

*

*

*

*

50 Depth (m)

35.7 35.6

0.5 1.0

100 19° 150

18° 17°

200

0.5

16°

15°

10 km Figure 7 A north–south transect across the LC south of Western Australia near 1171 E in June 1987. Upper panel: surface temperature and salinity across the Leeuwin Current. Lower panel: current speed from acoustic Doppler current profiler (ADCP) measurements, the temperature structure from expendable bathythermograph (XBT) casts, and Richardson numbers calculated from the current and temperature structure. Dark shading, o0.25 (the region of most active overturning); light shading, o0.5. The current speed reached 1.5 m s 1. Reproduced from Cresswell GR and Peterson JL (1993) The Leeuwin Current south of Western Australia. Australian Journal of Marine Freshwater Research 44: 285–303.

(c) 2011 Elsevier Inc. All Rights Reserved.

LEEUWIN CURRENT

14

15

16

17

18

19

451

20 126

–34

–36

–38

–40

NOAA 1 1 SST 07 Jun 1994 0832Z Figure 8 An SST image showing the sea south of WA between Cape Leeuwin (left) and the western Great Australian Bight (top right). There is a 1 1  1 1 latitude–longitude grid; land is black; the shelf edge is marked by a thin black line; clouds are white; and the temperature scale (1C) is at the top. Copyright 2000, CSIRO Division of Marine Research, Hobart.

western side and the reverse on the eastern side (Figure 9). Cyclonic eddies form on the western sides of the seaward offshoots. The eddies drift westward, with the anticyclonic ones repeatedly interacting with the LC until they are west of the longitude of Cape Leeuwin. At times there can be three anticyclonic eddies between the Recherche Archipelago and Cape Leeuwin. Individual eddies have been followed using satellite altimetry for over a year from their first interaction with the LC at the Recherche Archipelago to beyond Cape Leeuwin and out into the Indian Ocean. The structure of an anticyclonic eddy south of WA is shown in Figure 10. The north and south current speed maxima were 0.5 m s 1. The eddy vertical structure suggested that it had mixed to over 300-m depth during the winter and then, with the onset of summer, had been warmed near the surface. The depression of the water structure extended at least to 1000 m. The eddy was low in oxygen and nutrients as compared with the surrounding waters. This contrasts with the anticyclonic eddies formed by the LC west of

Australia: there the formation process entrains relatively nutrient-rich waters from the continental shelf.

The Leeuwin Current on the Continental Shelf An acoustic Doppler current profiler (ADCP) moored a few meters above the bottom at the 70-m isobath out from Perth gave hourly data at 4-m depth intervals for little over an year (Figure 11). In December, there was strong northward alongshore flow due to the winddriven Capes Current that lowered the temperature at the instrument by 21C. The onset of the LC was quite sudden in late March and it was present in the lower half of the water column until the end of the record. Nearer the surface the LC flow was often weaker, perhaps due to those waters being more susceptible to northward wind forcing from passing storms. The lower shelf waters, particularly from April to August, were drained out across the topography with the offshore flow peaking at over 0.1 m s 1 in May. Whether

(c) 2011 Elsevier Inc. All Rights Reserved.

452

LEEUWIN CURRENT

20°

Sea surface temperature

18°

Salinity

35.9 35.8 35.7 35.6

16° XBT:

48

*

49

Southward

*

50

51

Northward

*

*

52

*

>0.7

–0.5 0.5

50 Depth (m)

0.0 19° 100

18° 17°

150 16° 200

25 km Figure 9 An east–west transect south of Western Australia at 1171 E in Jun. 1987 across an offshoot connecting the LC to an anticyclonic eddy. Upper panel: surface temperature and salinity. Lower panel: the north–south current components from acoustic Doppler current profiler (ADCP) measurements, the temperature structure from expendable bathythermograph (XBT) casts, and Richardson numbers calculated from the current and temperature structure. Dark shading, o0.25; light shading, o0.5. Reproduced from Cresswell GR and Peterson JL (1993) The Leeuwin Current south of Western Australia. Australian Journal of Marine Freshwater Research 44: 285–303.

Offshoot eddy at 37° S, Dec. 1994, stns. 32:42

Depth

0

>35.6

35.5

–200 35 –400 Salinity 115

116

117

Depth

0

118

119

34.5 120

>16

16 14

–200

12 10

–400 Temperature 115

116

117

118

119

120

8 0.4 0.2

–100

0 –0.2

–200

–0.4

N–S velocity –300 115

116

117

118

119

120

Longitude Figure 10 A transect of salinity (practical salinity units), temperature (1C), and north–south velocity component (m s 1) from 381 S, 1151 E to 371 S, 1201 E across an anticyclonic eddy south of WA in Dec. 1994. The velocity component plot has been limited to the 300-m depth reached by the acoustic instrument taking the measurements.

(c) 2011 Elsevier Inc. All Rights Reserved.

(a)

Monthly PVDs Nov. 00 to Nov. 01 for selected depths 800

600 Nov.

Dec.

Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov. 2001

Progression along shore (km)

400

200

0

17 m

–200

33 m

–400

49 m

–600

65 m

–800

–1000

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

Progression across shore (km) (b)

Alongshore NCEP wind stress at 32° S, 114° E

Wind stress (N m−2)

0.2

0.1

0 −0.1 −0.2 Nov.

Dec.

2000 (c)

Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

2001 ADCP temperature

23

Temperature (°C)

22 21 20 19 18 17 Nov.

Dec.

Jan.

Feb.

Mar.

Apr.

May

Jun.

Figure 11 (a) The monthly progressive vector diagrams, or inferred water trajectories, for selected depths (every fourth depth bin) from Nov. 2000 to Nov. 2001. Only data from every fourth depth bin are shown. The trajectories are arranged in columns, with successive 200-km offsets, and by row, for the different depths. The axes are across and along the bottom topography. (b, c) The alongshore wind stress near the mooring site and the temperature recorded by the near-bottom instrument moored at the 70-m isobath.

(c) 2011 Elsevier Inc. All Rights Reserved.

454

LEEUWIN CURRENT

this was due to an Ekman bottom boundary layer beneath the LC, or to near-shore waters cooling and becoming denser and moving out across the shelf, or both, is not clear. Note that the near-bottom temperaturve (Figure 11(c)) was highest in the austral autumn/winter (March/July) due to the influence of the warm LC waters from the north.

See also Current Systems in the Indian Ocean. Current Systems in the Southern Ocean. Drifters and Floats. East Australian Current. Ekman Transport and Pumping. El Nin˜o Southern Oscillation (ENSO). Energetics of Ocean Mixing. Inverse Modeling of Tracers and Nutrients. Meddies and Sub-Surface Eddies. Mesoscale Eddies. Moorings. Satellite Remote Sensing of Sea Surface Temperatures. Sea Surface Exchanges of Momentum, Heat, and Fresh Water Determined by Satellite Remote Sensing. Single Point Current Meters. Upper Ocean Mixing Processes. Upwelling Ecosystems. Water Types and Water Masses.

Further Reading Andrews JC (1977) Eddy structure and the West Australian Current. Deep-Sea Research 24: 1133--1148. Cresswell GR and Golding TJ (1980) Observations of a south-flowing current in the southeastern Indian Ocean. Deep-Sea Research 27A: 449--466. Cresswell GR and Griffin DA (2004) The Leeuwin Current, eddies and sub-Antarctic waters off south-western Australia. Marine and Freshwater Research 55: 267--276. Cresswell GR and Peterson JL (1993) The Leeuwin Current south of Western Australia. Australian Journal of Marine Freshwater Research 44: 285--303. Domingues CM, Wijffels SE, Maltrud ME, Church JA, and Tomczak M (2006) Role of eddies in cooling the Leeuwin current. Geophysical Research Letters 33: L05603 (doi:10.1029/2005GL025216). Feng M, Meyers G, Pearce A, and Wijffels S (2003) Annual and interannual variations of the Leeuwin Current at 321S. Journal of Geophysical Research 108(C11): 3355 (doi:10.1029/2002JC001763).

Fieux M, Molcard R, and Morrow R (2005) Water properties and transport of the Leeuwin Current and eddies off Western Australia. Deep-Sea Research I 52: 1617--1635. Godfrey JS and Ridgway KR (1985) The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: Longshore steric height patterns, wind stresses and geostrophic flow. Journal of Physical Oceanography 15: 481--495. Holloway PE and Nye HC (1985) Leeuwin Current and wind distributions on the southern part of the Australian North West Shelf between January 1982 and July 1983. Australian Journal of Marine and Freshwater Research 36: 123--137. Middleton JF and Platov G (2003) The mean summertime circulation along Australia’s southern shelves: A numerical study. Journal of Physical Oceanography 33: 2270--2287. Morrow R, Birol F, Griffin D, and Sudre J (2004) Divergent pathways of cyclonic and anti-cyclonic ocean eddies. Geophysical Research Letters 31: L24311 (doi:10.1029/ 2004GL020974). Morrow R, Fang FX, Fieux M, and Molcard R (2003) Anatomy of three warm-core Leeuwin Current eddies. Deep-Sea Research II 50: 2229--2243. Pearce A and Pattiaratchi C (1999) The Capes Current: A summer countercurrent flowing past Cape Leeuwin and Cape Naturaliste, Western Australia. Continental Shelf Research 19: 401--420. Pearce AF and Phillips BF (1988) ENSO events, the Leeuwin Current, and larval recruitment of the western rock lobster. Journal Du Conseil 45(1): 13--21. Ridgway KR and Condie SA (2004) The 5500-km long boundary flow off western and southern Australia. Journal of Geophysical Research 109: C04017 (doi:10.1029/2003JC001921). Schodlok MP, Tomczak M, and White N (1997) Deep sections through the South Australian Basin and across the Australian–Antarctic discordance. Geophysical Research Letters 22: 2785--2788. Smith RL, Huyer A, Godfrey JS, and Church JA (1991) The Leeuwin Current off Western Australia, 1986–1987. Journal of Physical Oceanography 21: 323--345. Waite AM, Thompson PA, Pesant S, et al. (2007) The Leeuwin Current and its eddies: An introductory overview. Deep-Sea Research II 54: 789--796.

(c) 2011 Elsevier Inc. All Rights Reserved.

LONG-TERM TRACER CHANGES F. von Blanckenburg, Universita¨t Bern, Bern, Switzerland Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1492–1503, & 2001, Elsevier Ltd.

Introduction Ocean tracers that record long-term changes preserve certain water column information within the sediment. This information comprises (1) the tracers’ fluxes in the past, such as erosional input from the continents, hydrothermal activity at mid-ocean ridges, input of extraterrestrial material, or carbonate recycling; (2) the distribution of water masses in the past and the state of the past global thermohaline circulation. Inorganic isotope tracers whose isotope ratios are modified by radioactive decay in their source materials are ideally suited for these studies. Their original water column values can be measured in materials such as biogenic carbonates, ferromanganese crusts and nodules, and the authigenic phase of deep-sea sediments. Studies of tracer fluxes in the past are favored by those tracers whose residence time in the ocean (t, defined below) is long relative to the turnover time of the thermohaline circulation (1500 y), such as Sr and Os. Tracers of which t is of the order of, or shorter than the oceans turnover time (Nd, Hf, Pb, Be) offer the ability to label water masses isotopically. In this case, long-term isotope changes of these intermediate-t tracers are potentially caused by variations of the thermohaline circulation. However, secular variations of these isotope tracers can also be caused by regional variations in these tracers’ fluxes, mostly resulting from changes in weathering. It is not always straightforward to distinguish between these two causes of tracer variations. Certainly the globally uniform seawater isotope evolution of Sr, Os, and potentially also Be, offer excellent tools for isotope stratigraphy on long (My) timescales.

Definitions and Concepts Long-term tracers are those elements whose isotopic compositions provide information on the physical and chemical state of the oceans on timescales of several thousands of years to millions of years (My). For example, paleo-oceanographers aim to

reconstruct past water mass distributions and the mode of the thermohaline circulation. For this purpose it would be desirable to reconstruct past oceanographic water mass characteristics such as salinity, temperature, silica, or phosphorus content from the sedimentary record. Similarly, the reconstruction of the past land–sea transfer of certain tracers is desirable in order to reconstruct changes in the weathering history of the continents. However, these present-day tracers are usually not conserved in the sedimentary record. Even if they were precipitated chemically and stored in sediments, their changes in concentration as measured in a sedimentary column back through time cannot be directly related to past water mass properties. This is because the tracers’ concentrations in sediments depend on factors such as sedimentation rate, diagenesis, partitioning into a certain phase, uptake by organisms, and additions of the same element by detrital hemipelagic or aeolian material. Therefore ocean chemists make use of proxy tracers which are not routinely analyzed in surveys of present-day water masses, because their measurement presents a considerable effort compared with tracers such as salinity and temperature. However, their characteristics can be directly related to those well-known oceanographic seawater tracers. The conditions that need to be met for an element to be of use as a proxy tracer are that (1) it conserves a characteristic chemical or isotopic property when transferred from the water column into the sediment; (2) the elements or their isotopic composition can be extracted from the sediment; (3) the age of the sediment is known so that changes of the tracer over time can be reconstructed. One such proxy makes use of element ratios. For example, the ratio of Cd to Ca in foraminiferal tests is a proxy for the PO4 content of the past water mass in which the foraminifera formed and therefore provides information on the past thermohaline circulation. This tracer is explained in detail in the article on trace elements in foraminiferal tests. Similarly, the ratio of the intermediate uranium decay products 231Pa and 230Th, measured in bulk sediment, may under certain conditions provide information on the advection of water masses in the overlying water column in the past (see Cosmogenic Isotopes and Uranium-Thorium Series Isotopes in Ocean Profiles). Isotope ratios are ideally suited as long-term proxy tracers. Some of these isotope ratios are characteristic of certain seawater properties and

(c) 2011 Elsevier Inc. All Rights Reserved.

455

456

LONG-TERM TRACER CHANGES

can be measured in sediments, regardless of the actual tracer partitioning, concentration, or location of precipitation. A property describing the behavior of a tracer in sea water is the residence time t. If the tracer’s fluxes in and out of an ocean basin are invariant with time, the tracer is at steady state and the residence time can be calculated from the tracer’s ocean inventory: t¼

Inventory Inventory ¼ Fluxin Fluxout

A tracer suitable as a water mass tracer has a short global residence time (t) relative to the ocean’s mixing time. This ensures that isotope ‘fingerprints’ characteristic of a certain water mass are prevented from being completely dispersed by the global thermohaline circulation. While the global ocean mixing time is difficult to assess, a meaningful quantity is the time it takes for one turnover of the global deep water circulation, which is c. 1500 y. Tracers with t of this order have the potential to preserve distinct water mass labels. It can be assumed that a conservative (i.e. nonreactive) tracer would be almost perfectly homogenized within 10 000–20 000 years. Tracers with t in excess of this period will only Table 1

record changes in the global flux of this tracer, regardless of the water mass, the location of the input, or the location of the samples taken.

Isotope Tracers Used Much use is made of the stable isotopes of carbon as a paleo-water mass isotope ‘fingerprint’. The 13C/12C ratio in the tests of foraminifera depends on the relative position of the overlying water mass within the thermohaline circulation system. However, these isotope ratios are modified during the incorporation into organisms, depend on availability of nutrients, and like the isotopes of oxygen, also depend on seawater temperature. (These tracers are dealt with in the relevant articles; please refer to the See also section.) Isotope ratios of inorganic trace metals which are the topic of this chapter are not modified when incorporated into the sediment (note that some minor isotope fractionation might occur on incorporation into the sediment, but usually such shifts are either smaller than analytical precision or they are removed by the internal correction procedures of the techniques used). The variation in isotope ratios only varies

Long-term isotope tracers currently in use

Tracer

Isotopes

Sources

Average deep-water concentration

Global deep-water residence time

Strontium (Sr)

87

Sr (stable) ’87Rb (T1/2 ¼ 48.8 Gy) 86 Sr (stable, primordial)

7.6 mg g1

2–4 My

Osmium (Os)

187

10 fg g1

8000–40 000 y

Neodymium (Nd)

Nd (stable) ’ 147Sm (T1/2 ¼ 106 Gy) 144 Nd (stable, primordial) 176 Hf (stable) ’ 177Lu (T1/2 ¼ 37.3 Gy) 177 Hf (stable, primordial)

Mostly chemical weathering of the continental crust and carbonates Hydrothermal solutions from midocean ridges Dissolution of marine carbonates Erosion of the continental crust (chemical weathering important) Leaching of abyssal peridotites Cosmic dust and spherules Erosion of the continental crust

4 pg g1

B1000–2000 y

Erosion of the continental crust

0.18 pg g1

B1000–2000 y?

Hydrothermal solutions at mid-ocean ridges Erosion of the continental crust

1 pg g1

40 y (Atlantic)

Hafnium (Hf)

Lead (Pb)

Be Beryllium

Os (stable) ’ 187Re (T1/2 ¼ 43 Gy) 188 Os (stable, primordial) 143

Pb (stable) ’ 232Th (T1/2 ¼ 14.0 Gy) 207 Pb (stable) ’ 235U (T1/2 ¼ 0.704 Gy) 206 Pb (stable) ’ 238U (T1/2 ¼ 4.47 Gy) 204 Pb (stable, primordial) 10 Be (cosmogenic, T1/2 ¼ 1.5 My) 9 Be (stable, primordial) 208

Hydrothermal solutions at mid-ocean ridges (minor) Today: industrial Pb

10

Be: atmospheric precipitation by rain 9 Be: erosion of the continental crust

(c) 2011 Elsevier Inc. All Rights Reserved.

80–200 y (Pacific)

1000 atoms/g

250 y (Atlantic)

0.25 pg g1

600 y (Pacific)

LONG-TERM TRACER CHANGES

by radioactive decay of the parent isotope of at least one isotope in the tracer’s sources or cosmogenic production. The elements currently in use for paleooceanography are given in Table 1. As apparent from the properties listed in Table 1, ocean chemists have a variety of tracers at hand, covering a range of residence times and chemical behaviors. Those tracers varying due to radioactive decay have distinct isotopic compositions in their various source materials (Table 2). This makes them particularly useful both as water mass tracers, and to reconstruct the flux from these various sources into the oceans. It may be surprising to find the cosmogenic nuclide 10Be in this list of otherwise radiogenic tracers. The reason is that Be behaves very similarly to the other tracers in that the ratio 10Be/9Be is distinct in different water masses. Given that 10Be is the only tracer of which the flux into the oceans is known, t can be calculated precisely from its water column concentration. Further, the continent-derived isotope 9Be is the only tracer of which the flux into the oceans can be calculated from the 10Be/9Be ratio. Examples of the isotopes of Nd and Be as water mass labels are shown in Figure 1a and b. The isotope variations of Nd are so small that the 143 Nd/144Nd ratio is reported normalized to a ratio typically found in chondritic meteorites (‘CHUR’): 143

eNd ¼

Nd=144 Ndsample

143 Nd=144 Nd

! 1

 104

CHUR

The salinity contours in Figure 1 define water masses, such as North Atlantic Deep Water (NADW), Table 2

Isotope ratios of source materials

Isotope ratio

Pacific mid-ocean ridges

Average upper continental crust

Cosmic dust

87

0.7028 0.125 (abyssal peridotites) 0.5132 þ 10 þ 20 0.2834 18.5 15.5 38.0

0.72 1.26

N/A 0.126

Sr/86Sr Os/188Os

187

143

Nd/144Nd

eNd 176 Hf/177Hf eNd 206 Pb/204Pb 207 Pb/204Pb 208 Pb/204Pb

0.5121  11.4  10 0.2825 19.3 15.7 39.1

N/A N/A

143

Nd=144 Ndsample

143 Nd=144 Nd

CHUR

 1

* 10

4

176 eHf ¼

Antarctic Intermediate Water (AAIW), and Antarctic Bottom Water (AABW). Note that eNd is –13.5 in NADW, and 10Be/9Be is c. 0.5  107. In the southern circumpolar water eNd is –9, and 10Be/9Be is 1  107. 10Be/9Be, and in particular eNd, mimic the shape of salinity. Incorporation of these tracers into the sediment at a given location potentially provides information on the distribution and mixing of water masses at this location back through time. The schematic global distribution of deep-water isotope ratios of all tracers discussed here is shown in Figure 4A–F. Note that the variability decreases with increasing residence time. 87Sr/86Sr is perfectly homogenized (Figure 2A). The only location worldwide at which a different Sr isotope ratio has been measured in sea water is the restricted Baltic Sea, where riverine dilution halves the open-ocean salinity and leads to a distinct 87Sr/86Sr only just detectable by modern analytical methods. 187Os/188Os, with an estimated t of 8000–40 000 y, shows only a minute difference between the Atlantic and the other oceans (Figure 2B) show clear gradients between Atlantic and Pacific deep water. This is because the Atlantic receives the highest flux of continental erosion products (aeolian dust, river particulate matter, river dissolved matter) per unit open-ocean area. Furthermore, all this material is derived from old continental crust with an isotope composition distinct from younger rocks (Table 2). Labrador Sea water, for example, receives erosion products from Archean cratons with a unique isotope composition. In contrast, the Pacific receives most of its tracer input from the surrounding volcanic arcs, which have isotope compositions different from the continental crust surrounding the Atlantic. The Indian Ocean has ratios intermediate between the Atlantic and the Pacific for all of these tracers. Whether this is due to mixing of Atlantic water masses (advected through the circumpolar current) and Pacific water (advected via the Indonesian throughflow), or due to internal sources unique to the Indian Ocean is currently not known. 10Be/9Be ratios are lower in the Atlantic because the North Atlantic receives a higher flux of terrigenous 9Be. This keeps the 9Be concentration uniform worldwide, whereas 10Be increases along the advective flow path as expected from a nutrient-type tracer.

N/A

eNd and eHf are 143Nd/144Nd and 176Hf/177Hf ratios, respectively, normalized to a chondritic value CHUR. 143Nd/144NDCHUR ¼ 0.512638; 176Hf/177HfCHUR ¼ 0.282772; (N/A)L: Not Available. eNd ¼

457

Hf=177 Hfsample

176 Hf=177 Hf

CHUR

 1

* 10

4

Materials and Methods used in Long-term Tracer Studies It is important that sedimentary materials chosen for long-term tracer studies are true chemical or biogenic precipitates formed in the water column. Contamination by terrestrial detrital material (fine clays from

(c) 2011 Elsevier Inc. All Rights Reserved.

458

LONG-TERM TRACER CHANGES Nd Station 315 61°S/62°W _14 _10 _ 6

Station 271 49°S/33°W _6 _12

Station 302 33°S/42°W _6 _12

Station 30 Station 11 36°N/62°W 52°N/47°W _12 _16 _8 _20

Station 63 7°N/40°W _15 _11

0 34

34 5

3 4 .7

5

34. 45

34.50

34 .5 5 34 .6 0

34 .3 5

34 .9 5

34. 85

0

2000

35

0 3 4 .8 .85 34

.0

0

34 .7

3000

5

Depth (m)

.2

34 .3 0

.6

1000

34 .9 5

34 3 4..5 5 60

34 .85

4000 5000 60°S

40°S

20°S

0° Latitude

(A) 10

20°N

0.8

60°N

_

Be/ 9 Be (x10 7) Station 64 Station 22XC Station 8 34°N/63°W 41°N/63°W 45°N/41°W

Station 114 24°S/38°W 0.0

40°N

1.6

0.0

0.4

0.4

0.8

0.4

0.8

0 5

34.45

.6

34.50

34 .5 5 3 4 .6

5

34 .3 5

34 .9 5

34.85

0

2000

0

0

34 .3 0 3 4 .7

35

0 3 4 .8 5 3 4 .8

.0

0

34 .7

3000

5

Depth (m)

.2

34 .9 5

34

1000

34

3 4 .9

34 3 4. 5 5 .6 0

34 .85

4000 5000 60°S

40°S

20°S

0°

20°N

40°N

60°N

Latitude

(B)

Figure 1b (A) Salinity contours of Atlantic sea water with superimposed dissolved Nd isotope compositions. Stippled line gives the typical composition of NADW (eNd ¼ 13). Note the pronounced tongue of NADW with intermediate salinities and eNd of  13 spreading south, and the tongues of AABW and AAIW with lower salinities and eNd of  9 spreading north. (Reprinted with permission from von Blanckenburg F (1999) Tracing past ocean circulation? Science 286: 1862–1863. Copyright 1999 American Association for the Advancement of Science.) (B) Salinity contours of Atlantic sea water with superimposed dissolved Be isotope compositions. Stippled line gives the typical composition of NADW (10Be/9Be ¼ 0.6  107). Note the pronounced tongue of AAIW with lower salinities and 10Be/9Be of 1  107 spreading north. (Data reproduced with permission from Ku et al., 1990; Xu, 1994; Measures et al., 1996).

aeolian or hemipelagic sources) and material affected by chemical alteration through diagenetic processes has to be avoided. The isotopes of Sr are usually measured on carbonates or barite, while those of Be, Nd, or Hf are extracted from chemical sediments, such as ferromanganese (Fe-Mn) crusts, manganese nodules, the authigenic phase of deep-sea sediments, or marine phosphorites, argillites, and glauconites.

Results Sr

Sr isotopes represent the best-studied long-term tracer, as well-dated carbonate sequences are readily available, the extraction and analysis is simple, and the long residence time ensures worldwide homogenization. Therefore, a single isotope evolution curve has

(c) 2011 Elsevier Inc. All Rights Reserved.

459

LONG-TERM TRACER CHANGES

87

86

Hf

Sr/ Sr

+8

+1

+6

+1

+5 0.70918

0.70918

+2 +3

0.70918

+3

+4

(A)

+6

(D)

187

Os/188Os

206

Pb/

204

Pb

18.8 18.7

1.06

19.2 18

.8

18.6

19

1.03

1.03

.1

18.8

1.03 18

.9

19.

0

18.8

18 .8

(E)

(B) Nd

10

_ _ 8 13

x10 _7

_7

_7

10

_7

10

_7

0.9x10

8x

1x10

0.

_8

5x

_ 12 _9

1.3

_ 13

_ 6

_7

0.5x10

3x 10 _ 7

0.

_9

1.

_7

_ 4

10

_ 6

1x

_ 5

_20

Be/ 9Be

_1 0

_7

1x10

(C)

(F)

Figure 2A–F (A) Map of modern 87Sr/86Sr ratios in sea water (relative to 0.710248 for the Sr isotope standard SRM 987 (McArthur, 1994). Note that the long t of Sr (several My) allows for perfect homogenization and uniform isotope ratios in all basins. (B) 187 Os/188Os isotope ratios in modern sea water as measured on the surface of hydrogenous Fe-Mn crusts (Burton et al., 1999b). t of a few tens of thousands of years just allows for a small difference between the North Atlantic (with its rich input of old continental material) and the other oceans. (C) Dissolved Nd isotope compositions in modern deep sea water as measured in Mn nodules (reproduced with permission from Albare`de and Goldstein, 1992), adjusted for more recent measurements of Nd in deep sea water (see references in von Blanckenburg, 1999). A t of c. 1000 y allows for distinct gradients between basins, depending on the age and Sm/Nd ratio of their surrounding continental erosion sources. (D) eHf measured in the surface layer of hydrogenetic Fe-Mn crusts and nodules (Albare`de et al., 1998). Similar to Nd, distinct gradients exist between basins. (E) Pre-anthropogenic 206Pb/204Pb ratios measured in the surface layer of hydrogenetic Fe-Mn crusts and nodules. (Abouchami and Goldstein, 1995. Reprinted from Geochimica et Cosmochimica Acta, 60, von Blanckenburg F, O’Nions RK, Hein JR, Distribution and sources of pre-anthropogenic lead isotopes in deep ocean water as derived from Fe-Mn crust, 4957–4963, Copyright (1996), with permission from Elsevier Science.). Pre-anthropogenic Pb cannot be measured in modern sea water which is dominated by industrial Pb. The short t (40–200 y.) results in distinct signatures between and within basins. (F) 10Be/9Be ratios in modern sea water from both the dissolved phase and the surface layer of hydrogenetic Fe-Mn crusts. (Reprinted from Earth and Planetary Letters, 141, von Blanckenburg F, O’Nions RK, Belshaw NS, Gibb A, Hein JR, Global distribution of Beryllium isotopes in deep ocean water as derived from Fe-Mn crusts, 213–226, Copyright (1996) with permission from Elsevier Science.) 10Be is a cosmogenic nuclide that enters the ocean by precipitation from the atmosphere, whereas 9Be is stable and enters the oceans by erosion. The global deep-water t of 10Be is 600 y, allowing for preservation of distinct gradients between the basins.

(c) 2011 Elsevier Inc. All Rights Reserved.

460

LONG-TERM TRACER CHANGES

emerged dating back through the entire Phanerozoic (Figure 3), and at a much higher resolution for the Cenozoic (Figure 4). The curve was time-calibrated using biostratigraphy and, in some cases, geochron-

ology of intercalated ash layers. This makes Sr isotopes a very useful tool for stratigraphy. Ages of high confidence can be determined for those periods where 87 Sr/86Sr underwent strong changes. For some periods

87Sr/ 86Sr

0.710

0.708

0.706

Q

Tertiary

Cretaceous Jurassic

0

Triassic

Permian Carboniferous Devonian Silurian Ordovician

200

Cambrian

400

600

Age (Ma) Figure 3 87Sr/86Sr variations for the Phanerozoicum based on samples of brachiopods, belemnites, conodonts, foraminifera, and samples of micritic matrix. (Adapted from Chemical Geology, 161, Veizer J, Ala D, Azmy K et al., 87Sr/86Sr, d13C, d13O evolution of Phanerozoic sea water, 59–88, Copyright (1999), with permission from Elsevier Science.)

0.7092

87

Sr/86Sr

0.7088

0.7084

0.7080

0.7076 0

10

20

30

40

50

60

70

Age (Ma) Figure 4 Cenozoic evolution of marine 87Sr/86Sr, based primarily on analyses of foraminifera from all oceans. Data sources are as in McArthur (1994).

(c) 2011 Elsevier Inc. All Rights Reserved.

LONG-TERM TRACER CHANGES

unique ages cannot be assigned. For example, the slope in 87Sr/86Sr in the Early Tertiary period is too low to allow for high-resolution stratigraphy. The changes in the 87Sr/86Sr ratio are controlled by several processes. These are (1) the mid-ocean ridge flux, which is in turn controlled by the spreading rates of the seafloor; (2) the rate of chemical weathering, in particular that of feldspar and calcite; (3) the areal extent of the continents above sea level; (4) changes in the carbonate compensation depth (CCD). Seawater Sr isotope ratios vary within a small range over time. This is because of the long and efficient mixing of Sr, and also because dissolution of continental carbonate buffers seawater Sr isotope compositions within a narrow range. Some of the trends visible in Figure 3 are a relatively slow decrease in 87Sr/86Sr ratios from the Cambrian to the Jurassic period, upon which are superimposed a number of relatively large fluctuations. The sharp decline and following rise at the Permian/Triassic boundary are spectacular, and are thought to reflect either an extreme climate and weathering change, or the sudden mixing of a

461

previously stratified ocean. A second main trend is a relatively rapid increase in 87Sr/86Sr ratios from the Jurassic to the present, upon which a number of relatively small fluctuations are superimposed. Much discussion has been stimulated by the strong Cenozoic increase in 87Sr/86Sr (Figure 4) that has been linked by some workers to the uplift of the Himalayas and the ensuing delivery of high 87Sr/86Sr by Himalayan rivers. This view was challenged by the recent observation that 187Os/188Os (Figure 5), showing a similar and simultaneous increase, cannot be attributed to the dissolved flux draining the rising Himalayas. Therefore there must be a different cause for the rise in Sr too, possibly a worldwide increase in weathering rate. Similar attention was focused on the pronounced rise over the past 2.5 My. One possibility is that the latter can be explained simply by changes in sea level during the glaciations. However, calculations have shown that sea level variations of 200–300 m would be required – far in excess of the c. 100–150 m of change believed to have taken place during the Quarternary period. Therefore a much more plausible explanation is a change in the

1.0 0.9

0.8

0.6

187

Os/

188

Os

0.7

0.5

0.4

0.3 0.2 0.1 0

10

20

30

40

50

60

70

80

Age (Ma) Figure 5 Marine 187Os/188Os record for the past 80 My in all oceans from H2O2-leached metalliferous and hydrogenetic sediments. Note that the pronounced excursion to low ratios at the K/T boundary (65 My) is explained by a meteorite impact. (Reprinted from Geochimica et Cosmochimica Acta, 63, Pegram WJ, Turekian KK. The osmium isotopic composition change of Cenozoic sea water as inferred from a deep-sea core corrected for meteoritic contributions, 4053–4088, Copyright (1999), with permission from Elsevier Science.)

(c) 2011 Elsevier Inc. All Rights Reserved.

462

LONG-TERM TRACER CHANGES

terrigenous Sr input supplied by rivers to the oceans. Either a change in the dissolved Sr flux may have occurred or a change in the isotopic composition of rivers, or a combination of both. The increased erosion and availability of weatherable mineral surfaces resulting from the build up of glaciers during the northern hemisphere glaciation might have provided these changes. Os

Because of the exceedingly small differences in Os/188Os between the different ocean basins, Os has a potential similar to Sr to serve as an isotope

187

stratigraphic tool. It is particularly valuable for carbonate-poor pelagic clays and metalliferous sediments, from which Os is extracted by leaching techniques. The Os isotope curve (Figure 5) bears many similarities to the Sr isotope curve (Figure 4), with the exception of three excursions to low 187 Os/188Os ratios. The spectacular drop at the Cretaceous/Tertiary boundary is probably of meteorite impact origin. The second, mid-Paleocene decrease and also the slow recovery following the impact can be explained by exposure of coastal sediments imprinted by the meteoritic Os from the K/ T boundary. The Eocene-Oligocene excursion (B33 Ma) has been explained by an increased supply of

_3 Pacific

_5

_7

Nd (+)

Indian

_9 NW Atlantic

_ 11

_ 13 0

10

20

30

40

50

60

Age (Ma) Figure 6 Nd isotope variations in Cenozoic sea water based on the analyses of hydrogenetic Fe-Mn crusts. Note that despite the assumed t of more than 1000 y the oceans have kept their Nd isotope provinciality observed today throughout the last 50 My. (Reprinted from Earth and Planetary Science Letters, I55, O’Nions RK, Frank M, von Blanckenburg F, Ling HF. Secular variations of Nd and Pb isotopes in ferromanganese crusts from the Atlantic, Indian, and Pacific Oceans, 15–28, Copyright (1998) with permission from Elsevier Science.)

(c) 2011 Elsevier Inc. All Rights Reserved.

LONG-TERM TRACER CHANGES

nonradiogenic Os from peridotite weathering (Table 2). As is the case for 87Sr/86Sr, there is a strong increase of 187Os/188Os towards more radiogenic values over the past B14 My. As stated above, this has been linked to the rise of the Himalayas, but recent analyses of Himalayan river waters for Os isotope compositions do not support this view. A more likely possibility is the weathering of ancient crystalline terranes exposed by physical erosion, or the weathering of black shales. These organic-rich sediments have a high Re/Os ratio. Therefore old black shales have the potential to supply Os with a very high 187 Os/188Os ratio to sea water.

19.0

back as 55 Ma. It is thought that despite the large isotopic variability in source materials, the t of Nd is sufficiently long to allow for efficient intra-basin homogenization. This produces the basins’ characteristic Nd isotope blend. Significant variations in eNd are mainly observed for the past 5 My. In the Pacific a decrease in eNd over the past 5 My might be due to an increased flow of AABW (with low eNd, Figure 2C), a rearrangement of the thermohaline circulation following the opening of the Indonesian throughways for exchange of thermocline waters, or an increase in dust input. The strong decrease in north-west Atlantic eNd over the past 3–4 My has been linked to a strengthening of NADW production following closure of the Panama gateway (suppressing northward flow of AABW and AAIW high in eNd ). However, a pronounced decrease of eNd in a shallow ferromanganese crust off Florida has ocurred as early as 8–5 Ma. This has been ascribed to a decreasing inflow of Pacific water through the narrowing Panama gateway. Thus, a change in the amount and style of weathering associated with the onset of northern hemisphere glaciation at 3 Ma is a more likely explanation for the Pleistocene decrease in eNd . In particular the Labrador Sea, a major source of NADW, is surrounded by ancient rocks with eNd as low as  40 and is supplying deep water with eNd of  20 to NADW (Figure 2C). An increase in weathering of this component has the potential to drive the Nd in NADW towards lower compositions.

18.9

Pb

Nd

Nd isotopes, analyzed with low-time resolution in Fe-Mn crusts and given as eNd units, are presented in Figure 6. The most outstanding feature is that the provinciality, observed in eNd of the modern oceans (Figure 2C), has been a feature prevailing as far

NW Atlantic Indian

19.2

Pacific

Pb/ 204Pb

19.1

206

463

18.8

18.7

18.6

18.5 0

10

20

30

40

50

60

Age (My) Figure 7 Pb isotope variations in Cenozoic sea water based on the analyses of hydrogenetic Fe-Mn crusts. Because of the short t of Pb the oceans have maintained distinct isotope signals throughout the past 50 My. There is more intra-basin variability with time because the short t does not allow such efficient lateral homogenization within basins as is the case for Nd. Therefore, local changes in erosion are much more visible in Pb isotope variations. (Reprinted from Geochimica et Cosmochimica Acta, 63, Frank M, O’Nions RK, Hein JR, Banaker VK, 1689–1708, Copyright (1999) with permission from Elsevier Science.)

No information can be obtained on natural Pb from modern sea water, because of the strong contamination by industrial Pb. Therefore, the pre-anthropogenic Pb distribution has to be obtained from chemical sediments. 206Pb/204Pb time-series, analyzed in Fe-Mn crusts (Figure 7), show patterns of changes that are less clear than those of Nd. Relative differences even within ocean basins are much larger than those observed for Nd. This may be expected from the short residence time of Pb (Table 1), which does not allow for lateral within-basin homogenization to the same degree as Nd. Therefore, local sources dominate the natural Pb budget, and their changes in flux introduce strong isotope variability. For example, in the Indian Ocean a crust located close to the circumpolar current shows a distinctly different history from the more northerly one, experiencing strong changes. The pronounced increase in north-west Atlantic 206Pb/204Pb can be attributed, as eNd , to a change in NADW production, but is more likely due to a change in weathering of the

(c) 2011 Elsevier Inc. All Rights Reserved.

464

LONG-TERM TRACER CHANGES

_7

3.0×10 Pacific Fe _ Mn crusts

_7

Initial

Pacific10Be/9Be

_7

10Be/9Be Meas

2.0×10

10

_8 _7

10Be/9Be Init

10

1.0×10

_9

10

0.0 0

10 20 Depth (mm)

(A)

0

30

2

(B)

4

6

8

10

12

Age (My) _7

_7

10

Initial Pacific10Be/9Be

_7

2.0×10 _8

10

_7

10Be/9Be Init

10Be/9Be Meas

3.0×10

NW Atlantic Fe _ M n crusts

1.0×10 _9

10

C

d

0.0 0 (C)

5

10 Depth (mm)

15

20

1 (D)

3 5 Age (My)

7

Figure 8 10Be/9Be ratios in Pacific (A, B) and Atlantic (C, D) Fe-Mn crusts (Ling et al., 1997; von Blanckenburg and O’Nions, 1999). The smooth exponential decrease with depth usually observed in the measured data (as fitted following the stippled lines in A, C, where some individual outliers are attributed to alteration) is compatible with radioactive decay of 10Be only (T1/2 ¼ 1.5 My). A kink in the fits to the data is due to changes in growth rates. Changes in the initial ratio would be visible in the form of offsets in the data. At the time resolution of the samples taken (several hundred thousand years) these are not observed. Therefore, the derived growth rates can be used to time-correct the 10Be/9Be ratios for radioactive decay and to calculate initial ratios. The results shown in B and D indicate that the 10Be/9Be ratios in both the Pacific and the Atlantic have been within the range of modern sea water, for the last 7–10 My, despite presumably widely varying erosional input into the ocean.

glaciated areas or a change in the provenance of the erosional products. Be 10

Be/9Be has been analyzed in Fe-Mn crusts as chronometer for the past 10 My. The smooth logarithmic decrease in 10Be/9Be (Figure 8A and C) is compatible with radioactive decay of 10Be. This allows for calculation of growth rates and reconstruction of the initial 10Be/9Be ratios. Quite unexpectedly these initial 10Be/9Be ratios have been within the present-day range of the respective ocean basins through the past 7–10 My. Since the modern ocean basins do display strongly different 10Be/9Be

ratios (Figure 2F), and since a t of c. 600 y favors exchange of Be between basins, the relative constancy of these ratios with time suggests that large-scale changes in deep-water circulation have not taken place within the measured period.

Discussion and Conclusion Changes in the style of weathering have the potential to change the isotope composition of tracers released to the sea. Leaching experiments on fresh mechanically (glacially) weathered rocks have shown a strongly incongruent release (meaning the release of tracers with varying isotope compositions depending

(c) 2011 Elsevier Inc. All Rights Reserved.

LONG-TERM TRACER CHANGES

Table 3

465

Incongruent release of isotopes from strongly mechanically weathered continental rocks

Isotope ratio

Isotope change

Original fresh material

Experimental evidence

87

0.725-0.795

Young granitic moraine, Wind River Range, Wyoming, USA

1.5-9.5

Young granitic moraine, Wind River Range, Wyoming, USA

Ammonium acetate leach River water composition HCl leach (Blum and Erel, 1995) HCl leach

Sr/86Sr

187

Os/188Os

eNd

 26-  42

206

15.2-22.0

Pb/204Pb

Greenland river bedload Baffin Bay deep-sea sediment Greenland river bedload Baffin Bay deep-sea sediment

on the mode of liberation from rocks) of radiogenic Sr, Os, and Pb, and nonradiogenic Nd (up to 16 eNd units, Table 3). Strongly chemically weathered rocks do not show such an incongruent release. Therefore, some of the long-term changes seen in the isotope evolution of seawater Sr, Os, Nd, and Pb might be attributable to these effects. Certainly the evolution of Sr and Os is only controlled by weathering, and the relative contribution of various sources, such as continental rocks, MORB (Mid Ocean Ridge Basalt), carbonate recycling, peridotite weathering, and cosmic dust. The uniformity between basins makes Sr and Os isotopes reliable stratigraphic tools. 10 Be/9Be appears to be a robust dating tool for the past B10 My, due to its relatively constant observed initial ratio. It remains to be demonstrated whether Nd, Hf, and Pb have real value as tracers of past ocean circulation, as expected from their distinct water mass compositions, or whether their changes back through time merely record local changes in weathering. Clearly the time resolution achievable from Fe-Mn crust studies is not sufficient to answer these questions. Much more insight will be obtained when these radiogenic tracers and Be isotopes are applied to sediments allowing tracer change studies at the resolution of a few thousand years. This will allow more reliable studies on the relationship between climate change, ocean circulation, and continental weathering.

Suggested Reading The topic of radiogenic seawater tracers is too novel to be covered by a single monograph. All information is spread between numerous publications in international journals. Faure (1986) gives a general introduction into radiogenic isotope techniques.

Young terrestrial Fe-Mn coatings Peucker-Ehrenbrink and Blum (1998) HCl leach (own work) HCl and HBr leach (own work)

Analytical methods are summarized in a monograph by Potts (1987). Broecker and Peng (1982) provide a much-cited introduction into the topic of tracers in the sea. Ferromanganese crusts have recently been summarised by Hein et al. (1999). McArthur (1994) has reviewed the material suitable for Sr isotope analysis in carbonates, covering all ages of deposits from recent to the Precambrian. A summary of the suitability of marine clay minerals for isotope analyses is given by Stille et al. (1992). A brief summary on radiogenic seawater tracers, containing useful cross-references, has been published by the author (von Blanckenburg, 1999).

See also Authigenic Deposits. Carbon Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Cenozoic Oceans – Carbon Cycle Models. Cosmogenic Isotopes. Large Marine Ecosystems. Mid-Ocean Ridge Geochemistry and Petrology. Rare Earth Elements and their Isotopes in the Ocean. Redfield Ratio. River Inputs. Sediment Chronologies. Stable Carbon Isotope Variations in the Ocean. Uranium-Thorium Series Isotopes in Ocean Profiles. Water Types and Water Masses.

Further Reading Broecker WS and Peng TH (1982) Tracers in the Sea. Palisades: Lamont-Doherty Geological Observatory. Faure G (1986) Principles of Isotope Geology. John Wiley & Sons. Hein JR, Koschinsky A, Bau M, Manheim FT, Kang JK, and Roberts L (1999) Cobalt-rich ferromanganese crusts in the Pacific. In: Cronan DS (ed.) Handbook of Marine Mineral Deposits, pp. 239--279. Boca Raton: CRC Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

466

LONG-TERM TRACER CHANGES

McArthur JM (1994) Recent trends in strontium isotope stratigraphy. Terra Nova 6: 331--358. Potts PJ (1987) A Handbook of Silicate Rock Analysis. Blackie. Stille P, Chaudhuri S, Kharaka YK, and Clauer N (1992) Neodymium, strontium, oxygen and hydrogen isotope

compositions of waters in present and past oceans: a review. In: Clauer N and Chaudhuri S (eds.) Isotopic Signatures and Sedimentary Rocks, p. 555. Berlin: Springer Verlag. von Blanckenburg F (1999) Tracing past ocean circulation? Science 286: 1862--1863.

(c) 2011 Elsevier Inc. All Rights Reserved.

MACROBENTHOS J. D. Gage, Scottish Association for Marine Science, Oban, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1505–1515, & 2001, Elsevier Ltd.

Introduction The macrobenthos is a size-based category that is the most taxonomically diverse section of the benthos. Only in shallow water does the macrobenthos include both plants and animals. Here, attached macrophytes (including various algae and green vascular plants) may make up a large part of the benthic biomass in coastal areas. Meadows of sea grass (which root into and stabilize coarser sediments) and forests of macroalgae (usually attached to hard bottoms) provide habitat for smaller plant and animal species. In warm, shallow water, stony corals, which flourish as a consequence of symbiotic algae living in their tissues, overgrow large areas and these reefs provide a biogenic habitat for a wealth of other species, and their broken-down skeletons provide much of the sediment in adjacent areas. But the importance of such areas declines rapidly with declining potential for photosynthesis as light quickly vanishes with increasing depth, and the macrobenthos becomes the exclusive domain of heterotrophic life (fueled by breakdown of complex organic material) in soft sediments. Only at deep-sea hydrothermal vents is there an exception to this. These support lush concentrations of benthic biomass that relies not on photosynthetic production ultimately derived from the surface but on the activity of chemoautotrophic bacteria exploiting the emissions of reduced sulfur-containing inorganic compounds. The animal macrobenthos may be attached or may be able to move over hard surfaces provided by exposed bedrock, or may use as habitat the much larger and quantitatively important areas covered by soft sediment. Areas of rock, exposed as a consequence of water currents and turbulence (or, in the case of the newly formed sea floor at the spreading centers along the mid-ocean ridges, rock that has not had time to become covered in sediment settling from above) provide habitat for epibenthos, or epifauna. Even if epibenthos, both plant and animal, looks conspicuous between the tides (and is certainly important

with respect to fouling of colonizers of submerged hard surfaces made by man, such as ships and jetties), it is only a vanishingly small proportion of the huge area of the benthic habitat covering more than half the globe that is not covered by soft sediment. The term infauna has been used to categorize the organisms inhabiting soft sediment. But many epifauna, such as sea stars, are motile and can forage over the surface of sediments. The activities of the animals of the so-called infauna are usually focused on the sediment–water interface where their detrital food is concentrated, but this should not be taken to imply that sediment fauna is always burrowed out of sight. None the less, some species, particularly among larger crustacea, are capable of burrowing even a meter or more deep into the sediment. There are also many species closely related to epifaunal groups of hard substrata, such as sponges and Cnidaria (including sea pens, soft and stony corals and sea anemones), that anchor into the sediment for a sedentary life style, catching small particles from the bed flow.

Global Pattern in Macrobenthic Biomass Extensive Russian sampling after World War II established the precipitous decline in benthic biomass with increasing depth into the abyss. This is caused largely by mid-water consumption of particles escaping from the euphotic zone. Globally, the amount leaving the euphotic zone should be equivalent to the so-called ‘new production’ of roughly 3.4–4.7  109 tonnes C y 1, about 10% of total surface primary production. But this is distributed very unevenly. Perhaps 25–60% is exported in shallow seas, while in the deep ocean only 1–10% reaches the bottom; this is also influenced by latitude-related differences in depth of the mixed layer and intensity of seasonality in the upper ocean. Figure 1 shows how influences such as upwelling and inshore surface productivity will also affect local values of benthic biomass. Overall, these range from highs of more than 500 g m 2 in shallow, productive waters just tens of meters deep, to less than 0.05 g m 2 (equivalent to 2 mg C m2) on the abyssal plains. Trenches are deeper still but, by acting as sumps for material washed in from nearby island arcs and land mass, can support higher than expected biomass.

(c) 2011 Elsevier Inc. All Rights Reserved.

467

468

MACROBENTHOS

Upw elli

ng

Euphotic zone Intertidal zone

Co

ast

al

Oce

Biomass

anic

Euphotic Coast shelf

Aphotic Abyss

Slope

The lower size limit of macrobenthos was determined by Mare as that part retained in a 1 mm sieve, but later became reduced downwards to just 0.5 mm as it was realized not only that small, juvenile sizes were being lost in numbers but also that smaller species of groups already being sampled and taken as part of the macrofauna were not always retained adequately. Mare recognized that the limit might depend on habitat and deep-sea benthic biologists found they had to use even finer-meshed screens to collect the same sorts of animals that characteristically make up the macrobenthos in shallow waters. In the 1970s, Hessler found that he had to use a 297 mm mesh sieve to catch sufficient animal macrobenthos for study in box cores from the abyssal central North Pacific (a very oligotrophic area – nutrient-poor and therefore thin in plankton). Sieves with meshes of just 250 mm are now standard in recent large European studies on the deep-sea macrobenthos.

Depth Figure 1 Conceptual model of food availability and benthic biomass in relation to depth. Upwelling areas provide nutrients for enhanced coastal productivity. The ‘coastal’ curve refers to shelf areas supporting high productivity (usually wider shelves with land inputs, e.g., rivers) compared to the ‘oceanic’ curve where oceanic effects prevail (usually narrow shelves with little land input). The mismatch in the intertidal between biomass and high food availability is explained by the co-occurrence with the latter of high hydrodynamic disturbance by waves and currents. (Modified from Pearson and Rosenberg (1987).)

History and Size Limits of Macrobenthos The term macrobenthos dates from the early 1940s when Molly Mare published a study of an area of coastal soft sediment off Plymouth, England. In recognizing that the benthic ecosystem is fueled by a detrital rain of particles derived from photosynthetic production by macrophytes or phytoplankton, she identified the potential importance of the smaller size classes of metazoan and single-celled organism, right down to bacteria, in the decomposition cycle and food chains in the sediment. She differentiated the benthos into several subcategories based on size, or biomass. Before this time a distinct category for macrobenthos was unnecessary because the only part of the benthos generally thought to be worth studying was the animal life large enough to be eaten by fish. We now differentiate these from the very small metazoans and single-celled algae that Mare named meiobenthos. Another category is the hyperbenthos – small metazoan life that can swim off the bottom and form a distinct community in the benthic boundary layer.

Sources of Food and Feeding Types Patterns in feeding of macrobenthos have often been used to distinguish ecological zones. Although exact definition of feeding category for individual organisms has been controversial, the simplest classification is into suspension and deposit feeders, carnivores, and herbivores. More detailed categorization has proved difficult because of overlap and behavioral flexibility. Although most macrobenthos feed on detrital particles settling from the water column, such as feces, molts, and dead bodies of plankton, this passive sinking is augmented by currents that may resuspend particles periodically from the bottom. Macrobenthos may gather these particles either by catching them from bottom flow or by ingesting the sediment itself as deposit feeders, either in bulk or more selectively for the most nutritious particles. Where currents vary periodically, some animals can feed on both suspended and settled particles by simply changing the way they use their feeding appendages. Just as the particles caught by suspension feeders may range from inert floating detritus up to small swimming organisms, deposit feeding shades into predation where the particles encountered include smaller living benthos. Whether macrofauna can utilize dissolved organic matter in the sediment porewaters to any great extent is still unclear. Wildish has provided the most satisfactory classification of macrofaunal feeding types related to environment. This keeps all three categories but separates deposit feeders into surface and burrowing deposit feeders. Each of the five groups is subdivided

(c) 2011 Elsevier Inc. All Rights Reserved.

MACROBENTHOS

in terms of motility and also in terms of food-gathering technique, such as use of jaws and particleentangling structures. These may be arranged along an environmental gradient, such as that illustrated in Figure 2, to allow insight into the causal basis of previously described composition of macrofaunal communities. The relation of feeding to small body size in deepsea macrobenthos may be important. Thiel thought that small body size is a result of a balance between limited food and metabolic rate that makes larger size more efficient than smaller, and of the effects of small population size on reproductive success. Being small allows organisms to maintain higher population densities that increase the chance of encountering the opposite sex and so of reproducing and maintaining the population. The extent of faunal miniaturization is still debated and, surprisingly, not readily summarized by simply taking the total bulk of

the sample and dividing by the number of animals present. The exceptions seem be those organisms that have overcome the reproductive problem by being highly motile scavengers and that need also to be large enough to allow them to forage for the large food falls that occur very sporadically on the deep ocean bed. This scavenger community is quite well developed and includes close relatives of typical macrofaunal organisms in shallow water that in the deep sea grow to a relatively enormous size (Figure 3).

Size Spectra If the sizes of all individuals from an area of sediment are measured and plotted as frequencies along a logarithmic size axis, a pattern of peaks shows up corresponding to the micro-, meio- and macrobenthic size classes (Figure 4). This supports practical intuition but does not explain why such peaks occur (no

50 % 0

HMJ Herbivores HDJ

469

50

FSC FSX

Suspension feeders

FST FMC

FDC SDT

Surface deposit feeders

SDC SMX SST BSX BDX

Burrowers

BMX BDT CDJ

Carnivores Shallow vegetation a

b c

Shallow Offshore Offshore Offshore Offshore sand sandyrock mud gravel mud f g h e d g k l n p r s u i m o t j

CMJ Deep mud

v

x

w

Figure 2 Distribution of functional groups in boreal coastal macrobenthos (compiled from genera listed by N. S. Jones for the Irish Sea) along an environmental gradient of decreasing food availability and water turbulence and increasing depth and sedimentation. Functional groups: H, herbivore; F, suspension feeder; S, surface deposit feeder; B, burrowing deposit feeder; C, carnivore. Motility: M, motile; D, semi-motile; S, sessile. Feeding habit: J, jawed; C, ciliary mechanisms; T, tentaculate; X, other types. In the upper panel, width of line representing each functional group along the gradient indicates proportional composition at that depth. The lower panel gives a diagrammatic representation of typical feeding position of taxa representative of various groups relative to the sediment–water interface along each gradient. Key to taxa: (a) macroalgae; (b) sea urchins, e.g., Echinus (HMJ); (c) limpets, e.g., Patella (HDL); (d) Barnacles, e.g., Balanus (FSX); (e), (f) serpulids, sabellids (FST); (g) epifaunal bivalves, e.g., Mytilus (FSC); (h) brittle stars, e.g., Ophiothrix (FMC); (i) Venus (FSX); (j) Mya (FSC); (k) Cardium (FDC); (l) Tellinba (FDT); (m) Turritella (SMX); (n) Lanice (SST); (o) Abra (SDC); (p) Spio (SST); (r) Amphiura (FDT); (s) Echinocardium (BMX); (t) Ampharete (SST); (u) Maldane (BSX); (v) Glycera (CDJ); (w) Thyasira (BDX); (x) Amphiura (SDT). From Pearson and Rosenberg (1987).

(c) 2011 Elsevier Inc. All Rights Reserved.

470

MACROBENTHOS

1 cm Figure 3 Gigantism in a scavenging amphipod (family Lysianassidae), the cosmopolitan deep-sea species Eurythenes gryllus, compared to the size of a typical shallow-water lysiannasid, Orchomene nana (bottom left), a northern European shallow-water species. (Redrawn from Gage and Tyler (1991) and Hayward PJ and Ryland JS (1995) Handbook of the Marine Fauna of North-West Europe. Oxford: Oxford University Press.)

3

_2

Log10 [biomass (mm m )]

6

4

2

0 _2

1

10

2

10

4

Equivalent spherical diameter (µm)

Figure 4 Macrobenthic size spectra measured from an intertidal inlet in Nova Scotia, subtidal Bay of Fundy, and abyssal sediment from the Nares Abyssal Plain south of Bermuda. The median lines (dotted for intertidal and inshore, and double-dashed for abyssal plain) and range (continuous or solid lines) show a coherent pattern with biomass peaks at 1256 and 8192 mm equivalent spherical diameter (ESD). Downwardpointing arrows indicate minimum detectable biomass. Abundance in the bacterial (leftmost) and meiofaunal (middle) peaks averages 5  103 mm3 m 2. The macrofaunal biomass peak is an order of magnitude higher, but shows greater variability. Biomass in the troughs is about 2–3 orders of magnitude less than adjacent peaks. Sediment was a fluid siltclay. (From Schwinghamer P (1985) Observations on sizestructure and pelagic coupling of some shelf and abyssal benthic communities. In: Gibbs PE (ed) Proceedings of the Nineteenth European Marine Biology Symposium, Plymouth, Devon, U.K. 16–21 September 1984, pp. 347–359. Cambridge: Cambridge University Press.)

to microbenthos and bacteria each particle is a little world on which to attach and grow. Warwick provided a complementary explanation that the peaks also reflect size-related adaptation in the way the life history of the organism is optimized to its environment. For example, larvae of macrofauna exploit the trough between macro- and meiofauna to escape from meiofaunal predators, and thereafter quickly grow into the size range of the ‘macrofaunal’ peak. This is generally lower and less defined than the meiofaunal peak, where organism longevity is just a few weeks at most and there is therefore a narrow range in size, while individual macrofauna might grow over several years so that population size distributions are wider. However, subsequent studies have not found that clear peaks in size spectra occur everywhere. In the deep sea, body size miniaturization does not destroy this pattern, even if the trough at 512–1024 mm between meio- and macrofauna may be less than in coastal sediment (Figure 4). It seems more likely that low food supply has become more important than anything else, so that macrofauna, although settling at roughly the same size, simply do not grow anything like as large as similar coastal species, rather than their somehow perceiving the sediment environment differently from typical macrofaunal organisms in shallow water. We cannot therefore reject the idea that size-based differentiation of the benthos occurs; but is it sensible to stick rigidly to the strict size-based divisions that define the macrofauna as only those organisms within a given range of size, or is it better to compare like with like on the basis of higher taxa determining limits rather than size? With the former definition, the lower limit of the macrobenthos will be determined by size at 1.0 or 0.5 mm, even if this excludes smaller specimens belonging to the same higher taxon, or even much lower-level taxa. This assumes that a size-based functional distinction operates that for the purposes of the study (perhaps environmental impact assessment) will be more important in determining variability than taxonomic affinity. The former function-based definition has been referred to as macrofauna sensu stricto, while the latter, taxonomic one as macrofauna sensu lato.

Composition and Succession such peaks occur, for example, in pelagic communities). Schwinghamer thought that these peaks reflect the way the organism perceives its sediment environment: macrofauna as a continuous medium on, or in, which to move and burrow; meiofauna as a series of interstices between sediment particles; while

Macrobenthos characteristically includes a huge range of phyla (the major divisions of the animal kingdom). In fact, most higher-level taxa are marine and benthic, with most of these part of the macrobenthos. Of the 35 or so known phyla (the major divisions of the animal kingdom), 22 are exclusively

(c) 2011 Elsevier Inc. All Rights Reserved.

MACROBENTHOS

marine, with 11 restricted to the benthic environment. Virtually every known phylum is represented in the macrobenthos except for one, the Chaetognatha, or arrow worms (arguably found only in the plankton, although one bottom-living genus is known). This contrasts with the land (including freshwater environments) where only 12 phyla are found (there is only one small, obscure phylum of worm-like animals, the Onychophora, known only on land). This reflects the marine origins of life and the much shorter time of occupation for life on land (barely 400 million years), compared with 800 million years since metazoan organisms first appeared in the ancient ocean. Only five metazoan phyla are normally regarded as part of the next size group down, the meiofauna. The proportional representation of major taxa is conservative, and seems to vary little worldwide with depth, latitude, or productivity regime. It is only in stressed soft-sediment environments, such as those with high organic loading and depleted oxygen (often occurring together), where major departures to this pattern are found (Figure 5). Inshore studies on effects of pollution have contributed to a concept whereby such stress leads to a modified macrobenthos with fewer species, and these characterized by opportunist forms, mainly polychaetes. In tracing recovery after pollution events, it is not clear to what extent a predictable succession occurs. The modern consensus is that there is a random component imposed on a facultative succession in which ‘opportunist’ species pioneer colonization and bring about amelioration in sediment conditions. This allows a more diverse set of species that are more highly tuned to particular habitats to become established through progressively deeper and more extensive bioturbation (Figure 6).

How Many Macrobenthic Species are There? Up to a few years ago it was thought that of the 1.4 to 1.8 million or so species recorded on earth there are perhaps only 160 000 or so known marine species, about 10% of the total. A large-scale sampling programme in deep water off the eastern United States has thrown this into doubt. Along a 180 km section of the continental slope at about 2000 m depth, Grassle and Maciolek found 58% of the species – especially among polychaete (bristle) worms and peracarids (small, sandhopper sized crustaceans) – new to science. The curve of the accumulation of species plotted against increasing area sampled showed no sign of tailing off; the steady increment of new (but rare) species encouraging an

471

extrapolation that this will apply over the wider area of the deep ocean. Depths below the shelf edge cover about 90% of the domain of macrobenthic infauna. But an area of only about 0.5 km2 out of the almost 335  106 km2 area below 200 m depth has yet been adequately sampled for macrofauna using grabs or corers. Because of this huge unexplored area, the actual number of marine macrobenthic species present today is unknown. It must be vastly greater than earlier estimates based on shallow seas, and according to Grassle and Maciolek is conservatively greater than one million, and more likely to rise to 10 million as more of the deep sea is sampled. The overwhelming taxonomic challenge of describing these new species, the painstaking work of sorting samples from the sediment, and the difficulty of seabed experimentation have perhaps slowed progress in understanding the deep-water sediment community. In contrast, much more has been achieved in biological knowledge of hydrothermal vents (and to a lesser extent cold seep communities) since their discovery in 1977.

Large-scale Patterns in Macrobenthic Diversity Large-scale patterns, other than that for biomass, remain controversial. On the basis largely of sampling of the continental shelf, Thorson pointed out that species richness of the epifauna, occupying less than 10% of the total area, and maximally developed intertidally, rises steeply from low levels in the ice-scoured shallows in the Arctic to high values in the tropics. In contrast, the sediment macrofauna he referred to as ‘infauna,’ usually found deeper and unaffected by ice and meltwater, show much less change. This lack of a latitudinal gradient is supported in some other studies. However, latitudinal comparisons by Sanders in the 1960s found depressed diversity in shallow boreal macrobenthos stressed by wide seasonal temperature change compared to the tropics. Thorson through there were about four times more epibenthic than infaunal species, the microhabitat complexity and consequently high species diversification of the epibenthic habitat being much less obvious in sediments. The sameness of this habitat regardless of latitude led Thorson to his concept of parallel level-bottom communities related to sediment type. But several recent studies indicate shallow tropical and deep-sea sediments do not follow this pattern, with much more species-rich communities developing, albeit including lots of ‘rare’ species, in these habitats.

(c) 2011 Elsevier Inc. All Rights Reserved.

472

MACROBENTHOS

Macrofaunal abundances Central North Pacific (5600 m)

Tagus Abyssal Plain (5000 m)

Rockall Trough (2875 m)

Firth of Lorne (50 m)

Loch Crevan (50 m)

Setubal Canyon, Portugal (3400 m)

Inner Loch Etive (20 m)

Rame Mud (4 m)

Oxygen minimum zone off Oman

OMZ (1250 m) Bottom-water oxygen: 0.52 ml l

Mid OMZ (400 m)

Lower OMZ (850 m)

_1

0.20 ml l

_1

0.13 ml l

_1

Macrofaunal species

Firth of Lorne (50 m)

Annelida

Rockall Trough (2875 m)

Inner Loch Etive (20 m)

Crustacea

Mollusca

Echinodermata

Central North Pacific (5600 m)

Others

Figure 5 Proportional representation of the major taxonomic groups of macrofauna in differing soft sediment habitats and depths worldwide. The upper diagrams show this in terms of the relative abundance in these groups; the lower ones in terms of the number of species represented. A broadly similar pattern is shown between both sets although Crustacea are often more abundant in the deep sea rather than shallow-water macrobenthos. Representation of Annelida (mostly polychaete worms) shows most obvious variation in relation to organic carbon loading and oxygen, the three samples from the oxygen minimum zone (OMZ) in the Arabian Sea off Oman showing a pattern of increasing dominance by Annelida, and eventually complete loss of all other groups except Crustacea, with increasing oxygen depletion. (Data from Mare (1942); Gage (1972) Community structure of the benthos in Scottish sea-lochs. I. Introduction and species diversity. Marine Biology 14: 281–297, Gage J (1977) Structure of the abyssal macrobenthic community in the Rockall Trough. In: Keegan BF, O’Ceidigh P and Boaden PJS (eds) Biology of Benthic Organisms (11th European Marine Biology Symposium), pp. 247–260. Oxford: Pergamon Press; Levin LA, Gage JD, Martin C and Lamont PA (2000) Macrobenthic community structure within and beneath the oxygen minimum zone, NW Arabian Sea. Deep-Sea Research II 47: 189–226.)

(c) 2011 Elsevier Inc. All Rights Reserved.

1 Aerobic Sediment

2 Anaerobic Sediment

3

Depth in sediment (cm)

473

MACROBENTHOS

4 Normal

Transitory

Typical Macrofauna Dominants

Nucula Amphiura Terebellides Rhodine Echinocardium Nephrops

Labidoplax Corbula Goniada Thyasira Pholoe

Chaetozone Anaitides Pectinaria Myriochele Ophiodromus

Polluted

Grossly polluted

Capitella Scolelepis

No macrofauna Surface covered, by fibre blanket

,

Zone

Figure 6 Changes in macrobenthic fauna along an enrichment–disturbance gradient, such as that associated with pollution. The gradient can be replaced by time in tracing recovery along the x-axis to a ‘normal’ community (left) after a severe pollution event (right). Note the shift in body size (reflecting change from quick-growing and fast-turnover ‘pioneer’ species to slower-growing, longer-lived population species) from right to left, as well as the increased depth and extent of bioturbation in the sediment. There is also a concomitant increase in macrobenthic diversity. (From Pearson TH and Rosenbeg R (1978).)

Depth-related patterns in macrobenthic composition have been studied, particularly for larger benthic invertebrates (technically megabenthos, see discussion earlier) and demersal (bottom-living) fish. Rates of macrofaunal turnover, and clinal variation in individual species, correspond to the rate of change in depth. It is highest in upper bathyal and slowest and most subtle in the abyssal. Many changes in species composition can be related to trophic strategies along a gradient in food and hydrodynamic energy (see Figure 2), but changes in the sort and intensity of biological interactions, such as predation, varying with depth, may also be important, as can life-history characteristics (such as the incidence of planktotropic larval development). Recent studies have also established a degree of pressure adaptation during early development that will further limit vertical range. Such ecological processes must be considered in concert with processes at the evolutionary timescale for understanding of zonation patterns. These processes have been summarized using multivariate statistics. While helping in formulating ideas on causal factors, these may obscure the underlying complexity, which is best understood as the sum of the range and adaptation and evolutionary history of individual species. Rex postulated a mod-slope peak in macrobenthic diversity from studies of sled samples taken throughout the Atlantic by Sanders. However, there is high

variability among individual sample values compared and there are conflicting results from other sites worked in the north-eastern Atlantic. That deep-sea macrobenthos has high species diversity seems well founded, but the extent to which this contrasts with shallow water is unclear. In his original study off the north-eastern United States, Sanders showed an impoverished species richness in samples of macrofauna compared to the adjacent slope and rise. Gray has pointed out that on the outer continental shelf off Norway macrofaunal diversity may be comparable to that found in Sanders’ deep-sea samples, and it is considerably higher still off south-eastern Australia. This suggests not only that the inshore shelf off New England is rather poor in species richness but that the North Atlantic as a whole may be atypical, with perhaps historical factors operating there to restrict macrobenthic diversity compared to the southern hemisphere. Such factors may determine the differing response shown in a comparison of deep-sea macrobenthic diversity among sites throughout the Atlantic where the depression at high latitudes is absent south of the equator. The reduced levels at high latitudes may simply reflect Quaternary glaciation so that the deep Norwegian Sea, isolated by shallow sills from deep water to the south, has a much more recently diverged and quite distinct macrofauna from that in the Atlantic.

(c) 2011 Elsevier Inc. All Rights Reserved.

474

MACROBENTHOS

Small-scale Pattern In sampling macrobenthos from a ship it is easy to assume that the animals in this apparently homogeneous habitat are randomly distributed, but sample replicates may not always provide good estimates of the population mean and its sampling error. When samples are mapped over the sediment, or variability is analysed in large numbers of replicates, clumped distributions of some kind are commonplace (Figure 7). Nonrandomly even (regular) dispersions have been detected, but only at the centimeter scale, suggesting that they are actively defined by the ambit (such as the area swept by feeding tentacles) of individual animals. To describe rather than just detect such nonrandom spatial pattern has been a challenging task, not least because most macrofauna are not readily visible in seabed photographs and the very analysis of samples by sieving and mud will destroy fine-scale pattern. Spatial pattern is usually envisaged in the horizontal plane because macrobenthic organisms concentrate their activity on the sediment–water interface in feeding, movement, and reproduction. Pattern is a dynamic expression of this and consequently may change through time but marine sediments provide a three-dimensional habitat so that vertical as well as horizontal spatial patterns may occur. The latter may be best developed at the small scale where smaller macrofauna (and meiofauna) are concentrated around irrigatory or feeding burrows of larger species (Figure 8). The problem is that to

No. of individuals

(A)

(B)

20

20

10

10

analyze this pattern it is difficult not also to disrupt the habitat. Yet in order to understand the basis of such pattern it is vital to analyze and map pattern over a range of scales in conjunction with variability in the sediment habitat. Some of the most revealing studies have examined dispersions of individual species over plots measuring tens of meters square. These may reveal the two aspects of pattern, intensity and form. Intensity can relatively easily be measured by the ratio of variance to mean. This will distinguish distributions that are clumped, regular, or not statistically distinguishable from random. The form of pattern is an aspect that classical statistical tests of nonrandomness do not address. Yet a clumped pattern may be very different in form from that shown by another species that shows similar intensity of aggregation. Although a nuisance for the easy interpretation of sample statistics, an understanding of the biological basis of patterns will provide important insight into the processes maintaining macrobenthic communities.

Altered erosion

Small burrows

Increased stability

Geochemical fluxes

100

Aerobic burrow lining

100 100

100

Sediment_ water interface

0

0

Animal burrow

Moran's I

Distance (m) 0.8

2 1

0.4

0 _1

0.0 _ 0.4 0

20

40

60

0

20

40

Fecal pellets

60

Distance (m)

10 cm Figure 7 Upper, three-dimensional plots of abundance showing spatial dispersion patterns of two infaunal bivalve molluscs Nucula hartivigiana (A) and Soletellna siliqua (B) in a 9000 m2 area of mid-tide sandflat with no obvious gradients in physicochemical conditions. Both species show quite different spatial patterns. The lower plots show spatial correlograms with significant autocorrelation coefficients (measured as Moran’s I) denoted by filled circles (From Hall et al. (1984). Thrush et al., (1989).)

Anaerobic sediment

Figure 8 Benthic biological activity and seabed sediment structure. The diagram shows some of the ways macrobenthos, in conjunction with other size classes, influence sediment fabric, physicochemical properties and solute fluxes (see also Table 1). (After Meadows PS (1986) Biologica activity and seabed sediment structure. Nature 323: 207.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MACROBENTHOS

Functional Importance of Macrobenthos Using a grab as a quantitative sampler in the early years of the twentieth century, C. J. J. Petersen hoped to be able to work out from his thousands of samples taken in the North Sea how much food was available to fish such as flounder. Such links were well supported from finds of large numbers of benthic animals in fish guts, even if later work showed that fish are by no means as important predators of macrobenthos as are invertebrates such as sea stars. Petersen also noticed that characteristic and uniform assemblages of macrobenthos were found that could be related to sediment type and that provided him with statistical units that Thorson later used as his descriptive units in his concept of ‘parallel level bottom communities.’ Ecologists debated whether these existed as anything more than assemblages responding to similar conditions (as originally inferred by Petersen) or reflected functional units or ‘biocoenoses’ where biological interactions play an important, if unknown, role. However, the importance of biological interactions in the subtidal community is difficult to address experimentally, and most data are available from intertidal mudflats and sandflats where ecological gradients related to tidal exposure pose additional complexity. Manipulative experiments on the effects of predation by caging small areas show that predators like shore crabs can have big effects on prey densities, while other studies show competitive exclusion between worms with different burrowing styles that can be reflected in clumping patterns. This contrasts with the importance of grazers and predators in preventing dominance by fast-growing competitive superior species on rocky shores. Such biological interaction cascades down through the community – so-called ‘top-down’ control. But, in sediments, effects such as predation are not so marked overall. Perhaps the three-dimensional structure, the uneven distribution of food and irrigatory flows, and the often intense stratification of chemical processes reduce competition. Indirect effects, such an bioturbation, may take the place of competition. By bulk processing large quantities of sediment, large macrofaunal deposit feeders rework the sediment down to the greatest ocean depths and thus exert a major influence on benthic community structure. It has been suggested that this constant process of biogenic disturbance and alteration of the benthic environment by macrofauna (Figure 8, Table 1) encourages high species richness among smaller macrobenthos by reducing them to levels where competition is relaxed, so that more species can coexist. It is also argued that the constantly

475

changing micro-landscape created by other, larger species provides a rich niche variety for macrofauna. It is difficult to see this process operating on the vast abyssal plains where faunal densities, and therefore such biogenic effects, are so low but species richness is high. Grassle’s spatiotemporal mosaic theory sees the deep-sea bed having patchy and ephemeral food resources that create a relatively small, discrete, and widely separated patch structure promoting coexistence. Effects of larger-scale disturbances are more difficult to detect let alone manipulate in experiments with the sediment community. Yet the evidence is that physical disturbance such as that caused by storm-driven sediment scour and resuspension may have an important effect on assemblage structure and species richness on the exposed continental shelf and margin. The expectation that, just as on an exposed sandy shore, only a relatively small suite of species will be able to adapt to such conditions is confirmed in the deep sea on the continental rise off Nova Scotia, where benthic storms occur with relatively high frequency. Benthic storms may occasionally occur on the abyssal plains, so it should not be assumed that biogenic structure is simply longer-lasting there because it takes so long to be covered by the very low rate of natural sedimentation. Perhaps the most important determinant of the macrobenthic assemblage, or community, is the larval stage usually dispersed in the water column. Larvae can test the substratum and swim off until they find conditions suitable for settlement and metamorphosis. On rock this may involve a series of precise cues that can include presence of their own or other species. Less is known about settlement of infaunal species, but it is thought that positive cues such as microtopography may be much less important in sediment dwellers, while negative cues such as the presence of other species or unattractive sediment are more important. Nevertheless, a community will still be very largely constrained by supply of propagules. In a coastal area the access of larvae supply from adjacent breeding populations may be constrained by coastal topography and currents, not to mention barriers formed by features such as estuaries. The patch structure in the deep sea is maintained by water-borne dispersal stages with the resulting metapopulations spatially unautocorrelated (presence of an organism not dependent on other occurrences). In deep water the openness of the system may mean that the sediment is exposed to a much larger pool of species, even if they are at very low densities as larvae. In the tropics a similar effect results from the greater incidence of planktotrophy (larvae feeding in the plankton), when the longer

(c) 2011 Elsevier Inc. All Rights Reserved.

476

MACROBENTHOS

Table 1

Direct and indirect effects of macrofauna on soft sediments and their ecological consequences

larval life will ensure wider dispersal than that of the nonfeeding larvae prevalent in cooler waters. This may help to explain why so many, mostly rare, species can coexist in both environments.

Importance of Macrobenthos in Environmental Assessment Because the benthic community (unlike fish or plankton) is stationary or at best slow-moving over a small area of bottom, it is useful in monitoring environmental change caused by eutrophication and chemical contamination. Macrobenthos studies have defined the generic effects of such sources of stress by changing representation of major taxa, reduction in diversity, and increasing numerical dominance by small-sized opportunist species causing a downward shift in size structure. This seems to be accompanied by greater patchiness, reflected by increased variability in species abundances in sample replicates. It is also seen as greater variability in local species diversity caused by greater heterogeneity in species identities. This reflects subtle changes in abundance

and, particularly in the more species-rich communities, changes in presence/absence of rare species that might be detected earlier at less severe levels of disturbance. It is claimed that in bioassessments comparing species richness using samples of macrobenthos rare species should receive greater attention by taking larger samples because they contribute relatively more to diversity than the abundant community dominants. Other workers argue that very many species, especially rare ones, are interchangeable in the way they characterize samples. This question requires investigation of the way stressors impact the community, and whether it is the dominant or the rare species that are most sensitive, and therefore most rewarding for study in detecting impacts. Interpretation of impacts also has to proceed against a background of natural changes in benthic communities caused by little-understood, year-toyear differences in annual recruitment. In establishing a baseline there is a need also to take into account the little-understood effects of bottom trawling on coastal benthos. Such disturbance in parts of the North Sea may date back at least 100 years, and now

(c) 2011 Elsevier Inc. All Rights Reserved.

MACROBENTHOS

means that virtually every square meter of bottom is trawled over at least once a year. Such monitoring has in the past entailed costly benthic survey and tedious analysis of samples to species level. Consequently, there has been effort to see whether the effects of stress can be detected at higher taxonomic levels, such as families. Higher taxonomic levels may more closely reflect gradients in contamination than they do abundance of individual species because of the statistical noise generated from natural recruitment variability and from seasonal cycles such as reproduction. This hierarchical structure of macrobenthic response means that, as stress increases, the adaptability of first individual animals, then the species, and then genus, family, and so on, is exceeded so that the stress is manifest at progressively higher taxonomic level. Such new approaches, along with the nascent awareness of conservation of the rich benthic diversity, and with a need for improved environmental impact assessment on the deep continental margin, should ensure a continued active scientific interest in macrobenthos in the years to come.

See also Benthic Boundary Layer Effects. Benthic Foraminifera. Benthic Organisms Overview. Coral Reefs. Deep-Sea Fauna. Demersal Species Fisheries. Fiordic Ecosystems. Grabs for Shelf Benthic Sampling. Meiobenthos. Microphytobenthos. Phytobenthos. Pollution: Effects on Marine Communities. Rocky Shores. Sandy Beaches, Biology of

Further Reading Gage JD and Tyler PA (1991) Deep-sea Biology: A Natural History of Organisms at The Deep-sea Floor. Cambridge: Cambridge University Press.

477

Graf G and Rosenberg R (1997) Bioresuspension and biodeposition: a review. Journal of Marine Systems 11: 269--278. Gray JS (1981) The Ecology of Marine Sediments. Cambridge: Cambridge University Press. Hall SJ, Raffaelli D, and Thrush SF (1986) Patchiness and disturbance in shallow water benthic assemblages. In: Gee JHR and Giller PS (eds.) Organization of Communities: Past and Present, pp. 333--375. Oxford: Blackwell. Hall SJ, Raffaelli D, and Thrush SF (1994) Patchiness and disturbances in shallow water benthic assemblages. In: Giller PS, Hildrew HG, and Raffaelli DG (eds.) Aqautic Ecology: Scale, Patterns and Processes, pp. 333--375. Oxford: Blackwell Scientific Publications. Mare MF (1942) A study of a marine benthic community with special reference to the micro-organisms. Journal of the Marine Biological Association of the United Kingdom 25: 517–554. McLusky DS and McIntyre AD (1988) Characteristics of the benthic fauna. In: Postma H and Zijlstra JJ (eds.) Ecosystems of the World 27, Continental Shelves, pp. 131--154. Amsterdam: Elsevier. Pearson TH and Rosenberg R (1978) Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanography and Marine Biology: an Annual Review 16: 229--311. Pearson TH and Rosenberg R (1987) Feast and famine: structuring factors in marine benthic communities. In: Gee JHR and Giller PS (eds.) Organization of Communities: Past and Present, pp. 373--395. Oxford: Blackwell Scientific Publications. Rex MA (1997) Large-scale patterns of species diversity in the deep-sea benthos. In: Ormond RFG, Gage JD, and Angel MV (eds.) Marine Biodiversity: Patterns and Processes, pp. 94--121. Cambridge: Cambridge University Press. Rhoads DC (1974) Organism–sediment relations on the muddy sea floor. Oceanography and Marine Biology Annual Reviews 12: 263--300. Thorson G (1957) Bottom communities (sublittoral or shallow shelf). In: Hedgepeth JW (ed.) Treatise on Marine Ecology and Paleoecology, pp. 461--534. New York: Geological Society of America. Thrush S (1991) Spatial pattern in soft-bottom communities. Trends in Ecology and Evolution 6: 75--79.

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS F. J. Vine, University of East Anglia, Norwich, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1515–1525, & 2001, Elsevier Ltd.

intensity’), is the tesla (T). Because the magnitude of the Earth’s magnetic field and magnetic anomalies is very small compared to 1 T, they are usually specified in nanoteslas (nT) (1 nT ¼ 109 T). Fortunately, an old geophysical unit called the gamma is equivalent to 1 nT.

Introduction Since World War II it has been possible to measure the variations in the intensity of the Earth’s magnetic field over the oceans from aircraft or ships. In the 1950s the first detailed magnetic survey of an oceanic area, in the north-east Pacific, revealed a remarkable ‘grain’ of linear magnetic anomalies, quite unlike the anomaly pattern observed over the continents. In the 1960s it was realized that these linear anomalies result from a combination of seafloor spreading and reversals of the Earth’s magnetic field. Hence they provide a detailed record of both the evolution of the ocean basins, and the timing of reversals of the Earth’s magnetic field, during the past 160 million years. In addition, because of the dipolar nature of the field and the dominance of ‘fossil’ magnetization in the oceanic crust, the linear anomalies formed at midocean ridge crests, and the anomalies developed over isolated submarine volcanoes, can sometimes yield paleomagnetic information, such as the latitude at which these features were formed.

Units In the SI system the unit of magnetic induction, or flux density (which geophysicists refer to as ‘field

(A)

History of Measurement William Gilbert, one-time physician to Elizabeth I of England, is thought to have been the first person to realize that the form of the Earth’s magnetic field is essentially the same as that about a uniformly magnetized sphere. This is also equivalent to the field about a bar magnet (or magnetic dipole) placed at the center of the Earth, and aligned along the rotational axis (Figure 1). Certainly in terms of the written historical record he was the first person to propose this, in his Latin text De Magnete, published in 1600. Presumably, with the extension of European exploration to more southerly latitudes in the late fifteenth and the sixteenth centuries, mariners had problems with their compasses that Gilbert realized could be explained if the vertical component of the Earth’s magnetic field varies with latitude. Accurate measurements of the direction of the Earth’s magnetic field at London date from Gilbert’s time. Measurement of the strength or intensity of the field, however, was not possible until the early part of the nineteenth century. The equipment used then, and for the following one hundred years or so, included a delicate suspended magnet system and required accurate orientation and leveling before a measurement

N

N

S

S

(B)

Figure 1 The (A) normal (present-day) and (B) reversed states of the dipolar magnetic field of the Earth. Shaded area shows the Earth’s core; heavy arrows indicate directions of the field at the Earth’s surface.

478

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS

could be made. Measurements on a moving platform such as a ship were extremely difficult, therefore, and with the advent of iron and steel ships became impossible because of the magnetic fields associated with the ships themselves. Meanwhile, measurements on land had revealed that although the Earth’s magnetic field is, to a first approximation, equivalent to that about an axial dipole as envisaged by Gilbert, there is a considerable nondipole component, on average 20% of the dipole field. Moreover, the field is changing in time, albeit slightly and slowly, in both intensity and direction. As rather more than 70% of the Earth’s surface is covered by water and there were few magnetic measurements in these areas, detailed mapping of the field at the surface worldwide was seriously hampered. In 1929 the Carnegie Institution of Washington went to the length of building a wooden research ship, the Carnegie, to map the Earth’s magnetic field in oceanic areas. Similarly the then USSR commissioned a wooden research ship, the Zarya, in 1956. However, by this time new electronic instruments had been developed that were capable of making continuous measurements of the total intensity of the Earth’s magnetic field from aircraft and ships. Among the many projects instigated by the Allies in World War II, to counter the submarine menace, was the Magnetic Airborne Detector (MAD) project. The outcome was the development of the fluxgate magnetometer. To increase the sensitivity of the instrument, much of the Earth’s magnetic field is ‘backed off’ by a solenoid producing a biasing field. Initially it was not possible to produce a constant biasing field and the instruments tended to ‘drift,’ which meant that they were not ideal for scientific purposes. After the war these instruments were redeployed for use in conducting aeromagnetic surveys over land areas, in connection with oil and mineral exploration, and then modified for use from ships. In both instances the detector was housed in a ‘fish’ that was towed, so as to remove it from the magnetic fields associated with the ship or aircraft. In the 1950s the proton-precession magnetometer was developed, which had the advantage of achieving the same or somewhat better sensitivity (about 1 part in 50 000) without the problem of drift. Since 1970 even more sensitive magnetometers have been developed – the optical absorption magnetometers. Although in some ways superseded by these magnetometers based on proton and electron precession, fluxgates are still widely used because they have the advantage of measuring the component of the field directed along the axis of the detector rather than the total ambient field. This property makes them particularly useful in satellites, for example, where they

479

are often used as components in orientation systems, at the same time yielding measurements of the Earth’s magnetic field. Measurements of the Earth’s magnetic field in oceanic areas are, therefore, now relatively routine, whether they be from satellites, aircraft, ships, submersibles, or remotely operated or deeply towed vehicles.

Nature of the Earth’s Magnetic Field The differences between the measured magnetic field about the Earth and that predicted for a central and axially aligned dipole are considerable. Best-known is the difference between the magnetic poles – where the field is directed vertically – and the rotational, geographic, poles of the Earth. As a result, for most points on the Earth’s surface there is an angular difference between the directions to true north and to magnetic north. Mariners refer to this as the magnetic variation; scientists refer to it as the magnetic declination. The centered dipole that best fits the observed field predicts a field strength of 30 000 nT around the equator and a maximum value of 60 000 nT at both poles. The actual field departs considerably from this, as can be seen in Figure 2. The intensity and direction of the field, or any of their components (such as magnetic variation) at any one point, vary with time, typically by tens of nanoteslas and a few minutes of arc per year. This is known as the secular variation of the field. The form of the Earth’s magnetic field and its secular variation is thought to derive from the fact that it is generated in the outer, fluid core of the Earth, which is metallic (largely iron and nickel) and hence a good electrical conductor. Convective motions of this fluid conductor carrying electrical currents and interacting with magnetic fields produce a dynamo-like effect and an external magnetic field. The essential axial symmetry of this field is probably determined by the influence of the Coriolis force on the precise nature of the convective motions. Historical, archeomagnetic and paleomagnetic data suggest that although the field at any one time departs considerably from that predicted by a geocentric axial dipole, when averaged over several thousand years, the mean field is very close to that about such a dipole. On even longer, geological, timescales the field intermittently reverses its polarity completely (Figure 1), probably within a period of approximately 5000 years. The length of intervals of a particular polarity varies widely from a few tens of thousands of years to a few tens of millions of years.

(c) 2011 Elsevier Inc. All Rights Reserved.

480

MAGNETICS 80°N

60°N 40°N 20°N 0° 20°S 40°S 60°S

80°S 180°

140°W

100°W

60°W

20°W 0° 20°E

60°E

100°E

140°E

180°

Figure 2 Total intensity of the Earth’s magnetic field (in nT) at the Earth’s surface (for epoch 1980). (Reproduced from Langel RA, in Geomagnetism Vol. 1, JA Jacobs (ed.) ^ 1987 Academic Press, by permission of the publisher.)

The secular variation and changes in the polarity of the field result from dynamic processes deep in the Earth’s interior. There are also shorter-period variations of the field with time that are of external origin, essentially a result of the interaction of the solar wind with the Earth’s magnetic field. The most relevant of these in the present context is the daily or diurnal variation of the field. This is a smooth variation that is greatest during daylight hours and typically has an amplitude of a few tens of nanoteslas. It can be greater, however, near the magnetic equator and poles. Increased solar activity can produce higher-amplitude and more irregular variations, and intense sunspot activity produces global magnetic storms; high-amplitude, short-period variations during which magnetic surveying has to be discontinued.

Reduction of Magnetic Data Measurements of the Earth’s magnetic field over oceanic areas are corrected for the present-day spatial and time variations of the field described above in order to obtain residual ‘anomalies’ in the field. These are caused by magnetization contrasts at or near the Earth’s surface, i.e., in the upper lithosphere. Such anomalies should therefore yield information on the magnetization and structure of the oceanic crust. For measurements made at or near the Earth’s surface, i.e. from ships, aircraft, or submersibles, the secular and diurnal variation of the field can often be

ignored because these effects are very small compared to the amplitudes of the anomalies being mapped. However, should a very accurate survey be required, secular variation has to be taken into account if surveys made at different times are being combined and the observations should be corrected for diurnal variation using records from nearby land stations or moored buoys. If there are sufficient ‘cross-overs’ during the survey (i.e. repeat measurements at the same point), it may also be possible, indeed preferable, to use these to correct for diurnal variation. It may also be necessary to correct for any magnetic effect of the moving platform itself. If present, this effect will vary according to the direction of travel. For many purposes, however, the above corrections are so small that they can be ignored. The final correction, the removal of the main or ‘regional’ field of the Earth, that is, the field generated in the Earth’s core, must always be applied. In theory this should be simple. The depth to the core–mantle boundary, 2900 km, means that the field originating in the core should have a smooth, long-wavelength variation at or above the Earth’s surface. The magnetization contrasts within a few tens of kilometers of the Earth’s surface will produce anomalies of much shorter wavelength. Between these two source regions any magnetic minerals in the mantle are at a temperature above their Curie temperature and are effectively nonmagnetic. In practice, because of its complexity and because it is changing with time, it

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS

has proved difficult to accurately define the main field of the Earth, i.e. that originating in the core, and the way in which it is changing with time. The first attempt to define a global ‘International Geomagnetic Reference Field’ was made in the 1960s and, although revised and greatly improved at fiveyear intervals since then, its level, if not its gradients, can still be seen to be slightly incorrect for certain oceanic areas. As a result, particularly in the past, the regional field for particular profiles or surveys has been obtained by fitting a smooth long-wavelength curve or surface to the observed data. Once the long-wavelength ‘regional field’ has been removed from magnetic data, the resulting ‘residual’ or ‘total-field’ anomalies are assumed to result from magnetization contrasts within the upper, magnetic, part of the Earth’s lithosphere (Figure 3).

25°W

24°W

23°W

22°W 18°N

A Santo ~ Antao

17°N

10

Sal

m

>1000 m

C 00

10

~ Nicolau Sao >1000 m

m

00

D

m

2000

B

16°N

37 000

Field (nT)

B

36 000

A C Observed total magnetic field

35 000

D

34 000

Assumed regional field

Field (nT)

+ 800 + 400 0 _ 400 _ 800

Water depth (m)

33 000

0 2000 4000 6000

Total field magnetic anomaly

Cape Verde Islands

0

200

400

600

Dakar

800 1000 1200 km

Figure 3 Magnetic profile recorded by a fluxgate magnetometer between the Cape Verde islands and Dakar, Senegal. The ship’s track close to the islands is shown in the upper part of the diagram. The dashed line indicates the regional field used to calculate total field magnetic anomalies. The highamplitude, short-wavelength anomalies close to the Cape Verde Islands reflect the presence of highly magnetic volcanic rocks at shallow depth. (Reproduced from Heezen BC, et al., in Deep-Sea Research, Vol. 1 ^ 1953 Elsevier Science, by permission of the publisher.)

481

Magnetization of Ocean Floor Rocks All minerals, and hence all rocks, exhibit magnetic properties. However, the magnetization that most rock types acquire in the relatively weak magnetic field of the Earth is insufficient to produce significant anomalies in the Earth’s magnetic field, particularly when this is measured at some distance from the rocks, as is typically the case in oceanic areas. For rocks to be capable of producing appreciable anomalies they must contain more than a few percent by volume of ‘ferromagnetic’ minerals, that is, certain oxides and sulfides containing iron, notably magnetite (Fe3O4). Most sediments and ‘acid’ (silicarich) igneous rocks, such as granite, do not meet this criterion. Basic (silica-deficient) igneous rocks such as basalts, and the coarser-grained but chemically equivalent gabbros, and ultrabasic rocks such as peridotite do contain a higher proportion of iron oxides and are capable of producing anomalies. Metamorphic rocks, formed when preexisting rocks are subjected to high temperatures and/or pressures, are typically weakly magnetized except for some formed from basic or ultrabasic igneous rocks. Apart from its sedimentary veneer, the upper part of the oceanic lithosphere consists almost entirely of basic and ultrabasic rocks, i.e. basalts, gabbros, and peridotites. This is a consequence of the way in which it is formed by the process of seafloor spreading. At midocean ridge crests the ultrabasic peridotite of the Earth’s mantle undergoes partial melting, producing basic magma that rises and collects as a magma chamber within oceanic crust. Solidification of such magma chambers ultimately forms the main crustal layer of gabbro, but not before some ultrabasic rocks have formed at the base of the chamber from the accumulation of first formed crystals, and magma has been extruded through near vertical fissures to form pillow basalts on the seafloor. Solidification of the magma in these fissures forms a layer consisting of dikes between the gabbro and the basalts. Thus, because of the rock types present, the oceanic crust and upper mantle are relatively strongly magnetized and capable of producing largeamplitude anomalies in the Earth’s magnetic field, even when measured at sea level. The thickness of the magnetic layer is determined by the depth to the Curie point isotherm. Because of the way in which the oceanic lithosphere is formed, by spreading about ridge crests, this varies from a few kilometers depth within the crust at ridge crests to a depth of approximately 40 km in oceanic lithosphere that is 100 million years, or more, in age. In places the thermal regime associated with seafloor spreading is modified by mantle ‘hot spots.’ As a result there is an

(c) 2011 Elsevier Inc. All Rights Reserved.

482

MAGNETICS

enhanced degree of partial melting of the mantle and the larger volumes of magma produced extrude on to the seafloor to form seamounts, oceanic islands and, exceptionally, oceanic plateaux. These all involve an appreciable thickening of the oceanic crust, but the rock types involved are all essentially basic and potentially strongly magnetic (Figure 3).

Observed Anomalies The marked contrast in the way in which continental and oceanic crust are formed, and hence in the 135°W

predominant rock types in each setting, gives rise to a striking difference in the character of the total field anomalies developed over the two types of lithosphere. Within the continents the variety of rock types in mountain belts and their juxtaposition by folding and faulting produces magnetization contrasts and anomalies that delineate the general trend of the belt. Areas of igneous activity that include basic igneous rocks are characterized by very largeamplitude and typically short-wavelength anomalies, and sedimentary basins and extensive areas of granite are quiet magnetically. This pattern of

130°W

125°W

50°N

50°N

45°N

45°N

40°N

40°N 135°W

130°W

125°W

Figure 4 Linear magnetic anomalies in the north–east Pacific. Areas of positive anomaly are shown in black. Straight lines indicate faults offsetting the anomaly pattern; arrows, the axes of three short ridge lengths in the area – from north to south, the Explorer, Juan de Fuca and Gorda ridges. (Based on Figure 1 of Raff AD and Mason RG in Bull. Geol. Soc. Amer., Vol 72. ^ 1961 Geological Society of America. Reproduced by permission of the publisher.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS

anomalies is characteristic of the continental shelves out to the continental slope, the true geological boundary between the continents and the oceans, although, in that the shelves are typically underlain by a great thickness of sediments, they are often magnetically ‘quiet.’ Deep sea areas, that is, those underlain by oceanic lithosphere, are characterized by remarkably linear and parallel anomalies that extend for hundreds of kilometers and are truncated and offset by fracture zones (Figure 4). The fracture zones are formed by the transform faults that offset the crest of the midocean ridge system. Thus the linear magnetic anomalies parallel the ridge crests. The anomalies are remarkable for their linearity, their high amplitude and the steep magnetic gradients that separate highs from lows. Any explanation of them in terms of linear structures and/or lateral variations in rock type within the oceanic crust is extremely improbable. It transpires that they result from a combination of sea floor spreading and reversals of the Earth’s magnetic field. As new oceanic crust and upper mantle form at a ridge crest they acquire a permanent (remanent) magnetization which parallels the ambient direction of the field. If, as spreading occurs, the Earth’s magnetic field reverses, then the ribbon of newly formed oceanic lithosphere along the whole length of the spreading ridge system acquires a remanent magnetisation in the opposite direction. It is these contrasts between normally and reversely magnetized material which produce the high amplitude linear anomalies and the steep gradients between them. Rates of seafloor spreading vary greatly for different ridges and the interval between reversals of the Earth’s magnetic field is also very variable throughout geological time. However, rates of spreading are typically a few tens of millimeters per year and the average polarity interval is about 0.5 million years. Thus typical linear anomalies are 10–20 km in width. Initially, spreading rates could only be reliably determined for the past 3.5 million years. For this period the reversal timescale had been independently determined from measurements of the age and polarity of remanent magnetization of both subaerial lava flows and deep-sea sediments, and it is clearly reproduced in the anomalies recorded across midocean ridge crests (Figure 5). With the dating of older oceanic crust by the international Deep Sea Drilling Program it became possible to deduce spreading rates at earlier times and to calibrate the timescale of reversals of the Earth’s magnetic field implied by the older linear anomalies. In this way the geomagnetic reversal timescale for the past 160 million years has been deduced (Figure 6).

JUAN DE FUCA RIDGE

46°N : 130°W

(A) RIDGE MODEL km LAYER 2 LAYER 3

120

90

3 cm y 60

30

0

30

60

90

_1

120

S.L. 5 10 km 500°C ?

MOHO

MANTLE

483

BASALT H2O ?

(B) ANOMALY MAP

LINE OF

PROFILE (C)

(C) OBSERVED PROFILE

500 nT

(D) SIMULATION

120

60

0

60

4

2

0

2

120 km

S.L. 5 km 4

My

Figure 5 (A) Schematic crustal model for the Juan de Fuca Ridge, south-west of Vancouver Island. Shaded material in layer 2 is normally magnetized; unshaded material is reversely magnetized. SL ¼ sea level. (B) Part of the summary map of magnetic anomalies recorded over the Juan de Fuca Ridge (Figure 4). (C) Total field magnetic anomaly profile along the line indicated in (B). (D) Computed profile assuming the model and reversal timescale for the past 3.5 million years. (Reproduced from Vine FJ in The History of the Earth’s Crust. RA Phinney (ed.). ^ 1968 Princeton University Press, by permission of the publisher.)

In recording the times at which the Earth’s magnetic field has reversed its polarity, the linear magnetic anomalies also serve as time or growth lines that reveal the evolution of the ocean basins in terms of seafloor spreading (Figure 7). Thus it is possible to accurately reconstruct the most recent phase of continental drift (during the past 180 million years) when a former supercontinent was split up to form the present-day continents and the Atlantic and Indian Oceans. The Pacific Ocean was formed during the

(c) 2011 Elsevier Inc. All Rights Reserved.

484

MAGNETICS

0

PLIOCENE 5

60

1 2A 2 3

120

26 BARREMIAN

PALEOCENE

M1

27 M3

28 29

3A 4 4A

10

Ma ANOMALY

Ma ANOMALY

Ma ANOMALY Q PLEISTOCENE

MAESTRICHTIAN 70

5

HAUTERIVIAN

30

M5 M8

130

M10

5A 32

20

5C 5D 5E 6 6A 6B 6C

33

30

80

140 BERRIASIAN

M16

34

M18

SANTONIAN

8 9 10

TITHONIAN

CONIACIAN TURONIAN

90

M20 150

11 12

TERTIARY

M12 M14

CAMPANIAN

7

OLIGOCENE

M11 VALANGINIAN

5B

MIOCENE

M21 M22

KIMMERIDGIAN

CENOMANIAN

M23

13 M25

18 19 ALBIAN

50

21 22 23 24

CRETACEOUS

20

EOCENE

PM26

OXFORDIAN 100

JURASSIC

40

15 16 17

160

PM29

CALLOVIAN

110

170 BATHONIAN

APTIAN MO

25 60

120

180

Figure 6 Geomagnetic polarity timescale for the past 160 million years. (Reproduced from Jones EJW, Marine Geophysics. ^ 1999 John Wiley and Sons, by permission of the publisher.)

same period, and in the process dispersed marginal fragments of the supercontinent around its rim where they are now recognized as ‘suspect terranes.’ In contrast to this very detailed record of spreading and drift for the past 180 million years there is no such record for the previous 96% of geological time, and earlier phases of drift and mountain building have to be deduced from the more complex and fragmentary geological record within the remaining 40% of the Earth’s surface that is covered by continental crust.

Paleomagnetic Information Contained in Oceanic Magnetic Anomalies As a result of the dipolar nature of the Earth’s magnetic field, whereby its inclination to the horizontal varies systematically from the equator to the

poles (Figure 1), anomalies in the total field over a relatively simple and symmetrical feature such as the central, normally magnetized ribbon of crust at a midocean ridge crest typically have an asymmetry (Figure 8). The exceptions occur at the poles and across ridges trending exactly north–south, where the anomaly is a symmetrical high, and over an east– west trending ridge at the Equator, where the anomaly is a symmetrical low. For all other latitudes, and all orientations other than north–south, the degree of asymmetry, or the ‘phase-shift’ of the anomaly, is a function of the latitude and orientation. As a result of spreading, all older linear anomalies, unless formed about a north–south trending, east– west spreading ridge, will now be at a different latitude from that at which they were formed, and the direction of their remanent magnetization will be

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS

60°W

0°

60°E

60°N

60°N

30°N

30°N

0°

0°

30°S

30°S

60°S

60°S

(A)

485

60°W

0°

Active spreading center Fossil spreading center

60°E Discontinuity in spreading / breakup anomaly Fracture zone

Figure 7 A compilation of the trends of linear magnetic anomalies in (A) the Atlantic and Indian Oceans and (B) the Pacific Ocean. (Reproduced from Jones EJW, Marine Geophysics. ^ 1999 John Wiley and Sons, by permission of the publisher.)

different from the direction of the ambient magnetic field. As a consequence, the asymmetry of the anomaly is different from what one would predict for similarly directed remanence and ambient field. This difference between the observed and predicted phase shift reflects the latitudinal change. Although the interpretation of such data is somewhat ambiguous if the orientation of the anomaly is thought to have changed since the time of formation, it does provide paleolatitude, i.e. paleomagnetic, information for oceanic areas, which is otherwise rather sparse. As with all paleomagnetic data, they can be used to test independently derived models for the ‘absolute’ motion of plates and plate boundaries, such as ridge crests, across the face of the Earth.

In theory the ambiguity of paleolatitude determination mentioned above can be removed by carrying out the analysis on anomalies of the same age on either side of a particular ridge. An intriguing result of such studies, however, is that in some cases different latitudes of formation are deduced for the same anomaly on either side of the ridge. In that they were formed at the same time and at the same ridge crest, this cannot be so, and the result is giving us yet more information on the geometry or magnetization of the source region. Such a result could be produced by a decay in the intensity of the Earth’s magnetic field or an increase in the number of magnetic excursions or very short-lived reversals as a particular polarity interval progresses, and/or a more complex

(c) 2011 Elsevier Inc. All Rights Reserved.

MAGNETICS

487

Ridge crest

90°

0 1 2 km 3

Lavas Dikes Gabbro

4 5

Normal Reversed

500 nT

0 2 4 6 8 10 km 60° 30°

0°

S 50

0

N 50 km

(A)

Figure 9 Postulated geometry of normal/reverse magnetization contrasts in the oceanic crust, which could explain the anomalous phase shift of certain linear anomalies.

spreading and volcanic activity, it records both the history of reversals of the Earth’s magnetic field, and the lateral and latitudinal displacement of the crust during the past 160 million years. This record can be played back by measuring the anomalies in the intensity of the Earth’s magnetic field over the oceans at the present day.

500 nT

See also N N 30° E N 60° E

50

0

E 50 km

Deep-Sea Drilling Results. Geomagnetic Polarity Timescale. Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Propagating Rifts and Microplates. Seamounts and Off-Ridge Volcanism. Seismic Structure

(B)

Further Reading

Figure 8 Variation of the magnetic anomaly pattern (A) with geomagnetic latitude (all profiles are N–S; angles refer to magnetic inclination; no vertical exaggeration) and (B) with direction of the profile at fixed latitude (magnetic inclination is 451 in all cases; no vertical exaggeration). (Reproduced from Kearey P and Vine FJ. Global Tectonics. ^ 1996 Blackwell Science, by permission of the publisher.)

Bullard EC and Mason RG (1963) The magnetic field over the oceans. In: Hill MN (ed.) The Sea, vol. 3, pp. 175--217. London: Wiley-Interscience. Harrison CGA (1981) Magnetism of the oceanic crust. In: Emiliani C (ed.) The Sea, vol. 7, pp. 219--239. Wiley: New York. Jones EJW (1999) Marine Geophysics. Chichester: Wiley. Kearey P and Vine FJ (1996) Global Tectonics. Oxford: Blackwell Science. Vacquier V (1972) Geomagnetism in Marine Geology. Amsterdam: Elsevier.

England seamounts of the north–west Atlantic, and the Musician seamounts of the central Pacific. These yield paleomagnetic pole positions that are consistent with other results for the mid-Cretaceous for the North American and Pacific plates, respectively.

Conclusion Thus, because of the dominance of remanent magnetization in the oceanic crust, and the relative simplicity of the way in which it is formed by seafloor

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGANESE NODULES D. S. Cronan, Royal School of Mines, London, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1526–1533, & 2001, Elsevier Ltd.

Introduction Manganese nodules, together with micronodules and encrustations, are ferromanganese oxide deposits which contain variable amounts of other elements

Table 1

488

(Table 1). They occur throughout the oceans, although the economically interesting varieties have a much more restricted distribution. Manganese nodules are spherical to oblate in shape and range in size from less than 1 cm in diameter up to 10 cm or more. Most accrete around a nucleus of some sort, usually a volcanic fragment but sometimes biological remains. The deposits were first described in detail in the Challenger Reports. This work was co-authored by J. Murray and A. Renard, who between them initiated

Average abundances of elements in ferromanganese oxide deposits

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGANESE NODULES

the first great manganese nodule controversy. Murray believed the deposits to have been formed by submarine volcanic processes whereas Renard believed that they had precipitated from continental runoff products in sea water. This controversy remained unresolved until it was realized that nodules could obtain their metals from either or both sources. The evidence for this included the finding of abundant nodules in the Baltic Sea where there are no volcanic influences, and the finding of rapidly grown ferromanganese oxide crusts associated with submarine hydrothermal activity of volcanic origin on the Mid-Atlantic Ridge. Subsequently, a third source of metals to the deposits was discovered, diagenetic remobilization from underlying sediments. Thus marine ferromanganese oxides can be represented on a triangular diagram (Figure 1), the corners being occupied by hydrothermal (volcanically derived), hydrogenous (seawater derived) and diagenetic (sediment interstitial water derived) constituents. There appears to be a continuous compositional transition between hydrogenous and diagenetic deposits, all of which are formed relatively slowly at normal deep seafloor temperatures. By contrast, although theoretically possible, no continuous compositional gradation has been reported between hydrogenous and hydrothermal deposits, although mixtures of the two do occur. This may be partly because (1) the growth rates of hydrogenous and hydrothermal deposits are very different with the latter accumulating much more rapidly than the former leading to the incorporation of only limited amounts of the more slowly accumulating hydrogenous material in them, and (2) the temperatures of

489

formation of the deposits are different leading to mineralogical differences between them which can affect their chemical composition. Similarly, a continuous compositional gradation between hydrothermal and diagenetic ferromanganese oxide deposits has not been found, although again this is theoretically possible. However, the depositional conditions with which the respective deposits are associated i.e., high temperature hydrothermal activity in mainly sediment-free elevated volcanic areas on the one hand, and low-temperature accumulation of organic rich sediments in basin areas on the other, would preclude much mixing between the two. Possibly they may occur in sedimented active submarine volcanic areas.

Internal Structure The main feature of the internal structure of nodules is concentric banding which is developed to a greater or lesser extent in most of them (Figure 2). The bands represent thin layers of varying reflectivity in polished section, the more highly reflective layers being generally richer in manganese than the more poorly reflective ones. They are thought to possibly represent varying growth conditions.

mix dic

wn kno ing

mix

Sp ora

No

ing

Hydrothermal Mn/Fe >10 high growth rates

Continuous mixing Hydrogenous Mn/Fe ~1 low growth rates

Diagenetic Mn/Fe > 2.5 intermediate growth rates

Figure 1 Triangular representation of marine ferromanganese oxide deposits.

Figure 2 Concentric banding in a manganese nodule. (Reproduced by kind permission of CNEXO, France.)

(c) 2011 Elsevier Inc. All Rights Reserved.

490

MANGANESE NODULES

On a microscopic scale, a great variety of structures and textures are apparent in nodules, some of them indicative of postdepositional alteration of nodule interiors. One of the most commonly observed and most easily recognizable is that of collomorphic globular segregations of ferromanganese oxides on a scale of tenths of a millimeter or less, which often persist throughout much of the nodule interior. Often the segregations become linked into polygons or cusps elongated radially in the direction of growth of the nodules. Several workers have also recognized organic structures within manganese nodules. Furthermore, cracks and fissures of various sorts are a common feature of nodule interiors. Fracturing of nodules is a process which can lead to their breakup on the seafloor, in some cases as a result of the activity of benthic organisms, or of bottom currents. Fracturing is an important process in limiting the overall size of nodules growing under any particular set of conditions.

Growth Rates It is possible to assess the rate of growth of nodules either by dating their nuclei, which gives a minimum rate of growth, or by measuring age differences between their different layers. Most radiometric dating techniques indicate a slow growth rate for nodules, from a few to a few tens of millimeters per million years. Existing radiometric and other techniques for nodule dating include uranium series disequilibrium methods utilizing 230Th 231Pa, the 10Be method, the K-Ar method, fission track dating of nodule nuclei, and hydration rind dating. In spite of the overwhelming evidence for slow growth, data have been accumulating from a number of sources which indicate that the growth of nodules may be variable with periods of rapid accumulation being separated by periods of slower, or little or no growth. In general, the most important factor influencing nodule growth rate is likely to be the rate at which elements are supplied to the deposits, diagenetic sources generally supplying elements at a faster rate than hydrogenous sources (Figure 1). Further, the tops, bottoms and sides of nodules do not necessarily accumulate elements at the same rate, leading to the formation of asymmetric nodules in certain circumstances (Figure 3). Differences in the surface morphology between the tops, bottoms and sides of nodules in situ may also be partly related to growth rate differences. The tops receive slowly accumulating elements hydrogenously supplied from seawater and are smooth, whereas the bottoms receive more rapidly accumulating elements diagenetically supplied

Fe, Co from sea water

Sediment surface Mn, Ni, Cu from interstitial waters 1cm Figure 3 Morphological and compositional differences between the top and bottom of a Pacific nodule. (Reproduced with permission from Cronan, 1980.)

from the interstitial waters of the sediments and are rough (Figure 3). The ‘equatorial bulges’ at the sediment–water interface on some nodules have a greater abundance of organisms on them than elsewhere on the nodule surface, suggesting that the bulges may be due to rapid growth promoted by the organisms. It is evident therefore that nodule growth cannot be regarded as being continuous or regular. Nodules may accrete material at different rates at different times and on different surfaces. They may also be completely buried for periods of time during which it is possible that they may grow from interstitial waters at rates different from those while on the surface, or possibly not grow at all for some periods. Some even undergo dissolution, as occurs in the Peru Basin where some nodules get buried in suboxic to reducing sediments.

Distribution of Manganese Nodules The distribution and abundance of manganese nodules is very variable on an oceanwide basis, and can also be highly variable on a scale of a kilometer or less. Nevertheless, there are certain regional regularities in average nodule abundance that permit some broad areas of the oceans to be categorized as containing abundant nodules, and others containing few nodules (Figure 4), although it should always be borne in mind that within these regions local variations in nodule abundance do occur. The distribution of nodules on the seafloor is a function of a variety of factors which include the presence of nucleating agents and/or the nature and age of the substrate, the proximity of sources of elements, sedimentation rates and the influence of organisms. The presence of potential nuclei on the seafloor is of prime importance in determining nodule distribution. As most nodule nuclei are volcanic in origin, patterns of volcanic activity and the

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGANESE NODULES

491

Figure 4 Distribution of mangancese nodules in the oceans (updated from Cronan, 1980 after various sources.) , Areas of nodule coverage; , areas where nodules are locally abundant.

subsequent dispersal of volcanic materials have an important influence on where and in what amounts nodules occur. Other materials can also be important as nodule nuclei. Biogenic debris such as sharks’ teeth, can be locally abundant in areas of slow sedimentation and their distribution will in time influence the abundance of nodules in such areas. As most nuclei are subject to replacement with time, old nodules have sometimes completely replaced their nuclei and have fractured, thus providing abundant nodule fragments to serve as fresh nuclei for ferromanganese oxide deposition. In this way, given sufficient time, areas which initially contained only limited nuclei may become covered with nodules. One of the most important factors affecting nodule abundance on the seafloor is the rate of accumulation of their associated sediments, low sedimentation rates favoring high nodule abundances. Areas of the seafloor where sedimentation is rapid are generally only sparsely covered with nodules. For example, most continental margin areas have sedimentation rates that are too rapid for appreciable nodule development, as do turbiditefloored deep-sea abyssal plains. Low rates of sedimentation can result either from a minimal sediment supply to the seafloor or currents inhibiting its deposition. Large areas in the centers of ocean basins receive minimal sediment input. Under these

conditions substantial accumulation of nodules at the sediment surface is favored. Worldwide Nodule Distribution Patterns

Pacific Ocean As shown in, nodules are abundant in the Pacific Ocean in a broad area, called the Clarion–Clipperton Zone, between about 61N and 201N, extending from approximately 1201W to 1601W. The limits of the area are largely determined by sedimentation rates. Nodules are also locally abundant further west in the Central Pacific Basin. Sediments in the northern part of the areas of abundant nodules in the North Pacific are red clays with accumulation rates of around 1 mm per thousand years whereas in the south they are siliceous oozes with accumulation rates of 3 mm per thousand years, or more. Nodule distribution appears to be more irregular in the South Pacific than in the North Pacific, possibly as a result of the greater topographic and sedimentological diversity of the South Pacific. The nodules are most abundant in basin environments such as those of the south-western Pacific Basin, Peru Basin, Tiki Basin, Penrhyn Basin, and the CircumAntarctic area. Indian Ocean In the Indian Ocean the most extensive areas of nodule coverage are to the south

(c) 2011 Elsevier Inc. All Rights Reserved.

492

MANGANESE NODULES

of the equator. Few nodules have been recorded in the Arabian Sea or the Bay of Bengal, most probably because of the high rates of terrigenous sediment input in these regions from the south Asian rivers. The equatorial zone is also largely devoid of nodules. High nodule concentrations have been recorded in parts of the Crozet Basin, in the Central Indian Ocean Basin and in the Wharton Basin. Atlantic Ocean Nodule abundance in the Atlantic Ocean appears to be more limited than in the Pacific or Indian Oceans, probably as a result of its relatively high sedimentation rates. Another feature which inhibits nodule abundance in the Atlantic is that much of the seafloor is above the calcium carbonate compensation depth (CCD). The areas of the Atlantic where nodules do occur in appreciable amounts are those where sedimentation is inhibited. The deep water basins on either side of the MidAtlantic Ridge which are below the CCD and which accumulate only limited sediment contain nodules in reasonable abundance, particularly in the western Atlantic. Similarly, there is a widespread occurrence of nodules and encrustations in the Drake Passage–Scotia Sea area probably due to the strong bottom currents under the Circum-Antarctic current inhibiting sediment deposition in this region. Abundant nodule deposits on the Blake Plateau can also be related to high bottom currents. Buried nodules Most workers on the subject agree that the preferential concentration of nodules at the sediment surface is due to the activity of benthic organisms which can slightly move the nodules. Buried nodules have, however, been found in all the oceans of the world. Their abundance is highly variable, but it is possible that it may not be entirely random. Buried nodules recovered in large diameter cores are sometimes concentrated in distinct layers. These layers may represent ancient erosion surfaces or surfaces of nondeposition on which manganese nodules were concentrated in the past. By contrast, in the Peru Basin large asymmetrical nodules get buried when their bottoms get stuck in tenacious suboxic sediment just below the surface layer.

Compositional Variability of Manganese Nodules Manganese nodules exhibit a continuous mixing from diagenetic end members which contain the mineral 10A˚ manganite (todorokite) and are enriched in Mn, Ni and Cu, to hydrogenous end members which contain the mineral d MnO2

(vernadite) and are enriched in Fe and Co. The diagenetic deposits derive their metals at least in part from the recycling through the sediment interstitial waters of elements originally contained in organic phases on their decay and dissolution in the sediments, whereas the hydrogenous deposits receive their metals from normal sea water or diagenetically unenriched interstitial waters. Potentially ore-grade manganese nodules of resource interest fall near the diagenetic end member in composition. These are nodules that are variably enriched in Ni and Cu, up to a maximum of about 3.0% combined. One of the most striking features shown by chemical data on nodules are enrichments of many elements over and above their normal crustal abundances (Table 1). Some elements such as Mn, Co, Mo and Tl are concentrated about 100-fold or more; Ni, Ag, Ir and Pb are concentrated from about 50- to 100-fold, B, Cu, Zn, Cd, Yb, W and Bi from about 10 to 50-fold and P, V, Fe, Sr, Y, Zr, Ba, La and Hg up to about 10-fold above crustal abundances.

Regional Compositional Variability

Pacific Ocean In the Pacific, potentially ore-grade nodules are generally confined to two zones running roughly east–west in the tropical regions, which are well separated in the eastern Pacific but which converge at about 1701–1801W (Figure 5). They follow the isolines of intermediate biological productivity, strongly suggestive of a biological control on their distribution. Within these zones, the nodules preferentially occupy basin areas near or below the CCD. Thus they are found in the Peru Basin, Tiki Basin, Penrhyn Basin, Nova Canton Trough area, Central Pacific Basin and Clarion– Clipperton Zone (Figure 5). Nodules in all these areas have features in common and are thought to have attained their distinctive composition by similar processes. The potentially ore-grade manganese nodule field in the Peru Basin, centered at about 71–81S and 901W (Figure 5), is situated under the southern flank of the equatorial zone of high biological productivity on a seafloor composed of pelagic brown mud with variable amounts of siliceous and calcareous remains. Nodules from near the CCD at around 4250 m are characterized by diagenetic growth and are enriched in Mn, Ni and Cu, whereas those from shallower depth are characterized mainly by hydrogenous growth. The Mn/Fe ratio increases from south to north as productivity increases, whereas the Ni and Cu contents reach maximum values in the middle of the area where Mn/Fe ratios are about 5.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGANESE NODULES

180°

140°

160°

120°

100°

493

80°

10 0

0

20

20°

Central Pacific Basin

Clarion _ Clipperton Zone

0° Peru Basin

Nova-Canton Trough Area Penrhyn Basin Tiki Basin

20

150

0

10 0

20°

50 Figure 5 Approximate limits of areas of nickel- and copper-rich nodules in the subequatorial Pacific referred to in the text (productivity isolines if g C m2 y1).

In the Tiki Basin there is also an increase in the Mn/Fe ratio of the nodules from south to north. All Ni þ Cu values are above the lower limit expected in diagenetically supplied material. The Penrhyn Basin nodules fall compositionally within the lower and middle parts of the Mn/Fe range for Pacific nodules as a whole. However, nodules from the northern part of the Basin have the highest Mn/Fe ratios and highest Mn, Ni and Cu concentrations reflecting diagenetic supply of metals to them, although Ni and Cu decrease slightly as the equator is approached. Superimposed on this trend are variations in nodule composition with their distance above or below the CCD. In the Mn-, Ni-, and Cu-rich nodule area, maximum values of these metals in nodules occur within about 200 m above and below the CCD. The latititudinal variation in Mn, Ni and Cu in Penrhyn Basin nodules may be due to there being a hydrogenous source of these metals throughout the Basin, superimposed on which is a diagenetic source of them between about 21 and 61S at depths near the CCD, but less so in the very north of the Basin (0–21S) where siliceous sedimentation prevails under highest productivity waters. In the Nova Canton Trough area, manganese concentrations in the nodules are at a maximum between the equator and 2.51S, where the Mn/Fe ratio is also highest. Manganese shows a tendency to

decrease towards the south. Nickel and copper show similar trends to Mn, with maximum values of these elements being centered just south of the equator at depths of 5300–5500 m, just below the CCD. In the central part of the Central Pacific Basin, between the Magellan Trough and the Nova Canton Trough, diagenetic nodules are found associated with siliceous ooze and clay sedimentation below the CCD. Their Ni and Cu contents increase southeastwards reaching a maximum at about 2.51–31N and then decrease again towards the equator where productivity is highest. The Clarion–Clipperton Zone deposits rest largely on slowly accumulated siliceous ooze and pelagic clay below the CCD. The axis of highest average Mn/ Fe ratio and Mn, Ni and Cu concentrations runs roughly southwest–northeast with values of these elements decreasing both to the north and south as productivity declines respectively to the north and increases towards the equatorial maximum in the south. Indian Ocean In the Indian Ocean, Mn-, Ni-, and Cu-rich nodules are present in the Central Indian Ocean Basin between about 51 and 151S. They are largely diagenetic in origin and rest on siliceous sediments below the CCD under high productivity waters. The deposits show north–south compositional

(c) 2011 Elsevier Inc. All Rights Reserved.

494

MANGANESE NODULES

variability with the highest grades occurring in the north.

Atlantic Ocean In the Atlantic Ocean, diagenetic Mn-, Ni-, and Cu-rich nodules occur most notably in the Angola Basin and to a lesser extent in the Cape/Agulhas Basin and the East Georgia Basin. These three areas have in common elevated biological productivity and elevated organic carbon contents in their sediments, which coupled with their depth near or below the CCD would help to explain the composition of their nodules. However, Ni and Cu contents are lower in them than in areas of diagenetic nodules in the Pacific and Indian Oceans.

Economic Potential Interest in manganese nodules commenced around the mid-1960s and developed during the 1970s, at the same time as the Third United Nations Law of the Sea Conference. However, the outcome of that Conference, in 1982, was widely regarded as unfavorable for the mining industry. This, coupled with a general downturn in metal prices, resulted in a lessening of mining company interest in nodules. About this time, however, several governmentbacked consortia became interested in them and this work expanded as evaluation of the deposits by mining companies declined. Part 11 of the 1982 Law of the Sea Convention, that part dealing with deepsea mining, was substantially amended in an agreement on 28 July 1994 which ameliorated some of the provisions relating to deep-sea mining. The Convention entered into force in November 1994. During the 1980s interest in manganese nodules in exclusive economic zones (EEZs) started to increase. An important result of the Third Law of the Sea Conference, was the acceptance of a 200-nauticalmile EEZ in which the adjacent coastal state could claim any mineral deposits as their own. The nodules found in EEZs are similar to those found in adjacent parts of the International Seabed Area, and are of greatest economic potential in the EEZs of the South Pacific. At the beginning of the twenty-first century, the out-look for manganese nodule mining remains rather unclear. It is likely to commence some time in this century, although it is not possible to give a precise estimate as to when. The year 2015 has been suggested as the earliest possible date for nodule mining outside of the EEZs. It is possible, however, that EEZ mining for nodules might commence earlier

if conditions were favorable. It would depend upon many factors; economic, technological, and political.

Discussion A model to explain the compositional variability of nodules in the Penrhyn Basin can be summarized as follows. Under the flanks of the high productivity area, reduced sedimentation rates near the CCD due to calcium carbonate dissolution enhance the content of metal-bearing organic carbon rich phases (fecal material, marine snow, etc.) in the sediments, the decay of which drives the diagenetic reactions that in turn promote the enrichment of Mn, Ni, and Cu in the nodules via the sediment interstitial waters. Away from the CCD, organic carbon concentrating processes are less effective. Further south as productivity declines, there is probably insufficient organic carbon supplied to the seafloor to promote the formation of diagenetic nodules at any depth. Under the equator, siliceous ooze replaces pelagic clay as the main sediment builder at and below the CCD, and when its rate of accumulation is high it dilutes the concentrations of organic carbon-bearing material at all depths to levels below that at which diagenetic Mn, Ni, and Cu rich nodules can form. To a greater or lesser extent, this model can account for much of the variability in nodule composition found in the other South Pacific areas described, although local factors may also apply. In the Peru Basin, as in the Penrhyn Basin, diagenetic Mn-, Ni-, and Cu-rich nodules are concentrated near the CCD and their Ni and Cu contents reach a maximum south of the highest productivity waters. In the Tiki Basin, the greatest diagenetic influences are also found in the north of the Basin. As the South Pacific basins deepen to the west, the areas of diagenetic nodules tend to occur below the CCD as, for example, in the Nova Canton Trough area. This may be because the settling rates of large organic particles are quite fast in the deep ocean. Probably only limited decay of this material takes place between it settling through the CCD and reaching the seafloor, and enough probably gets sedimented to extend the depth of diagenetic nodule formation to well below the CCD under high productivity waters where there is limited siliceous sediment accumulation. In the North Pacific, the trends in nodule composition in relation to the equatorial zone are the mirror image of those in the south. Thus in both the Central Pacific Basin and the Clarion–Clipperton Zone the highest nodule grades occur in diagenetic nodules on the northern flanks of the high

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGANESE NODULES

productivity area and decline both to the north and south. The general model erected to explain the Penrhyn Basin nodule variability thus probably applies, at least in part, to these areas also. The model also has some applicability in the Indian Ocean but less in the Atlantic. In the Indian Ocean, diagenetic nodules associated with sediments containing moderate amounts of organic carbon occur resting on siliceous ooze to the south of the equatorial zone in the Central Indian Ocean Basin. Farther to the south these nodules give way to hydrogenous varieties resting on pelagic clay. However, in the north the changes in nodule composition that might be expected under higher productivity waters do not occur, probably because terrigenous sedimentation becomes important in those areas which in turn reduces the Mn, Ni, and Cu content of nodules. In the Atlantic, the influence of equatorial high productivity on nodule composition that is evident in the Pacific is not seen, mainly because the seafloor in the equatorial area is largely above the CCD. Where diagenetic nodules do occur, as in the Angola, Cape and East Georgia Basins, productivity is also elevated, but the seafloor is near or below the CCD leading to reduced sedimentation rates.

Conclusions Manganese nodules, although not being mined today, are a considerable resource for the future. They consist of ferromanganese oxides variably enriched in Ni, Cu, and other metals. They generally accumulate around a nucleus and exhibit internal layering on both a macro- and microscale. Growth rates are generally slow. The most potentially economic varieties of the deposits occur in the subequatorial Pacific under the flanks of the equatorial

495

zone of high biological productivity, at depths near the CCD. Similar nodules occur in the Indian Ocean under similar conditions.

See also Authigenic Deposits. Hydrothermal Vent Biota. Hydrothermal Vent Ecology. Hydrothermal Vent Fluids, Chemistry of.

Further Reading Cronan DS (1980) Underwater Minerals. London: Academic Press. Cronan DS (1992) Marine Minerals in Exclusive Economic Zones. London: Chapman and Hall. Cronan DS (ed.) (2000) Handbook of Marine Mineral Deposits. Boca Raton: CRC Press. Cronan DS (2000) Origin of manganese nodule ‘ore provinces’. Proceedings of the 31st International Geological Congress, Rio de Janero, Brazil, August 2000. Earney FC (1990) Marine Mineral Resources. London: Routledge. Glasby GP (ed.) (1977) Marine Manganese Deposits. Amsterdam: Elsevier. Halbach P, Friedrich G, and von Stackelberg U (eds.) (1988) The Manganese Nodule Belt of the Pacific Ocean. Stuttgart: Enke. Nicholson K. Hein J, Buhn B, Dasgupta S (eds.) (1997) Manganese Mineralisation: Geochemistry and Mineralogy of Terrestrial and Marine Deposits. Geological Society Special Publication 119, London. Roy S (1981) Manganese Deposits. London: Academic Press. Teleki PG, Dobson MR, Moore JR, and von Stackelberg U (eds.) (1987) Marine Minerals: Advances in Research and Resource Assessment. Dordrecht: D. Riedel.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGROVES M. D. Spalding, UNEP World Conservation Monitoring Centre and Cambridge Coastal Research Unit, Cambridge, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1533–1542, & 2001, Elsevier Ltd.

highly varied salinities, often approaching hypersaline conditions. Soils may be shallow, but even where they are deep they are usually anaerobic within a few millimeters of the soil surface. Many mangrove species show one or more of a range of physiological, morphological or life-history adaptations in order to cope with these conditions.

Definition

Coping with Salt

The term mangrove is used to define both a group of plants and also a community or habitat type in the coastal zone. Mangrove plants live in or adjacent to the intertidal zone. Mangrove communities are those in which these plants predominate. Other terms for these communities include coastal woodland, intertidal forest, tidal forest, mangrove forest, mangrove swamp and mangal. The word mangrove can be clearly traced to the Portugese word ‘mangue’ and the Spanish word ‘mangle’, both of which are actually used in the description of the habitats, rather than the plants themselves, but still have been joined to the English word ‘grove’ to give the word ‘mangrove.’ It has been suggested that the original Portugese word has been adapted from a similar word used locally by the people of Senegal, however an alternative derivation may be the word ‘manggimanggi’, which is still used in parts of eastern Indonesia to describe one genus (Avicennia).

All mangroves are able to exclude most of the salt in sea water from their xylem. The exact mechanisms for this remain unclear, but it would appear to be an ultrafiltration process operating at the endodermis of the roots. One group, which includes Bruguiera, Lumnitzera, Rhizophora and Sonneratia, is highly efficient in this initial salt exclusion and shows only minor further mechanisms for salt secretion. A second group, which includes Aegialitis, Aegiceras and Avicennia appear to be less efficient at this initial salt exclusion and hence also need to actively secrete salt from their leaves. This is done metabolically, using special salt glands on the leaf surface. The salt evaporates, leaving crystals which may be washed or blown off the leaf surface. In these latter species such exuded salt is often visible on the leaf surface (Figure 1).

Mangrove Species Mangrove plants are not a simple taxonomic group, but are largely defined by the ecological niche where they live. The simplest definition describes a shrub or tree which normally grows in or adjacent to the intertidal zone and which has developed special adaptations in order to survive in this environment. Using such a definition a broad range of species can be identified, coming from a number of different families. Although there is no consensus as to which species are, or are not, true mangroves, there is a core group of some 30–40 species which are agreed by most authors. Furthermore, these ‘core’ species are the most important, both numerically and structurally, in almost all mangrove communities. Table 1 lists a large range of mangrove species (of tree, shrub, fern, and palm), and highlights those which might be regarded as core species. All of these plants have adapted to a harsh environment, with regular inundation of the soil and

496

Anaerobic Soils

The morphological feature for which mangroves are best known is the development of aerial roots. These have developed in most mangrove species in order to cope with the need for atmospheric oxygen at the absorbing surfaces and the impossibility of obtaining such oxygen in an anaerobic and regularly inundated environment. Various types of roots are illustrated in Figure 1. The stilt root, exemplified by Rhizophora (Figure 1B) consists of long branching structures which arch out away from the tree and may loop down to the soil and up again. Such stilt roots also occur in Bruguiera and Ceriops although in older specimens they fuse to the trunk as buttresses. They also occur sporadically in other species, including Avicennia. A number of unrelated groups have developed structures known as pneumatophores which are simple upward extensions from the horizontal root into the air above. These are best developed in Avicennia and Sonneratia (Figure 1C), the former typically having narrow, pencil-like pneumatophores, the

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGROVES

Table 1

497

List of mangrove species: species in bold typeface are those which are considered ‘core’ species

Family

Species

Pteridaceae

Acrostichum aureum Acrostichum danaeifolium Acrostichum speciosum Aegialitis annulata Aegialitis rotundifolia Pelliciera rhizophorae Camptostemon philippensis Camptostemon schultzii Heritiera fomes Heritiera globosa Heritiera littoralis Diospyros ferrea Aegiceras corniculatum Aegiceras floridum Cynometra iripa Mora oleifera Conocarpus erectus Laguncularia racemosa Lumnitzera littorea Lumnitzera racemosa Lumnitzera  rosea Pemphis acidula Osbornia octodonta Sonneratia alba Sonneratia apetala Sonneratia caseolaris Sonneratia griffithii Sonneratia lanceolata Sonneratia ovata Sonneratia  gulngai Sonneratia  urama Bruguiera cylindrica Bruguiera exaristata Bruguiera gymnorrhiza Bruguiera hainesii

Plumbaginaceae Pellicieraceae Bombacaceae Sterculiaceae

Ebenaceae Myrsinaceae Caesalpiniaceae Combretaceae

Lythraceae Myrtaceae Sonneratiaceae

Rhizophoraceae

Family

latter with secondary thickening so that they can become quite tall and conical. One adaptation on the theme of pneumatophores is that of root knees where more rounded knobs are observed to extend upwards from the roots. In Xylocarpus mekongensis these are simply the result of localized secondary cambial growth, but in Bruguiera (Figure 1D) and Ceriops they are the result of a primary looping growth. In these species branching may also occur on these root knees. Buttress roots are a common adaptation of many tropical trees, but in Xylocarpus granatum (Figure 1E) and to some degree in Heritiera such flange-like extensions of the trunk continue into plank roots which are vertically extended roots with a sinuous plank-like form extending above the soil. The surfaces of the aerial roots are amply covered with porous lenticels to enable gaseous exchange, and the internal structure of the roots is highly

Euphorbiaceae Meliaceae

Avicenniaceae

Acanthaceae Bignoniaceae Rubiaceae Arecaceae

Species Bruguiera parviflora Bruguiera sexangula Ceriops australis Ceriops decandra Ceriops tagal Kandelia candel Rhizophora apiculata Rhizophora harrisonii Rhizophora mangle Rhizophora mucronata Rhizophora racemosa Rhizophora samoensis Rhizophora stylosa Rhizophora  lamarckii Rhizophora  selala Excoecaria agallocha Excoecaria indica Aglaia cucullata Xylocarpus granatum Xylocarpus mekongensis Avicennia alba Avicennia bicolor Avicennia germinans Avicennia integra Avicennia marina Avicennia officinalis Avicennia rumphiana Avicennia schaueriana Acanthus ebracteatus Acanthus ilicifolius Dolichandrone spathacea Tabebuia palustris Scyphiphora hydrophyllacea Nypa fruticans

adapted, with large internal gas spaces, making up around 40% of the total root volume in some species. It is further widely accepted that there must be some form of ventilatory mechanism to aid gaseous exchange. A system of tidal suction is the probable mechanism in most species: during high tides, oxygen is used by the plant, while carbon dioxide is readily absorbed in the sea water, leading to reduced pressure within the roots. As the tide recedes and the lenticels open, water is then sucked into the roots. Seeds and Seedlings

Establishment of new mangrove plants in the unstable substrates and regular tidal washing of the mangrove environment presents a particular evolutionary challenge. All mangroves are dispersed by water and particular structures in the seed or the fruit are adapted to support flotation. In a number of

(c) 2011 Elsevier Inc. All Rights Reserved.

498

MANGROVES

groups a degree of vivipary is observed which is unusual in most nonmangroves. The Rhizophoraceae have developed this to its fullest extent and here the embryo grows out of the seed coat and then out of the fruit while still attached to the parent plant, so that the propagule which is eventually released is actually a seedling rather than a seed (Figure 1F). In a number of other groups, including Aegiceras, Avicennia, Nypa and Pelliciera cryptovivipary exists in which the embryo emerges from the seed coat, but not the fruit, prior to abcission. Longevity of seedlings is clearly important for many species. Most species are able to survive (float and remain viable) for over a month, whereas some Avicennia propagules have been shown to remain viable for over a year while in salt water.

Distribution and Biogeography As a result of their restriction to intertidal areas, mangroves are limited in global extent (Figure 2), and are, in fact, one of the most globally restricted of all forest types. Figure 2 clearly shows the absolute limits to mangrove distribution. Mangroves are largely confined to the regions between 301 north and south of the equator, with notable extensions beyond this to the north in Bermuda (321200 N) and Japan (311220 N), and to the south in Australia (381450 S), New Zealand (381030 S) and South Africa (321590 S). Within these confines they are widely distributed, although their latitudinal development is restricted along the western coasts of the Americas and Africa. In the Pacific Ocean natural mangrove

(A)

(C)

(B)

(D)

Figure 1 Mangrove adaptations: (A) salt crystals secreted onto the surface of a leaf, Avicennia; (B) stilt roots of Rhizophora; (C) pneumatophores in Sonneratia; (D) root knees in Bruguiera; (E) plank roots in Xylocarpus; (F) Rhizophora propagule.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGROVES

499

1. There is a distinct region of very high mangrove diversity, sometimes referred to as the ‘diversity anomaly’, centered over south-east Asia. 2. Away from this high diversity region mangroves generally show relatively even levels of low diversity, although there is smaller peak of diversity around southern Central America. 3. There is a wide area of the central and western Pacific Ocean from 120 to 1601W where mangroves do not occur. 4. Even in the area of highest mangrove diversity there is a very rapid latitudinal decline in species numbers away from the tropics. One further observation, which is not fully illustrated in the figures, concerns the division of the global mangrove flora into two highly distinct subregions. An eastern group (sometimes known as the Indo-West Pacific) forms one vast and contiguous block stretching from the Red Sea and East Africa to the central Pacific. This group has a totally different species composition from the western group (the Atlantic–East Pacific or Atlantic–Caribbean–East Pacific), which includes both Pacific and Atlantic shores of the Americas, the Caribbean and the shores of West Africa. A number of these patterns are explored more fully, below. Latitudinal Patterns

Mangroves limits are closely correlated to minimum temperature requirements. There is only one genus (Avicennia) which survives in environments where frosts may occur, but many species appear to have their latitudinal limits set by less extreme cold temperatures; air temperatures of 51C appear inimical to most mangrove species. Sea-surface temperatures may be more important than air temperatures for some species. The 241C mean annual isotherm appears to be the minimum water temperature tolerated by mangroves in most areas, although this

Figure 1 Continued

communities are limited to western areas, and they are absent from many Pacific islands. In all, an estimated 114 countries and territories have mangroves, however for many nations the total area is very small indeed, and the total global area of these forests is only 181 000 km2. Table 2 provides a summary of total mangrove areas by region. Although these statistics suggest a relatively wide distribution, the distribution of individual species within these areas is clearly far more restricted, and Figure 3 provides a plot of mangrove biodiversity patterns. A number of points of particular interest are clearly illustrated.

Table 2

Total mangrove area by region

Region

South and south-east Asia Australasia The Americas West Africa East Africa and the Middle East Total area

Area (km2)

75 18 49 27 10

173 789 096 995 024

Proportion of global total 41.5% 10.4% 27.1% 15.5% 5.5%

181 077

Data calculated from best available national sources in Spalding et al. (1997).

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved.

5

4

5 6 9 9

5

6

3

7

12

5

5

11

5

4

6

4

2

3

7

6

4

5

7

4 5

7

1

3 7 6 4 6 5 4 6

6

4

8 8 2 7 2 7

6

6

3

10

2

9

3

8

1

3

9

6

2

1 1

2

9

1

2 4

9

5

2

20 23

24 20

21 24

35

5

36

29 29

7

45

23

30

11

29

13

37

44

14

10

3

8

4

15 16

22

2

1

9

2

4

8

3 1

3

Figure 3 A global map of mangrove diversity plotting contours of equal diversity (1–5 species, 6–10, 11–15, 16–20, 21–25, 26–30, 31–35, 36–40, and 41–45). (Reproduced with permission from Spalding (1998).)

5

Figure 2 The global distribution of mangrove forests (data kindly provided by the UNEP World Conservation Monitoring Centre).

500 MANGROVES

MANGROVES

minimum is closer to 271C on the north Atlantic coasts of America and Africa, and may be much lower in areas such as southern Japan. The relatively low-latitude limits to mangrove in Peru and Angola are probably related to the cold water currents which affect these coastlines. Eastern and Western Floras

The division of mangroves into two distinct floras is almost complete at the species level. Of the species listed in Table 1, three genera, but only one species, are shared between the two regions. In addition to having distinctive floras, the overall niche-space occupied by the two floras differs, with western mangroves restricted to higher intertidal and downstream estuarine locations than those of the eastern group. None of these differences can be related to contemporary ecology, and they are clearly of historical origin. Mangroves have a considerable known history, with the oldest of the modern taxa, Nypa, being recorded from the Cretaceous (69 million years BP) and Pellicera and Rhizophora dating back to the Eocene (30 million years BP). Information on the centers of origin and subsequent distribution patterns of mangroves is still unclear, and it is likely, given their disparate taxonomic origins, that mangroves evolved independently in a number of localities. Despite this, a number of authors have suggested that the majority of mangrove species have an eastern Tethys Sea origin with dispersal north and westwards (through a proto-Mediterranean) into the Atlantic and then via the Panama gap into the eastern Pacific. Whatever mechanisms may have operated, the climatic conditions, which were once suitable for a pan-Tethyan flora, changed. With the cooling and closure of the Mediterranean from the Tethys Sea the mangrove floras were separated. Divergence of the two communities then occurred through one or more of a number of mechanisms, including natural process of genetic drift and separation, possible extinction and radiation. It is clear that the Atlantic Ocean and the isthmus of Panama now represent insurmountable barriers to mangrove dispersal, however the closeness of the floras on either side of these barriers reflect the relatively short geological period over which these barriers have been in place. The Diversity Anomaly

Apart from having quite distinctive faunas, the eastern mangroves have a much greater diversity than the western group. Of the species listed in Table 1, only 13 are found in the western group, whereas 59 are found in the eastern group. This

501

‘diversity anomaly’ is reflected not only in regional statistics, but also at local scales, with individual sites in the west typically having lower species counts than equivalent sites in the east. A number of theories have been propounded to explain this. It has been suggested, for example, that if most mangroves had originated in the eastern Tethys Sea the western flora may be depauperate simply as a result of being an immigrant flora. Alternatively the harsh environmental conditions during the Pleistocene, with significant temperature and sea-level fluctuation may have driven the extinction of a number of western mangroves. By contrast the Indo-West Pacific with its long and complex coastline is known to have had at least pockets of benign climatic conditions over geological timescales. These refugia may have allowed for further allopatric speciation events during periods of isolation from other areas, followed by periods of recombination with other areas as conditions ameliorated. The relatively rapid tailing off of diversity westwards from Southeast Asia has been related to the relatively harsh climatic conditions which still prevail over much of this area, and the very large distances between more suitable localities for mangroves preventing recolonization. Similarly the absence of mangroves from the central and eastern Pacific is related to the very long distances between areas of suitable habitat. There is some evidence that mangroves may once have been more widespread in the Pacific, but, if this is the case, their disappearance from certain islands is probably explained by the climatic and eustatic changes of the Pleistocene. Given the good dispersal ability of many mangrove species, distances must be very large indeed to prevent colonization, but it has been suggested, given the relatively short time since the beginning of the last interglacial, that mangrove communities may currently be in a state of expansion.

Biodiversity Patterns at Finer Resolutions: Zonation and Succession Numerous localized ecological factors influence the occurrence and growth-patterns of mangroves. In addition to factors which affect the majority of plant species, such as water, nutrients, drainage, and soiltype, significant further influence is produced by salinity and tidal influence. Considerable efforts have been made to define patterns of zonation in mangrove communities, and although such patterns do occur in many communities, the enormous variation in local

(c) 2011 Elsevier Inc. All Rights Reserved.

502

MANGROVES

conditions makes the preparation of simple summaries of ‘typical’ zonation patterns very difficult. In many tidal areas the regular drying out of the soil, often coupled with patterns of restricted water circulation, or high rates of evaporation, serves to increase salinities to considerably higher levels than the surrounding sea water. This is particularly the case in areas of back mangrove where tidal flushing is less frequent and water circulation may be more restricted. It is further exacerbated in arid regions. In many areas this leads to the development of wide areas of stunted mangroves, or even bare salt-pans where mangroves cannot grow. This situation is diminished, or even reversed in areas where the freshwater input is more considerable, either from high rainfall or terrestrial runoff, or in some estuarine environments. The tides also exert influences in other ways, most notably through inundation, but also through their influence on soils. Different mangrove species show quite different tolerances to inundation. Species such as Avicennia and Rhizophora, which are relatively tolerant of frequent and quite high tidal waters, typically form the most seaward zone of the mangrove system. Tides also influence the soil through the delivery or removal of nutrients, and also the resorting of sediments. Typically finer sediments are found at higher locations in the tidal frame, whereas coarser sediments tend to be deposited or re-distributed lower down. Once again, the complexity of interactions is highly varied between localities. In many cases mangrove communities may follow a succession and this has been linked to the process of terrestrial advancement (coastal progradation). The patterns shown in zonation often provide a spatial model for such a temporal succession, starting with the more inundation and salt-tolerant species. These are able to bind nutrients and sediments, gradually raising their position in the tidal frame such that they are then replaced by those species requiring slightly less saline and inundated conditions, and then by mangrove associates and then nonmangrove species. Such successional processes occur in many areas; in parts of Southeast Asia where there is a high input of allochthonous material, rates of coastal advancement have been recorded at 120–200 m year1. In other areas, however, the notion of mangroves ‘creating land’ is clearly not valid and mangroves show a range of responses to differing impacts of waves, climate, and sediments. In the Florida Everglades there is considerable evidence for the movement of mangrove communities both landwards and seawards, depending on sea-level changes and it may be more accurate to regard mangroves in these areas as

opportunistic followers of sedimentation and substrate or elevation changes.

Humans and Mangroves Humans have lived in close contact with mangrove communities for millennia and in many cases have made considerable use of this association. Archaeological sites have been located which demonstrate human presence in mangrove areas in Venezuela dating back 5000–6000 years, and there is an Egyptian inscription dating back to the time of King Assa (3580–3536 BC) which mentions mangroves. Countries of the Middle East began a vigorous trade in mangrove timber from about the ninth century, largely for boat-building, exporting from outposts along the shores of East Africa. The European nations became involved in the utilization of mangrove bark as a source of tannins, particularly from the Americas from the sixteenth century. The earliest record of mangrove protection dates to an edict from the King of Portugal in 1760 who restricted the cutting of mangroves for timber in Brazil unless their bark was also used for tannins. Despite such early concerns, the overexploitation of mangroves began in earnest towards the middle of the twentieth century and is continuing, and in many areas accelerating at the present time. In many areas mangroves are highly productive and their location on the coastline places them in a zone where many other human activities have, until recently, been somewhat restricted. At the same time they often exist in close proximity to centers of human population, and can be relatively easily approached by sea or land. This makes their utilization inevitable in many areas, although the degree of sustainability of such use is highly variable. Utilization of Mangroves

Timber and wood products One of the commonest uses of mangroves is as a source of wood. Mangrove wood is often used for fuel either directly or after conversion into charcoal. The former is widespread among artisanal communities worldwide, the latter often for commercial purposes. Mangrove wood is also used for timber; the relatively small size of mangrove trees in many areas has meant that the primary usage of timber is the preparation of timber poles for fencing, housing construction, making of fish-traps and other activities. Larger trees can be utilized for preparation of planking, and indeed some species have a very high value associated with their dense wood and resistance to rot, which is important for construction of houses and boats (both for local

(c) 2011 Elsevier Inc. All Rights Reserved.

MANGROVES

and commercial use). Further industrial use of mangrove wood is in the production of wood-pulp for the paper industry, and chipboard. Fisheries Mangroves and the associated channels which run between them, are important areas of fish productivity. Numerous species inhabit mangrove areas and form the basis of artisanal and commercial fisheries, including crab, prawn, and mollusk fisheries. Mangrove areas are also widely used by a number of offshore fish species which are of commercial importance. These species, which include some highly profitable shrimp species, use mangrove areas for spawning or as a nursery ground and loss of mangrove areas has severe negative impacts on fishery productivity. Cagebased fisheries have been established in many of the wider channels, and mangrove areas are widely used for the capture of juvenile prawns for transfer to aquaculture ponds. In recent years, wide areas of mangrove forest have been cut down in the development of intertidal aquaculture ponds, particularly in south-east Asia. Although this is a highly profitable industry, poor planning has led to the rapid and virtually irrevocable degradation of many of these ponds after only a few years. Rehabilitation of these lands is rarely undertaken with the result that local communities lose a source of valuable natural resources, and the shrimp pond developers move on to new areas. Coastal protection The important role which mangroves play in the stabilization of coastal sediments and the reduction of coastal erosion has already been mentioned. This role is frequently overlooked until such time as the mangroves are removed and major storm events hit coastlines. The massive and devastating cyclones which regularly impact the coastline of the Bay of Bengal have drawn particular attention to these issues and in a number of localities around the globe there are now efforts to establish mangrove plantations precisely to stabilize sediments and reduce the impact of storm surges. Alongside these three key areas of human importance, mangroves are regularly utilized for other purposes, a number of which are outlined in Table 3. It is highly difficult to place values on many of these uses and functions of mangroves. Apart from direct utilization of wood products, the link between particular products or functions and the mangrove communities which provide them is rarely made. Furthermore, for numerous communities the value in economic terms is greatly enhanced by the social value, providing a source of employment, protein and protection for some of the world’s poorest communities.

Table 3

503

Minor or regionally restriced uses of mangroves

Honey production Fodder Recreation

Thatch and matting

Tannin extraction

Traditional medicine Food

An important economic activity in some countries For cattle, camels and goats, notably in India and Pakistan Walkways, boat-based tours and other visiting facilities have been established for tourists and local communities in some areas, notably Trinidad, Bangladesh, and Australia Primarily from the leaves of the mangrove palm Nypa fruticans in south-east Asia and from introduced populations in West Africa Formerly widespread, this activity has become less significant as synthetic products have become available Still widespread in many traditional communities Nypa fruticians is widely used for the production of sugar, alcohol and vinegar. Fruits of Avicennia, Kandelia and Bruguiera are used as a source of food in some countries

Overexploitation and Loss

Mention has already been made of the widespread loss of mangrove communities worldwide. Apart from conversion into aquaculture ponds, much of this is related to land reclamation activities for agriculture and for urban and industrial development, and large areas have also been severely degraded or removed by commercial timber companies or through overexploitation by local communities. Some further degradation or loss has been related to human-induced changes to the water regime (including upstream dams leading to reductions in sedimentation at river mouths), pollution (mangroves are particularly sensitive to oil spills), and conversion into salt pans for industrial salt production. To date there is no globally available figure for total mangrove loss, however national loss statistics are available for a number of countries. In south-east Asia, for example, the loss figures for four countries are: Malaysia, 12% from 1980 to 1990; the Philippines, a 60% loss from 4000 km2 originally to 1600 km2 today; Thailand, a 55% loss from 5500 km2 in 1961 to 2470 km2 in 1986; and Vietnam, a 37% loss from 4000 km2 originally to 2525 today. These figures alone suggest a total of some 7445 km2 of mangrove loss, representing over 4% of the current global total. The four countries concerned have certainly suffered significant mangrove loss, but they are not alone. Sea-level rise associated with global climate change must also be considered as a significant threat

(c) 2011 Elsevier Inc. All Rights Reserved.

504

MANGROVES

to mangrove ecosystems. It is important to note that the impacts of proposed changes (most models predict rises in sea-level of 30–100 cm by 2100) are relatively insignificant in some areas where high levels of sediment movement and deposition will counter such rises, or where other eustatic changes, such as those associated with tectonic movements, will remove or further enhance changing sea-level effects. Furthermore, mangrove species and communities are highly opportunistic and will colonize new areas with some rapidity. Sea-level rise remains a problem, however, as mangrove communities in many areas may become squeezed out as sea-level rise forces mangrove communities landwards, but human use prevents landward migration.

ing. The Matang Mangrove Reserve in Malaysia is perhaps the best-known example. Studies have shown combined benefits arising from timber and fuelwood products (notably charcoal), but even more importantly from a large nearshore fishery (directly or indirectly providing employment for over 4000 people), from aquaculture on the mud flats below the mangroves, and from tourism. It is rare that such holistic studies have been carried out. Often the human benefits provided by mangrove fall between several sectors of the economy, fisheries, forestry, tourism, and coastal protection, and their combined benefits are not realized. A better perception of these benefits would undoubtedly lead to much wider-scale protection for mangroves globally.

Protection and Plantation

Despite the massive losses which mangrove communities have gone through in the past decades there have also been concerted efforts to protect them in some areas, and the growing realization of their value has led to widespread efforts to utilize mangroves in a more sustainable manner, and in some places large areas of mangrove plantations have now been established. Worldwide, there are currently an estimated 850 protected areas with mangroves spread between 75 countries, which are managed for conservation purposes. These cover over 16 000 km2 of mangrove, or 9% of the global total. Although this is a far higher proportion than for many other forest types, active protection is absent from many of these areas, and the remaining unprotected sites are probably more threatened than many other forest types because of their vulnerability to human exploitation. Increasing recognition of the various values of mangrove forests is leading to widescale mangrove plantation in some areas, for coastal defence, as a source of fuel, or for fisheries enhancement. Plantations in Bangladesh, Vietnam, and Pakistan now cover over 1700 km2, and Cuba is reported to have planted some 257 km2 of mangroves. Overall, however, when weighed against the statistics of mangrove loss, the area of such plantations remains insignificant. Active management of these and other existing mangrove areas for economic production is increas-

See also Coastal Topography, Human Impact on. Coastal Zone Management. Crustacean Fisheries. Fisheries Overview. Sea Level Change.

Further Reading Chapman VJ (1976) Mangrove Vegetation. Vaduz, Germany: J Cramer. Field CD (1995) Journey Amongst Mangroves. Okinawa, Japan: International Society for Mangrove Ecosystems. Field CD (ed.) (1996) Restoration of Mangrove Ecosystems. Okinawa, Japan: International Society for Mangrove Ecosytems. Robertson AI and Alongi DM (eds.) (1992) Coastal and Estuarine Studies, 41: Tropical Mangrove Ecosystems. Washington, USA: American Geophysical Union. Saenger P, Hegerl EJ, and Davie JDS (eds.) (1983) IUCN Commission on Ecology Papers, 3: Global Status of Mangrove Ecosystems. Gland, Switzerland: IUCN (The World Conservation Union). Spalding MD (1998) Biodiversity Patterns in Coral Reefs and Mangrove Forests: Global and Local Scales. PhD Dissertation, University of Cambridge. Spalding MD, Blasco F, and Field CD (eds.) (1997) World Mangrove Atlas. Okinawa, Japan: International Society for Mangrove Ecosystems. Tomlinson PB (1986) The Botany of Mangroves. Cambridge, UK: Cambridge University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, DEEP WATER H. Hotta, H. Momma and S. Takagawa, Japan Marine Science & Technology Center, Japan Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1542–1549, & 2001, Elsevier Ltd.

Introduction Deep-ocean underwater investigations are much more difficult to carry out than investigations on land or in outer space. This is because electromagnetic waves, such as light and radio waves, do not penetrate deep into sea water, and they cannot be used for remote sensing and data transmission. Moreover, deep-sea underwater environments are physically and physiologically too severe for humans to endure the high pressures and low temperatures. First of all, pressure increases by 1 atmosphere for every 10 meters depth because the density of water is 1000 times greater than that of air. Furthermore, as we have no gills we can not breathe under water. Water temperature decreases to 11C or less in the deep sea and there is almost no ambient light at depth because sunlight can not penetrate through more than a few hundred meters of sea water. These are several of the reasons why we need either manned or unmanned submersibles to work in the deep sea. A typical manned submersible consists of four major components: a pressure hull, propellers (thrusters), buoyant materials, and observational instruments. The pressure hull is a spherical shell made of high-strength steel or titanium. The typical internal diameter of the hull is approximately 2 m, which allows up to three people to stay at one atmosphere for 8–12 h during underwater operations. In case of emergency, a life-support system enables a stay of three to five days. Several thrusters are usually installed on the body of the submersible to give maneuverability. The buoyant material is syntactic foam, which is made of glass microballoons and an adhesive matrix. Its specific gravity is approximately 0.5 gf ml 1. Observational instruments such as cameras, lights, sonar, CTD (conductivity, temperature, and depth sensors) etc., are also very important for gathering information on the deep-sea environment. It should be mentioned that the power consumption of the lights can reach as much as 15% of the total power consumption of the submersible.

History of Deep Submersibles The first modern deep diving by humans, to a depth of 923 m, was achieved in 1934 by William Beebe, an American zoologist, and Otis Burton using the bathysphere, which means ‘deep sphere’. The bathysphere was a small spherical shell made of cast iron, 135 cm in inside diameter designed for two observers. The bathysphere had an entrance hatch and a small glass view port. As the sphere was lowered by a cable and lacked thrusters, it was impossible to maneuver. The next advance, using a free-swimming vehicle, occurred after World War II, in 1947. The bathyscaph FNRS II was invented by Auguste Piccard, who had been studying cosmic rays using a manned balloon in Switzerland. The principle of the bathyscaph was the same as that of a balloon. Instead of hydrogen or helium gas, gasoline was used as the buoyant material. During descent, air ballast tanks were filled with sea water, and for ascent, iron shot ballast was released. The pressure hull was made of drop-forged iron hemispheres, 2 m in inside diameter and 90 mm in thickness, allowing for two crew members. It was able to maneuver around the seafloor by thrusters driven by electric motors. Later, the second bathyscaph, Trieste, was sold to the US Navy, and independently at the same time, the French Navy developed the bathyscaph FNRS III, and later Archimede. In 1960, the Trieste made a dive into the Challenger Deep in the Mariana Trench, to a depth of 10 918 m. This historic dive was conducted by Jacques Piccard, son of Auguste Piccard, and Don Walsh from the US Navy. The bathyscaph was the first generation of deep-diving manned submersibles. It was very big and slow as it needed more than 100 kiloliter capacity gasoline tanks to provide flotation for the 2 m diameter pressure hull. In 1964, the second generation of deep submersibles began. Alvin was funded by the US Navy under the guidance of the Woods Hole Oceanographic Institution (WHOI). At first, its depth capability was only 1800 m. It was small enough to be able to put on board the R/V Lulu, which became its support ship. Instead of gasoline flotation, syntactic foam was used. Alvin had horizontal and vertical thrusters to maneuver freely in three dimensions. Scientific instruments, including manipulators, cameras, sonar and a navigation system, were installed. Three observation windows were available for the three crew members. In France, the two-person 3000

(c) 2011 Elsevier Inc. All Rights Reserved.

505

506

MANNED SUBMERSIBLES, DEEP WATER

m-class submersible Cyana was built. These two vehicles typified submersibles during the 1960s. At present, the depth capability of the Alvin has been increased to 4500 m by replacing the high-strength steel pressure hull with a titanium alloy sphere in 1973. In the 1980s, 6000 m-class submersibles, such as the Nautile from France, the Sea Cliff of the US Navy, the Mir I and Mir II from Russia and the Shinkai 6500 from Japan, were built. They were theoretically able to cover more than 98% of the world’s ocean floor. What will the third generation of deep submersibles be like? Manned submersibles of the third generation, which would be capable of exceeding 10 000 m depth, have not yet been developed at the time of this report. One possibility is a small and highly maneuverable one- or two-person submersible with a transparent acrylic or ceramic pressure hull. Another possibility is a deep submergence laboratory, which would be able to carry several scientists and crew long distances and long durations without the assistance of a mother ship. This would be the realization of the dream like ‘Nautilus’ in 20 000 Leagues Under The Sea by French novelist Jules

Verne. Strong scientific and/or social goals would be needed for such a submersible design to be pursued. And there is a third possibility that the next generation will be evolutionary upgrades of existing second-generation submersibles.

Principles of Modern Submersibles Descent and Ascent

There are several methods to submerge vehicles into the deep sea. The simplest way is to suspend a sphere by a cable, known as a bathysphere. Mobility, however, is greatly limited. A second method relies on powerful thrusters to adjust vertical position in relatively shallow water. The submersible Deep Flight is a high-speed design which uses thrust power coupled with fins for motion control like the wings of a jet fighter. It descends and ascends obliquely in the water column at speeds up to 10 knots. Most submersibles employ a third method that, while using weak thrusters to control attitude and horizontal movement, relies principally on an adjustable buoyancy system for descent and ascent (Figure 1).

Vent air and fill sea water in ballast tank to descend

Variable ballast tank (Air and sea water) Drop weights Sea water Blow out sea water from ballast tank Descent

Ascent

Jettison drop weights partially for neutral buoyancy

Jettison all the drop weight to ascend Sea water

Weight control by variable ballast tank (Air and sea water)

Figure 1 Principle of descent and ascent for a modern deep submersible.

(c) 2011 Elsevier Inc. All Rights Reserved.

)

MANNED SUBMERSIBLES, DEEP WATER

Water density (gf ml

_1

1.08

1.05

1.02

Water pressure (MPa) 50

100

Water depth (m)

0 Water pressure is affected by compression of sea water 5000 Simple calculation of water pressure (proportional to water depth)

10 000

Figure 2 Relations between water depth, pressure and water density at a water temperature of 01C and salinity of 34.5%.

When on the surface, the submersible’s air ballast tank is filled with air creating positive buoyancy, hence it floats. When the dive begins, air is vented from the ballast tank and filled with sea water, thus creating negative buoyancy and sinking the vehicle. As the submersible dives deeper, buoyancy increases modestly due to the increasing water density created by the increasing pressure. Thus the submersible slows slightly as it dives deeper (Figure 2). When the submersible approaches the seafloor (50–100 m in altitude, i.e., height above the bottom), a portion of its ballast (usually lead or some other heavy material) is jettisoned to achieve neutral buoyancy. Perfect neutral buoyancy occurs when the positively buoyant materials (things which tend to float) on the submersible balance the negatively buoyant materials (things which tend to sink). This allows the vehicle to hover weightless in position and move freely about. As perfect neutral buoyancy is difficult to maintain, most submersibles have auxiliary weight-adjusting (trim and ballast) systems. This consists of a sea-water pumping system to draw in or expel water, thus adjusting the buoyancy of the submersible. Upon completing its mission, the remaining ballast is jettisoned and the submersible now with positive buoyancy begins ascending. When resurfaced, air from a high-pressure bottle is blown into the air ballast tank to give enough draft to the submersible for the recovery operation. Water Pressure

Water pressure increases by 0.1 MPa per 10 m depth. Thus every component sensitive to pressure must be

507

isolated from intense pressure changes. First and foremost are the passengers which are protected against great ambient pressure by a pressure hull or pressure vessel, maintained at surface pressure. The ambient pressure exerts strong compressional force on the pressure hull which is therefore designed to avoid any tensile stress. The strongest geometric shape against outside pressure relative to volume and hence weight is a sphere, followed by a cylinder (capped at both ends). However, it is not easy to arrange instruments inside a sphere effectively. In order to increase mobility, it is important to make submersibles small and light. The pressure hull is one of the largest and heaviest components of the submersible. The hull must be as small (and light) as possible, while affording appropriate strength against external pressure. Thus for deep-diving submersibles, a spherical pressure hull is employed whereas shallower vehicles can use a cylindrical shape if so desired. The material used for the pressure hull is critical. In earlier vehicles, steel was used. Later, titanium alloy was the material of choice. Titanium alloy has very high tensile strength, and is resistant to corrosion and relatively light (specific gravity B60% that of steel). Recently, the trend in submersible construction is to use nonmetallic materials, such as fiber- or graphite-reinforced plastics (FRP or GRP), or ceramics. Components not sensitive to pressure or saline conditions need no special consideration. Though those devices which require electrical insulation need to be housed in oil-filled compartments called oil-filled pressure compensation systems (Figure 3). These systems do not require heavy pressure hulls and thus reduce the weight of the submersible overall. Electric motors, hydraulic systems, batteries, wiring, and power transistors are all housed in pressure compensation systems. Technology is being developed to apply ambient pressure to electronic devices such as integrated circuits (ICs) and large scale ICs (LSIs).

Buoyancy

With the exception of some shallow-water submersibles, the total weight of the essential systems is larger than the total buoyancy. This means that extra buoyancy is needed to balance the excess weight. Wood or foam-rubber cannot be used for this purpose because they shrink under increasing water pressure. The material providing buoyancy must have a relatively small specific gravity while remaining strong under high-pressure conditions.

(c) 2011 Elsevier Inc. All Rights Reserved.

508

MANNED SUBMERSIBLES, DEEP WATER

Pressure resistant housing (thick wall) Rotary seal (high friction)

Valve

Electric motor . .

. .

Oxygen bottle

Fan

Rotary seal (low friction)

Pressure compensated housing (thin wall) Desiccant CO2 absorbent

Electric motor . .

. Figure 4 Life support system for a deep submersible.

Oil bladder (for compensation of pressure and the oil volume) Figure 3 Pressure resistant and pressure compensated housings for electric motors.

Historically, gasoline was used to provide buoyancy in bathyspheres as it did not lose buoyancy under pressure. However, its specific gravity was too large for practical use – huge volumes are needed to offset the weight. With the invention of syntactic foam, a superior material for deep-diving submersibles became available. Syntactic foam consists of tiny microscopic spheres of glass embedded in resin. These microballoons are 40–200 mm in diameter, and are closely packed with resin filling in the surrounding spaces. Proper selection of the balloons and resin allows the proper pressure tolerance and specific gravity to be created. For example, the syntactic foam used by the Shinkai 6500 is tolerant up to 130 MPa with a specific gravity of 0.54 gf ml 1 and the foam used by the ROV Kaiko is tolerant up to 160 MPa with a specific gravity of 0.63 gf ml 1.

Life Support

The pressure hull is a very small space where crew members must stay for up to 20 h, depending on their mission. Since the pressure hull is maintained at ambient pressure, no decompression of the occupants is needed. High-pressure oxygen bottles provide oxygen within the pressure hull, while carbon scrubbers absorb carbon-dioxide (Figure 4). Extra life-support is required with varying standards depending on the country (from 32 h to 120 h)

Energy

Energy for deep-diving submersibles is supplied by rechargeable (secondary) batteries. There are several types including: lead acid, nickel cadmium, nickel hydrogen, oxidized silver zinc. Batteries which contain a higher density of energy are preferable to reduce the weight and volume of the submersible. However, such batteries are very expensive. These batteries are housed in oil-filled pressure-compensated systems to reduce weight. Recent developments in fuel cell technology offer the promise of higher density energy coupled with higher efficiency. This has great potential for submersible applications. Instrumentation

There are many scientific and observational instruments employed on research submersibles. Due to the limited payload, these instruments must be as light and small as possible. For example, bulky camera bodies must be streamlined and aligned with lenses in a compact manner, thus reducing the size and weight of their pressure housing (see Figure 5). Furthermore, physical conditions such as extremely cold temperatures or the differential absorption of white light must be considered. Thus, for example, engineers must consider both the compactness of video cameras and their color sensitivity. Manipulator

Pressure hulls, thrusters, batteries, and buoyancy materials are all essential parts of the modern submersible. One of the most important tools is the manipulator arm. Manipulators extend the arms and hands of the pilot, allowing sample collection and

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, DEEP WATER

509

The support ship not only supports the diving operation, including launch, recovery, communication, and positioning, but also provides a place to conduct on-board research during a cruise. Accordingly, research laboratories, computers, instruments for onboard data analysis, multi-narrow beam echo sounders, etc., are necessary. Also, remotely operated vehicles (ROVs) or autonomous underwater vehicles (AUVs), can be operated during the nighttime, or in case the manned submersible cannot be operated because of bad weather.

Video camera

Video camera

Deep Submersibles in the World Alvin (USA)

Figure 5 Examples of inside alignments of the pressure case.

deployment of experimental equipment. Most manipulators are driven by hydraulic pressure. The most advanced manipulators operate in a masterslave system. The operator handles a master arm (controller) which imitates human arm and hand movements, and the motions are translated to the slave-unit (manipulator) which follows precisely the motions of the master arm. There are usually one or two manipulators on a research submersible. Navigation

Underwater navigation is one of the most crucial elements of deep-sea submersible researches. Usually, long base line (LBL) or super short base line (SSBL) acoustic navigation systems are used depending on the accuracy required. The positioning error of LBL systems is 5–15 m, and it is approximately 2–5% of slant range for SSBL systems. An advantage of SSBL systems is that seafloor transponders are not necessary, whereas at least two seafloor transponders are necessary for LBL systems. In both systems, absolute or geodetic position is determined by surface navigation systems, such as Differential Global Positioning System (DGPS). Although the position of the submersible is usually reported from the mother ship by voice, the Shinkai 6500 has an automatic LBL navigation system. Surface Support (the Mother Ship)

The R/V Lulu, the original support ship of the submersible Alvin, was retired and replaced by the R/V Atlantis II in 1983. In 1997 Atlantis II was replaced by the newly commissioned Research Vessel Atlantis.

Alvin (Figure 6) was built in 1964 by Litton Industries with funds of the US Navy and operated by WHOI. Its original depth capability was 1800 m with a steel pressure hull. Later, the pressure hull was replaced by titanium alloy to increase depth capability up to 4500 m. Alvin’s size and weight are: length 7.1 m; width 2.6 m; height 3.7 m and weight 17 tf in air. The outside diameter of the pressure hull is 2.08 m with a wall thickness of 49 mm. It is equipped with three view ports, 120 mm in inside diameter, two manipulators with seven degrees of freedom. The original catamaran support ship, R/V Lulu, was replaced by the R/V Atlantis II. Launch and recovery take place at the stern A-frame. It has been the leading deep submersible in the world. Nautile (France)

Nautile (Figure 7) was built in 1985 by IFREMER (French Institution for Marine Research and Development) in France. Depth capability of 6000 m was aimed to cover 98% of the world’s ocean floor. It is 8 m long, 2.7 m wide, 3.81 m in height and weighs 19.3 tf in air. It is equipped with three view ports, a manipulator, a grabber and a small companion ROV Robin. The position of the submersible is directly calculated by interrogating the seafloor transponders. Still video images are transmitted to the support ship through the acoustic link. Sea Cliff (USA)

Sea Cliff was originally built in 1968 as a 3000 mclass submersible by General Dynamics Corp. for the US Navy as a sister submersible for the Turtle. In 1985, the Sea Cliff was converted into a 6000 mclass deep submersible. It is 7.9 m long, 3.7 m wide, 3.7 m high and weighs 23 tf in air. (Sea Cliff and Turtle are currently out of commission.)

(c) 2011 Elsevier Inc. All Rights Reserved.

510

MANNED SUBMERSIBLES, DEEP WATER

Figure 6 US submersible Alvin.

Mir I and Mir II (Russia)

The 6000 m-class submersibles Mir I and Mir II (Figure 8) were built in 1987 by Rauma Repola in Finland for the P.P. Shirshov Institute of Oceanology in Russia. They are 7.8 m long, 3.8 m wide, 3.65 m

Figure 7 French 6000 m-class submersible Nautile.

high and weigh 18.7 tf in air. Inside diameter of the pressure hull, which is made of high-strength steel, is 2.1 m with a wall thickness of 40 mm. Launch and recovery take place by an articulated crane over the side of the support ship, the R/V Academik Mistilav Keldysh. If necessary, both Mir I and Mir II are

Figure 8 Russian 6000 m-class submersible Mir I or Mir II.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, DEEP WATER

launched simultaneously to carry out cooperative or independent research. Another characteristic feature of the Mir is a powerful secondary battery, 100 kWh of total energy, which allows it to stay more than 20 hours underwater, or to carry out more than 14 h of continuous operation on the bottom. Shinkai 6500 (Japan)

The Shinkai 6500 (Figure 9) was built in 1989 by Mitsubishi Heavy Industries and operated by the Japan Marine Science Technology Center (JAMSTEC). It is 9.5 m long, 2.71 m wide, 3.21 m high and weighs 25.8 tf in air. The pressure hull is made of titanium alloy, 73.5 mm in thickness, and has an inside diameter of 2 m. It is equipped with three view ports, two manipulators with seven degrees of freedom. Position of the submersible is calculated and displayed in real time by directly interrogating the seafloor transponders. Still color video images are transmitted automatically at 10 s intervals to the support ship, the R/V Yokosuka, through the acoustic link during the diving operation. Launch and recovery take place at the stern A-frame of the R/V Yokosuka.

Major Contributions of Deep Submersibles The dive to the Challenger Deep in the Mariana Trench by the bathyscaph Trieste in 1960 was one of the most spectacular achievements of the twentieth century. However, the dive was mainly for adventure rather than for science. In 1963, the US nuclear

511

submarine Thresher sank in 2500 m of water off Cape Cod in New England. After an extensive search for the submarine, the bathyscaph Trieste made dives to inspect the wreck in detail and recover small objects. The operation demonstrated the importance of using deep submersibles and advanced deep ocean technology to increase knowledge of the deep ocean. In 1966, hydrogen bombs were lost with a downed US B-52 bomber off Palomares, Spain. The Alvin showed the great utility of deep submersibles by locating and assisting in the recovery of lost objects from the sea. Between 1973 and 1974, project FAMOUS (French–American Mid-Ocean Undersea Study) was conducted in the Mid-Atlantic Ridge off the Azores using the French bathyscaph Archimede and the US submersible Alvin. The project was the first systematic and successful use of deep submersibles for science. They discovered and sampled fresh pillow lavas and lava flows at 3000 m deep in the rift valley, where the oceanic crusts were being created, providing visual evidence of Plate Tectonics. In 1977, Alvin discovered a hydrothermal vent and vent animals in the East Pacific Rise off the Galapagos Islands at a depth of 2450 m. Discovery of these chemosynthetic animals, which were not dependent on photosynthesis, had a profound impact on biology in the twentieth century. Manned submersibles now compete with unmanned submersibles, such as ROVs and AUVs. Because of the expense of operation and maintenance, national funding is necessary for manned submersibles. However, ROVs and AUVs can be operated by private companies or institutions. In

Figure 9 Japanese 6000 m-class submersible Shinkai 6500.

(c) 2011 Elsevier Inc. All Rights Reserved.

512

MANNED SUBMERSIBLES, DEEP WATER

spite of the costs, the ability of the human observer to rapidly process information to make decisions provides an advantage and justifies continued use of manned submersibles.

See also Deep Submergence, Science of. Drifters and Floats. Manned Submersibles, Shallow Water. Moorings. Platforms: Autonomous Underwater Vehicles. Remotely Operated Vehicles (ROVs). Rigs and offshore Structures.

Further Reading Beebe W (1934) Half Miles Down. New York: Harcourt Brace. Busby RF (1990) Undersea Vehicles Directory – 1990–91, 4th edn. Arlington, VA: Busby Associates. Funnel C (ed.) (1999) Jane’s Underwater Technology, 2nd edn. UK: Jane’s Information Group Limited. Kaharl VA (1990) Water Baby – The Story of Alvin. New York: Oxford University Press. Piccard A (1956) Earth Sky and Sea. New York: Oxford University Press. Piccard J and Dietz RS (1961) Seven Miles Down. New York: G.P. Putnam.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, SHALLOW WATER T. Askew, Harbor Branch Oceanographic Institute, Ft Pierce, FL, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1550–1555, & 2001, Elsevier Ltd.

Introduction Early man’s insatiable curiosity to look beneath the surface of the sea in search of natural treasures that were useful in a primitive, comfortless mode of living were a true test of his limits of endurance. The fragile vehicle of the human body quickly discovered that most of the sea’s depths were unapproachable without some form of protection against the destructive hostilities of the ocean. Modern technology has paved the way for man to conquer the hostile marine environment by creating a host of manned undersea vehicles. Called submersibles, these small engineering marvels carry out missions of science, exploration, and engineering. The ability to conduct science and other operations under the sea rather than from the surface has stimulated the submersible builder/operator to further develop the specialized tools and instruments which provide humans with the opportunity to be present and perform tasks in relative comfort in ocean bottom locations that would otherwise be destructive to human life. Over the past fifty years a depth of 1000 m has surfaced as the transition point for shallow vs deep water manned submersibles. During the prior 100 years any device that enabled man to explore the ocean depths beyond breath-holding capabilities would have been considered deep.

History While it is difficult to pinpoint the advent of the first submersible it is thought that in 1620 Cornelius van Drebel constructed a vehicle under contract to King James I of England. It was operated by 12 rowers with leather sleeves, waterproofing the oar ports. It is said that the craft navigated the Thames River for several hours at a depth of 4 m and carried a secret substance that purified the air, perhaps soda lime? In 1707, Dr Edmund Halley built a diving bell with a ‘lock-out’ capability. It had glass ports above to provide light, provisions for replenishing its air, and crude umbilical-supplied diving helmets which

permitted divers to walk around outside. In 1776, Dr David Bushnell built and navigated the first submarine employed in war-like operations. Bushnell’s Turtle was built of wood, egg-shaped with a conning tower on top and propelled horizontally and vertically by a primitive form of screw propeller after flooding a tank which allowed it to submerge. In the early 1800s, Robert Fulton, inventor of the steamship, built two iron-formed copper-clad submarines, Nautilus and Mute. Both vehicles carried out successful tests, but were never used operationally. The first ‘modern submersible’ was Simon Lake’s Argonaut I, a small vehicle with wheels and a bottom hatch that could be opened after the interior was pressurized to ambient. While there are numerous other early submarines, the manned submersible did not emerge as a useful and functional means of accomplishing underwater work until the early 1960s. It was during this same period of time that the French-built Soucoupe, sometimes referred to as the ‘diving saucer’, came into being. Made famous by Jacques Cousteau on his weekly television series, the Soucoupe is credited with introducing the general population to underwater science. Launched in 1959, the diving saucer was able to dive to 350 m. The USS Thresher tragedy in 1963 appears to have spurred a movement among several large corporations such as General Motors (Deep Ocean Work Boat, DOWB), General Dynamics (Star I, II, III, Sea Cliff, and Turtle), Westinghouse (Deepstar 2000, 4000), and General Mills (Alvin), along with numerous other start-up companies formed solely to manufacture submersibles. Perry Submarine Builders, a Florida-based company, started manufacturing small shallow-water, three-person submersibles in 1962, and continued until 1980 (Figures 1 and 2). International Hydrodynamics Ltd (based in Vancouver, BC, Canada) commenced building the Pisces series of submersibles in 1962. The pressure hull material in the 1960s was for the most part steel with one or more view ports. The operating depth ranged from 30 m to 600 m, which was considered very deep for a free-swimming, untethered vehicle. The US Navy began design work on Deep Jeep in 1960 and after 4 years of trials and tribulations it was commissioned with a design depth of 609 m. A two-person vehicle, Deep Jeep included many features incorporated in today’s submersibles, such as a dropable battery pod, electric propulsion motors that operate in silicone oil-filled housings, and shaped resin blocks filled with glass micro-balloons

(c) 2011 Elsevier Inc. All Rights Reserved.

513

514

MANNED SUBMERSIBLES, SHALLOW WATER

Figure 1 Perry-Link Deep Diver, 1967. Owned by Harbor Branch Oceanographic Institution. Length 6.7 m, beam 1.5 m, height 2.6 m, weight 7485 kg, crew 1, observers 2, duration 3–5 hours.

used to create buoyancy. Deep Jeep was eventually transferred to the Scripps Institution of Oceanography in 1966 after a stint searching for a lost ‘H’ bomb off Palomares, Spain. Unfortunately, Deep Jeep was never placed into service as a scientific submersible due to a lack of funding. The missing bomb was actually found by another vehicle, Alvin. Alvin did get funding and proved to be useful as a scientific tool. The Nekton series of small two-person submersibles appeared in 1968, 1970, and 1971. The Alpha, Beta, and Gamma were the brainchildren of Doug Privitt, who started building small submersibles for recreation in the 1950s. The Nektons had a depth capability of 304 m. The tiny submersibles conducted hundreds of dives for scientific purposes as well as for military and oilfield customers. In 1982, the Nekton Delta, a slightly larger submersible with a depth rating of 365 m was

unveiled and is still operating today with well over 3000 dives logged. A few of the submersibles were designed with a diver ‘lock-out’ capability. The first modern vehicle was the Perry-Link ‘Deep Diver’ built in 1967 and able to dive to 366 m. This feature enabled a separate compartment carrying divers to be pressurized internally to the same depth as outside, thus allowing the occupants to open a hatch and exit where they could perform various tasks while under the supervision of the pilots. Once the work was completed, the divers would re-enter the diving compartment, closing the outer and inner hatches; thereby maintaining the bottom depth until reaching the surface, where they could decompress either by remaining in the compartment or transferring into a larger, more comfortable decompression chamber via a transfer trunk. Acrylic plastic was tested for the first time as a new material for pressure hulls in 1966 by the US Navy.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, SHALLOW WATER

515

Figure 2 Perry PC 1204 Clelia, 1976. Owned and operated by Harbor Branch Oceanographic Institution. Length 6.7 m, beam 2.4 m, height 2.4 m, weight 8160 kg, crew 1, observers 2, duration 3–5 hours.

The Hikino, a unique submersible that incorporated a 142 cm diameter and a 0.635 cm thick hull, was only able to dive to 6 m. This experimental vehicle was used to gain experience with plastic hulls, which eventually led to the development of Kumukahi, Nemo (Naval Experimental Manned Observatory), Makakai, and Johnson-Sea-Link. The Kumukahi, launched in 1969, incorporated a unique 135 cm acrylic plastic sphere formed in four sections. It was 3.175 cm thick and could dive to 92 m. The Nemo, launched in 1970, and Makakai, launched in 1971, both utilized spheres made of 12 curved pentagons formed from a 6.35 cm flat sheet of PlexiglasTM. The pentagons were bonded together to make one large sphere capable of diving to 183 m. The Johnson-Sea-Link, designed by Edwin Link and built by Aluminum Company of America (ALCOA), utilized a Nemo-style PlexiglasTM sphere, 167.64 cm in diameter, 10.16 cm thick, and made of 12 curved pentagons formed from flat sheet and bonded together. This new thicker hull had an operational depth of 304 m.

Present Day Submersibles The submersibles currently in use today are for the most part classified as either shallow-water or deepwater vehicles, the discriminating depth being approximately 1000 m. This is where the practicality of using compressed gases for ballasting becomes

impractical. The deeper diving vehicles utilize various drop weight methods; most use two sets of weights (usually scrap steel cut into uniform blocks). One set of weights is released upon reaching the bottom, allowing the vehicle to maneuver, travel, and perform tasks in a neutral condition. The other set of weights is dropped to make the vehicle buoyant, which carries it back to the surface once the dive is complete. The shallow vehicles use their thrusters and/or water ballast to descend to the bottom and some of the more sophisticated submersibles have variable ballast systems which allow the pilot to achieve a neutral condition by varying the water level in a pressure tank. One advantage of the shallow vehicles is that the view ports (commonly called windows) can be much larger both in size and numbers, and where an acrylic plastic sphere is used the entire hull becomes a window. Since the late 1960s and early 1970s acrylic plastic pressure hulls have emerged as an ideal engineering solution to create a strong, transparent, corrosionresistant, nonmagnetic pressure hull. The limiting factor of the acrylic sphere is its ability to resist implosion from external pressure at great depths. Its strength comes from the shape and wall thickness. Therefore, the greater the depth the operator aims to reach, the thicker the sphere must be, which results in a hull that is much too heavy to be practical for use on a small submersible designed to go deeper than 1000 m.

(c) 2011 Elsevier Inc. All Rights Reserved.

516

MANNED SUBMERSIBLES, SHALLOW WATER

These shallow-water submersibles, once quite numerous because of their usefulness in the offshore oilfield industry, are now limited to a few operators and mostly used for scientific investigations. The Johnson-Sea-Links (J-S-Ls) stand out as two of the most advanced manned submersibles (Figure 3). J-S-L I, commissioned by the Smithsonian Institution in January 1971, was named for designer and donors Edwin A. Link and J. Seward Johnson. Edwin Link, responsible for the submersible’s unique design and noted for his inventions in the aviation field, turned his energies to solving the problems of undersea diving, a technology then still in its infancy. One of his objectives was to carry out scientific work under water for lengthy periods. The Johnson-Sea-Link was the most sophisticated diving craft he had created for this purpose, and it promised to be one of the most effective of the new generation of small submersible vehicles that were being built to penetrate the shallow depths of the continental shelf (183 m or 100 fathoms). Originally designed for a depth of 304 m, the vessel’s unique features include a two-person transparent acrylic sphere, 1.82 m in diameter and 10.16 cm thick, that provides panoramic underwater visibility to a pilot and a scientist/observer. Behind the sphere, there is a separate 2.4 m long cylindrical, welded, aluminum alloy lock-out/lock-in compartment that will enable scientists to exit from its bottom and collect specimens of undersea flora and fauna. The acrylic sphere and the aluminum cylinder are enclosed within a simple jointed aluminum tubular frame, a configuration that makes the vessel resemble a helicopter rather than a conventionally

shaped submarine. Attached to the frame are the vessel’s ballast tanks, thrusters, compressed air, mixed gas flasks, and battery pod. The aluminum alloy parts of the submersible, lightweight and strong, along with the acrylic capsule which was patterned after the prototype used by the US Navy on the Nemo, had extraordinary advantages over traditional materials like steel. They were most of all immune to the corrosive effects of sea water. The emphasis in engineering of the submersible was on safety. Switches, connectors, and all operating gear were especially designed to avoid possible safety hazards. The rear diving compartment allows one diver to exit for scientific collections while tethered for communications and breathing air supply, while the other diver/tender remains inside as a safety backup. Once the dive is completed and the submersible is recovered by a special deckmounted crane on the support ship, the divers can transfer into a larger decompression chamber via a transfer trunk which is bolted to the lock-out/lock-in compartment. Now, 30 years later, the Johnson-Sea-Links with a 904 m depth rating, remain state of the art underwater vehicles. Sophisticated hydraulic manipulators work in conjunction with a rotating bin collection platform which allow 12 separate locations to be sampled and simultaneously documented by digital color video cameras mounted on electric pan and tilt mechanisms and aimed with lasers. Illumination is provided by a variety of underwater lighting systems utilizing zenon arc lamps, metal halide, and halogen bulbs. Acoustic beacons provide real time position

Figure 3 Johnson-Sea-Link I and II, 1971 and 1976. Owned and operated by Harbor Branch Oceanographic Institution. Length 7.2 m, beam 2.5 m, height 3.1 m, weight 10400 kg, crew 2, observers 2, duration 3–5 hours.

(c) 2011 Elsevier Inc. All Rights Reserved.

MANNED SUBMERSIBLES, SHALLOW WATER

and depth information to shipboard computer tracking systems that not only show the submersible’s position on the bottom, but also its relationship to the ship in latitude and longitude via the satellite-based global positioning system (GPS). The lock-out/lock-in compartment is now utilized as an observation and instrumentation compartment, which remains at one atmosphere. Today’s shallow-water submersibles (average dive 3–5 h) require a support vessel to provide the necessities that are not available due to their relatively small size. The batteries must be charged, compressed air and oxygen flasks must be replenished. Carbon dioxide removal material, usually soda lime or lithium hydroxide, is also replenished so as to provide maximum life support in case of trouble. Most submersibles today carry 5 days of life support, which allows time to effect a rescue should it become necessary. The support vessel also must have a launch/recovery system capable of safely handling the submersible in all sea conditions. Over the last 30 years, the highly trained crews that operate the ship’s handling systems and the submersibles, have virtually made the shallow-water submersibles an everyday scientific tool where the laboratory becomes the ocean bottom.

Operations The two Johnson-Sea-Links have accumulated over 8000 dives for science, engineering, archaeology, and training purposes since 1971. They have developed into highly sophisticated science tools. Literally thousands of new species of marine life have been photographed, documented by video camera and collected without disturbing the surrounding habitat. Behavioral studies of fish, marine mammals and invertebrates as well as sampling of the water column and bottom areas for chemical analysis and geological studies are everyday tasks for the submersibles. In addition, numerous historical shipwrecks from galleons to warships like the USS Monitor have been explored and documented, preserving their legacy for future generations. Johnson-Sea-Links I and II (J-S-Ls) were pressed into service to assist in locating, identifying, and ultimately recovering many key pieces of the ill-fated Space Shuttle Challenger. This disaster, viewed by the world via television, added a new dimension to the J-S-Ls’ capabilities. Previously only known for their pioneering efforts in marine science, they proved to be valuable assets in the search and recovery operation. The J-S-Ls completed a total of 109 dives,

517

including mapping a large area of the right solid rocket booster debris at a depth of 365 m. The vehicles proved their worth throughout the operation by consistently performing beyond expectations. They were launched and recovered easily and quickly. They could work on several contacts per day, taking NASA engineers to the wreckage for firsthand detailed examination of debris while video cameras recorded what was being seen and said. Significant pieces were rigged with lifting bridles for recovery. The autonomous operation of the J-S-Ls, a dedicated support vessel, and highly trained operations personnel made for a successful conclusion to an operation that had a significant impact on the future of the US Space Program.

Summary There is no question that the manned submersible has earned its place in history. Much of what are now cataloged as new species were discovered in the last 30 years with the aid of submersibles. The ability to conduct marine science experiments in situ led to the development of intricate precision instruments, sampling devices for delicate invertebrates and gelatinous organisms that previously were only seen in blobs or pieces due to the primitive methods used to collect them. While some suggest that remotely operated vehicles (ROVs) could, and have replaced the manned submersible, in reality they are complementary. There is no substitute for the autonomous, highly maneuverable submersible that can approach and collect without contact delicate zooplankton, while observing behavior and measuring the levels of bioluminescence, or probing brine pools and cold seep regions in the Gulf of Mexico for specialized collections of biological, geological, and geochemical samples. Tubeworms are routinely marked for growth rate studies and collected individually, along with other biological species that thrive in these chemosynthetic communities. Sediments and methane ice (gas hydrates) are also selectively retrieved for later analysis. Some new vehicles are still being produced, but have limited payloads, which restricts them to specific tasks such as underwater camera platforms or observation. Some are easily transportable but are small and restricted to one occupant; they can be carried by smaller support vessels and are more economical to operate. Man’s desire to explore the lakes, oceans, and seas has not diminished. New technology will only enable, not reduce, the need for man’s presence in these hostile environments.

(c) 2011 Elsevier Inc. All Rights Reserved.

518

MANNED SUBMERSIBLES, SHALLOW WATER

See also Deep Submergence, Science of. Manned Submersibles, Deep Water. Remotely Operated Vehicles (ROVs).

Further Reading Askew TM (1980) JOHNSON-SEA-LINK Operations Manual. Fort Pierce: Harbor Branch Foundation. Busby F (1976, 1981) Undersea Vechicles Directory. Arlington: Busby & Associates.

Forman WR (1968) KUMUKAHI Design and Operations Manual. Makapuu, HI: Oceanic Institute. Forman W (1999) The History of American Deep Submersible Operations. Flagstaff, AZ: Best Publishing Co. Link MC (1973) Windows in the Sea. Washington, DC: Smithsonian Institute Press. Stachiw JD (1986) The Origins of Acrylic Plastic Submersibles. American Society of Mechanical Engineers, Asme Paper 86-WA/HH-5. Van Hoek S and Link MC (1993) From Sky to Sea; A Story of Ed Link. Flagstaff, AZ: Best Publishing Co.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE DISEASES AND HEALTH A. E. Ellis, Marine Laboratory, Aberdeen, Scotland, UK

and to restrict their spread within a farm in a variety of ways.

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1555–1560, & 2001, Elsevier Ltd.

Introduction As with all forms of intensive culture where a single species is reared at high population densities, infectious disease agents are able to transmit easily between host individuals and large economic losses can result from disease outbreaks. Husbandry methods are designed to minimize these losses by employing a variety of strategies, but central to all of these is providing the cultured animal with an optimal environment that does not jeopardize the animal’s health and well-being. All animals have innate and acquired defenses against infectious agents and when environmental conditions are good for the host, these defense mechanisms will provide protection against most infections. However, animals under stress have less energy available to combat infections and are therefore more prone to disease. Although some facilities on a farm may be able to exclude the entry of pathogens, for example hatcheries with disinfected water supplies, it is impossible to exclude pathogens in an open marine situation. Under these conditions, stress management is paramount in maintaining the health of cultured animals. Even then, because of the close proximity of individuals in a farm, if certain pathogens do gain entry they are able to spread and multiply extremely rapidly and such massive infectious burdens can overcome the defenses of even healthy animals. In such cases some form of treatment, or even better, prophylaxis, is required to prevent crippling losses. This article describes some of the management strategies available to fish and shellfish farmers in avoiding or reducing the losses from infectious diseases and some of the prophylactic measures and treatments. The most important diseases encountered in mariculture are summarized in Table 1.

Health Management Facility Design

Farms and husbandry practices can be designed in such a way as to avoid the introduction of pathogens

Isolate the hatchery Infectious agents can be excluded from hatcheries by disinfecting the incoming water using filters, ultraviolet lamps or ozone treatments. It is also important not to introduce infections from other parts of the farm that may be contaminated. The hatchery then should stand apart and strict hygiene standards applied to equipment and personnel entering the hatchery. Some diseases cause major mortalities in young fry while older fish are more resistant. For example, infectious pancreatic necrosis virus (IPNV) causes mass mortality in halibut fry, but juveniles are much more resistant. It is vitally important therefore, to exclude the entry of IPNV into the halibut hatchery and as this virus has a widespread distribution in the marine environment, disinfection of the water supply may be necessary. Hygiene practice Limiting the spread of disease agents on a farm include having hand nets for each tank and disinfecting the net after each use, disinfectant foot-baths at the farm entrance and between buildings, and restricted movement of staff, their protective clothing and equipment. Prompt removal of dead and moribund stock is essential as large numbers of pathogens are shed into the water from such animals. In small tanks this can easily be done using a hand net. In large sea cages, lifting the net can be stressful to fish and divers are expensive. Special equipment such as air-lift pumps, or specially designed cages with socks fitted in the bottom in which dead fish collect and which can be hoisted out on a pulley are more practical. Proper disposal of dead animals is essential. Methods such as incineration, rendering, ensiling and, on a small scale, in lime pits are recommended. Husbandry and Minimizing Stress

Animals under stress are more prone to infectious diseases. However, it is not possible to eliminate all the procedures that are known to induce stress in aquaculture animals, as many are integral parts of aquaculture, e.g., netting, grading, and transport. Nevertheless, it is possible for farming practices to minimize the effects of these stressors and others, e.g., overcrowding and poor water quality, can be avoided by farmers adhering to the recommended

(c) 2011 Elsevier Inc. All Rights Reserved.

519

520

MARICULTURE DISEASES AND HEALTH

Table 1

Principal diseases of fish and shellfish in mariculture

Disease agent

Host

Prevention/treatment

Shellfish pathogens Protozoa Bonamia ostreae Marteilia refringens

European flat oyster Oyster, mussel

Exclusion Exclusion

Bacteria: Aerococcus viridans (Gaffkaemia).

Lobsters

Improve husbandry

Salmon, turbot, halibut, cod, sea bass, yellowtail Atlantic salmon Atlantic salmon Turbot Stripped jack, Japanese flounder, barramundi, sea bass, sea bream, turbot, halibut

Sanitary precautions, vaccinate

All marine species Salmon, trout, cod Salmon Salmon

Vaccinate, antibiotics Vaccinate, antibiotics Avoid stress, vaccinate Vaccinate, antibiotics

Turbot, halibut, flounder, salmon

Vaccinate, antibiotics, avoid stress

Sole, flounder, turbot, salmon, sea bass Atlantic salmon

Avoid stress, antibiotics Vaccinate

Pacific salmon

Exclude

Salmonids, sea bass, sea bream, stripped bass, cod, red drum, tilapia Yellowtail, sea bass, sea bream

Sanitary

Salmonids

Sanitary

Most species

Avoid stress, sanitary, chemical baths, e.g., formalin

Salmonids Salmonids, sea bass, sea bream

Low salinity bath, H2O2 In-feed insecticides, H2O2, cleaner fish

Fin-fish pathogens Viruses Infectious pancreatic necrosis Infectious salmon anemia Pancreas disease Viral hemorrhagic septicemia Viral nervous necrosis

Bacteria Vibriosis (Vibrio anguillarum) Cold water vibriosis (Vibrio salmonicida) Winter ulcers (Vibrio viscosus) Typical furunculosis (Aeromonas salmonicida) Atypical furunculosis (Aeromonas salmonicida achromogenes) Flexibacter maritimus Enteric redmouth (Yersinia ruckeri) Bacterial kidney disease (Renibacterium salmoninarum) Mycobacteriosis (Mycobacterium marinum) Pseudotuberculosis (Photobacterium damselae piscicida) Piscirickettsiosis (Piscirickettsia salmonis) Parasites Protozoan: many species; Amyloodinium, Cryptobia, Ichthyobodo, Cryptocaryon, Tricodina Paramebic gill disease Crustacea: Sea lice

Sanitary precautions, eradicate Avoid stressors, vaccinate Sanitary precautions Sanitary precautions

Vaccinate, sanitary

limits for stocking densities, water flow rates and feeding regimes. In cases where stressors are unavoidable, farmers can adopt certain strategies to minimize the stress.

Use of anesthesia Although anesthetics can disturb the physiology of fish, light anesthesia can have a calming effect on fish during handling and transport and so reduce the stress resulting from these procedures.

Withdrawal of food prior to handling Following feeding the oxygen requirement of fish is increased. Withdrawal of food two or three days prior to handling the fish will therefore minimize respiratory stress. It also avoids fouling of the water during transportation with fecal material and regurgitated food.

Avoidance of stressors at high temperatures High temperatures increase the oxygen demand of animals and stress-induced mortality can result from respiratory failure at high temperatures. It is therefore safer to carry out netting, grading and transport at low water temperatures.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE DISEASES AND HEALTH

Avoidance of multiple stressors and permitting recovery The effects of multiple stressors can be additive or even synergistic so, for instance, sudden changes of temperature should be avoided during or after transport. Where possible, recovery from stress should be facilitated. Generally, the duration of the recovery period is proportional to the duration of the stressor. Thus, reducing the time of netting, grading or transport will result in recovery in a shorter time. The duration required for recovery to occur may be from a few days to two weeks. Selective breeding In salmonids, it is now established that the magnitude of the stress response is a heritable characteristic and programs now exist for selecting broodstock which have a low stress response to handling stressors. This accelerates the process of domestication to produce stocks, which are more tolerant of aquaculture procedures with resultant benefits in increased health, survival and productivity. Such breeding programs have been conducted in Norway for some years and have achieved improvements in resistance to furunculosis and infectious salmon anemia virus (ISAV) in Atlantic salmon. Management of the Pathogen

Breaking the pathogen’s life cycle If a disease is introduced on to a farm, it is important to restrict the horizontal spread especially to different year classes of fish/shellfish. Hence, before a new year class of animals is introduced to a part of the farm, all tanks, equipment etc. should be thoroughly cleaned and disinfected. In sea-cage sites it is a useful technique to physically separate year classes to break the infection cycle. Many pathogens do not survive for long periods of time away from their host and allowing a site to be fallow for a period of time may eliminate or drastically reduce the pathogen load. This practice has been very effective in controlling losses from furunculosis in marine salmon farms and also significantly reduces the salmon lice populations particularly in the first year after fallowing. Eliminate vertical transmission Several pathogens may persist as an asymptomatic carrier state and be present in the gonadal fluids of infected broodstock and can infect the next generation of fry. In salmon farming IPNV may persist in or on the ova and disinfection of the eggs with iodine-based disinfectants (Iodophores) is recommended immediately after fertilization. The testing of gonadal fluids for the presence of IPNV can also be carried out and batches of eggs

521

from infected parents destroyed. IPNV has been associated with mass mortalities in salmon and halibut fry and has been isolated from a wide range of marine fish and shellfish, including sea bass, turbot, striped bass, cod, and yellowtail. Avoid infected stock Many countries employ regulations to prevent the movement of eggs or fish that are infected with certain ‘notifiable’ diseases from an infected to a noninfected site. These policies are designed to limit the spread of the disease but require specialized sampling procedures and laboratory facilities to perform the diagnostic techniques. By testing the stock frequently they can be certified to be free from these diseases. Such certification is required for international trade in live fish and eggs but within a country it is widely practiced voluntarily because stock certified to be ‘disease-free’ command premium prices. Eradication of infected stock Commercially this is a drastic step to take especially when state compensation is not usually available even when state regulations might require eradication of stock. This policy is usually only practiced rarely and when potential calamitous circumstances may result, for instance, the introduction of an important exotic pathogen into an area previously free of that disease. This has occurred in Scotland where the European Commission directives have required compulsory slaughter of turbot infected with viral hemorrhagic septicemia virus (VHSV) and Atlantic salmon infected with ISAV.

Treatments Viral Diseases

There are no treatments available for viral diseases in aquaculture. These diseases must be controlled by husbandry and management strategies as described above, or by vaccination (see below). Bacterial Diseases

Antibiotics can be used to treat many bacterial infections in aquaculture. They are usually mixed into the feed. Before the advent of vaccines against many bacterial diseases of fin fish, antibiotic treatments were commonly used. However, after a few years, the bacterial pathogens developed resistance to the antibiotics. Furthermore, there was a growing concern that the large amounts of antibiotics being used in aquaculture would have damaging effects on the environment and that antibiotic residues in fish flesh may have dangerous consequences for consumers by

(c) 2011 Elsevier Inc. All Rights Reserved.

522

MARICULTURE DISEASES AND HEALTH

promoting the development of antibiotic resistant strains of human bacterial pathogens. These concerns have led to many restrictions on the use of antibiotics in aquaculture, especially in defining long withdrawal periods to ensure that carcasses for consumption are free of residues. These regulations have made the use of antibiotic treatments impractical for fish that are soon to be harvested for consumption but their use in the hatchery is still an important method of controlling losses from bacterial pathogens. For most bacterial diseases of fish, vaccines have become the most important means of control (see below) and this has led to drastic reductions in the use of antibiotics in mariculture. Parasite Diseases

Sea lice The most economically important parasitic disease in mariculture of fin fish is caused by sea lice infestation of salmon. These crustacean parasites normally infest wild fish and when they enter the salmon cages they rapidly multiply. The lice larvae and adults feed on the mucus and skin of the salmon and heavy infestations result in large haemorrhagic ulcers especially on the head and around the dorsal fin. These compromise the fish’s osmoregulation and allow opportunistic bacterial pathogens to enter the tissues. Without treatment the fish will die. A range of treatments are available and recently very effective and environmental friendly in-feed treatments such as ‘Slice’ and ‘Calicide’ have replaced the highly toxic organophosphate bath treatments. Hydrogen peroxide is also used. As a biological control method, cleaner fish are used but they are not a complete method of control.

many nonspecific defense mechanisms. These can increase disease resistance levels but only for a short period of time. Thus, in their capacity to induce such responses they are also used in shellfish culture, especially of shrimps. Current Status of Vaccination

In Atlantic salmon mariculture, vaccination has been very successful in controlling many bacterial diseases and has almost replaced the need for antibiotic treatments. In recent years vaccination of sea bass and sea bream has become common practice. Most of the commercial vaccines are against bacterial diseases because these are relatively cheap to produce. Obviously the cost per dose of vaccine for use in aquaculture must be very low and it is inexpensive to culture most bacteria in large fermenters and to inactivate the bacteria and their toxins chemically (usually with formalin). It is much more expensive to culture viruses in tissue culture and this has been a major obstacle in commercializing vaccines against virus diseases of fish. However, modern molecular biology techniques have made it possible to transfer viral genes to bacteria and yeasts, which are inexpensive to culture and produce large amounts of viral vaccine cheaply. A number of vaccines against viral diseases of Atlantic salmon are now becoming available. Currently available commercial vaccines for use in mariculture are summarized in Table 2. Methods of Vaccination

There are two methods of administering vaccines to fish: immersion in a dilute suspension of the vaccine or injection into the body cavity. For practical

Vaccination

Table 2

Use of Vaccines

Vaccines against

Vertebrates can be distinguished from invertebrates in their ability to respond immunologically in a specific manner to a pathogen or vaccine. Invertebrates, such as shellfish, only possess nonspecific defense mechanisms. In the strict definition, vaccines are used only as a prophylactic measure in vertebrates because a particular vaccine against a particular disease induces protection that is specific for that particular disease and does not protect against other diseases. Vaccines induce long-term protection and in aquaculture a single administration is usually sufficient to induce protection until the fish are harvested. However, vaccines also have nonspecific immunostimulatory properties that can also activate

Vaccines available for fish species in mariculture

Bacterial diseases Vibriosis Winter ulcers Coldwater Vibriosis Enteric redmouth Furunculosis Pseudotuberculosis Viral diseases Infectious pancreatic necrosis (IPNV) Pancreas disease Infectious salmon anemia (ISAV)

(c) 2011 Elsevier Inc. All Rights Reserved.

Maricultured species

Salmon, sea bass, sea bream, turbot, halibut Salmon Salmon Salmon Salmon Sea bass

Salmon Salmon Salmon

MARICULTURE DISEASES AND HEALTH

reasons the latter method requires the fish to be over about 15 g in weight. Immersion vaccination is effective for some, but not all vaccines. The vaccine against the bacterial disease vibriosis is effective when administered by immersion. It is used widely in salmon and sea bass farming and probably could be administered by this route to most marine fish species. The vaccine against pseudotuberculosis can also be administered by immersion to sea bass. With the exception of the vaccine against enteric redmouth, which is delivered by immersion to fish in freshwater hatcheries, all the other vaccines must be delivered by injection in order to achieve effective protection. Injection vaccination induces long-term protection and the cost per dose is very small. However, it is obviously very labor intensive. Atlantic salmon are usually vaccinated several months before transfer to sea water so that the protective immunity has time to develop before the stress of transportation to sea and exposure to the pathogens encountered in the marine environment.

Conclusions It is axiomatic that intensive farming of animals goes hand in hand with culture of their pathogens. The mariculture of fish and shellfish has had severe problems from time to time as a consequence of infectious diseases. During the 1970s, Bonamia and Marteilia virtually eliminated the culture of the European flat oyster in France and growers turned to production of the more resistant Pacific oyster. In Atlantic salmon farming, Norway was initially plagued with vibriosis diseases and Scotland suffered badly from furunculosis in the late 1980s. These bacterial diseases have been very successfully brought under control by vaccines. However, there are still many diseases for which vaccines are not available and the susceptibility of Pacific salmon to bacterial kidney disease has markedly restricted the development of the culture of these fish species on the Pacific coast of North America. As new industries grow, new diseases come to the foreground, for instance piscirikettsia in Chilean salmon culture, paramoebic gill disease in Tasmanian salmon culture, pseudotuberculosis in Mediterranean sea bass and sea bream and Japanese yellowtail culture. Old

523

diseases find new hosts, for example IPNV long known to affect salmon hatcheries, has in recent years caused high mortality in salmon postsmolts and has devastated several halibut hatcheries. To combat these diseases and to ensure the sustainability of aquaculture great attention must be paid to sanitation and good husbandry (including nutrition). In some cases these are insufficient in themselves and the presence of certain enzootic diseases, or following their introduction, have made it impossible for certain species to be cultured, for example, the European flat oyster in France. The treatment of disease by chemotherapy, which was performed widely in the 1970s and 1980s, resulted in the induction of antibiotic-resistant strains of bacteria and chemoresistant lice. Furthermore, the growing concern for the environment and the consumer about the increasing usage of chemicals and antibiotics in aquaculture, led to increasing control and restrictions on their usage. This stimulated much research in the 1980s and 1990s into development of more environmentally and consumer friendly methods of control such as vaccines and immunostimulants. These have achieved remarkable success and the pace of current research in this area using biotechnology to produce vaccines more cheaply, suggests that this approach will allow continued growth and sustainability of fin-fish mariculture into the future.

See also Crustacean Fisheries. Mariculture Overview. Molluskan Fisheries. Salmon Fisheries, Atlantic. Salmonids.

Further Reading Bruno DW and Poppe TT (1996) A Colour Atlas of Salmonid Diseases. London: Academic Press. Ellis AE (ed.) (1988) Fish Vaccination. London: Academic Press. Lightner DV (1996) A Handbook of Pathology and Diagnostic Procedures for Diseases of Cultured Penaeid Shrimps. The World Acquaculture Society. Roberts RJ (ed.) (2000) Fish Pathology, 3rd edn. London: W.B. Saunders.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF AQUARIUM FISHES N. Forteath, Inspection Head Wharf, TAS, Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1560–1567, & 2001, Elsevier Ltd.

Introduction Marine fishes and invertebrates have been kept in aquaria for decades. However, attempts to maintain marine species in a captive environment have been dependent on trial and error for the most part but it has been mainly through the attention and care of aquarists that our knowledge about many marine species has been obtained. This is particularly true of the charismatic syngnathids, which includes at least 40 species of sea horses. During the past 30 years, technological advances in corrosion resistant materials together with advances in aquaculture systems have brought about a rapid increase in demand for large public marine aquarium displays, oceanariums and hobby aquaria suitable for colorful and exotic ornamental species. These developments have led to the establishment of important export industries for live fishes, invertebrates and so-called ‘living rocks’. Attempts to reduce dependence on wild harvesting through the development of marine fish and invertebrate hatcheries met with limited success. The availability of equipment, which greatly assists in meeting the water quality requirements for popular marine organisms, has turned the attention of aquarists towards maintaining increasingly complex living marine ecosystems and more exotic species. A fundamental requirement for the success of such endeavors is the need to understand species biology and interspecific relationships within the tank community. Modern marine aquarists must draw increasingly on scientific knowledge and this is illustrated below with reference to sea horse and coral reef aquaria, respectively.

History The first scientific and public aquarium was built in the London Zoological Gardens in 1853. This facility was closed within a few years and another attempt was not undertaken until 1924. By the 1930s several public aquaria were built in other European capitals but by the end of World War II only that of

524

Berlin remained. During the 1970s the scene was set for a new generation of public aquaria, several specializing in marine displays, and others becoming more popularly known as oceanariums due to the presence of marine mammals, displays of large marine fish and interactive educational activities. Themes have added to the public interest. For example, Monterey Bay Aquarium exhibits a spectacular kelp forest, a theme repeated by several world class aquaria and Osaka Aquarium sets out to recreate the diverse environments found around the Pacific Ocean. Some of these aquaria have found that exhibits of species native to their location alone are not successful in attracting visitor numbers. The New Jersey Aquarium, for example, has been forced to build new facilities and tanks housing over 1000 brightly colored marine tropical fish with other ventures having to rely on the lure of sharks and touch pools. The history of public aquaria has evolved from stand alone tank exhibits to massive 2–3 million liter tanks through which pass viewing tunnels. Once, visitors were content to be mere observers of the fishes and invertebrates but by the end of the millennium the emphasis changed to ensuring the public became actual participants in the aquarium experience. The modern day visitors seek as near an interactive experience as possible and hope to be transformed into the marine environment and witness for themselves the marine underwater world. The concept of modern marine aquarium-keeping in the home has its origins in the United Kingdom and Germany. The United States is now the world’s most developed market in terms of households maintaining aquaria, especially those holding exotic marine species: there are about 2.5 million marine hobby aquariums in the USA. In Holland and Germany, the emphasis has been on reef culture, a hobby which is becoming more widespread. The manufacture of products designed specifically for the ornamental fish trade first began in 1954 and scientific and technical advances during the 1960s brought aquarium keeping a very long way from the goldfish and goldfish bowl. The development of suitable materials for marine aquaria has been hampered by the corrosive nature and the toxicity of materials when immersed in sea water. One of the first authoritative texts on materials and methods for marine aquaria was written by Spotte in 1970, followed by Hawkins in 1981. The volume of marine ornamental fish involved in the international trade is difficult to calculate

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF AQUARIUM FISHES

accurately since records are poor. Current estimates are between 100 and 200 tonnes per annum which probably corresponds to more than 20 million individual specimens. The trade is highly dependent on harvesting from the wild. In Sri Lanka, Indonesia, the Philippines, Fiji and Cook Islands, the export of tropical reef species is now one of the most important export industries employing significant numbers of village people. Members of the family Pomacentridae, in particular clown fishes, Amphiprion spp., and blue– green chromis, Chromis viridis, are central to the industry and cleaner wrasse, Labroides dimidiatus, flame angels, Centropyge loriculus, red hawks, Neocirrhites armatus, tangs, Acanthus spp., and seahorses, Hippocampus spp. are important. The fire shrimp, Lysmata debelius, is a major species with respect to invertebrates. Historically, culture of marine ornamentals has not been in competition with the wild harvest industry. However, recent advances in aquaculture technology will undoubtedly enable more marine ornamentals to be farmed. One European hatchery already has an annual production of Amphiprion spp. which is 15 times that exported from Sri Lanka, whereas in Australia seahorse farming is gaining momentum. In both Europe and USA, commercial production of fire shrimp, L. debelius, is being attempted. To date the greatest impediment to culture lies in the fact that many popular marine ornamentals produce tiny, free-floating eggs and the newly hatched larvae either do not accept traditional prey used for rearing such as rotifers and brine shrimp, or the prey has proved nutritionally inadequate. During the 1980s, the Dutch and German aquarists pioneered the development of miniature reef aquaria. Their success among other things, depended on efficient means of purifying the water. The Dutch developed a ‘wet and dry’ or trickle filter. These filters acted as both mechanical and biological systems. More compact and efficient trickle filter systems have been developed with the advent of Dupla bioballs Table 1

525

during the 1990s. These aquaria are filled with socalled ‘living rock’ which is initially removed from coral reefs. Coral culture per se has recently been established in the Philippines, Solomon Islands, Palau, Guam, and the United States and is aimed at reducing the need to remove living coral from natural reefs. The upsurge in popularity in marine ornamentals over the past 30 years for both public and hobby aquaria has raised serious conservation concerns. There are calls for sustainable management of coral reefs worldwide and even bans on harvesting of all organisms including ‘living rock’. The sea horses in particular have received attention from aquarists and conservationists. Sea horses are one of the most popular of all marine species and are probably responsible for converting more aquarists to marine aquarium-keeping than any other fish. Unlike most other marine ornamentals sea horses have been bred in captivity for many years. In 1996, the international conservation group TRAFFIC (the monitoring arm of the World Wide Trust for Nature) claimed sea horses were under threat from overfishing for use in traditional medicines, aquaria, and as curios. According to TRAFFIC at least 22 countries export sea horses, the largest known exporters being India, the Philippines, Thailand and Vietnam. Importers for aquaria include Australia, Canada, Germany, Japan, The Netherlands, United Kingdom, and the United States. The species commonly in demand are shown in Table 1. The greatest problem for both sides in the debate over the trade in wild-caught marine ornamental fishes in general is a historical lack of scientific data on the biology of these animals both in their natural and captive environments. It is true to say that most of the information about marine ornamentals is derived from intelligence gathering by aquarists and scientific rigor has been applied in the case of only a few species. One such species is the pot-bellied sea horse, Hippocampus abdominalis, which has been studied both in the wild and captive environment for several years. Data on this species serve as useful

Sea horse species commonly kept in aquaria

Aquarium common name

Species name

Geographic reference

Size (cm)

Color

Dwarf sea horse Northern giant Spotted sea horse Short-snouted Mediterranean Golden or Hawaiian Pot-bellied sea horse

Hippocampus zosterae H. erectus H. reidi H. hippocampus H. ramulosus H. kuda H. abdominalis

Florida coast New Jersey coast Florida coast Mediterranean Mediterranean Western Pacific Southern Australian coast

5 20 18 15 15 30 30

Green, gold, black, white Mottled, yellow, red, black, white Mottled, white, black Red/black Yellow/green Golden Mottled, gold, white, black

(c) 2011 Elsevier Inc. All Rights Reserved.

526

MARICULTURE OF AQUARIUM FISHES

comparative tools for knowledge about other ornamental fishes.

Table 2

Parameters affecting life support in marine aquaria

Parameter

The Pot-bellied Sea Horse and Other Aquarium Species Species Suitability for Aquaria

Table 2 sets out major parameters affecting life support in marine aquaria. Factors about a species that it would be advantageous to know prior to selection for the aquarium are:

• • • • • • • •

water temperature range, water quality requirements, behavior and habitat (territorial, aggressive, cannibalistic, pelagic, benthic), diet, breeding biology, size and age, ability to withstand stress, health (resistance and susceptibility).

Physical Temperature Salinity Particular matter

Light

Water motion

Chemical pH and alkalinity Gases

Nutrients

Water Temperature Range

Organic compounds

The pot-bellied seahorse H. abdominalis has a broad distribution along the coastal shores of Australia, being found from Fremantle in Western Australia eastwards as far as Central New South Wales, all around Tasmania and also much of New Zealand. Within its range water temperatures may reach 281C for several months and fall to 91C for a few weeks. Acclimation trials in the laboratory and in home aquaria have shown that this species will live at water temperatures as high as 301C and as low as 81C. Unlike many marine ornamental fishes, H. abdominalis is eurythermal. Many marine aquaria are kept between 24 and 261C which is considered satisfactory for a number of marine ornamentals, however some species require higher temperatures, for example some butterfly fishes (Chaetodontidae) survive best at 291C. Many tropical coral reef species are stenothermous and are difficult to maintain in temperate climates without accurate thermal control. The more eurythermous a species, the easier it will be to acclimate to a range of water temperature fluctuations.

Toxic compounds

Water Quality Requirements

Salinity The salinity of sea water may alter due to freshwater run-off or evaporation. Some marine species are less tolerant than others to salinity changes, and it is important to determine whether or not a species will survive even relatively minor changes. H. abdominalis is euryhaline, growing and

Factor

Composition Size Concentration Artificial/natural Photoperiod Spectrum Intensity Surge Laminar Turbulence

Total gas pressure Dissolved oxygen Un-ionized ammonia Hydrogen sulfide Carbon dioxide Nitrogen compounds Phosphorus compounds Trace metals Biodegradable Nonbiodegradable Heavy metals Biocides

Biological Bacteria Virus Fungi Others

breeding at salinities between 15–37 parts per thousand (%). It is a coastal dwelling species being recorded at depths of 1–15 m and may be present in a range of habitats from estuaries, open rocky substrates and artificial harbors. Often, pot-bellied sea horses can be found attached to nets and cages used to farm other fish species. The somewhat euryecious behavior of this sea horse has probably resulted in its broad tolerance of salinities. Many marine aquaria depend on artificial sea water which is purchased as a salt mixture and added to dechlorinated fresh water. Natural sea water is a complex chemical mixture of salts and trace elements. It has been shown that some artificial seawater mixtures are unsuitable for marine plants and even particular life stages of some fishes. Particular attention must be paid to trace elements in coral reef aquaria when using either natural or artificial salt water. Coral reef aquaria are also more sensitive to salinity changes than general fish aquaria. Table 3 gives a useful saltwater recipe.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF AQUARIUM FISHES

Table 3

The Wiedermann–Kramer saltwater formula

In 100 liters of distilled water: Sodium chloride (NaCl) Magnesium sulfate crystals (MgSO4  7H2O) Magnesium chloride crystals (MgCl2  6H2O) Calcium chloride crystalsa (Cacl2  6H2O) Potassium chloridea (KCl) Sodium bicarbonate (NaHCO3) Sodium bromide (NaBr) Sodium bicarbonate (NaCO3) Boric acid (H3BO3) Strontium chloride (SrCl2) Potassium iodate (KlO3)

2765 g 706 g 558 g 154 g 69.7 g 25 g 10 g 3.5 g 2.6 g 1.5 g 0.01 g

a The potassium chloride should be dissolved separately with some of the 100 liters of distilled water as should the calcium chloride. Add these after the other substances have been dissolved.

Other Water Quality Guidelines

Various tables have been provided setting out water quality guidelines for mariculture. Table 4 is useful for well-stocked marine fish and coral reef aquaria but is possibly too rigid for lightly stocked fish tanks. The toxicity of several parameters given in Table 3 may be reduced by ensuring the water is always close to saturation with respect to dissolved oxygen concentration (DOC). Studies on H. abdominalis have indicated that this species becomes stressed when DOC falls below 85% saturation. Furthermore, the species is tolerant of much higher un-ionized ammonia (NH3-N) levels when the water is close to DOC saturation. A combination of low dissolved oxygen (o80%) and NH3N greater than 0.02 mg l1 may result in high mortalities. Ammonia is the major end product of protein metabolism in sea horses and most aquatic animals. It is toxic in the un-ionized form (NH3). Ammonia concentration expressed as the NH3 compound is converted into a nitrogen basis by multiplying by 0.822. The concentration of un-ionized ammonia depends on total ammonia, pH, temperature, and salinity. The concentration of un-ionized ammonia is equal to:

Table 4

527

Water quality levels for the aquarium

Parameter

Level

Dissolved oxygen Ammonia Nitrite Hydrogen sulfide Chlorine residual pH Copper Zinc

90–100% saturation (> 6 mg l1) o0.02 mg l1 NH3-N o0.1 mg l1 NO2-N o0.001 mg l1 as H2S o0.001 mg l1 7.8–8.2 o0.003 mg l1 o0.0025 mg l1

Table 5 Temp (1C)

20 25 30 a

Mole fraction of un-ionized ammonia in sea watera pH 7.8

7.9

8.0

8.1

8.2

0.0136 0.0195 0.0274

0.0171 0.0244 0.0343

0.0215 0.0305 0.0428

0.0269 0.0381 0.0532

0.0336 0.0475 0.0661

Modified from Huguenin and Colt (1989).

subgravel filter which removes ammonia and nitrite by nitrification and Figure 1(B) shows the configuration of a power filter using both nitrification and absorption to remove ammonia. Both methods have been used to maintain water quality in marine ornamental aquaria. Unfortunately, information on toxicity levels of ammonia for marine ornamentals is poorly documented but Table 4 is probably a useful guide given the data on cultured species. The pot-bellied sea horse is intolerant to even low levels of hydrogen sulfide (H2S) (Table 4). This gas is difficult to measure at low levels thus care is required to avoid anoxic areas in aquaria. H2S is almost certainly toxic to other marine ornamentals also, particularly reef dwellers. Chlorine, copper and zinc have all proved toxic to H. abdominalis at levels exceeding those shown in Table 4. Chorine and copper are often used in  aquaria: the former to sterilize equipment and the Un-ionized ammonia mgl1 as NH3 -NÞ ¼ ðaÞðTANÞ latter in treatment of various diseases. Furthermore, chlorine may be present in tap water when mixed where a ¼ mole fraction of un-ionized ammonia and with artificial seawater mixtures. Great care is required in the use of these chemicals. Available TAN ¼ total ammonia nitrogen (mg l1 as N). Table 5 gives the mole fraction for given tem- aquarium test kits are seldom sensitive enough to detect chronic chlorine concentrations. Often 1–5 mg peratures and pH in sea water. 1 The concentration of un-ionized ammonia is about l sodium thiosulfate or sodium sulfite are used to 40% less in sea water than fresh water, but its toxi- remove chlorine but for some marine species these city is increased by the generally higher pH in the too may prove toxic. Bioassays for chlorine toxicity former. Figure 1(A) shows the operation of a simple using marine ornamentals have not been carried out.

(c) 2011 Elsevier Inc. All Rights Reserved.

528

MARICULTURE OF AQUARIUM FISHES

active species feeding both in the water column and over the substratum. However, at night the fish ‘roosts’ often in association with other specimens. Furthermore, this species is remarkably gregarious and stocking levels as high as 10 fish, 6 cm in length per liter, have been regularly maintained in hatchery trials. Although sea horses may tolerate the presence of several of their own species, their slow feeding behavior puts them at a competitive disadvantage in a mixed species aquarium, where faster feeders will ingest the sea horses’ food. Predation is a serious problem in marine aquaria. Sea horses are known to be prey for other fishes both in the wild and in aquaria. Members of the antennariids, particularly the sargassum fish, Histrio histrio, are known to feed on sea horses as are groupers (Serranidae) and trigger Fishes (Balistinae), flatheads (Platycephalidae) and cod (Moridae). Territorial species are common among the coraldwelling fishes and such behavior makes them difficult to keep in mixed-species aquaria. The blue damsel, Pomacentrus coelestris, shoals in its natural habitat but becomes pugnacious in the confines of the aquarium. Several marine ornamentals seek protection or are cryptic. The majority of clown or anemone fish (Amphiprion) retreat into sea anemones if threatened, in particular, the anemones Stoichaetis spp., Radianthus spp., and Tealia spp., whereas some wrasse dive beneath sand when frightened. Cryptic coloration is seen in the sea horses and color changes have often been reported. The cleaner wrasse, Labroides dimidiatus, lives in shoals over reefs but is successfully maintained singly in the aquarium, where even the most aggressive fish species welcome its attention. Other wrasses mimic the coloration and shape of L. dimidiatus simply to lure potential prey towards them. Behavior and habitat of many marine ornamentals has been gleaned from observation only, but failure to understand these factors make fish-keeping difficult. Figure 1 (A) Undergravel filtration within an aquarium tank. (B) Canister filter with power head and filter media chambers.

Copper toxicity can be significantly reduced with the addition of 1–10 mg l1 of EDTA. EDTA is also a good chelating agent for zinc. Behavior and Habitat

It is widely believed that sea horses spend their time anchored by their prehensile tails to suitable objects. This is not necessarily true. H. abdominalis is an

Diet

The dietary requirements are poorly known for marine ornamentals and many artificial feeds may do little more than prevent starvation without live or frozen feed supplements. Furthermore, a given species may require different foods at various life stages. Several stenophagic species are known in the aquarium mainly consisting of algal and live coral feeders, for example the melon butterfish Chaetodon trifasciatus. The diet of others may not even be known in spite of such fish being sold for the aquarium. The

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF AQUARIUM FISHES

regal angelfish, Pygoplites diacanthus, seldom lives for long in the aquarium and dies from starvation. Sea horses are easy to feed but require either live or frozen crustacea or small fish. The pot-bellied seahorse has been reared through all growth stages using diets of enriched brine shrimp, Artemia salina, live or frozen amphipods, small krill species, and fish fry. The ready acceptance and good growth rates recorded in hatchery-produced pot-bellied sea horses using 48-h-old enriched brine shrimp have resulted in significant numbers of sea horses being raised in at least one commercial farm. Apart from hatcheries for clown fishes and pot-bellied sea horses, the intensive culture of marine ornamentals has proved difficult due to a lack of suitable prey species for fry. Breeding Behavior

Breeding behavior can be induced in several ornamental fishes with appropriate stimuli and environments. The easier species are sequential hermaphrodites such as serranids. The pomacentrids of the genus Amphiprion are a further good example. However, sea horses have been extensively studied. The sea horses are unique in that the male receives, fertilizes and broods the eggs in an abdominal pouch following a ritualized dance. Much has been made of monogamy but as further studies are undertaken scientific support for such breeding behavior is being questioned. H. abdominalis is polygamous both in the wild and captivity. The pot-bellied sea horse in captivity, at least, shows breeding behavior as early as four months of age and males may give birth to a single offspring; by one year of age males may give birth to as many as 80 fry and at two years of age 500 fry. Precocity in other marine ornamentals is not recorded but may exist. Size and Age

The size and age of aquarium fish have been seldom studied scientifically but has been observed. Groupers (Serranidae) and triggerfishes (Balistinae) quickly outgrow aquaria, and some angelfishes and emperors may show dramatic color changes with age, becoming more or less pleasing to aquarists. Longevity likewise is unknown for most marine ornamentals but the pygmy sea horse H. zosterae lives for no more than two years, whereas H. abdominalis may live for up to nine years. Stress

Stress probably plays a pivotal role in the health of marine ornamentals but scientific studies have not

529

been undertaken. The aquarist would do well to remember that stress suppresses aspects of the immune response of fishes and that studies on cultured species demonstrate that capture, water changes, crowding, transport, temperature changes, and poor water quality induce stress responses. Furthermore, stress can be cumulative and some species may be more responsive than others. Farmed species tend to show a higher stress threshold than wild ones. The potential advantage of purchasing hatchery-reared ornamentals (if available) are obvious, since survival in farmed stock should be greater than in wild fish held in aquaria. Health Management

Good health management results from an understanding of the biological needs of a species. Treatment with chemicals is a short-term remedy only and the use of antibiotics may exacerbate problems through bacterial resistance. A considerable number of pathogens have been recorded in marine aquarium fishes and include viruses, bacteria, protozoa, and metazoa. Most diseases have been shown not to be peculiar to a given species but epizootic. For example, several of the disease organisms recorded in sea horses, in particular Vibrio spp., protozoa and microsporidea, are known to infect other fish species also. In coral reef aquaria, nonpathogenic diseases due to poor water quality may be common and in-depth knowledge pertaining to the husbandry of such systems is essential for their well-being.

Coral Reef Aquaria The Challenge

The coral reef is one of the best-adapted ecosystems to be found in the world. Such reefs are biologically derived and the organisms which contribute substantially to their construction are hermatypic corals although ahermatypic species are present. Coral reefs support communities with a species diversity that far exceeds those of neighboring habitats and the symbiotic relationship between zooxanthellae and the scleractinian corals are central to the reef’s wellbeing. Zooxanthellae are also present in many octocorallians, zoanthids, sea anemones, hydrozoans and even giant clams. As zooxanthellae require light for photosynthesis, the reef is dependent on clear water. Coral reefs are further restricted by their requirement for warm water at 20–281C, and the great diversity of life demands a plentiful supply of oxygen. The challenge for the aquarist lies in the need to

(c) 2011 Elsevier Inc. All Rights Reserved.

530

MARICULTURE OF AQUARIUM FISHES

match the physical parameters of the water in the aquarium as closely as possible with sea water of the reef itself. Physical Considerations

Temperature The recommended water temperature for coral reef aquaria is a stable 241C. The greater the temperature fluctuations the less the diversity of life the aquarium will support. At temperatures less than 181C the reef will die and above 301C increasing mortalities among zooxanthellae will occur leading to the death of hermatypic corals. Light Water bathing coral reefs has a blue appearance which has been called the color of ocean deserts. Most of the primary production is the result of photosynthesis by benthic autotrophs (zooxanthellae) rather than drifting plankton. Photosynthetic pigments of zooxanthellae absorb maximally within the light wavelength bands that penetrate furthest into sea (400–750 mm) and therefore clear oceanic water is essential. In the aquarium, both fluorescent and metal halides are available which will supply light at the correct wavelength. Actimic-03 fluorescents combined with white fluorescents are suitable. Typically three tubes, two actinic and one white, will be needed for a 200 liter tank. Metal halide lamps cannot be placed close to the tank because of heat and such lamps produce UV light which will destroy some organisms. A glass sheet placed over the aquarium will prevent UV penetration. One 175 W lamp is recommended per 60 cm of aquarium length.

biological components of the reef itself to produce an autotrophic system, most employ protein skimmers, mechanical filters and biological filters to prevent poisoning of the system. Figure 2 represents nutrient cycling over a coral reef and Figure 1 shows a suitable filter for reef aquaria. Removal of nitrate can be achieved through the use of specialized filters which grow denitrifying bacteria. Living rock Living rock is dead compacted coral which has been colonized by various invertebrates. In addition, there will be algae and bacteria. Different sources of living rock will provide different populations of organisms. Over time the organisms which survived the transfer from reef to aquarium become established

Nitric acid and nitrates

Nitrous acid and nitrites Rebuilt into plants N2O Bacterial oxidation

Ammonia (NH4OH + NH2)

N2 Free nitrogen

Assimilation

Eaten by animals

Water movement Coral reefs are subjected to various types of water movements, namely surges, laminar currents, and turbulence. These water motions play an essential role by bringing oxygen to the corals and plants, removing detritus and, in the case of ahermatypic corals, transporting their food. In the aquarium these necessary water movements must be present. Power head filters are available which produce acceptable surges and currents.

Plant and animal protein

Figure 2 Nutrient cycling over a coral reef.

Table 6 aquaria

Biological Considerations

Nutrients Coral reefs are limited energy and nutrient traps: rather than being lost to deep water sediments, some organic compounds and nutrients are retained and recycled. However, water movements rid the reef of dangerous excess nutrients which might promote major macrophyte growth. The coral reef aquarium soon becomes a nutritional soup if both organic compounds and nutrients are not recycled or removed. Although skilled and knowledgeable aquarists are able to use the

Some additional faunal components for coral reef

Niche

Common name

Detritus feeder

Anemone shrimp Shrimp Fiddler crab Algal feeder Tiger cowrie Money cowrie Starfish Plankton feeders Mandarin fish Psychedelic fish Midas blenny

(c) 2011 Elsevier Inc. All Rights Reserved.

Scientific name Periclimenea brevicarpalis P. pedersoni Uca spp. Cypraea tigris C. moneta Patiria spp. Synchiropus splendidus S. picturatus Escenius midas

MARICULTURE OF AQUARIUM FISHES

and the aquarium is ready for corals. By the time the corals are placed in the aquarium all filter systems must be operating efficiently. Ahermatypic corals should be introduced first and placed in the darker regions of the tank since they feed on plankton. Once hermatypic corals are introduced appropriate blue light for at least 12 hours each day must be available, and strict water quality maintained (Table 4). Nitrate must not rise above 15 mg l1 and pH fall below 8. Additional faunal components The living rock will introduce various invertebrates. Additional species must be selected carefully and on a scientific rather than an esthetic basis. The introduction of detritus and algal feeders will probably be essential and coral eaters must be avoided. Fish species require high protein diets which will necessitate further reductions of ammonia, nitrite, and nitrate from the aquarium. The species given in Table 6 might be considered but there are many others. Selection will depend on the inhabitants.

531

Further Reading Barnes RSK and Mann KH (1995) Fundamentals of Aquatic Ecology. Oxford: Blackwell Scientific. Emmens CW (1995) Marine Aquaria and Miniature Reefs. Neptune City: T.F.H. Publications. Hawkins AD (1981) Aquarium Systems. New York: Academic Press. Huguenin JE and Colt J (1989) Design and Operating Guide for Aquaculture Seawater Systems. Amsterdam: Elsevier. Lawson TB (1995) Fundamentals of Aquacultural Engineering. New York: Chapman & Hall. Spottle S (1979) Seawater Aquariums – the Captive Environment. New York: John Wiley. Timmons MB and Losordo TM (1994) Aquaculture Water Reuse Systems. Engineering Design and Management. Amsterdam: Elsevier. Untergasser D (1989) Handbook of Fish Diseases. Neptune City: T.F.H. Publications.

See also Corals and Human Disturbance. Coral Reefs. Coral Reef and Other Tropical Fisheries. Mariculture Diseases and Health. Mariculture Overview.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF MEDITERRANEAN SPECIES G. Barnabe´, Universite´ de Montpellier II, France F. Doumenge, Muse´e Oce´anographique de Monaco, Monaco Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1567–1572, & 2001, Elsevier Ltd.

Basic Requirements Obtaining Stock for Ongrowing

The starting point of any farming operation is the acquisition of stock for rearing; these may be spat for mollusks or alevins, fry or juveniles for fish. From the wild For the Mediterranean mussel (Mytilus galloprovincialis), spat is always collected from the wild, from rocky shores or shallow harbors where they are abundant. Conditions are less favorable for oyster culture; the native (flat) oyster (Ostrea edulis) is captured only in the Adriatic and the other remnants of natural stocks are unable to support intensive culture. Spat from Japanese (cupped) oysters (Crassostrea gigas) has to be imported from the Atlantic coast. Clam culture utilizes both spat from Mediterranean species (Tapes decussatus) and a species originating in Japan (Tapes philippinarum) which has spread very rapidly, especially in the Adriatic. Juvenile marine fish such as eels (Anguilla anguilla), mullets (Mugil sp.), gilthead sea bream (Sparus aurata) and sea bass (Dicentrarchus labrax) are traditionally captured in spring in the mouths of rivers or in traps or other places in protected lagoons in Italy; these form the basis of the valliculture of the northern Adriatic, a type of extensive fish culture. In practice, elvers and yellow eels are supplied mainly from fisheries in other Mediterranean lagoons, especially in France. Bluefin tuna (Thunnus thynnus) for ongrowing are taken from the spring and summer fishery for juveniles weighing a few tens of kilograms by Spanish seine netters in the western Mediterranean and Croatians in the Adriatic. Controlled reproduction in hatcheries The transition to intensive fish culture has been made possible by the control of reproduction in sea bass and sea bream (see Mariculture Overview). In 1999 around 100 hatcheries produced 209 million sea bass fry and 242 million gilthead sea bream fry; this represents respectively a doubling and a 50%

532

increase on production in 1997. Hatcheries are very different from the structures used for ongrowing; France exports tens of millions of juveniles to all the Mediterranean countries.

Access to Technology

As an activity, aquaculture is becoming increasingly complex with regard to the technology associated with rearing as well as the interactions with the physical and socio-economic environment. It requires more specialized training than can be obtained in the traditional workplace. Management of a hatchery, monitoring of water quality, and genetic research are examples of the new requirements for training. The transition to cage culture took place using cages designed and manufactured for salmonid culture. Manufacturing feed granules has developed only recently in countries such as Greece.

Feeding

Mollusk culture exploits the natural production of plankton and thus follows natural changes in productivity. The extensive valliculture systems used in northern Italy in managed protected lagoons are based on natural production improved through control of the water (fish trapped in the channels communicating with the sea, overwintering in deep areas) and input of juveniles. Production remains low (20 kg ha1 y1). Intensive fish culture in cages uses pelleted proteinrich diets where fishmeal comprises 60% of the dry weight. However, bluefin tuna only consume whole fish or fresh or frozen cephalopods during ongrowing. Their conversion rate is exceptionally poor, of the order of 15–20% depending on water temperature.

Transport

The aquaculture cycle requires dependable and rapid transport to take juveniles to ongrowing facilities and, especially, to move the final production output which is generally valuable and perishable to the consumer market, and must remain chilled at all times. For isolated sites, particularly small islands, there is always the problem that the transport of feed and various items of equipment is expensive.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF MEDITERRANEAN SPECIES

533

Manpower Requirements

Unfavorable

Hatcheries require a specialized and motivated workforce, as the production cycle for fry must not be interrupted. Sea sites are becoming ever more highly mechanized, even in countries such as Turkey where labor is cheap, because the volumes to be handled are increasing continuously.

Eutrophication, toxic blooms, bacterial pollution As elsewhere in the world, Mediterranean aquaculture cannot escape problems of disease. Nodovirus has hit fish in cages and two protozoans have decimated cultured stocks of the native oyster, Ostrea edulis. Harmful blooms frequently affect lagoons where shellfish are cultured and the presence of serious pathogens (cholera, salmonella) has prohibited the use of certain zones for production.

Capital

Mediterranean aquaculture requires heavy investment both for setting up and for operating. It is dependent on international capital. Scandinavian, British, and Japanese interests play a considerable part, alongside local entrepreneurs.

Distinctive aspects Favorable

Sheltered waters There are numerous large areas of sheltered water around the Mediterranean, the shoreline having been submerged and straightened out by the rapid rise in sea level following the last glaciation, 12 000 years ago. Nowadays, many bays provide big expanses of water which are both deep and sheltered (Carthagena, Toulon, La Spezia, Gaete, Naples, Trieste, Thessaloniki, etc.). These waters are productive, but are threatened by pollution. Channels between the islands of the Dalmatian archipelago or in the Aegean Sea or the large gulfs surrounding the Hellenic peninsular (Patras, Corinth, Thessaloniki, Argolis, Evvoia, etc.), and to the north of Entail are perfectly suited to the types of aquaculture systems that benefit from a rapid turnover of water. Above all, the input of sediment from deltas or the effects of littoral erosion have led to the build-up of sand bars forming lido-type complexes of impounded lagoons where the waters experience strong variations in salinity and high productivity because of the movements through gaps in the barriers. High demand from local markets The Latin countries of the north-west Mediterranean and Adriatic have a culinary tradition based on the consumption of large quantities of seafood. Markets in Spain, Mediterranean France, and Italy pay high prices for the fresh products of mollusk culture and marine fish farming. There is also the strong seasonal demand from more than 100 million tourists who travel around all the coasts and particularly the islands of the Mediterranean, except where there are political problems.

Extreme temperatures In sheltered lagoons or enclosed waters (e.g., the Bougara Sea, Tunisia), temperatures may exceed 301C in summer, causing harmful blooms which are prejudicial to the success of aquaculture. In the north-west Mediterranean and the northern Adriatic, the water temperature in the lagoons may drop to below þ 51C and ice may develop on the surface of the lagoons. In general, thermal conditions favor the eastern Mediterranean. Competition for space (urbanization, tourism) Littoral space is under pressure from several users; tourism, navigation, and especially the extension of industrial and urban developments. Land bases are essential for mariculture. Lack of sites is further aggravated by complex regulations which are applied rigorously. Finally, the tendency to designate large areas of wetland and similar areas of sea for nature conservation removes significant areas from the expansion of aquaculture. Dependence on fisheries When juvenile mollusks or fish for mariculture rearing are taken from the natural environment they are supplied by fisheries. This dependence is a major risk to the effective operation of the production process. In the absence of industrial pelagic fisheries there is no fishmeal production in the Mediterranean. In order to satisfy the demand linked to the expansion of fish culture, almost all components of the feed must be imported. This places production further under external control. Substitution of plant for animal protein in the diet is becoming necessary. Cultural limits of the markets While the culture of Roman Catholic Christianity provides for a boosted local market, the Orthodox Christian culture is far less demanding in terms of seafood products. The Muslim countries of the southern coast and the Levantine Basin do not have the monetary resources nor, significantly, a dietary tradition adapted to the products of marine aquaculture. In contrast, the Mediterranean bluefin tuna achieves high prices on the Japanese sashimi market.

(c) 2011 Elsevier Inc. All Rights Reserved.

534

MARICULTURE OF MEDITERRANEAN SPECIES

Production Systems Mollusk Culture

The oyster reared everywhere (although sometimes in small quantities) is the cupped or Japanese oyster (Crassostrea gigas) introduced to Europe from Japan in the 1960s to replace the Portuguese oyster (Crassostrea angulata), stocks of which had been drastically reduced by disease. One single major center, the Etang de Thau, produces around 12 000 tonnes each year using spat taken from the Atlantic coast. The native mussel (Mytilus galloprovincialis) is reared in small quantities around the Spanish coast (Ebro Delta, Mar Menor). In contrast, in Italy annual production is between 100 000 and 130 000 tonnes; the main sites are the Gulf of Taranto and the northern and mid-Adriatic, as well as several bays in the Tyrrhenean Sea (La Spezia, Gaete, and Naples). Along the French coast, only the Etang de Thau has a production of a few thousand tonnes. In Greece, the Gulf of Thessaloniki has a production of the same order. For around 20 years, Japanese-type long lines have been installed in the open sea off Languedoc Rousillon in waters between 15 and 35 m deep. These structures resist the forces of the sea and production can reach tens of thousands of tonnes of mussels, but is strongly limited by storms and predation by sea bream. Oysters and mussels are always cultured in suspension on ropes either in lagoons or in the sea. The spat of oyster is captured on the Atlantic coast on various substrates and is often attached with the help of quick-setting cement on the rearing bars, as this technique yields the best results. Other farmers leave the spat to develop on the mollusc shells where they have attached; the shells are placed in nets, spaced regularly, on the rearing ropes. Mussel spat, collected from the wild, is placed in a tubular net, which is then attached to the rearing ropes. Mussels attach themselves to the artificial substrate with the aid of their byssus. It is therefore necessary to detach and clean the mussels once or twice during their growth. Clams are especially abundant around the Italian Adriatic coast. The Japanese species, introduced as hatchery-reared spat at the beginning of the 1980s to seed protected areas, has spread very rapidly and invaded neighboring sectors. The density of these mollusks can reach 4000 individuals per square meter and production is around 40 000 tonnes per annum. A large part of the production is exported to Spain.

to the demand for high quality aquatic products that cannot be supplied on a regular basis from fisheries based on wild stocks. This production is based on hatcheries (see Mariculture Overview). Broodstock can be held in cages or in earth ponds. In such ponds, control of the photoperiod and temperature allows fertilized eggs to be obtained through almost the whole year. Larvae are reared in the same way as those of other marine fish (see Marine Fish Larvae) although sea bass can be fed on Artemia nauplii from first feeding; this does away with the burden of rearing rotifers in these hatcheries. The trend is towards enlargement to bring about economies of scale; some hatcheries produce 20 million juveniles each year; those producing fewer than 1–2 million juveniles per year are unlikely to be profitable. Water is generally recycled within the hatchery to save energy and to maintain the stability of physicochemical characteristics. The quality of the water available in the natural environment determines the suitability of a site for a hatchery. Almost all ongrowing is carried out in sea cages. All Mediterranean countries have contributed to the development and there is a tendency to move further and further east. Cages have been installed in sheltered coastal waters, but the scarcity of such sites and progress in cage technology are encouraging the development of exposed sites in the open sea. Techniques used in this type of farming resemble those used in the cage farming of salmonids. (see Salmonid Farming) and identical cages are used. These have a diameter of up to 20 m and are up to 10 m deep. Dry granular feed is used and the conversion rate is continuously improving (between 1.3 and 2). Growth rate is determined by water temperature; in Greece a sea bass reaches a weight of 300 g in 11 months, but would take twice as long to reach the same weight on the French coast. This demonstrates the advantage of the eastern Mediterranean. Other species, for which larval rearing is more difficult, are reared on a small scale using different techniques. These include the dentex (Dentex dentex), common (Couch’s) sea bream (Pagrus pagrus) produced in mesocosms in Greece, and the greater amberjack (Seriola dumerilii), ongrown in Spain. Diversification of species offers the potential for increasing markets; this is becoming true for the meagre (Argyrosomos regius) and the red drum (Scianops ocellatus), and many other trials are taking place. Ongrowing Bluefin Tuna

Culture of Sea Bass and Gilthead Sea Bream

Control and management of the spawning of sea bass and sea bream has made it possible to respond

Japanese attempts at rearing larval bluefin tuna have not yet produced economically viable results. Mediterranean aquaculture depends on juveniles or

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OF MEDITERRANEAN SPECIES

sub-adults captured in the spring fishery, which are transported in nets supported by rafts to large cages where they will be kept for a few months. The towage, which must be at speeds of between 1 and 2 knots, may take several weeks. The fish are harvested at the end of autumn and beginning of winter, before periods of low temperature and bad weather. In general, the tuna double their weight after 6–7 months of ongrowing and their flesh acquires the color and quality which puts them in demand for the Japanese sashimi market.

Regionalization During the last 20 years, regional specialization has developed progressively, based on a balance of favorable and unfavorable factors for each of the types of mariculture. The development of Mediterranean aquaculture production has thus been subject to an evolutionary process which has taken account not only of the major factors previously described but also conditions peculiar to each nation. This has produced contrasting situations, as decribed below. France and Spain – Relatively Low Level of Aquaculture Development

This is due to a coincidence of relatively high levels of aquaculture development on the Atlantic coasts of both countries (mussels and fish in Galicia, mussels and oysters in the Bay of Biscay and the English Channel) and large quantities of imports supplementing regional production. Aquaculture developments remain small in size and are very spread out. One exception, demonstrating the possibilities that exist, is the success of the ongrowing of bluefin tuna in Spain in the Bay of Carthagena, from which 6000–7000 tonnes were sold in 1999–2000, for a revenue of over US $100 000 000. The Size of the Italian Market

In the year 2000 only one third of the market was supplied by national production, in spite of rapid increases; production of sea bass (8800 tonnes) and gilthead sea bream (6200 tonnes) have both doubled from 1997. Italy absorbs almost all of the Maltese production (600 tonnes of sea bass and 1600 tonnes of sea bream in 2000); most of this comes from stock from eastern Mediterranean hatcheries. The 40 000 tonnes of clams produced in the Adriatic saturate the market and the excess is exported to Spain. In spite of health problems, Italian mussel culture supports an annual market of 100 000–130 000 tonnes. In addition, the upper Adriatic

535

has preserved the tradition of exploiting the valli and the lagoons. In 2000 the Italian aquaculture market (by value) was supplied 40% by sea bass and sea bream (200 million lire), 33% by clams (165 million lire), and 27% by mussels (135 million lire). The Dalmatian archipelago belonging to Croatia retains a sector of high quality traditional oyster culture. However, new possibilities have opened up with the transfer of technology and finance from Croatian emigrants to south Australia who have developed the ongrowing of bluefin tuna in the Ile de Kali, using the techniques they practiced in Port Lincoln. However, biological and logistic constraints are currently holding back development. Production from the Croatian businesses in 1999 and 2000 was limited to 1000 tonnes of small tuna (20–25 kg each); these receive relatively poor prices in Japan.

The Pioneering Front for Culture of Sea Bass and Sea Bream in the Eastern Mediterranean

This has reached Greece where production has gone from 1600 tonnes in 1990 to 6000 tonnes in 1992, 18 000 tonnes in 1996, 36 000 tonnes in 1998 and 56 000 tonnes in 2000. The movement then reached neighbouring Turkey, going from 1200 tonnes in 1992 to 3500 tonnes in 1994, reaching 12 500 tonnes in 1998 and 14 000 tonnes in 2000. Cyprus has also recently developed production which reached 1500 tonnes in 2000. As part of this conquest of space, the first Greek seafish farms were installed between 1985 and 1995 in the west, in the Ionian islands and in Arcadia (center of the Peloponese). Then, via the Gulf of Corinth, they were joined from 1990 to 1995 by an active center in Argolis (east of the Peloponese). In addition, suitable sites for cages were found in the bays behind the barrier of the Isle of Evvoia. Finally, since 1990, developments have progressed to the Aegean around the archipelagos fringing the Peloponese. Between 1995 and 2000 Greek sites appear to have become saturated and mariculture has moved to the Turkish coast and the Anatolian bays of the Aegean coast.

Perspectives and Problems Lack of sites

The area dedicated to intensive cage mariculture remains modest: the whole of the French marine fish culture, around 6000 tonnes annual production, occupies no more than 10 ha of sea and 5 ha of land.

(c) 2011 Elsevier Inc. All Rights Reserved.

536

MARICULTURE OF MEDITERRANEAN SPECIES

Pollution Ascribed to Aquaculture

It is politically correct to speak of the pollution derived from intensive cage rearing. When this alleged pollution (from fish excreting mainly nitrogen and phosphorus) is ejected into oligotrophic open waters such as the Mediterranean it can be seen as a benefit rather than a nuisance. The FAO, elsewhere, has suggested that significant increases in fish catches occur in areas where human-derived wastes have increased in the Mediterranean. Health Limits and Shellfish

value. This type of rearing has expanded eastwards from the European Mediterranean countries but has not yet reached the southern shore. Markets, particularly the huge European market of 360 million inhabitants, are not yet saturated. Diversification of the species produced may open up new markets. The expansion of cage-based mariculture has not yet finished, while progress in technology is unpredictable. The major missing element in Mediterranean aquaculture is the rearing of penaeids, in spite of several sporadic but insignificant attempts at production in Southern Italy and Morocco.

Production of mussels fluctuates widely: pollution of coastal and lagoon waters (toxic plankton and bacterial pollution) regularly prevents their sale. Regular consumption of mussels could be dangerous as the species concentrates okadaic acid (a strong carcinogen) produced by toxic phytoplankton.

See also

Marketing Problems

Further Reading

Overproduction of sea bass and gilthead sea bream has led to a periodic collapse in prices. This phenomenon appears to be due to lack of planning, organization, and commercial astuteness by the producers, as well as competition between countries operating within different economic frameworks (cost of manpower). The salmon market is characterized by identical examples. For bluefin tuna, dependence on a single, distant market (the sashimi market in Japan) makes ongrowing a risky activity but very profitable in economic terms.

Barnabe G (1974) Mass rearing of the bass Dicentrarchus labrax L. In: Blaxter JHS (ed.) The Early Life History of Fish, pp. 749--753. Berlin: Springer-Verlag. Barnabe G (1976) Ponte induite ET e´levage des larves du Loup Dicentrarchus labrax (L.) et de la Dorade Sparus aurata (L.). Stud. Rev. C.G.P.M. (FAO) 55: 63--116. Barnabe G (1990) Open sea aquaculture in the Mediterranean. In: Barnabe G (ed.) Aquaculture, pp. 429--442. New York: Ellis Horwood. Barnabe G (1990) Rearing bass and gilthead bream. In: Barnabe G (ed.) Aquaculture, pp. 647--683. New York: Ellis Horwood. Barnabe G and Barnabe-Quet R (1985) Avancement et ame´lioration de la ponte induite chez le Loup Dicentrarchus labrax (L) a` l0 aide d0 un analogue de LHRH injecte´. Aquaculture 49: 125--132. Doumenge F (1991) Medite´rrane´e. In: Encyclopaedia Universalis pp. 871–873. Doumenge F (1999) L0 aquaculture des thons rouges et son de´veloppement e´conomique. Biol Mar Medi 6: 107--148. Ferlin P and Lacroix D (2000) Current state and future development of aquaculture in the Mediterranean region. World Aquaculture 31: 20--23. Heral M (1990) Traditional oyster culture in France. In: Barnabe´ G (ed.) Aquaculture, pp. 342--387. New York: Ellis Horwood.

Conclusions Traditional mollusk culture maintains its position but is encountering problems of limited availability of water. Transfer out to the open sea which is less polluted has been piloted in Languedoc (France) for two decades, but has demonstrated neither the suitability of the techniques nor their profitability, and production has stagnated. The explosive growth of the production of marine fish in cages can be said to demonstrate the true revolution in Mediterranean mariculture. This is based entirely on species with a high commercial

Mariculture Overview. Salmonid Farming.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OVERVIEW M. Phillips, Network of Aquaculture Centres in Asia-Pacific (NACA), Bangkok, Thailand & 2009 Elsevier Ltd. All rights reserved.

Introduction: Mariculture – A Growing Ocean Industry of Global Importance Global production from aquaculture has grown substantially in recent years, contributing in evermore significant quantities to the world’s supply of fish for human consumption. According to FAO statistics, in 2004, aquaculture production from mariculture was 30.2 million tonnes, representing 50.9% of the global total of farmed aquatic products. Freshwater aquaculture contributed 25.8 million tonnes, or 43.4%. The remaining 3.4 million tonnes, or 5.7%, came from production in brackish environments (Figure 1). Mollusks and aquatic plants (seaweeds), on the other hand, almost evenly make up most of mariculture at 42.9% and 45.9%, respectively. These statistics, while accurately reflecting overall trends, should be viewed with some caution as the definition of mariculture is not adopted consistently across the world. For example, when reporting to FAO, it is known that some countries report penaeid shrimp in brackish water, and some in mariculture categories. In this article, we focus on the culture of aquatic animals in the marine environment. Nevertheless, the amount of aquatic products from farming of marine animals and plants is substantial, and expected to continue to grow. The overall production of mariculture is dominated by Asia, with seven of the top 10 producing countries within Asia. Other regions of Latin America and Europe however

Brackish water culture 6%

Mariculture 51% Freshwater culture 43%

also produce significant and growing quantities of farmed marine product. The largest producer of mariculture products by far is China, with nearly 22 million tonnes of farmed marine species. A breakdown of production among the top 15 countries, and major commodities produced, is given in Table 1.

Commodity and System Descriptions A wide array of species and farming systems are used around the world for farming aquatic animals and plants in the marine environment. The range of marine organisms produced through mariculture currently include seaweeds, mollusks, crustaceans, marine fish, and a wide range of other minor species such as sea cucumbers and sea horses. Various containment or holding facilities are common to marine ecosystems, including sea-based pens, cages, stakes, vertical or horizontal lines, afloat or bottom set, and racks, as well as the seabed for the direct broadcast of clams, cockles, and similar species. Mollusks

Many species of bivalve and gastropod mollusks are farmed around the world. Bivalve mollusks are the major component of mariculture production, with most, such as the commonly farmed mussel, being high-volume low-value commodities. At the other end of the value spectrum, there is substantial production of pearls from farming, an extremely low-volume but high-value product. Despite the fact that hatchery production technologies have been developed for many bivalves and gastropods, much bivalve culture still relies on collection of seedstock from the wild. Artificial settlement substrates, such as bamboo poles, wooden stakes, coconut husks, or lengths of frayed rope, are used to collect young bivalves, or spat, at settlements. The spat are then transferred to other grow-out substrates (‘relayed’), or cultured on the settlement substrate. Some high-value species (such as the abalone) are farmed in land-based tanks and raceways, but most mollusk farming takes place in the sea, where three major systems are commonly used:

•

Figure 1 Aquaculture production by environment in 2004. From FAO statistics for 2004.

Within-particulate substrates – this system is used to culture substrate-inhabiting cockles, clams, and other species. Mesh covers or fences may be used to exclude predators.

(c) 2011 Elsevier Inc. All Rights Reserved.

537

538

MARICULTURE OVERVIEW

Table 1

Top 15 mariculture producers

Country

Production (tonnes)

Major species/commodities farmed in mariculture

China

21 980 595

Seaweeds (kelp, wakame), mollusks (oysters, mussels, scallops, cockles, etc.) dominate, but very high diversity of species cultured, and larger volumes of marine fish and crustaceans Seaweeds (Kappaphycus), with small quantities of fish (milkfish and groupers) and mollusks Seaweeds, mollusks, marine fish Seaweeds, mollusks, marine fish Atlantic salmon, other salmonid species, smaller quantities of mollusks, and seaweed Atlantic salmon, other salmonid species Seaweeds, mollusks

Philippines Japan Korea, Republic of Chile Norway Korea, Dem. People’s Rep. Indonesia Thailand Spain United States of America France United Kingdom Vietnam Canada

1 273 598 1 214 958 927 557 688 798 637 993 504 295

Total (all countries)

30 219 472

420 919 400 400 332 062 242 937

Seaweeds, smaller quantities of marine fish, and pearls Mollusks (cockle, oyster, green mussel), small quantities of marine fish (groupers) Blue mussel dominates, but high diversity of fish and mollusks Mollusks (oysters) with small quantities of other mollusks and fish (salmon)

198 625 192 819 185 235 134 699

Mollusks (mussel and oyster) with small quantities of other mollusks and fish (salmon) Atlantic salmon and mussels Seaweeds (Gracilaria) and mollusks Atlantic salmon and other salmonid species, and mollusks

From FAO statistics for 2004.

•

•

On or just above the bottom – this culture system is commonly used for culture of bivalves that tolerate intertidal exposure, such as oysters and mussels. Rows of wooden or bamboo stakes are arranged horizontally or vertically. Bivalves may also be cultured on racks above the bottom in mesh boxes, mesh baskets, trays, and horizontal wooden and asbestos-cement battens. Surface or suspended culture – bivalves are often cultured on ropes or in containers, suspended from floating rafts or buoyant long-lines.

Management of the mollusk cultures involves thinning the bivalves where culture density is too high to support optimal growth and development, checking for and controlling predators, and controlling biofouling. Mollusk production can be very high, reaching 1800 tonnes per hectare annually. With a cooked meat yield of around 20%, this is equivalent to 360 tonnes of cooked meat per hectare per year, an enormous yield from a limited water area. Farmed bivalves are commonly sold as whole fresh product, although some product is simply processed, for example, shucked and sold as fresh or frozen meat. There has been some development of longer-life products, including canned and pickled mussels. Because of their filter-feeding nature, and the environments in which they are grown, edible bivalves are subject to a range of human health concerns, including accumulation of heavy metals, retention of human-health bacterial and viral pathogens,

and accumulation of toxins responsible for a range of shellfish poisoning syndromes. One option to improve the product quality of bivalves is depuration, which is commonly practiced with temperate mussels, but less so in tropical areas. Seaweeds

Aquatic plants are a major production component of mariculture, particularly in the Asia-Pacific region. About 13.6 million tonnes of aquatic plants were produced in 2004. China is the largest producer, producing just less than 10 million tonnes. The dominant cultured species is Japanese kelp Laminaria japonica. There are around 200 species of seaweed used worldwide, or which about 10 species are intensively cultivated – including the brown algae L. japonica and Undaria pinnatifida; the red algae Porphyra, Eucheuma, Kappaphycus, and Gracilaria; and the green algae Monostrema and Enteromorpha. Seaweeds are grown for a variety of uses, including direct consumption, either as food or for medicinal purposes, extraction of the commercially valuable polysaccharides alginate and carrageenan, use as fertilizers, and feed for other aquaculture commodities, such as abalone and sea urchins. Because cultured seaweeds reproduce vegetatively, seedstock is obtained from cuttings. Grow-out is undertaken using natural substrates, such as long-lines, rafts, nets, ponds, or tanks.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OVERVIEW

Production technology for seaweeds is inexpensive and requires only simple equipment. For this reason, seaweed culture is often undertaken in relatively undeveloped areas where infrastructure may limit the development of other aquaculture commodities, for example, in the Pacific Island atolls. Seaweeds can be grown using simple techniques, but are also subject to a range of physiological and pathological problems, such as ‘green rot’ and ‘white rot’ caused by environmental conditions, ‘ice-ice’ disease, and epiphyte growth. In addition, cultured seaweeds are often consumed by herbivores, particularly sea urchins and herbivorous fish species, such as rabbitfish. Selective breeding for specific traits has been undertaken in China to improve productivity, increase iodine content, and increase thermal tolerance to better meet market demands. More recently, modern genetic manipulation techniques are being used to improve temperature tolerance, increase agar or carrageenan content, and increase growth rates. Improved growth and environmental tolerance of cultured strains is generally regarded as a priority for improving production and value of cultured seaweeds in the future. Seaweed aquaculture is well suited for small-scale village operations. Seaweed fisheries are traditionally the domain of women in many Pacific island countries, so it is a natural progression for women to be involved in seaweed farming. In the Philippines and Indonesia, seaweed provides much-needed employment and income for many thousands of farmers in remote coastal areas. Marine Finfish

Marine finfish aquaculture is well established globally, and is growing rapidly. A wide range of species is cultivated, and the diversity of culture is also steadily increasing. In the Americas and northern Europe, the main species is the Atlantic salmon (Salmo salar), with smaller quantities of salmonids and species. Chile, in particular, has seen the most explosive growth of salmon farming in recent years, and is poised to become the number-one producer of Atlantic salmon. In the Mediterranean, a range of warmer water species are cultured, such as seabass and seabream. Asia is again the major producer of farmed marine fish. The Japanese amberjack Seriola quinqueradiata is at the top of the production tables, with around 160 000 tonnes produced in 2004, but the region is characterized by the extreme diversity of species farmed, in line with the diverse fish-eating habits of

539

the people living in the region. Seabreams are also common, with barramundi or Asian seabass (Lates calcarifer) cultured in both brackish water and mariculture environments. Grouper culture is expanding rapidly in Asia, driven by high prices in the live-fish markets of Hong Kong SAR and China, and the decreasing availability of wild-caught product due to overfishing. Southern bluefin tuna (Thunnus mccoyii) is cultured in Australia using wild-caught juveniles. Although production of this species is relatively small (3500–4000 tonnes per annum in 2001–03), it brings very high prices in the Japanese market and thus supports a highly lucrative local industry in South Australia. The 2003 production of 3500 tonnes was valued at US$65 million. Hatchery technologies are well developed for most temperate species (such as salmon and seabream) but less well developed for tropical species such as groupers where the industry is still reliant on collection of wild fingerlings, a concern for future sustainability of the sector. The bulk of marine fish are presently farmed in net cages located in coastal waters. Most cultured species are carnivores, leading to environmental concerns over the source of feed for marine fish farms, with most still heavily reliant on wild-caught socalled ‘trash’ fish. Excessive stocking of cages in coastal waters also leads to concern over water and sediment pollution, as well as impacts from escapes and disease transfer on wild fish populations. Crustaceans

Although there is substantial production of marine shrimps globally, this production is undertaken in coastal brackish water ponds and thus does not meet the definition of mariculture. There has been some experimental culture of shrimp in cages in the Pacific, but this has not yet been commercially implemented. Tropical spiny rock lobsters and particularly the ornate lobster Panulirus ornatus are cultured in Southeast Asia, with the bulk of production in Vietnam and the Philippines. Lobster aquaculture in Vietnam produces about 1500 tonnes valued at around US$40 million per annum. Tropical spiny rock lobsters are cultured in cages and fed exclusively on fresh fish and shellfish. In the medium to long term, it is necessary to develop hatchery production technology for seedstock for tropical spiny rock lobsters. There is currently considerable research effort on developing larval-rearing technologies for tropical spiny rock lobsters in Southeast Asia and in Australia. As in the case of tropical marine fish farming, there is also a need to develop less-wasteful and less-polluting diets

(c) 2011 Elsevier Inc. All Rights Reserved.

540

MARICULTURE OVERVIEW

to replace the use of wild-caught fish and shellfish as diets. Other Miscellaneous Invertebrates

There are a range of other invertebrates being farmed in the sea, such as sea cucumbers, sponges, corals, sea horses, and others. Farming of some species has been ongoing for some time, such as the well-developed sea cucumber farming in northern China, but others are more recent innovations or still at the research stage. Sponge farming, for example, is generating considerable interest in the research community, but commercial production of farmed sponges is low, mainly in the Pacific islands. This farming is similar to seaweed culture as sponges can be propagated vegetatively, with little infrastructure necessary to establish farms. The harvested product, bath sponges, can be dried and stored and, like seaweed culture, may be ideal for remote communities, such as those found among the Pacific islands.

Environmental Challenges Environmental Impacts

Mariculture is an important economic activity in many coastal areas but is facing a number of environmental challenges because of the various environmental ‘goods’ and ‘services’ required for its development. The many interactions between mariculture and the environment include impacts of: (1) the environment on mariculture; (2) mariculture on the environment; and (3) mariculture on mariculture. The environment impacts on mariculture through its effects on water, land, and other resources necessary for successful mariculture. These impacts may be negative or positive, for example, water pollution may provide nutrients which are beneficial to mariculture production in some extensive culture systems, but, on the other hand, toxic pollutants and pathogens can be extremely damaging. An example is the farming of oysters and other filter-feeding mollusks which generally grow faster in areas where nutrient levels are elevated by discharge of wastewater from nearby centers of human population. However, excessive levels of human and industrial waste cause serious problems for mollusk culture, such as contamination with pathogens and toxins from dinoflagellates. Aquaculture is highly sensitive to adverse environmental changes (e.g., water quality and seed quality) and it is therefore in the long-term interests of mariculture farmers and governments to work toward protection and enhancement of environmental quality. The effects of global climate change, although poorly understood in the fishery

sector, are likely to have further significant influences on future mariculture development. The impacts of mariculture on the environment include the positive and negative effects farming operations may have on water, land, and other resources required by other aquaculturists or other user groups. Impacts may include loss or degradation of natural habitats, water quality, and sediment changes; overharvesting of wild seed; and introduction of disease and exotic species and competition with other sectors for resources. In increasingly crowded coastal areas, mariculture is running into more conflicts with tourism, navigation, and other coastal developments. Mariculture can have significant positive environmental impacts. The nutrient-absorbing properties of seaweeds and mollusks can help improve coastal water quality. There are also environmental benefits from restocking of overfished populations or degraded habitats, such as coral reefs. For example, farming of high-value coral reef species is being seen as one means of reducing threats associated with overexploitation of threatened coral reef fishes traditionally collected for food and the ornamental trade. Finally, mariculture development may also have an impact on itself. The rapid expansion in some areas with limited resources (e.g., water and seed) has led to overexploitation of these resources beyond the capacity of the environment to sustain growth, followed by an eventual collapse. In mariculture systems, such problems have been particularly acute in intensive cage culture, where self-pollution has led to disease and water-quality problems which have undermined the sustainability of farming, from economic and environmental viewpoints. Such problems emphasize the importance of environmental sustainability in mariculture management, and the need to minimize overharvesting of resources and hold discharge rates within the assimilative capacity of the surrounding environment. The nature and the scale of the environmental interactions of mariculture, and people’s perception of their significance, are also influenced by a complex interaction of different factors, such as follows:

•

The technology, farming and management systems, and the capacity of farmers to manage technology. Most mariculture technology, particularly in extensive and semi-intensive farming systems, such as mollusk and seaweed farming, and well-managed intensive systems, is environmentally neutral or low in impact compared to other food production sectors.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OVERVIEW

•

•

•

•

The environment where mariculture farms are located (i.e., climatic, water, sediment, and biological features), the suitability of the environment for the cultured animals and the environmental conditions under which animals and plants are cultured. The financial and economic feasibility and investment, such as the amount invested in proper farm infrastructure, short- versus long-term economic viability of farming operations, and investment and market incentives or disincentives, and the marketability of products. The sociocultural aspects, such as the intensity of resource use, population pressures, social and cultural values, and aptitudes in relation to aquaculture. Social conflicts and increasing consumer perceptions all play an important role. The institutional and political environment, such as government policy and the legal framework, political interventions, plus the scale and quality of technical extension support and other institutional and noninstitutional factors.

These many interacting factors make both understanding environmental interactions and their management (as in most sectors – not just mariculture) both complex and challenging.

•

•

Environmental Management of Mariculture The sustainable development of mariculture requires adoption of management strategies which enhance positive impacts (social, economic, and environmental impacts) and mitigate against environmental impacts associated with farm siting and operation. Such management requires consideration of: (1) the farming activity, for example, in terms of the location, design, farming system, investment, and operational management; (2) the ‘integration’ of mariculture into the surrounding coastal environment; and (3) supporting policies and legislation that are favorable toward sustainable development. Technology and Farming Systems Management

The following factors are of crucial importance in environmental management at the farm level:

•

Farm siting. The sites selected for aquaculture and the habitat at the farm location play one of the most important roles in the environmental and social interactions of aquaculture. Farm siting is also crucial to the sustainability of an investment; incorrect siting (e.g., cages located in areas with unsuitable water quality) often lead to increased

•

541

investment costs associated with operation and amelioration of environmental problems. Farms are better sited away from sensitive habitats (e.g., coral reefs) and in areas with sufficient water exchange to maintain environmental conditions. Problems of overstocking of mollusk culture beds are recognized in the Republic of Korea, for example, where regulations have been developed to restrict the areas covered by mollusk culture. For marine cage culture, one particularly interesting aspect of siting is the use of offshore cages, and new technologies developed in European countries are now attracting increasing interest in Asia. Farm construction and design features. Farm construction and system design has a significant influence on the impact of mariculture operations on the environment. Suitable design and construction techniques should be used when establishing new farms, and as far as possible seek to cause minimum disturbance to the surrounding ecosystems. The design and operation of aquaculture farms should also seek to make efficient use of natural resources used, such as energy and fuel. This approach is not just environmentally sound, but also economic because of increasing energy costs. Water and sediment management. Development of aquaculture should minimize impacts on water resources, avoiding impacts on water quality caused by discharge of farm nutrients and organic material. For sea-based aquaculture, where waste materials are discharged directly into the surrounding environment, careful control of feed levels and feed quality is the main method of reducing waste discharge, along with good farm siting. In temperate aquaculture, recent research has been responsible for a range of technological and management innovations – low-pollution feeds and novel self-feeding systems, lower stocking densities, vaccines, waste-treatment facilities – that have helped reduce environmental impacts. Complex models have also been developed to predict environmental impacts, and keep stocking levels within the assimilative capacity of the surrounding marine environment. In mariculture, there are also examples of integrated, polyculture, and alternate cropping farming systems that help to reduce impacts. For example in China and Korea, polyculture on sea-based mollusk and seaweed farms is practiced and for more intensive aquaculture operations, effluent rich in nutrients and microorganisms, is potentially suitable for culturing fin fish, mollusks, and seaweed. Suitable species and seed. A supply of healthy and quality fish, crustacean, and mollusk seed is

(c) 2011 Elsevier Inc. All Rights Reserved.

542

•

•

•

MARICULTURE OVERVIEW

essential for the development of mariculture. Emphasis should be given to healthy and quality hatchery-reared stock, rather than collection from the wild. Imports of alien species require import risk assessment and management, to reduce risks to local aquaculture industries and native biodiversity. Feeds and feed management. Access to feeds, and efficient use of feeds is of critical importance for a cost-effective and environmentally sound mariculture industry. This is due to many factors, including the fact that feeds account for 50% or more of intensive farming costs. Waste and uneaten feed can also lead to undesirable water pollution. Increasing concern is also being expressed about the use of marine resources (fish meal as ingredients) for aquaculture feeds. One of the biggest constraints to farming of carnivorous marine fish such as groupers is feed. The development of sustainable supplies of feed needs serious consideration for future development of mariculture at a global level. Aquatic animal health management. Aquatic animal and plant diseases are a major cause of unsustainability, particularly in more intensive forms of mariculture. Health management practices are necessary to reduce disease risks, to control the entry of pathogens to farming systems, maintain healthy conditions for cultured animals and plants, and avoid use of harmful disease control chemicals. Food safety. Improving the quality and safety of aquaculture products and reducing risks to ecosystems and human health from chemical use and microbiological contamination is essential for modern aquaculture development, and marketing of products on domestic and international markets. Normally, seafood is considered healthy food but there are some risks associated with production and processing that should be minimized. The two food-safety issues, that can also be considered environmental issues, are chemical and biological. The chemical risk is associated with chemicals applied in aquaculture production and the biological is associated with bacteria or parasites that can be transferred to humans from the seafood products. Increasing calls for total traceability of food products are also affecting the food production industry such that consumers can be assured that the product has been produced without addition of undesirable or harmful chemicals or additives, and that the environments and ecosystems affected by the production facilities have not been compromised in any way.

•

Economic and social/community aspects. The employment generated by mariculture can be highly significant, and globally aquaculture has become an important employer in remote and poor coastal communities. Poorly planned mariculture can also lead to social conflicts, and the future development and operation of mariculture farms must also be done in a socially responsible manner, as far as possible without compromising local traditions and communities, and ideally contributing positively. The special traditions of many coastal people and their relation with the sea in many places deserve particularly careful attention in planning and implementation of mariculture.

Planning, Policy, and Legal Aspects Integrated Coastal Area Management

Effective planning processes are essential for sustainable development of mariculture in coastal areas. Integrated coastal area management (ICAM) is a concept that is being given increasing attention as a result of pressures on common resources in coastal areas arising from increasing populations combined with urbanization, pollution, tourism, and other changes. The integration of mariculture into the coastal area has been the subject of considerable recent interest, although practical experience in implementation is still limited in large measure because of the absence of adequate policies and legislation and institutional problems, such as the lack of unitary authorities with sufficiently broad powers and responsibilities. Zoning of aquaculture areas within the coastal area is showing some success. In China, Korea, Japan, Hong Kong, and Singapore, there are now well-developed zoning regulations for water-based coastal aquaculture operations (marine cages, mollusks, and seaweeds). For example, Hong Kong has 26 designated ‘marine fish culture zones’ within which all marine fish-culture activities are carried out. In the State of Hawaii, ‘best areas’ for aquaculture have been identified, and in Europe zoning laws are being strictly applied to many coastal areas where aquaculture is being developed. Such an approach allows for mariculture to be developed in designated areas, reducing risks of conflicts with other coastal zone users and uses. Policy and Legal Issues

While much can be done at farm levels and by integrated coastal management, government involvement

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE OVERVIEW

through appropriate policy and legal instruments is important in any strategy for mariculture sustainability. Some of the important issues include legislation, economic incentives/disincentives, private sector/community participation in policy formulation, planning processes, research and knowledge transfer, balance between food and export earnings, and others. While policy development and most matters of mariculture practice have been regarded as purely national concerns, they are coming to acquire an increasingly international significance. The implication of this is that, while previously states would look merely to national priorities in setting mariculture policy, particularly legislation/standards, for the future it will be necessary for such activities to take account of international requirements, including various bilateral and multilateral trade policies. International standards of public health for aquaculture products and the harmonization of trade controls are examples of this trend. Government regulations are an important management component in maintaining environmental quality, reducing negative environmental impacts, and allocating natural resources between competing users and integration of aquaculture into coastal area management. Mariculture is a relative newcomer among many traditional uses of natural resources and has commonly been conducted within an amalgam of fisheries, water resources, and agricultural and industrial regulations. It is becoming increasingly clear that specific regulations governing aquaculture are necessary, not least to protect aquaculture development itself. Key issues to be considered in mariculture legislation are farm siting, use of water area and bottom in coastal and offshore waters; waste discharge, protection of wild species, introduction of exotic or nonindigenous species, aquatic animal health; and use of drugs and chemicals. Environmental impact assessment (EIA) can also be an important legal tool which is being more widely applied to mariculture. The timely application of EIA (covering social, economic, and ecological issues) to larger-scale coastal mariculture projects can be one way to properly identify environmental problems at an early phase of projects, thus enabling proper environmental management measures to be incorporated in project design and management. Such measures will ultimately make the project more sustainable. A major difficulty with EIAs is that they are difficult (and generally impractical) to apply to smaller-scale mariculture developments, common throughout many parts of Asia, and do not easily take account of the potential

543

cumulative effects of many small-scale farms. Strategic environmental assessment (SEA) can provide a broader means of assessing impacts.

Conclusion Mariculture is and will increasingly become an important producer of aquatic food in coastal areas, as well as a source of employment and income for many coastal communities. Well-planned and -managed mariculture can also contribute positively to coastal environmental integrity. However, mariculture’s future development will occur, in many areas, with increasing pressure on coastal resources caused by rising populations, and increasing competition for resources. Thus, considerable attention will be necessary to improve the environmental management of aquaculture through environmentally sound technology and better management, supported by effective policy and planning strategies and legislation.

See also Mariculture, Economic and Social Impacts.

Further Reading Clay J (2004) World Aquaculture and the Environment. A Commodity by Commodity Guide to Impacts and Practices. Washington, DC: Island Press. FAO/NACA/UNEP/WB/WWF (2006) International Principles for Responsible Shrimp Farming, 20pp. Bangkok, Thailand: Network of Aquaculture Centres in Asia-Pacific (NACA). http://www.enaca.org/uploads/ international-shrimp-principles-06.pdf (accessed Apr. 2008). Hansen PK, Ervik A, Schaanning M, et al. (2001) Regulating the local environmental impact of intensive, marine fish farming-II. The monitoring programme of the MOM system (Modelling-Ongrowing fish farmsMonitoring). Aquaculture 194: 75--92. Hites RA, Foran JA, Carpenter DO, Hamilton MC, Knuth BA, and Schwager SJ (2004) Global assessment of organic contaminants in farmed salmon. Science 303: 226--229. Joint FAO/NACA/WHO Study Group (1999) Food safety issues associated with products from aquaculture. WHO Technical Report Series 883. http://www.who. int/foodsafety/publications/fs_management/en/aquaculture.pdf (accessed Apr. 2008). Karakassis I, Pitta P, and Krom MD (2005) Contribution of fish farming to the nutrient loading of the Mediterranean. Scientia Marina 69: 313--321. NACA/FAO (2001) Aquaculture in the third millennium. In: Subasinghe RP, Bueno PB, Phillips MJ, Hough C, McGladdery SE, and Arthur JR (eds.) Technical

(c) 2011 Elsevier Inc. All Rights Reserved.

544

MARICULTURE OVERVIEW

Proceedings of the Conference on Aquaculture in the Third Millennium. Bangkok, Thailand, 20–25 February 2000, 471pp. Bangkok, NACA and Rome: FAO. Naylor R, Hindar K, Flaming IA, et al. (2005) Fugitive salmon: Assessing the risks of escaped fish from net-pen aquaculture. BioScience 55: 427--473. Neori A, Chopin T, Troell M, et al. (2004) Integrated aquaculture: Rationale, evolution and state of the art emphasizing sea-weed biofiltration in modern mariculture. Aquaculture 231: 361--391. Network of Aquaculture Centres in Asia-Pacific (2006) Regional review on aquaculture development. 3. Asia and the Pacific – 2005. FAO Fisheries Circular No. 1017/3, 97pp. Rome: FAO. Phillips MJ (1998) Tropical mariculture and coastal environmental integrity. In: De Silva S (ed.) Tropical Mariculture, pp. 17–69. London: Academic Press. Pillay TVR (1992) Aquaculture and the Environment, 158pp. London: Blackwell. Secretariat of the Convention on Biological Diversity (2004) Solutions for sustainable mariculture – avoiding the adverse effects of mariculture on biological diversity, CBD Technical Series No. 12. http://www. biodiv.org/doc/publications/cbd-ts-12.pdf (accessed Apr. 2008). Tacon AJC, Hasan MR, and Subasinghe RP (2006) Use of fishery resources as feed inputs for aquaculture

development: Trends and policy implications. FAO Fisheries Circular No. 1018. Rome: FAO. World Bank (2006) Aquaculture: Changing the Face of the Waters. Meeting the Promise and Challenge of Sustainable Aquaculture. Report no. 36622. Agriculture and Rural Development Department, the World Bank. http://siteresources.worldbank.org/INTARD/Resources/ Aquaculture_ESW_vGDP.pdf (accessed Apr. 2008).

Relevant Websites http://www.pbs.org – Farming the Seas, Marine Fish and Aquaculture Series, PBS. http://www.fao.org/fi – Food and Agriculture Organisation of the United Nations. http://www.cbd.int Jakarta Mandate, Marine and Coastal Biodiversity: Mariculture, Convention on Biological Diversity. http://www.enaca.org – Network of Aquaculture Centres in Asia-Pacific. http://www.seaplant.net – The Southeast Asia Seaplant Network. http://www.oceansatlas.org – UN Atlas of the Oceans.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS C. R. Engle, University of Arkansas at Pine Bluff, Pine Bluff, AR, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Mariculture is a broad term that encompasses the cultivation of a wide variety of species of aquatic organisms, including both plants and animals. These different products are produced across the world with a wide array of technologies. Each technology in each situation will entail various price and cost structures within different social contexts. Thus, each aquaculture enterprise will have distinct types and levels of economic and social impacts. Moreover, there are as many management philosophies, strategies, and business plans as there are aquaculture entrepreneurs. Each of these will result in different economic and social impacts. As an example, consider two shrimp farms located in a developing country that utilize the same production technology. One farm hires and trains local people as both workers and managers, invests in local schools and health centers, while paying local and national taxes. This farm will likely have a large positive social and economic impact. On the other hand, another farm that imports managers, pays the lowest wages possible, displaces local families through land acquisitions, pays few taxes, and exports earnings to developed countries may have negative social and perhaps even economic impacts. Economic and social structure, interactions, and impacts are complex and interconnected, even in rural areas with seemingly simple economies. This article discusses a variety of types of impacts that can occur from mariculture enterprises.

Economic Impacts Economic impacts begin with the direct effects from the sale of product produced by the mariculture operation. However, the impacts extend well beyond the effect of sales revenue to the farm. As the direct output of marine fish, shellfish, and seaweed production increases, the demand for supply inputs such as feed, fingerlings, equipment, repairs, transportation services, and processing services also increases.

These activities represent indirect effects. Subsequent increased household spending will follow. As the industry grows, employment in all segments of the industry also grows and these new jobs create more income that generates additional economic activity. Thus, growth of the mariculture industry results in greater spending that multiplies throughout the economy. Mariculture is an important economic activity in many parts of the world. Table 1 lists the top 15 mariculture-producing countries in the world, both in terms of the volume of metric tons produced and the value in 2005. Its economic importance can be measured in total sales volume, total employment, or total export volume for large aquaculture industry sectors. Macroeconomic effects include growth that promotes trade and domestic resource utilization. Fish production in the Philippines, for example, accounted for 3.9% of gross domestic product (GDP) in 2001. In India, it contributed 1.4% of national GDP and 5.4% of agricultural GDP. On the microlevel, incomes and livelihoods of the poor are enhanced through mariculture production. Small-scale and subsistence mariculture provides high-quality protein for household consumption,

Table 1 Top 15 mariculture-producing countries, with volumes (metric tons) produced in 2005 and value (in US) Country

Volume of production (metric ton)

Value of production ($1000 US)

China The Philippines Japan Korea, Rep. Indonesia Chile Korea, Dem. Norway Thailand France Canada Spain United Kingdom United States Vietnam Others

22 677 724 1 419 727 1 211 959 1 042 142 950 819 889 867 703 292 504 295 656 636 347 750 216 103 136 724 190 426 161 339 134 937 173 800

14 981 801 165 335 3 848 906 1 317 250 338 093 3 069 169 283 362 2 072 562 58 587 571 543 503 974 262 394 584 152 235 912 158 800 3 268 283

Total

31 417 540

31 720 123

Source: FishStat Plus (http://www.fao.org).

(c) 2011 Elsevier Inc. All Rights Reserved.

545

546

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

35 000 000

30 000 000

Metric ton

25 000 000

20 000 000

15 000 000

10 000 000

5 000 000

19

5 19 0 5 19 2 5 19 4 5 19 6 5 19 8 60 19 6 19 2 6 19 4 6 19 6 6 19 8 7 19 0 7 19 2 7 19 4 7 19 6 7 19 8 8 19 0 82 19 8 19 4 8 19 6 88 19 9 19 0 9 19 2 9 19 4 9 19 6 9 20 8 0 20 0 02 20 04

0

Figure 1 Growth of mariculture production worldwide, 1950–2005. Source: Food and Agriculture Organization of the United Nations.

generates supplemental income from sales to local markets, and can serve as a savings account to meet needs for cash during difficult financial times.

Table 2 Top 15 mariculture species cultured, volume (metric tons), and value ($1000 US), 2005 Species

Volume (metric ton)

Value ($1000 US)

1 216 791 436 924 391 210 713 846 280 267 2 880 687 4 911 256 1 387 990 4 496 196 183 575 866 383 238 331 2 739 753 985 667 1 239 811 8 448 853

4 659 841 436 772 385 131 589 836 26 679 2 358 586 2 941 148 1 419 130 3 003 831 762 707 121 294 131 188 1 101 507 394 257 1 677 870 11 710 346

31 417 540

31 720 123

Income from Sales Revenue

According to the Food and Agriculture Organization of the United Nations, mariculture production grew from just over 300 000 metric ton in 1950 to 31.4 million metric ton valued at $31.7 billion in 2005 (Figure 1). Figure 1 also shows that the total volume of mariculture production has tripled over the past decade alone. Sales revenue received by operators of mariculture businesses has increased rapidly over this same time period. The top mariculture product category in 2005, based on quantity produced, was Japanese kelp, followed in descending order of quantities produced by oysters, Japanese carpetshell, wakame, Yesso scallop, and salmon (Table 2). In terms of value, salmon was the top mariculture product produced in 2005, followed by oysters, kelp, and Japanese carpetshell. The top five marine finfish raised in 2005 (in terms of value) were Atlantic salmon, Japanese amberjack, rainbow trout, seabreams, and halibut. The increase in production and sales of mariculture products has resulted not only from the expansion of existing products and technologies (i.e., salmon, Japanese amberjack) but also from the rapid development of technologies to culture new species

Atlantic salmon Blood cockle Blue mussel Constricted tagelus Green mussel Japanese carpetshell Japanese kelp Laver (nori) Pacific cupped oyster Rainbow trout Red seaweeds Sea snails Wakame Warty gracilaria Yesso scallop Others Total

Source: FishStat Plus (http://www.fao.org).

(i.e., cobia, grouper, and tuna). New mariculture farms have created new markets and opportunities for local populations, such as the development of backyard hatcheries in countries such as Indonesia. In Bali, for example, the development of small-scale hatcheries has resulted in substantial increases in

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

income as compared to the more traditional coconut crops. The immediate benefit to a mariculture entrepreneur is the cash income received from sale of the aquatic products produced. This income then is spent to pay for cash expenditures related to feed, fingerlings, repairs to facilities and equipment, fuel, labor, and other operating costs. Payments on loans will be made to financial lenders involved in providing capital for the facilities, equipment, and operation of the business. Profit remaining after expenses can be spent by the owner on other goods and services, invested back into the business, or invested for the future benefit of the owner. Mariculture of certain species has begun to exceed the value of the same species in capture fisheries. Salmon is a prime example, in which the value of farmed salmon has exceeded that of wild-caught salmon for a number of years. A similar trend can be observed for the rapidly developing capture-based mariculture technologies. In the Murcia region of Spain, for example, the economic value of tuna farming now represents 8 times the value of regional fisheries. Employment

Mariculture businesses generate employment for the owner of the business, for those who serve as managers and foremen, and those who constitute the principal workforce for the business. Studies in China, Vietnam, Philippines, Indonesia, and Thailand suggest that shrimp farming uses more labor per hectare than does rice farming. Employment is generated throughout the market supply chain. Development of mariculture businesses increases demand for fry and fingerlings, feeds, construction, equipment used on the farm (trucks, tractors, aerators, nets, boats, etc.), and repairs to facilities and equipment. In Bangladesh, it is estimated that about 300 000 people derive a significant part of their annual income from shrimp seed collection. Mariculture businesses also create jobs for fish collectors, brokers, and vendors. These intermediaries provide important marketing functions that are difficult for smaller-scale producers to handle. With declines in catch of some previously important fish products, fish vendors need new fish products to sell. Increasing volumes from mariculture can keep these marketers in business. The employment stimulated by development of mariculture businesses is especially significant in economically depressed areas with high rates of unemployment and underemployment. Many mariculture activities develop in rural coastal communities

547

where jobs are limited and mariculture can constitute a critical source of employment. Many jobs in fishing communities have been reduced as fishing opportunities have become more scarce. Given its relationship to the fishing industry and some degree of similarity in the skills required, mariculture businesses are often a welcome alternative for the existing skilled workforce. Capture-based mariculture operations in particular provide job opportunities that require skill sets that are easily met by those who have worked in the fishing industry. The rapid development of capturebased tuna farms in the Mediterranean Sea provides a good example. In Spain, fishermen have become active partners in tuna farms. As a result, the number of specialized bluefin tuna boats has increased to supply the capture-based tuna farms in the area. These boats capture the small pelagic fish that are used as feed. In Croatia, trawlers transport live fish or feed to tuna farms off shore. This has generated important sources of employment in heavily depopulated Croatian islands. Tuna farming is labor-intensive, but offers opportunities for younger workers to develop new skills. Many of the employees on the tuna farms are young, 25–35 years old, who work as divers. The divers are used to crowd the tuna for hand-harvesting without stressing the fish, inspect for mortalities, and check the integrity of moorings. Working conditions on tuna farms are preferable to those on tuna boats because the hours and salaries are more regular. Workers can spend weekends on shore as compared to spending long periods of time at sea on tuna boats. The improved working conditions have improved social stability. Formal studies of economic impacts have measured the effects on employment for fish farm owners and from the secondary businesses such as supply companies, feed mills, processing plants, transportation services, etc. In the Philippines, fish production accounted for 4.4% of total national employment in 2001. In the United States, catfish production accounted for nearly half of all employment in one particular county. Tax Revenue

Tax policies vary considerably from country to country. Nevertheless, there is some form of tax structure in most countries that is used to finance local, regional, and national governments. Tax structures typically include some form of business tax in addition to combinations of property, income, value-added, or sales taxes. Tax revenue is the major source of revenue used for investments in roads and

(c) 2011 Elsevier Inc. All Rights Reserved.

548

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

bridges, schools, health, and public sanitation facilities. Mariculture businesses contribute to the tax base in that particular region or market. Moreover, the wages paid to managers and workers in the business are subject to income, sales, and property taxes. As the business grows, it pays more taxes. Increasing payrolls from increased employment by the company results in greater tax revenue from increased incomes, increased spending (sales taxes), and increased property taxes paid (as people purchase larger homes and/or more land). Export Revenue and Foreign Exchange

The largest volume of mariculture production is in developing countries whereas the largest markets for mariculture products are in the more developed nations of the world. Thus, much of the flow of international trade in mariculture products is from the developing to the developed world. This trade results in export revenue for the companies involved. Exporters in the live grouper trade in Asia have been shown to earn returns on total costs as high as 94%. Export volumes and revenue further generate foreign exchange for the exporting country. This is particularly important for countries that are dependent on other countries for particular goods and services. In order to import products not produced domestically, the country needs sufficient foreign exchange. Moreover, if the country exports products to countries whose currency is strong (i.e., has a high value relative to other forms of currency), the country can then import a relatively greater amount of product from countries whose currency does not have as high a value. Many low-income fooddeficient countries use fish exports to pay for imports of low-value food commodities. For example, in 2000, Indonesia, China, the Philippines, India, and Bangladesh paid off 63%, 62%, 22%, 54%, and 32%, respectively, of their food import bills by exporting fish and fish products. Economic Growth

Economic growth is generated from increased savings, investment, and money supply. The process starts with profitable businesses. Profits can either be saved by the owner or invested back into the business. Savings deposited in a bank or invested in stocks or bonds increase the amount of capital available to be loaned out to other potential investors. With increasing amounts of capital available, the cost of capital (the interest rate) is reduced, making it easier for both individuals and businesses to borrow capital. As a result, spending on housing,

vehicles, property, and new and expanded businesses increases. This in turn increases demand for housing, vehicles, and other goods which increases demand for employment. As the economy grows (as measured by the GDP), the standard of living rises as income levels rise and citizens can afford to purchase greater varieties of goods and services. The availability of goods and services also increases with economic growth. Economic growth increases demand which also tends to increase price levels. As prices increase, businesses become more profitable and wages rise. Economic growth is necessary for standards of living to increase. However, continuous economic growth means that prices also rise continuously. Continuous increases in prices constitute inflation that, if excessive, can create economic difficulties. It is beyond the scope of this article to discuss the mechanisms used by different countries to manage the economy on a national scale and how to manage both economic growth and inflation. Economic Development

Economic development occurs as economies diversify, provide a greater variety of goods and services, and as the purchasing power of consumers increases. This combination results in a higher standard of living for more people in the country. Economic development results in higher levels of education, greater employment opportunities, and higher income levels. Communities are strengthened with economic development because increasing numbers of jobs result in higher income levels. Higher standards of living provide greater incentives for young people to stay in the area rather than outmigrate in search of better employment and income opportunities. Studies of the economic impacts of aquaculture have highlighted the importance of backwardly linked businesses. These are the businesses that produce fingerlings, feed, and sell other supplies to the aquaculture farms. Studies of the linkages in the US catfish industry show that economic output of and the value added by the secondary businesses are greater than those of the farms themselves. Economic Multiplier Effects

Each time a dollar exchanges hands in an economy, its impact ‘multiplies’. This occurs because that same dollar can then be used to purchase some other item. Each purchase represents demand for the product being purchased. Each time that same dollar is used to purchase another good or service, it generates additional demand in the economy. Those businesses

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

that have higher expenditures, particularly those with high initial capital expenditures, tend to have higher multiplier effects in the economy. Aquaculture businesses tend to have high capital expenditures and, thus, tend to have relatively higher multipliers than other types of businesses, particularly other types of agricultural businesses. These high capital expenditures come from the expense of building ponds, constructing net pens or submerged cages, and the equipment costs associated with aerators, boats, trucks, tractors, and other types of equipment. As an example of how economic multipliers occur, consider a new shrimp farm enterprise. Construction of the ponds will require either purchasing bulldozers and bulldozer operators or contracting with a company that has the equipment and expertise to build ponds. The shrimp farm owner pays the pond construction company for the ponds that were built. The dollars paid to the construction company are used to pay equipment operators, fuel companies, for water supply and control pipes and valves, and repairs to equipment. The pond construction workers use the dollars received as wages to purchase more food, clothing for the family, school costs, healthcare costs, and other items. The dollars spent for food in the local market become income for the vendors who pay the farmers who raised the food. Those farmers can now pay for their seed and fertilizer, buy more clothing, and pay school and healthcare expenses, etc. In this expenditure chain, the dollar ‘turned over’ five times, creating additional economic demand each time. Several formal studies of economic impacts from aquaculture have been conducted. These studies have shown large amounts of total economic output, employment, and value-added effects from aquaculture businesses. On the local area, aquaculture can generate the majority of economic output in a given district. Multipliers as high as 6.1 have been calculated for aquaculture activities on the local level.

549

To the extent that a mariculture business creates an externality that increases costs for another business or community, there will be negative economic impacts. The primary potential sources of negative economic impacts from mariculture include: (1) discharge of effluents that may contain problematic levels of nutrients, antibiotics, or nonnative species; (2) spread of diseases; (3) use of high levels of fish meal in the diets fed to the species cultured; (4) clearing mangroves from coastal areas; and (5) consequences of predator control measures. The history of the development of mariculture includes cases in which unregulated discharge of untreated effluents from shrimp farms resulted in poor water-quality conditions in bays and estuaries. Since this same water also served as influent for other shrimp farms, its poorer quality stressed shrimp and facilitated spread of viral diseases. Economic losses resulted. Concerns have been noted over the increasing demand for fish meal as mariculture of carnivorous species grows. The fear is that this increasing use will result in a decline in the stocks of the marine species that serve as forage for wild stocks and that are used in the production of fish meal. If this were to occur, economic losses could occur from declines in capture fisheries. Loss of mangrove forests in coastal areas results in loss of important nursery grounds for a number of species and less protection during storms. The loss of mangrove areas is due to many uses, including for construction, firewood, for salt production, and others. If the construction of ponds for mariculture reduces mangrove areas, additional negative impacts will occur. Control of predators on net pens and other types of mariculture operations often involves lethal methods. The use of lethal methods raises concerns related to biodiversity and viability of wild populations. While there is no direct link to other economic activities in many of these cases, the loss of ecosystem services from reductions in biodiversity may at some point result in other economic problems.

Potential Negative Economic Impacts

Mariculture has been criticized by some for what economists call ‘externalities’. Externalities refer to an effect on an individual or community other than the individual or community that created the problem. Pollution is often referred to as an externality. For example, if effluents discharged from a mariculture business create water-quality problems for another farm or community downstream, that farm or community will have to spend more to clean up that pollution. Yet, those costs are ‘external’ to the business that created the problem.

Social Impacts Food Security

Aquaculture is an important food security strategy for many individuals and countries. Domestic production of food provides a buffer against interruptions in the food supply that can result from international disputes, trade embargoes, war, transportation accidents, or natural disasters. There are many examples of fish ponds being used to feed households during times of strife. Examples include

(c) 2011 Elsevier Inc. All Rights Reserved.

550

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

fish pond owners in Vietnam during the war with the United States and during the Contra War in Nicaragua, among others. Aquaculture has been shown to function as a food reserve for many subsistence farming households. Many food crops are seasonal in nature while fish in ponds or cages can be available year-round and used particularly during periods of prolonged drought or in between different crops. Pond levees also can provide additional land area to raise vegetable crops for household consumption and sale.

Poverty alleviation is accompanied by a number of positive social impacts. These include improved access to food (that results in higher nutritional and health levels), improved access to education (due to higher income levels and ability to pay for fees and supplies), and improved employment opportunities. In Sumatra, Indonesia, profits from grouper farming enable members of Moslem communities to make pilgrimages to Mecca, enriching both the individuals and the community. In Vietnam, income from grouper hatcheries contributes 10–50% of the annual income of fishermen.

Nutritional Benefits

Fish are widely recognized to be a high-quality source of animal protein. Fish have long been viewed as ‘poor people’s food’. Subsistence farmers who raise fish, shellfish, or seaweed have a ready source of nutritious food for home consumption throughout the year. Particularly in Asia, fish play a vital role in supplying inexpensive animal protein to poorer households. This is important because the proportion of the food budget allocated to fish is higher among low-income groups. Moreover, it has been shown that rural people consume more fish than do urban dwellers and that fish producers consume more fish than do nonproducers. Those farmers who also supply the local market with aquatic products provide a supply of high-quality protein to other households in the area. Increasing supplies of farmed products result in lower prices in seafood markets. Tuna prices in Japan, for example, have been falling as a result of the increased farmed supply. This has resulted in making tuna and its nutritional benefits more readily available to middle-income buyers in Japan. Health Benefits

In addition to the protein content, a number of aquatic products provide additional health benefits. Farmed fish such as salmon have high levels of omega-3 fatty acids that reduce the risk of heart disease. Products such as seaweeds are rich in vitamins. Mariculture can enable the poorest of the poor and even the landless to benefit from public resources. For example, cage culture of fish in public waters, culture of mollusks and seaweeds along the coast, and culture-based fisheries in public water bodies provide a source of healthy food for subsistence families. Poverty Alleviation

As aquaculture has grown, its development has frequently occurred in economically depressed areas.

Enhancement of Capture Fisheries

Marine aquaculture can improve the condition of fisheries through supplemental stockings from aquaculture. This will help to meet the global demand for fish products. Increased mariculture supply relieves pressure on traditional protein sources for species such as live cod and haddock. Given that fish supplies from capture fisheries are believed to have reached or be close to maximum sustainable yield, mariculture can reduce the expected shortage of fish. Potential Negative Social Impacts

Mariculture has been criticized for creating negative social impacts. These criticisms have tended to focus on displacement of people if large businesses buy land and move people off that land. Economic development is accompanied by social change and individuals and households frequently are affected in various ways by construction of new infrastructure and production and processing facilities. In developed countries, federal laws typically provide for some degree of compensation for land taken through eminent domain for construction of highways or power lines. However, developing countries rarely have similar mechanisms for compensating those who lose land. In situations in which the resources involved are in the public domain (such as coastal areas), granting a concession for a mariculture operation along the coast may result in reduced access for poor and landless individuals to collect shellfish and other foods for their households. Resource ownership or use rights in coastal zones frequently are ambiguous. Many of these areas traditionally have been common access areas, but are now under pressure for development for mariculture activities. Displaced and poor migrant people are frequently marginalized in coastal lands and often depend to some degree on common resources. However, the extent of conflict over the use of resources is variable. Surveys in Asia reported less than 10% of farms experiencing conflicts in most

(c) 2011 Elsevier Inc. All Rights Reserved.

MARICULTURE, ECONOMIC AND SOCIAL IMPACTS

countries, but higher incidents of conflict were reported in India (29%) and China (94%). In the Mediterranean, capture-based mariculture has resulted in conflicts between local tuna fishermen who fish with long lines, and cage towing operations. Other conflicts have occurred in Croatia due to the smell and pollution during the summer from bluefin tuna farms. Uncollected fat skims on the sea surface that results from feeding oily trash fish spread outside the licensed zones onto beaches frequented by tourists. On the positive side, tourism in Spain was enhanced by offering guided tours to offshore cages. Additional negative impacts could potentially include: (1) marine mammal entanglements; (2) threaten genetic makeup of wild fish stocks; (3) privatize what some think of as free open-access resource; (4) esthetically undesirable; and (5) negative impacts on commercial fishermen. For example, excessive collection of postlarvae and egg-laden female shrimp can result in loss of income for fishermen and reduction of natural shrimp and fish stocks.

Conclusions The economic and social impacts of mariculture are variable throughout the world and are based on the range of technologies, cultures, habitats, land types, and social, economic, and political differences. Positive economic impacts include increased revenue, employment, tax revenue, foreign exchange, economic growth and development, and multiplicative effects through the economy. Negative economic effects could occur through excessive discharge of effluents, spread of disease, and through excessive use of fish meal. Mariculture enhances food security of households and countries and provides nutritional and health benefits while alleviating poverty. However, if earnings are exported and access to public domain resources restricted to poorer classes, negative social impacts can occur.

See also Diversity of Marine Species. Dynamics of Exploited Marine Fish Populations. Global Marine Pollution.

551

Mariculture Diseases and Health. Mariculture of Aquarium Fishes. Mariculture of Mediterranean Species. Mariculture Overview. Marine Chemical and Medicine Resources. Marine Fishery Resources, Global State of. Marine Mammal Trophic Levels and Interactions. Marine Mammals, History of Exploitation. Marine Protected Areas. Network Analysis of Food Webs.

Further Reading Aguero M and Gonzalez E (1997) Aquaculture economics in Latin America and the Caribbean: A regional assessment. In: Charles AT, Agbayani RF, Agbayani EC, et al. (eds.) Aquaculture Economics in Developing Countries: Regional Assessments and an Annotated Bibliography. FAO Fisheries Circular No. 932. Rome: FAO. Dey MM and Ahmed M (2005) Special Issue: Sustainable Aquaculture Development in Asia. Aquaculture Economics and Management 9(1/2): 286pp. Edwards P (2000) Aquaculture, poverty impacts and livelihoods. Natural Resource Perspectives 56 (Jun. 2000). Engle CR and Kaliba AR (2004) Special Issue: The Economic Impacts of Aquaculture in the United States. Journal of Applied Aquaculture 15(1/2): 172pp. Ottolenghi F, Silvestri C, Giordano P, Lovatelli A, and New MB (2004) Capture-Based Aquaculture. The Fattening of Eels, Groupers, Tunas and Yellowtails, 308pp. Rome: FAO. Tacon GJ (2001) Increasing the contribution of aquaculture for food security and poverty alleviation. In: Subasinghe RP, Bueno P, Phillips MJ, Hough C, McGladdery SE, and Arthur JR (eds.) Aquaculture in the Third Millennium. Technical Proceedings of the Conference on Aquaculture in the Third Millennium, pp. 63–72, Bangkok, Thailand, 20–25 February 2000. Bangkok and Rome: NACA and FAO.

Relevant Websites http://www.fao.org – Food and Agriculture Organization of the United Nations. http://www.worldfishcenter.org – WorldFish Center.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE ALGAL GENOMICS AND EVOLUTION A. Reyes-Prieto, H. S. Yoon, and D. Bhattacharya, University of Iowa, Iowa City, IA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: The Genomic Perspective The understanding of oceanic biodiversity and ecology has been revolutionized by the study of complete nuclear genome sequences. Genomic approaches have helped to elucidate ecological-geochemical processes like the carbon cycle and the metabolism of microorganisms that are the key players in oceanic ecosystems. Genomics also has been used to understand gene expression patterns during massive algal ‘blooms’, the evolution, diversification, and ecology of photosynthetic eukaryotes, and to provide insights into niche adaptation at the genomic level. However, the vast majority of sequenced genomes of marine photosynthetic microorganisms comprises cyanobacteria (prokaryotes) and the genomes of only a handful of marine algae such as the diatoms Thalassiosira pseudonana and Phaeodactylum tricornutum, the pelagophyte Aureococcus anophagefferens, the haptophyte Emiliania huxleyi, and the green algae Ostreococcus tauri and Ostreococcus lucimarinus are available in the public databases. Genomes of nonmarine algae that have been sequenced include the thermophilic red alga Cyanidioschyzon merolae and the green algal genetic model Chlamydomonas reinhardtii. Numerous algal genome projects are however currently underway, such as for the red algae Galdieria sulphuraria, Porphyra purpurea, Chondrus crispus, the glaucophyte Cyanophora paradoxa, the green algae Volvox carteri, Dunaliella salina, Micromonas pusilla (two ecotypes), and Bathycoccus sp., the heterokont alga Ectocarpus siliculosus, the cryptophyte Guillardia theta, the chlorarachniophyte amoeba Bigelowiella natans, a number of marine diatoms (Pseudonitzchia multiseries and Fragilariopsis cylindrus), and the haptophytes Emiliania huxleyi, and Phaeocystis (two species). All of these taxa were chosen because of their fundamental roles in the marine phytoplankton, their toxicity, or central position in understanding algal evolution.

Photosynthesis in Eukaryotes: The Origin of Plastids It has now been clearly documented that the eukaryotic photosynthetic organelle (plastid) originated

552

through endosymbiosis, whereby a single-celled protist (host) engulfed and retained a photosynthetic cyanobacterium (endosymbiont). This primary endosymbiosis is believed to have given rise to the plastid in the common ancestor of the Plantae, which comprises red, green (including plants), and glaucophyte algae (Figure 1). Molecular divergence time estimations indicate that the primary endosymbiosis is an ancient event in eukaryote evolution, likely having occurred in the late Paleoproterozoic (c. 1.5  109 years ago). The establishment of the primary plastid involved critical evolutionary steps, such as the emergence of metabolic connections between the host cytoplasm and the endosymbiont. Molecular phylogenetic analyses suggest that some host proteins involved in the transport of nutrients across biological membranes (translocators) were relocated to the endosymbiont inner membrane and allowed the movement of photosynthesis products (sugars, nutrients) to the host. This was an essential step toward establishment of the symbiotic relationship (Figure 2). Another key step for plastid establishment was the

Red algae

Green algae

Glaucophytes

Plantae ancestor 'First alga'

Primary endosymbiosis

Heterotrophic protist

Cyanobacterium

Figure 1 Origin of the plastid in the common ancestor of the Plantae through primary endosymbiosis.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE ALGAL GENOMICS AND EVOLUTION

Translocators Endosymbiont metabolites Protein precursors

Cytosolic protein synthesis Host Genome

Mature proteins

Protein import Endosymbiont genome

Endosymbiotic gene transfer

Figure 2 Key processes that occur after primary plastid endosymbiosis include endosymbiotic gene transfer (EGT), the evolution of a plastid protein import machinery (TIC-TOC translocons), and a system (translocators) for the regulated movement of metabolites across the plastid membranes.

transfer and integration of genes from the cyanobacterium to the host nucleus (endosymbiotic gene transfer, EGT). As a result of EGT and the loss of superfluous genes, the endosymbiont genome was reduced dramatically to the point that modern plastid genomes generally contain less than 200 genes from c. 4000 that were likely present in its free-living cyanobacterial ancestor. What drove EGT? Evidently it was favored by natural selection to increase host fitness in response to Muller’s ratchet. This population genetics theory posits that genes encoded in nonrecombining genomes of small population size (such as in organelles) suffer the ratchet-like accumulation of degenerative mutations. The relocation of plastid genes to the nucleus where the recombination process, inherent to sexual reproduction, is present would act to counteract the ratchet. Given this process that drives EGT, why are genes still encoded in the organelle genome? There are two major theories in this respect. The first is that the remnant genes encode highly hydrophobic proteins (e.g., membrane-integral proteins) that, if nuclear-encoded and translated in the cytosol, would be difficult to transport (via the plastid protein import system) into the organelle. The second possibility is that the remaining genes encode important regulators of plastid activity in response to the cellular redox state and therefore must be encoded in the compartment where their gene products carry out their function. This need for co-localization of genes and their

553

function (the co-localization for redox regulation (CORR) hypothesis) may place limits on EGT for core redox regulatory proteins such as the reaction center subunits of photosystems I and II (e.g., plastidencoded psaA, psbA). Another fundamental issue that needs to be clarified to understand early algal genome evolution is quantifying cyanobacterial EGT to the nucleus of the first photosynthetic eukaryotes. The first systematic analysis of EGT was done using complete genome data from the flowering plant Arabidopsis thaliana. Molecular evolutionary analyses of the c. 25 000 predicted Arabidopsis nuclear proteins suggested that EGT left a sizeable mark on the Plantae that far exceeds the lateral transfer of photosynthetic capacity. Apparently, c. 18% (4500/25 000) of Arabidopsis nuclear genes originated from the ancestral cyanobacterium through EGT and, remarkably, many of these genes have evolved nonphotosynthetic or nonplastid functions. A recent comparative genomic analysis of partial genome data from the early diverging glaucophyte algae Cyanophora paradoxa, considered a ‘living fossil’ among Plantae, provided results regarding EGT that are markedly different than in Arabidopsis. In Cyanophora, only about 11% (c. 1500 genes) of the 12 000–15 000 estimated genes have a cyanobacterial (endosymbiotic) origin. Most of the identified cyanobacterial derived genes in Cyanophora have a plastid-related function (90%). These results indicate that the early cyanobacterial contribution to the nuclear genome was shaped by selective forces to retain the photosynthetic endosymbiont. A concomitant key event in endosymbiois was the evolution of the plastid import machinery to transport into the organelle the protein products of those endosymbiont genes relocated to the nucleus by EGT (Figure 2). It is likely that the first protein import mechanism evolved from the host endomembrane system (derived from the secretory pathway), and later in evolution the sophisticated and well-known translocon complexes of the modern plastids appeared (TIC-TOC). It seems that once these (and other) intricate events had occurred and the first populations of photosynthetic eukaryotes were established, the path was set for the rise and diversification of the algae.

Secondary Endosymbiosis and the Chromalveolate Algae The evolution of oxygenic photosynthesis in eukaryotes resulted in an enormous collection of descendant lineages that are critical for primary production in terrestrial and oceanic ecosystems.

(c) 2011 Elsevier Inc. All Rights Reserved.

554

MARINE ALGAL GENOMICS AND EVOLUTION

(a)

Haptophytes

(b) Stramenopiles

Alveolates

Cryptophytes

Photosynthetic chromalveolate ancestor

Host nucleus EGT

Secondary endosymbiosis

Secondary endosymbiont

Plastid-lacking chromalveolate ancestor

EGT Nucleus

Red alga Plastid

Green algae

Glaucophytes Plantae ancestor

Figure 3 Evolution of chromalveolate plastids. (a) Origin of the plastid in the putative common ancestor of chromalveolates through secondary endosymbiosis. (b) The different types of gene transfers that occurred after the origin of the secondary plastid in chromalveolates.

Primary plastid origin set the wheels in motion for another major event in eukaryotic evolution – red algal secondary endosymbiosis. This is a different type of plastid acquisition in which a nonphotosynthetic protist engulfed a eukaryotic red alga and retained the plastid (Figure 3(a)). This event is believed to have occurred c. 1.2  109 years ago in the photosynthetic ancestor of the Chromalveolata (see below). The chromalveolates are a putative monophyletic group of eukaryotes that originally comprised the following chlorophyll-c-containing algae and their nonphotosynthetic or plastid-lacking descendants: cryptophytes, haptophytes, stramenopiles, dinoflagellates, ciliates, and apicomplexans. Recent studies have however identified other protist lineages (e.g., katablepharids, telonemids, and Rhizaria) that also appear to be members of the Chromalveolata. Secondary endosymbiosis also involved EGT, but now with a new level of complexity: genes were transferred from the nucleus of the red alga (secondary endosymbiont) to the host genome (Figure 3(b)). The massive loss of unnecessary or redundant genes

and EGT to the host nucleus resulted in the diminution or complete disappearance of the red-algal endosymbiont nucleus. Only the cryptophyte algae retain a remnant of this diminished nucleus that is located between the plastid membranes and is referred to as the nucleomorph. The gain of photosynthesis by chromalveolates was a key point in the Earth’s history because many of these algae now comprise the ecologically dominant photosynthetic eukaryotes in the oceans (i.e., diatoms, haptophytes, and dinoflagellates).

The Diatoms Diatoms (Bacilliarophyta) with c. 20 000 described and as many as 100 000 estimated species are photosynthetic unicellular organisms characterized by the silica cell wall called a frustule. They are the most diverse group within stramenopiles and are phylogenetically related to brown algae, kelps, and nonphotosynthetic oomycetes. The diatoms play a fundamental role in energy webs in the worlds’

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE ALGAL GENOMICS AND EVOLUTION

oceans, with some studies indicating that they are responsible for c. 35–40% of primary production in marine ecosystems and c. 25% of total biosphere organic carbon fixation. Elucidation of diatom biology and ecology has provided a springboard to study the genomes in these critical taxa. The genome of the centric marine diatom T. pseudonana (411 000 genes) was the first that was fully sequenced from any marine alga. The data provided important insights into diatom biology, in particular photosynthesis and silica deposition. Given the major contribution of diatoms to primary production, a fundamental question is to elucidate diatom carbon fixation mechanisms because these cells live in habitats (oceans) that are characterized by low CO2 concentrations. It has been suggested that diatoms possess physical and/or biochemical CO2 concentration mechanisms that improve carbon fixation efficiency by limiting photorespiration (analogous to C4 photosynthesis in some land plants). C4 photosynthesis is a mechanism exploited by plants (often characterized by a Kranz leaf anatomy) to increase the relative concentration of CO2 to O2 to boost carbon fixation by ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO). However, the presence of a C4-like metabolism in diatoms is a subject of debate. The diatom Thalassiosira weissflogii apparently possesses a special form of C4-like photosynthesis that has a central role under low CO2 concentrations. The genomic data are however ambiguous in this respect because T. pseudonana contains the complete set of genes required for C4-like photosynthesis, but the localization of essential enzymes for concentrating CO2 in C4-type photosynthesis, the carbonic anhydrases (a and g types), in the cytoplasm is not consistent with typical C4 photosynthesis. Recent experimental studies suggest that photosynthetic metabolism in diatoms is diverse and some diatoms perform exclusively C3-like metabolism (more efficient in nonlimiting CO2 conditions) and others are capable of ‘intermediate’ C3–C4 metabolism. In this context, comparative genomics has provided some important clues that can be used to direct experiments aimed at clarifying the details of photosynthesis in this group of marine algae. Marine geosciences are also critical to understanding the history of C3–C4 metabolism. A typical indicator of photosynthetic activity is the carbon isotopic discrimination (fractionation) of the heavier but rare carbon isotopes 13C and 14C in contrast to the lightest and most abundant 12C. On average, the 13 12 C/ C ratio decreases in biologically fixed carbon compounds. Moreover, in C3 photosynthesis, the 13C depletion is significantly increased with respect to C4

555

metabolism. Therefore, the study of the 13C ratio in diatom fossil (biomarker) compounds such as highly branched isoprenoid alkenes that are present in marine sediments may help us understand the biochemical nature and history of diatom photosynthesis. Another important consideration in diatom biology is the physiological necessity for silica and the role of this element in biogeochemical recycling of minerals in the oceans. Diatoms can take up dissolved silicic acid and microprecipitate silica (SiO2) (biosilica) to generate the intricate frustules that surround the cell. Diatoms play a major role in the interconnected Si and C biogeochemical recycling in the oceans and silica metabolism appears to be rare among eukaryotes. Genomic approaches have advanced our understanding of silica utilization. Study of the Thalassiosira genome identified at least three silicic acid (soluble silicon source) membrane transporters and other proteins of the pathway such as spermine and spermidine synthases (involved in polyamine synthesis), the phosphoproteins silaffins (essentials for silica precipitation and formation of silica deposition vesicles), and frustulins (glycoproteins responsible for frustule casing). Another remarkable result from analysis of the Thalassiosira genome is the identification of the complete enzymatic repertoire for nitrogen metabolism involving the urea–ornithine cycle, suggesting that pyrimidine metabolism occurs in the cytoplasm as in heterotrophic eukaryotes. Recent advances and upcoming data in diatom genomics should significantly increase our understanding of the biological processes that govern primary production in the oceans.

The Haptophytes The Haptophyta comprises c. 500 mostly marine algal species, many of which are characterized by calcite (calcium carbonate) scales that cover the cell (coccoliths). In particular, E. huxleyi, which is a highly abundant haptophyte in the ocean, has been considered a critical component of marine environments because of its dual capacity to fix environmental carbon via biomineralization (calcium carbonate, calcite) and through photosynthesis. The haptophytes having coccoliths are responsible for c. 50% of the calcium carbonate precipitation in oceans. The impact from Emiliania blooms is great enough to have a possible influence on global climate. Occasionally these blooms reach a size of c. 100 000 km2 with cell accumulations that can affect local climate. This is due to the optical properties of coccoliths that can reflect a significant amount of sunlight and heat from the water. The

(c) 2011 Elsevier Inc. All Rights Reserved.

556

MARINE ALGAL GENOMICS AND EVOLUTION

study of coccolith mineralization (coccolithogenesis) has also been analyzed using gene expression patterns in E. huxleyi cultures raised under calcifying and noncalcifying conditions. This work has identified proteins that are putatively related to coccolithogenesis. Much work still has to be done with candidate proteins to verify their roles in coccolith formation because many lack identifiable homologs in the public sequence databases. Like other chromalveolates (e.g., diatoms), haptophytes contain a secondary plastid derived from a red algal endosymbiont with chlorophyll c as the principal photopigment. Despite their biological importance, genome-scale data have only now started to emerge with the E. huxleyi genome that is currently being sequenced to completion. This information will be critical for bioinformatically annotating Emiliania genes with their putative functions and testing these results using functional genomics approaches in this nascent algal model.

Dinoflagellates Dinoflagellates, along with ciliates and apicomplexans (intracellular parasites; e.g., Plasmodium) form the protist group Alveolata that is subsumed into the Chromalveolata. Ciliates have apparently lost outright the ancestral plastid they shared with other chromalveolates, whereas the apicomplexans lost the photosynthetic capacity but retain a vestigial plastid (apicoplast) as a compartment for carrying out other plastid functions such as fatty acid biosynthesis. The dinoflagellates comprise c. 2000 species and the vast majority are free-living cells, with autotrophic and heterotrophic lifestyles, although many parasitic taxa also exist. The photosynthetic dinoflagellates play a fundamental role as primary producers in coastal waters. One well-known example is the mutualistic association between some dinoflagellates of the genus Symbiodinium and the reef-forming corals. Under optimal environmental conditions, symbiotic dinoflagellates provide 490% of their total photosynthetic production to the coral host. It is believed that this mutualistic association influenced the diversification of cnidarian species since the late Triassic. The collapse of the symbiosis produces the so-called coral bleaching that is due to the loss of the dinoflagellate photopigments in the coral cells. Bleaching is a major cause of coral mortality, reduced fecundity, and increased disease vulnerability, with serious effects on reef ecosystems. The success of the symbiotic relationship is influenced by several environmental factors, and it has been suggested that

elevated temperature, irradiance, and pathogen infections disrupt the symbiosis. Understanding the physiological process that governs the dinoflagellate– coral symbiosis is of high priority in marine ecology and conservation biology. An obvious goal in the genomics era is to identify the key molecular processes that lead to the establishment, maintenance, and loss of the dinoflagellate–coral symbiosis, with the ultimate goal being to identify the genes that underlie this interaction. Preliminary results of gene expression profiles with the model anemone Anthopleura elegantissima suggest that the mutualism is not governed by a few symbiotic genes in the coral cell, but by particular expression patterns of common major cellular processes. Proteins mediating cell–cell interactions (e.g., sym32) have been proposed to play a role in the symbiosis by facilitating the engulfment of the dinoflagellate by cnidarian gastrodermal cells. What about the contribution of the symbiont? There are numerous questions to be addressed in this regard but these require extensive genomic data from dinoflagellates, which as we discuss below is not a trivial DNA sequencing task. Dinoflagellate marine species like Alexandrium and Karenia are infamous for producing potent neurotoxins in so-called ‘red tides’. Red tides are one form of harmful algal bloom (HAB) and pose serious threats to humans, marine mammals, seabirds, fish, and many other organisms that ingest the toxin. The presence of toxins is not however always related to photopigment concentration and high cell densities that can cause the red coloring of seawater. One of the most common dinoflagellate toxic compounds is the saxitoxin (1000 times more potent than cyanide) produced typically by members of the genus Alexandrium. The genetic and biochemical nature of the saxitoxin biosynthetic pathway is controversial because there are closely related toxic and nontoxic strains of Alexandrium. It has been proposed that the ability to produce saxitoxins stems from symbiotic bacteria and is not intrinsic to the dinoflagellate genic repertoire. However, saxitoxin production apparently remains when symbiotic bacteria are removed from culture. This issue that has remarkable ecological and economic consequences remains a fundamental unresolved aspect of dinoflagellate biology. Ongoing functional genomic analysis of Alexandrium, Karenia, and other HAB-forming dinoflagellates will ultimately clarify the environmental and genetic factors that underlie bloom formation and toxin production. However, any attempt to study dinoflagellates at the genomic level has to deal with the large amount of DNA present in their nucleus. Dinoflagellates contain 3–250 pg DNA/cell corresponding to approximately

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE ALGAL GENOMICS AND EVOLUTION

3000–215 000 megabases (millions of base pairs, Mb), which is 1 or 2 orders of magnitude larger than the typical model organism; for example, the haploid human genome is of 3180 Mb. It has been suggested that polyploidy or polyteny may account for the massive cellular DNA content in dinoflagellates, but molecular studies do not support this hypothesis. It is therefore unclear why these algae possess such large genomes, in particular because their genetic complexity does not appear to be commensurate with their DNA content. Flow cytometry investigations suggest some members of the genus Symbiodinium might be targets to genome sequencing in the future. Other promising windows for algal researchers are the picoeukaryote-size dinoflagellates (o3 mm), which probably have genomes more adequate for exhaustive sequencing. Finally, another reasonable approach to take with dinoflagellate genomics is to sequence single chromosomes. It is evident therefore that there are serious practical limitations to genomic studies in dinoflagellates. However, numerous smaller-scale (e.g., expressed sequence tags, ESTs) genomic approaches have been applied to these taxa and have provided important insights into dinoflagellate evolution. Dinoflagellate Plastid Evolution

Here we discuss some recent insights into dinoflagellate plastid evolution that have resulted from EST investigations. If we assume that there was a single ancestral red algal plastid that unites the chromalveolates, then this organelle has been independently lost in some lineages (e.g., ciliates, oomycetes, telonemids). Plastid loss is not rare in nature, but dinoflagellates show multiple examples of this phenomenon as well as the unique ability to recruit new algal plastids through tertiary endosymbiosis (see below). The most common type of plastid in dinoflagellates contains peridinin as the major carotenoid photopigment. Remarkably, the plastid genome of peridinin-containing dinoflagellates is highly reduced and broken up into minicircles typically containing a single gene. The minicircles encode core subunits of the photosynthetic machinery (atpA, atpB, petB, petD, psaA, psaB, psbA–E (consistent with the CORR hypothesis described above)) and other proteins of unknown function (ycf16, ycf24, rpl28, and rpl23) as well as rRNA and tRNA. This unique arrangement is in stark contrast to typical algal and plant genomes that encode 100–200 genes. Analysis of an EST data set from the peridinin-containing dinoflagellate Alexandrium tamarense identified 15 genes in the nucleus of this species that are always plastid-encoded in other photosynthetic eukaryotes.

557

Clearly these nuclear genes are the result of EGT from the plastid to the nuclear genome. The forces behind this unprecedented migration of universally conserved components of the plastid genome to the nucleus are not yet fully understood. Unlike other eukaryotes that have reduced plastid genomes because of a loss of photosynthesis due to a parasitic lifestyle, the peridinin dinoflagellates have drastically reduced their plastid genome but are generally freeliving and retain photosynthetic capacity. Equally surprising is the observation that a handful of Alexandrium genes encoding plastid-targeted proteins (hemB, tufA) trace their origin to green algae, and not to red algae as expected under the chromalveolate hypothesis. These data may indicate a ‘hidden’ green algal endosymbiosis in the dinoflagellates (and potentially other chromalveolates) or, alternatively, multiple independent cases of horizontal gene transfer (HGT) from green algae. The phagotrophic capacity of dinoflagellates would provide an explanation for the latter hypothesis. Under this scenario the dinoflagellates recruit genes via HGT from algal (or other) food sources that are brought into their cells. Ongoing comparative genomics and phylogenomic studies demonstrate that endosymbiosis and HGT have played a critical role in shaping the genomes of dinoflagellates. Although fascinating in terms of genome evolution, these processes complicate interpretation of gene data and necessitate great care when assigning dinoflagellate DNA sequences to their putative sources of origin. Plastid Replacement through Tertiary Endosymbiosis

The dinoflagellates are also unique in their ability to take up plastids through repeated endosymbioses. There are several dinoflagellates that have replaced the ancestral chromalveolate plastid of red algal origin with one from other chromalveolates (i.e., cryptophytes, diatoms, or haptophytes) or green algae through a process termed tertiary endosymbiosis (Figure 4). A prominent example of tertiary endosymbiosis is the red tide dinoflagellate Karenia brevis that has acquired a plastid from a haptophyte endosymbiont. A direct consequence of this plastid acquisition is that Karenia now has fucoxanthin (uniquely derived from the haptophyte) as a plastid photopigment, instead of the ancestral peridinin. Comparative genomic analysis of the fucoxanthincontaining dinoflagellates Karenia and Karlodinium micrum provide key insights into tertiary endosymbiosis. One important observation is that most (and perhaps all) of the nuclear genes that encode plastid-targeted proteins that were inherited from the

(c) 2011 Elsevier Inc. All Rights Reserved.

558

MARINE ALGAL GENOMICS AND EVOLUTION

Stramenopiles

Tertiary endosymbioses in dinoflagellates Peridinium

Haptophytes

E.g., Karenia (fucoxanthin) Cryptophytes Dinophysis

Chromalveolate ancestor

Dinoflagellates (peridinin)

Secondary endosymbiosis

Lepidodinium

Red algae Green algae

Glaucophytes Primary endosymbiosis Figure 4 Tertiary endosymbiosis in dinoflagellates occurred independently in different lineages and involved evolutionarily distantly related algal endosymbionts.

red algal secondary endosymbiont via EGT have been replaced with the haptophyte homologs. This underlines the dynamic evolutionary history of dinoflagellate genomes that may be an adaptation to generate genic diversity through their mixotrophic lifestyle. Picoeukaryotes: ‘Hidden’ Biodiversity in the World’s Oceans

In recent years it has become apparent that a vast unexplored diversity of algal forms exist in oceans as prokaryote-sized cells (0.2–3 mm of cell mean diameter) as part of the eukaryotic picoplankton (picoeukaryotes). Marine picoeukaryotes are found in all of the major algal groups (e.g., green algae, haptophytes, stramenopiles, and dinoflagellates) and have often been uncovered using environmental (meta)genomics or environmental polymerase chain reaction (PCR)

approaches in which DNA sequences are determined from environmental samples without culturing particular isolates. These data demonstrate the existence of an extraordinary biodiversity of oceanic picoeukaryotes although their potential contribution to major biogeochemical processes is largely unknown. Studies of the global distribution of the ubiquitous marine green algal picoeukaryotes Ostreococcus spp. and Micromonas pusilla (both are early-diverging prasinophyte green algae) suggest the existence of ecotypes (populations adapted to a particular ecological niche). These isolates are excellent models to study genome-wide adaptation to particular habitats with the added benefit that they have highly reduced genomes (i.e., resulting in reduced sequencing costs). Taking advantage of these aspects, the nuclear genome has been recently sequenced from three different Ostreococcus isolates. Ostreococcus tauri possesses a genome of c. 12.6 Mb, which is one of the smallest

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE ALGAL GENOMICS AND EVOLUTION

known for a free-living eukaryote, with 8166 predicted genes. The genetic repertoire of Ostreococcus includes the typical enzymes involved in C4-like photosynthesis, and as in the case of the diatom T. pseudonana, leaves open several questions about the details of this metabolism in a unicellular marine alga and its role in primary production under limiting light and CO2 conditions.

Summary Algae are an ancient polyphyletic assemblage of eukaryotes that have played a central role in shaping the Earth’s history. Given their diversity of forms, the study of genomes plays an increasingly important role in helping us understand the origin of algal lineages and their adaptation to different oceanic environments. Plastid endosymbiosis that was explored in most detail here is only one example of the application of genome data to resolve difficult evolutionary problems. This area of research will also be facilitated by the booming field of metagenomics that has as one aim to assemble complete genomes and large genome fragments from environmental DNA samples. This culture-independent approach has until now been most effectively applied to prokaryotic DNA. Once it becomes feasible to generate useful metagenome data from the much larger and complex algal genomes, it will be theoretically possible to reconstruct the network of interactions that define particular marine environments, inclusive of prokaryotes, eukaryotes, and viruses. The current acceleration in DNA-sequencing technologies that will soon provide individual researchers with hundreds of billions of base pairs of sequence information promises to make this dream a reality. Given therefore the breathtaking pace of advances in the field of genomics it is hard to predict where we will be 10 years from now. It is clear however that marine biologists and oceanographers will have unparalleled opportunities to study the ecology, diversification, and evolution of photosynthetic eukaryotes.

See also Marine Mammal Evolution and Taxonomy. Phytoplankton Blooms. Phytoplankton Size Structure.

559

Further Reading Armbrust EV, Berges JA, Bowler C, et al. (2004) The genome of the diatom Thalassiosira pseudonana: Ecology, evolution, and metabolism. Science 306: 79--86. Bachvaroff TR, Sanchez Puerta MV, and Delwiche CF (2005) Chlorophyll c-containing plastid relationships based on analyses of a multigene data set with all four chromalveolate lineages. Molecular Biology and Evolution 22: 1772--1782. Bhattacharya D, Yoon HS, and Hackett JD (2004) Photosynthetic eukaryotes unite: Endosymbiosis connects the dots. BioEssays 26: 50--60. Derelle E, Ferraz C, Rombauts S, et al. (2006) Genome analysis of the smallest free living eukaryote Ostreococcus tauri unveils many unique features. Proceedings of the National Academy of Sciences of the United States of America 103: 11647--11652. Falkowski PG, Katz ME, Knoll AH, et al. (2004) The evolution of modern eukaryotic phytoplankton. Science 305: 354--360. Harper JT and Keeling PJ (2003) Nucleus-encoded, plastidtargeted glyceraldehyde-3-phosphate dehydrogenase (GAPDH) indicates a single origin for chromalveolate plastids. Molecular Biology and Evolution 20: 1730--1735. Nozaki H (2005) A new scenario of plastid evolution: Plastid primary endosymbiosis before the divergence of the ‘Plantae’, emended. Journal of Plant Research 118: 247--255. Patron NJ, Rogers MB, and Keeling PJ (2004) Gene replacement of fructose-1,6-bisphosphate aldolase supports the hypothesis of a single photosynthetic ancestor of chromalveolates. Eukaryotic Cell 3: 1169--1175. Reyes-Prieto A, Weber AP, and Bhattacharya D (2007) The origin and establishment of the plastid in algae and plants. Annual Review of Genetics 41: 147--168. Rodriguez-Ezpeleta N, Brinkmann H, Burey SC, et al. (2005) Monophyly of primary photosynthetic eukaryotes: Green plants, red algae, and glaucophytes. Current Biology 15: 1325--1330. Takishita K, Ishida K, and Maruyama T (2004) Phylogeny of nuclear-encoded plastid-targeted GAPDH gene supports separate origins for the peridinin- and the fucoxanthin derivative-containing plastids of dinoflagellates. Protist 155: 447--458. Yoon HS, Hackett JD, Pinto G, and Bhattacharya D (2002) The single, ancient origin of chromist plastids. Proceedings of the National Academy of Sciences of the United States of America 99: 15507--15512.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE BIOTECHNOLOGY H. O. Halvorson, University of Massachusetts Boston, Boston, MA, USA F. Quezada, Biotechnology Center of Excellence Corporation, Waltham, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Definitions

Marine biotechnology is the application of biotechnological approaches to marine organisms in the harnessing of commercial products and services. Marine biotechnology is a multidisciplinary activity marrying three traditional marine disciplines – ocean science, marine biology, and marine engineering – with the more modern fields of molecular and cellular biology, genomics, and proteomics. Marine Biotechnology as an Academic and Research Discipline

The world’s oceans, which comprise two-thirds of the surface of planet, possess the largest habitats on Earth and contain the most ancient forms of life. Over time, marine microbes have changed the global climate and structured the atmosphere. The large and unexplored diversity of living species today is found mostly in the oceans. With the tools of molecular biology, these specific adaptations can be understood in detail and applied for the benefit of biomedical and industrial innovation, environmental remediation, food production, and fundamental scientific progress. Marine organisms have traditionally been good models for the study of communication, defense, adhesion, host interactions, disease, epidemics, nutrition, and adaptation to large, often extreme, variations in their environment. There are numerous examples of marine models contributing to understanding basic concepts within biology and medicine. The possibility to cultivate these organisms, and to study them in the laboratory at a cellular, molecular, or genetic level, has opened new perspectives. Professional Associations

The formation and establishment of professional associations in marine biotechnology reflects the growing worldwide interest in this area. The Japanese Society for Marine Biotechnology was organized in

560

1989. Since then, other counterpart associations were formed: the American Society for Molecular Marine Biology and Biotechnology, the Asian-Pacific Marine Biotechnology Society, and the Pan-American Marine Biotechnology Association. Networks of marine laboratories also came together. The National Association of Marine Laboratories, in the United States, and the European Network of Marine Laboratories, in Europe, were formed, with strong interests in marine biotechnology. In 2000, Canada initiated an interdisciplinary program (Aquanet) to support aquaculture in that country. In 1989, the Japanese Society for Marine Biotechnology organized the first International Marine Biotechnology Conferences (IMBC). The first European meetings on marine biotechnology were held at Montpellier (1992), Willemshaven (Germany) (1998), Edinburgh (1998), and Noordwijkerhout (the Netherlands) (1999). The IMBC meetings focused on new breakthroughs in basic and applied science, industrial applications, environmental science, industrial applications, commerce, and international issues of marine biotechnology. Two new marine journals were established: Journal of Marine Biotechnology (Japan) and Molecular Marine Biology and Biotechnology (USA). These were merged to form Marine Biotechnology, an international journal on the molecular and cellular biology of marine life and its technology applications.

Marine Biotechnology and Marine Conservation Tools for the Study of Marine Ecology

Oceans are highways of commerce for international trade, providing food, contributing to our energy supply (oil and gas), and serving as recreational areas. The health of our ocean ecosystems is threatened by overfishing and unintended bycatch, pollution, habitat loss, boating traffic, and climate change. This threat comes from a number of human activities: conversion of coastal habitats for other uses, nonpoint pollution, airborne pollution, waterborne diseases, marine debris, ballast water, and invasive species, among others. Some of the tools of marine biotechnology have helped with ongoing monitoring to assess the health of ocean and coastal ecosystems to guide management in decisions on the use of oceans and coastal waters and establishment of standard practices and procedures.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE BIOTECHNOLOGY

Multiple means are used to collect the desired information: aircraft, ships, moored instruments, drifters, gliders, submersibles, remotely operated vehicle (ROV), and satellites. The satellite communications infrastructure provides global broadband coverage to support ocean observations. Monitoring stations and buoys collect and transmit continuous data streams on climate, weather, air quality, temperature, surface pressure, ocean dynamics, and other selected variables. The Integrated Ocean Observing System (IOOS) and Global Ocean Observing System (GOOS) integrate the collected data. Additional sensors are being developed to monitor marine ecology. Optical and sonic probes are used to count populations of marine organisms. Quantification of plant and animal biodiversity can be a useful approach to follow the health of the habitat. Molecular techniques used to measure marine microbial diversity should be modified as marine sensors to detect marine pathogens and toxins. Sensors to measure marine pollutants should be expanded. Major challenges are collection, storage, and assimilation of the large volume of data as well as the ability to provide these data rapidly to the user community. Recent developments in molecular biology, ecology, and environmental engineering now offer opportunities to modify organisms so that their basic biological processes are more efficient and can degrade more complex chemicals and higher volumes of waste materials. Understanding Life Cycles of Marine Organisms

Life cycles in marine organisms are composed of stages that can vary dramatically in size, shape, mobility, and food preferences. Reproductive strategies abound in the ocean and include fission, budding, free-spawned gametes (eggs and sperm), external embryo hatching, internal embryo hatching, and live birth. A significant number of marine organisms are hermaphrodites, containing the reproductive organs of both sexes. Many species of fish develop from a fertilized egg through a juvenile phase and into an adult phase. Deep-sea fish have larval stages distinctly separate from the juvenile and adult stage. The life cycle of some commercial species, such as salmon and trout, have been very well documented. Marine organisms face challenges to fertilization of free-spawned gametes, nutrient acquisition during larval development, benthic site selection, juvenile survival, and adult reproductive location and timing. Understanding ecological and evolutionary processes at the single organism and population levels is essential.

561

Major Product Areas of Interest Medicines from the Sea

Marine molecules often belong to new classes without terrestrial counterparts, for example, halogenated compounds. In addition, marine microorganisms are a source of new genes. Secondary metabolites produced by marine bacteria and invertebrates have yielded pharmaceutical products such as novel anti-inflammatory agents, anticancer agents, and antibiotics. For example, isolation of the compound manoalide from a Pacific sponge has led to the development of more than 300 chemical analogs, with many of these going on to clinical trials as anti-inflammatory agents. Melanins have a range of chromophoric properties that can be exploited for sunscreens, dyes, and coloring. Marine sponges also sequester different kinds of organic compounds, including fungicides and antibiotics, which may allow them to act as slow-release agents. Other examples include the circulating cells (amoebocytes) of the horseshoe crab that contain molecules that react with the outer coats of Gram-negative bacteria, and thus have found use in the detection of early infection in humans and pyrogens in biotechnological products. Lectins, found in all invertebrates, are being used to target cancer cells, screen bacteria, and immobilize enzymes. Seaweed produces compounds like laminarin and fucoidans that are known to protect against radiation damage, lower cholesterol levels, and help in wound repair. They also have an anti-inflammatory action, as well as strong immunomodulatory effects. Seaweed-derived products increase resistance to bacterial, viral, and parasitic infections (including infection after surgery) and help with prevention of opportunistic infections in immunocompromised individuals (such as HIV sufferers and geriatric patients). Secondary metabolites, such as halogenated compounds, extracted from macroalgae (seaweeds) are of great promise as antibacterial and antiviral agents or antifouling agents. Extracts from certain red algae are used in bone replacement therapy. Biomaterials and Nutraceuticals

Historically, macroalgae (seaweeds) have been used as a subsidiary food and have been used extensively in medicine. Macroalgae make use of different photosynthesizing pigments that divide them roughly into three groups – the brown, green, and red algae – which is in contrast to the terrestrial environment with only the green plants. The biochemical diversity present in seaweeds provides the potential for a very large array of products. For example, polysaccharides

(c) 2011 Elsevier Inc. All Rights Reserved.

562

MARINE BIOTECHNOLOGY

from certain red algae or brown algae are widely used as thickening agents in the food industry, cosmetics, and even building materials. Seaweeds are used in aquaculture diets for urchin, abalone, and fish as an alternative food source to fishmeal and fish oil. These seaweeds contain all of the essential amino acids, polyunsaturated fatty acids, vitamins, and minerals, as well as a small amount of protein. Eelgrass produces an effective antifouling agent against bacteria, algal spores, and a variety of hard-fouling barnacles and tubeworms. Marine diatoms, mollusks, and other marine invertebrates generate elaborate mineralized structures on a nanometer scale that can have unusual and useful properties. Deep-sea hydrothermal vents now offer a new source of a variety of fascinating microorganisms well adapted to these extreme environments. This newly appreciated bacterial diversity includes strains that are able to produce a number of unusual microbial exopolysaccharides with interesting chemical and physical properties. Chitin and chitosan are associated with proteins in the exoskeleton of many invertebrate species, such as annelids, shellfish, and insects, and also in the envelope of many fungi, molds, and yeasts. Chitin polymers are natural, nontoxic, and biodegradable and have many applications in food and pharmaceuticals as well as cosmetics. ‘Antifreeze glycoproteins’ which inhibit ice-crystal formation are found in the tissues of fish in the Arctic and Antarctic (e.g., Arctic char) as well as microorganisms which live at low temperatures. These species may prove useful in industrial and medical cryopreservation processes. Antioxidant peptides have been isolated from extracts of prawn muscle and seaweeds. These have applications as food additives and in cosmetics.

Green fluorescent protein, a naturally fluorescent protein first found in jellyfish, is now widely employed as a sensitive fluorescent molecular ‘tag’ to identify and localize individual proteins within a cell or a subset of cells within a tissue and to follow gene expression in various systems. Aquatic extremophiles produce thermostable DNAmodifying enzymes used in research and industrial applications. Some enzymes produced by marine bacteria are salt resistant: extracellular proteases are used in detergents and industrial cleaning applications. Vibrio species have been found to produce a variety of extracellular proteases. Vibrio alginolyticus produces collagenase, an enzyme used in the dispersion of cells in tissue culture studies.

Novel Marine-derived Bioprocesses Bioremediation

Marine microorganisms, either as independent strains or as members of microbial consortia, express novel biodegradation pathways for breaking down a wide variety of organic pollutants. Marine microorganisms frequently produce environmentally friendly chemicals such as biopolymers and nontoxic biosurfactants that can also be applied in environmental waste management and treatment. Recent findings into the basis of cell–cell communication have shown that this process is involved in biofilm formation leading to environmental corrosion. This has generated a search for new bioactive molecules active in preventing such communication and controlling subsequent fouling. In addition, further understanding of the interaction of marine microbes with toxic heavy metals has suggested their application in various treatments of contaminated water systems.

Biotechnology Applications to Marine Aquaculture

Industrial Enzymes

Microorganisms provide the basis for development of sophisticated biosensors, and diagnostic devices for medicine, aquaculture, and environmental biomonitoring. Intact cell preparations and isolated enzyme systems for bioluminescence are used as biosensors. The genes encoding these enzymes have been cloned from marine bacteria and have been subsequently transferred successfully to a variety of organisms where they are expressed only under defined environmental conditions. Other bioluminescent proteins from marine organisms are currently under study in order to produce gene probes that can be employed to detect human pathogens in food, or fish pathogens in aquaculture systems.

Reproduction

Over the past two centuries, the practice of agriculture involved classical selective breeding of plants and animals to enhance a desirable characteristic by increasing the expression of a particular gene in the population over subsequent generations. In aquaculture, cells, molecules, and genes can now be modified directly to improve production efficiency through improvements in growth rates, food conversion, disease resistance, product quality, and composition. This same approach can help conserve wild species and genetic resources and provide unique models for biomedical research. Key hormones

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE BIOTECHNOLOGY

regulate the processes of molting, development, and reproduction in economically important crustaceans. Foreign genes controlling growth-enhancing hormones have now been introduced into scallops and a number of fish. Hormones have been used to develop monosex lines, agents for smolt enhancement, and enhanced feed-conversion efficiency. Microbial Diseases and Pathogen Control

Bacteria, viruses, and other pathogens are a natural part of any ecosystem and aquatic organisms have co-evolved with them. Fish diseases do not occur as a single caused event but are the end result of interactions of the disease, the fish, and the environment. In intensive culture, handling, crowding, transporting, drug treatments, undernourishment, fluctuating temperatures, and poor water quality continuously affect fish. These conditions impose considerable stress on fish, rendering them susceptible to a wide variety of pathogens. Bacteria cause the most severe disease problems in aquaculture. Specific bacterial pathogens are responsible for specific disease problems. Gram-negative bacteria are the most frequent causes of disease in finfish, whereas Gram-positive bacteria are most common in crustaceans. Viral diseases cause serious problems in aquaculture requiring quarantine and destruction of the infected stock. Viral disease in wild stock fish are formidable obstacles to raising fish in net-pen aquaculture in open waters. Fungal infections are frequently encountered when the water quality is poor, or in the presence of stress, inadequate nutrition, and skin trauma which provides a port of entry for molds. Saprolegniosis, one of the most common water molds encountered, occurs with handling, crowding, heavy feeding, and high organic loads. The use of chemotherapeutics in fish, as in land-based animals, is extremely controversial. Bacterial resistance to antibiotics has developed in animal systems, in which antibiotics are routinely used. Resistant bacteria may be carried on food products, and thus infect the consumer or may infect the handler directly. There are very limited numbers of chemical/antibiotic treatments approved for use in aquacultured fish. Drug development is expensive and pharmaceutical companies are not willing to invest the money necessary to directly approve drugs for use in aquacultured animals. The decline in the use of antibiotics as a defense for man and animals against pathogenic microorganisms provides the biomedical and biotechnology industry with new opportunities to design and rapidly produce vaccines.

563

Genetic Management and Genetic Engineering

Modern biotechnology approaches to aquaculture of marine species include the use of marker-assisted breeding techniques. These techniques allow hatcheries and breeders to select reproductive pathways on the basis of optimal levels of diversity in the cultivated population and to screen for certain traits. Genetic engineering, on the other hand, is seen as the more precise and effective method for genetic improvement of specific traits that are controlled by single genes. A transgenic organism is one whose genome contains DNA inserted from another organism. DNA is introduced into embryo stem cells, which are then merged with early-stage embryos. If the gene of interest is present in a germ line, they can be bred to homogeneity by future generations. Genes are activated to produce protein by adjacent genes that produce promoter sequences. The promoter sequences are responsible for switching on genes in specific areas of the body. This technology has been used successfully in numerous marine microalgae and animal systems to alter characteristics in a population by conferring temperature resistance to plants and marine animals, growth hormone, antifreeze protein, or disease resistance. An effective biological containment of some commercially important species must be developed. Cloning wild genotypes has benefits for stock enhancement, conservation, restoration, and hatchery production.

Improved Feed and Nutrition

Exclusion of traditional raw materials from aquaculture feeds and possible limitation in the use of those remaining impose a great risk to aquaculture sustainability. A certain reduction of production costs can be obtained through adequate larval nutrition, diversification of marine organisms used for the first feeding of larvae, and development of efficient artificial diets for early larval stages. These, together with feed-dispersing techniques, would add to the overall quality of larvae used. Other considerations include how the plant genome can now be manipulated to produce products economically for use in aquaculture. The use of genetically modified crops to eliminate toxic products and increase specific nutrients is now possible. Further understanding of the fish immune system and the relation of its condition to nutrition and stress would help in improvements in growth and disease resistance.

(c) 2011 Elsevier Inc. All Rights Reserved.

564

MARINE BIOTECHNOLOGY

Oceans and Human Health Public health authorities, tourism, and the seafood industry have been increasingly concerned over public health risks from the marine sources. The degradation of the oceans makes an impact on how we use the ocean and where we buy our seafood. The dramatic growth of the international seafood market increases the global potential for risk. The survival of human populations may depend on the preservation of healthy, diverse, and sustainable ocean and coastal systems. Molecular technology has provided an increasing range of diagnostic tools to clinicians and environmental regulators to measure risk assessment. The main effort of such risk assessment should be the protection of individuals and populations from harm. Including biomarkers in the risk assessment process provides a functional measure of exposure to chemicals, which stress an organism. In recent years, an increasing number of biomarkers have been used to measure molecular damage, developmental abnormality, and physiological impairment in invertebrate species. General biological stress biomarkers include cardiac activity in bivalve mollusks, dye retention in the hemocytes of bivalve mollusks, measurement of immunocompetence in invertebrates, apoptosis assay, and defecation assay with shrimp. Behavior biomarkers include swimming behavior in shrimp, burrowing behavior assay with amphipods, and a starfish-righting assay. Chemical biomarkers include acetylcholinesterase inhibition in crustaceans, florescence assay in urine samples of crabs, assay of endocrine disruptors in gastropod mollusks, and genotoxins. Developing nations dominate both the production and international trade in seafood. Since it is now recognized that most of the seafood-borne risk comes from the environment, a more global perspective for product protection and coastal waters is needed. A chain of custody/product traceability would assure critical information essential to a comprehensive view of seafood risk mitigation. Such a system would incorporate a set of protocols focusing on simplified determination of associated socioeconomic dynamics. Here again, advanced research in marine biotechnology can help to provide some of the analytical tools to address these seafood safety issues.

Policy Issues The role of government and public agencies in the development of marine biotechnology and its diverse

applications to the environment, commercial endeavors, scientific discovery, new products, and other aspects brings attention to several policy-related issues affecting this field. Among these are issues of access and benefit sharing, intellectual property rights, and ecological concerns. These and other considerations are highlighted below. Funding

One of these concerns the need for public support for research and human resource development in marine biotechnology. With continuing competition for research funds and dwindling allocations for higher education, those institutions involved in the field of marine biotechnology are faced with the need to highlight the relevance of this field to the high-priority problems of scientific, economic, and ecological nature. User Conflicts

The growing urbanization of coastal areas along with maritime commercial transport and recreational development of marine zones have created new and competing uses for the marine habitats as a natural environment. Additionally, offshore marine aquaculture itself has environmental impacts, which must be considered along with land-based runoff and related consequences. As groups sustain all of these activities with their respective interests and constituencies, appropriate mechanisms must be found to reach acceptable arrangements for all. Regulatory Framework

The regulatory provisions for field applications of marine biotechnology need to be in place in a clear and consistent framework at both the national and regional levels. This is a continuing challenge that will require a greater understanding of the delicate marine habitat, the tools of research, and their applications for public benefit through sustainable development of marine resources. Regulations must be science-based, transparent, and enforceable at every level of government. International harmonization of regulations will require continuing attention. Intellectual Property Protection

Commercialization of knowledge and technological breakthroughs generated by marine biotechnology research requires recognition and protection of the intellectual property involved. Countries that have put intellectual property regimes in place relevant to marine resources are also dealing with issues of

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE BIOTECHNOLOGY

assignment of benefits to researchers and to governmental agencies in charge of managing the resources. In the case of marine-protected areas, the intellectual property issues are widely debated. Recent National Policy Initiatives

The United States has shown increased interest in marine biotechnology during the past several years. The Oceans Act of 2000 established a 16-member Commission on Ocean Policy to undertake an 18month study and make recommendations for a national ocean policy for the United States, including stewardship of marine resources, pollution prevention, and commerce and marine science. In 2001, the US Department of Agriculture sponsored a workshop titled Biotechnology–Aquaculture Interface: The Site of Maximum Impact Workshop, which focused on the research and development opportunities of new technologies on genomics, biocomplexity, biocellular technology, biosecurity, and social issues. In 2002, the Ocean Studies Board of the National Academy of Sciences reported on two workshops that highlighted new developments and opportunities in environmental and biomedical applications of marine biotechnology. The report highlighted the need to: (1) develop a fundamental understanding of the genetic, nutritional, and environmental factors that control the production of primary and secondary metabolites in marine organisms, as a basis for developing new and improved products and (2) identify bioactive compounds and determine their mechanisms of action and natural function, to provide models for new lines of selectively active materials for application in medicine and the chemical industry.

Future Perspectives Marine-related biotechnological tools can now be used to determine the effect of natural and anthropogenic perturbations on commercial stocks, propose management tools, determine predator–prey relationships, and restore habitats. The development of predictive models for analyzing potential global climate changes depends on the acquisition of fundamental information on molecular regulatory mechanisms of plankton growth in the oceans. An increased understanding of marine ecological systems will allow the specification of the ‘normal’ baseline level of their biological function, the monitoring and prediction of potential changes, or biological impacts on the systems. To grow and sustain a vital marine biotechnology presence, the essential elements include a supportive

565

government infrastructure, public and private funding for research and development, support for high-risk start-up companies through venture capital funds, a strong intellectual property system, public acceptance through awareness and involvement, and a reliable qualified human resource pipeline. Ideally, all of these areas would be addressed simultaneously, but practical reality and limited resources mean that progress will not be made uniformly in all areas. Having a goal of moving forward in all of these areas is a reasonable expectation for the coming years.

Glossary anthropogenic Processes, objects, or materials that are derived from human activities. apoptosis Programmed cell death. aquatic extremophiles Organisms that live in extreme environments of salt, chemicals, pH, temperature, etc. benthic Living near the bottom of the sea. biomarkers Indicators of a particular state of an organism. bioremediation The use of living organisms to clean up or improve polluted or contaminated environments. biosurfactants Compounds produced on surfaces of living cells that reduce the surface tension. bycatch Species caught in a fishery intended to target another species. chemotherapeutics The branch of therapeutics that deals with the treatment of disease by means of chemical substances or drugs. cryopreservation A process used to preserve cells or tissues through cooling to subzero temperatures. genetic management The use of genetic means to control the characteristics of a population. genotoxins A chemical or other agent that damages cellular DNA, resulting in mutations or cancer. Gram-negative bacteria Bacteria that do not retain crystal violet dye in the Gram stain. marine ecology The relationship between marine organisms and their environment. pyrogens Compounds that produce fever. user conflicts Competing uses for usage of marine zones, including fishing, aquaculture, recreation, and marine transport.

See also Coastal Zone Management. Mariculture Diseases and Health. Mariculture Overview. Ocean Zoning.

(c) 2011 Elsevier Inc. All Rights Reserved.

566

MARINE BIOTECHNOLOGY

Further Reading Attaway DH and Zaborsky OR (1993) Marine Biotechology, Vol. 1: Pharmaceutical and Bioactive Natural Products, 524pp. New York: Springer. Board of Biology, Ocean Studies Board, National Research Council (2000) Opportunities for Environmental Applications of Marine Biotechnology, 187pp. Washington, DC: National Academy Press. Bowen RE, Halvorson H, and Depledge MH (2006) The oceans and human health. Marine Pollution Bulletin 53: 539--656. Fenical W, Greenberg M, Halvorson HO, and HunterCevera JC (1993) International Marine Biotechnology Conference, 2 vols. Dubuque, IA: Williams C. Brown Publishers. Le Gal Y and Halvorson HO (1998) New Developments in Marine Biotechnology, 343pp. New York: Plenum. Le Gal Y and Ulber R (1997) Marine Biotechnology II, 297pp. Berlin: Springer.

Matsunaga T (2004) Marine Biotechnology: Proceedings of Marine Biotechnology Conference 2003, Chiba, Japan. Marine Biotechnology 6, 554pp. National Research Council (2007) Marine Biotechnology in the Twenty-First Century. Problems, Promises and Products, 132pp. Washington, DC: National Academy Press.

Relevant Websites http://www.floridabiotech.org – Center of Excellence in Biomedical and Marine Biotechnology. http://www.umbi.umd.edu – Center of Marine Biotechnology, UMBI. http://www.esmb.org – European Society for Marine Biotechnology. http://www.research.noaa.gov – Marine Biotechnology, Oceans and Coastal Research, NOAA Research.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE CHEMICAL AND MEDICINE RESOURCES S. Ali and C. Llewellyn, Plymouth Marine Laboratory, Plymouth, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction The marine environment consists of several defined habitats ranging from the sea surface microlayer which encompasses the first few microns of the water column, through the bulk water column itself, down to the ocean floor and the subsurface sediments underneath which can be found hydrothermal vents, cold seeps, hydrocarbon seeps, and saturated brines, as well as a wide range of mineral and geological variation. It has become increasingly apparent that within all these oceanic layers there is a diversity of micro- and macroorganisms capable of generating a plethora of previously undescribed molecules through novel metabolic pathways which could be of value to both industry and the clinic. The biological diversity in some marine ecosystems may exceed that of the tropical rain forests and this is supported by the presence of 34 out of the 36 phyla of life. This biodiversity stems from the wide range of environmental conditions to which marine organisms have adapted for survival, including extremes of pH (acid and alkali), temperature (high and low), salinity, pressure, and chemical toxicity (complex polycyclic hydrocarbons, heavy metals). Marine organisms currently being exploited for biotechnology include sponges, tunicates, bryozoans, mollusks, bacteria, cyanobacteria, macroalgae (seaweeds), and microalgae. These organisms have produced compounds with good activities for a range of infectious and noninfectious disease with high specificity for the target molecule (usually an enzyme). Targets of marine natural products which may be clinically relevant include ion channels and G-proteincoupled receptors, protein serine-threonine kinases, protein tyrosine kinases, phospholipase A2, microtubule-interfering agents (of which the largest number identified are of marine origin), and DNA-interactive compounds. In addition to small organic molecules, marine organisms are increasingly being recognized as a potential source of novel enzymes which could be of industrial and pharmaceutical importance. More than 30 000 diseases have been clinically described, yet less than one-third of these can be treated based on symptoms and only a small number can be cured.

Thus, the potential market for novel marine compounds for clinical development is enormous. In addition to providing new molecules for direct clinical intervention, the marine environment is also rich in compounds which are finding uses as natural additives in foods, as nutritional supplements including color additives and antioxidants, and as vitamins, oils, and cofactors which enhance general well-being. Marine organisms are also increasingly providing new solutions to developments in such diverse fields as bioremediation, biocatalysis and chemistry, materials science, nanotechnology, and energy. Some of the potential uses of marine products are summarized in Figure 1. The oceans have long been a source of nutrients, additives, and medicines derived from marine mammals and fish; however, this article focuses on some of the potential which is harbored in predominantly microscopic organisms which are now being increasingly studied for novel bioactive compounds and chemicals and may provide a sustainable alternative source for new compounds and processes.

Novel Metabolites and Drug Discovery Marine organisms have long been recognized as a source of novel metabolites with applications in human disease therapy. Particular emphasis has been placed on the invertebrates such as sponges, mollusks, tunicates, and bryozoans, but more recently advances in genetics and microbial culture have led to a growing interest in cyanobacteria and marine bacteria. For example, a number of anticancer drugs have been derived from marine sources such as sponges which have proven difficult to cultivate and their metabolites display a structural complexity which often precludes total chemical synthesis as an option for potential drug candidates. In recent years, studies have suggested that many of these complex molecules may in fact be the product of microbes which live in a symbiotic relationship with the sponge and that some of these molecules may be the final product of reactions carried out by different organisms. A major challenge within marine biotechnology will be to ascertain the nature of the organisms present in the symbiotic relationship and to identify the pathways involved in metabolite production. A recent advance in molecular biology with the development of metagenomics has opened up the possibility of organismindependent cultivation of genetic material and subsequent screening and characterization of that

(c) 2011 Elsevier Inc. All Rights Reserved.

567

568

MARINE CHEMICAL AND MEDICINE RESOURCES

Enzymes Substrates Biopolymers

Ceramics Adhesives Nanostructures

Energy-rich oils Microbial batteries Hydrogen production

Biocatalysis Materials science

Biofuels

Algal feedstocks Surfactants Fluorescent molecules

Heavy metal recovery Desalination Pollution detection

Novel metabolites for chemistry

Marine organisms

Bioremediation

Sunscreens and cosmetics

Drug discovery

Anti-infectives Anticancer Metabolic diseases

Nutraceuticals

Vitamins Antioxidants Probiotics

Food additives

Collagens Antioxidants Revitalisers

Colorants Texturing agents Essential oils

Figure 1 Uses for marine organisms and their products in the chemical, pharmaceutical, energy, and environmental industries.

DNA for novel metabolic pathways and enzymes. This provides a powerful tool for accessing difficultto-culture microbes which exist in complex symbiotic relationships. Recent advances in the isolation and culture of marine bacteria using both flow cytometry and microencapsulation-based methods have yielded a vast array of previously unknown bacteria (Figure 2). Increasingly, these bacteria are being tested for the presence of bioactive compounds with activities against a diverse range of human and infectious diseases including cancer, HIV, hepatitis C, malaria, and those caused by the increasingly drug-resistant common bacterial pathogens (e.g., Staphylococcus aureus, Enterococcus faecalis, Mycobacterium tuberculosis). For example, the antibiotic abyssomicin C has been isolated from an actinomycete which

was cultured from marine sediment collected in the Sea of Japan. Abyssomicin has been shown to interfere with the synthesis of the essential cofactor folic acid in bacteria and is active against the methicillinresistant Staphylococcus aureus (MRSA) pathogen. Thus, marine bacteria represent a vast untapped source for novel compounds with the potential for development as novel drugs.

Marine-Derived Nutraceuticals and Food Additives In recent years, there has been an upsurge in the consumption of nutritional supplements such as vitamins and cofactors which are essential to cellular function. One traditional supplement has been

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE CHEMICAL AND MEDICINE RESOURCES

(a)

569

(b)

Figure 2 Examples of marine microbial cultures. (a) Bacteria isolated from the English Channel. (b) Microalgal cultures of chlorophytes, cryptophytes, and haptophytes.

cod-liver oil, which is rich in the omega-3 and omega-6 long-chain polyunsaturated fatty acids (PUFAs). Fatty acids have long been used as supplements in the aquaculture industry but in recent years their medicinal value to humans has been proposed. PUFAs have been implicated in enhanced blood circulation and brain development, particularly docosahexaenoic acid (omega-3), which plays an important role in early brain development and neurite outgrowth. Although PUFAs have traditionally been extracted from fish oils they do not naturally occur there but rather accumulate from the diet of the fish. The primary source of PUFAs are marine microbes, and in recent years the isolation and characterization of microbes which produce these fatty acids have allowed their production in other organisms. At present considerable effort is being expended on obtaining high-level production of PUFAs in plants using genes from microalgae with the aim of eventually producing PUFAs in plants which have traditionally been a source of natural oils (e.g., linseed oil, rapeseed oil). The technology to grow and extract oils from such plants in large scale already exists and the development of genetically modified plants which can produce PUFAs in a readily accessible and sustainable form for an increasing market is highly desirable. Consumer-led demand for naturally occurring food colorings and antioxidants has resulted in increased interest in photosynthetic pigments and in particular the carotenoids which occur within

marine microalgae (Figure 3). There is, for example, widespread use of the carotenoid astaxanthin; this is a pigmented antioxidant produced by many microalgae and is responsible for the red color often associated with crustaceans such as shrimps, crabs, and lobsters. Astaxanthin possesses an unusual antioxidant activity which has been implicated in a wide range of health benefits such as preventing cardiovascular disease, modulating the immune system as well as effects in cancer, diabetes, and ocular health. Its antioxidant activities may also have a neuroprotective effect. It has been used extensively in the feed of farmed fish as a nutritional supplement and is partly responsible for the strong coloration often observed in farmed salmon, a fish which naturally accumulates astaxanthin in the wild resulting in the pink hue of its flesh. Other carotenoids such as betacarotene and lutein are used widely as food coloring and antioxidants. Inclusion, for example, in the diets of chickens leads to a darkening of the egg yolk resulting in a rich yellow color. Another group of pigments of commercial importance is that of the phycobiliproteins; these are used as colorings, in cosmetics, and as fluorescent dyes for flow cytometry and in immunological assays. More recent research suggests that phycobiliproteins have anticancer and anti-inflammatory properties. A more unusual and unique pigment which is being used increasingly in personal care products and may also possess anticancer and antiHIV activity is ‘marennine’, a blue-green pigment

(c) 2011 Elsevier Inc. All Rights Reserved.

570

MARINE CHEMICAL AND MEDICINE RESOURCES

(a)

(b)

O OH

HO

Astaxanthin

O

OH

HO

Lutein

B,B-carotene Figure 3 (a) Any one species of microalgae contains an array of carotenoid pigments, as shown here with the fractionated isolates obtained from the chromatographic analysis of an Emiliania huxleyi extract. (b) Chemical structures of cartenoids widely used as color additives and antioxidants.

produced by Haslea ostrearia. Several species of microalgae and, in particular, Haematococcus, Dunaliella, and Spirulina are now grown on large commercial scale to accommodate the growing demand for natural pigments. In addition to being rich in phycobiliproteins, Spirulina, a filamentous cyanobacterium, contains a wide variety of nutrients including potentially beneficial proteins, lipids, vitamins, and antioxidants.

Spirulina is also reported to have various beneficial effects including antiviral activity, immunomodulatory effects, and a role in modulating metabolic function in humans which could be of value in managing diseases involving lipids and carbohydrates such as diabetes. Furthermore, studies indicate that pretreatment with Spirulina may reduce the toxic side effects observed with some drugs on mammalian organs such as the heart and kidneys.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE CHEMICAL AND MEDICINE RESOURCES

The marine environment is a rich source for naturally occurring antioxidants and pigments with a diverse range of microorganisms producing a unique and valuable resource. The potential for the discovery of new pigments and other additives which can be used to replace some of the existing artificial additives currently being used in the food industry is significant.

Sunscreens The continuous exposure of marine organisms to strong sunlight has resulted in some, primarily the macro- and microalgae, evolving compounds which provide a very good screen against ultraviolet (UV) light. These organisms have the ability to synthesize small organic molecules, called mycosporine-like amino acids (MAAs), which are capable of absorbing UV light very efficiently and thus prevent DNA damage. Over 20 different MAAs occur in nature with a wide range of marine organisms utilizing them, including corals, anemones, limpets, shrimp, sea urchins, and some vertebrates including fish and fish eggs. MAAs are widely distributed across the marine environment; however, they can only be synthesized by certain types of bacteria or algae. For example, red seaweeds and some bloom-forming phytoplankton species are a particularly rich source of MAAs. Studies have revealed that in addition to their screening ability, some MAAs, such as mycosporine-glycine, have antioxidant properties. The ability of naturally occurring compounds such as MAAs to act as effective sunscreens has resulted in some interest from the commercial sector as to the value of these compounds in creams and cosmetics.

Biocatalysis The ability of enzymes to synthesize complex chiral molecules with high efficiency and precision is of considerable interest within the pharmaceutical and chemical industries and marine bacteria present a new source for novel enzymes with not only unusual synthetic properties but also potentially valuable catalytic and structural properties. This arises from the ability of marine bacteria to grow under extreme conditions such as high and low temperature, high pressure (extreme depth), high salinity, and extremes of pH. This has opened up the potential to isolate naturally occurring small molecules which are difficult to synthesize in the laboratory but which could be of value in synthetic organic chemistry as intermediates. The existence of a large number of potentially novel enzymes in these same organisms

571

which are capable of performing diverse chemical modifications not readily amenable by standard chemical synthesis also opens the route to novel chemical modification of synthetic molecules using biocatalysis and biotransformation. Growth of microbes under these extreme conditions has led to proteins which possess different temperature optima and improved stability, which has been exploited in the development of new processes and methods such as the polymerase chain reaction, a method for amplifying specific fragments of DNA, and which depends on a thermostable DNA polymerase isolated from a thermophilic microorganism. It has been suggested that enzymes which display high salt tolerance may be of value in the development of enzyme reactions to be performed in organic solvents as they appear to be less prone to denaturing under dehydrating conditions. Marine bacteria make up the largest potential single source of novelty in the world’s oceans and of these the major component are the actinobacteria which includes the actinomycetes. Actinomycetes are readily isolated from the marine environment and consequently are the best studied of the actinobacteria but the other more difficult to culture members are now being identified using advanced culturing and molecular techniques. The actinomycetes in particular hold the promise of tremendous diversity and to date have been underexploited. Terrestrial actinomycetes are responsible for about half of the known bioactive molecules isolated from natural sources to date and include antibiotics, antitumor compounds, immunosuppressants, and novel enzymes. Consequently, the isolation of new organisms from the environment and their analysis for novel metabolites has been a cornerstone in drug discovery. In recent years, however, terrestrial organisms have divulged less novelty than before, and advances in microbiology and genetics have now made the exploitation of marine-derived actinomycetes more attractive. The recognition that the world’s oceans are rich in biological diversity and that extreme environmental conditions (e.g., high pressure and temperature at deep-sea hydrothermal vents) have not repressed the development of organisms to form distinct ecological niches suggests that these habitats will be a rich source of chemical novelty. Although much emphasis has been placed on isolating organisms from extreme environments in the search for novel biocatalysts, the general marine environment should not be ignored. Both micro- and macroalgae have been demonstrated to produce novel enzymes with possible applications in biocatalysis such as the haloperoxidases, enzymes capable of

(c) 2011 Elsevier Inc. All Rights Reserved.

572

MARINE CHEMICAL AND MEDICINE RESOURCES

introducing halogen atoms into metabolites. For example, two species of tropical red macroalgae produce halogenated compounds as a defense against predators and such compounds are being tested for medical applications. The availability of haloperoxidases with different catalytic functions would be of use in generating new types of halogenated molecules for the chemical and pharmaceutical industries. The realization that viruses are the most abundant biological agents in the marine environment and the discovery of highly diverse, ancient, giant viruses with genomes comparable in size to the smallest microbes opens up new sources of genetic diversity. Current indications are that the oceans contain a wide variety of both DNA and RNA viruses with survival strategies which mimic those of terrestrial viruses yet these marine viruses encode a great many proteins of unknown function. Most marine viruses are assumed to be bacteriophages because virus particles are most commonly detected in the vicinity of bacteria, and bacteria are the most abundant organisms in the oceans. Recent studies have revealed that marine viruses encode unexpected and novel proteins which would not be expected to occur within a virus genome. For example, the giant algal viruses have been shown to encode novel glycosylases, potassium pumps, and a pathway for the synthesis of complex sphingolipids. This biochemical diversity indicates that marine viruses could be a rich source for exploitation in the future for new types of carbohydrate and lipid as well as new proteins and enzymes.

ammonia could be of value in the treatment of wastewater, and an understanding of the anaerobic oxidation of ammonia could lead to the development of new chemical processes. The same organisms also possess unusual metabolic intermediates such as hydrazine and produce unusual lipids which could also be of value in the search for new chemical intermediates. Heterocyclic molecules containing sulfur, nitrogen, and oxygen are among the most potent pollutants and inevitably contaminate the marine environment to a considerable extent. The use of microbes to degrade and detoxify such compounds is gaining considerable interest as a process which is environmentally friendly and would represent a long-term solution to removing heterocyclic contaminants. A particularly rich source of such organisms is the marine environment where growth in close proximity to sulfur-rich hydrothermal vents or adjacent to hydrocarbon (oil) seeps on the ocean floor has produced a plethora of microorganisms with metabolisms adapted to the utilization of a wide variety of carbon-, nitrogen-, and sulfur-based chemistries. Again, these organisms are also a very rich source of enzymes with previously unknown characteristics such as unusual substrate specificity, which could be of great value to the chemicals industry where they could be utilized in the production of new or difficult-to-synthesize compounds because they can perform reactions which are difficult to duplicate using traditional synthetic chemistry methods.

Microbial Fuel Cells and Biofuels Bioremediation Pollution of the marine environment is a growing concern particularly with the continuous discharge of both industrial and domestic waste into rivers and estuaries leading to concerns about the impact such pollution could have on long-term human health. The discovery of marine microorganisms capable of detoxifying heavy metals and utilizing complex hydrocarbons as an energy source has provided a new impetus to develop natural solutions to the problems of environmental pollution. However, it should be remembered that toxic substances are not the only causes of marine distress and that the utilization of fertilizers and the disposal of sewage can also result in an imbalance in the marine ecology, resulting in the formation of large, often toxic, algal blooms which although not always a direct threat to human health do lead to widespread ecological damage. Thus, the discovery of microbes capable of growing in the presence of high concentrations of

The use of marine organisms to produce fuels has also been proposed. The generation of electricity through the degradation of organic matter has recently been demonstrated to occur in marine sediments and may be mediated by complex communities of marine microorganisms. These organisms degrade complex organic matter such as carbohydrates and proteins to simpler molecules such as acetate which are then used by electricity-generating bacteria to reduce metals such as iron and manganese. By replacing the naturally occurring metals with an anode these bacteria, under anoxic conditions, will supply the electrons needed to produce an electric current to a cathode linked to the anode by wires and exposed to the oxygen in the water column. It has been suggested that this type of system could be used to supply the electricity needed to operate equipment in regions where access is difficult and so eliminate the need to replace batteries. Microbial fuel cells would be selfsustaining, would not require the preprocessing of

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE CHEMICAL AND MEDICINE RESOURCES

fuels to function efficiently, and would not contribute net CO2 to the atmosphere nor produce toxic waste as with conventional batteries. Another area of intense study is the development of renewable biofuels with much focus being given to developing terrestrial plant species to produce the precursors to biodiesel. Microalgae present a potential alternative source of hydrocarbons for the generation of biofuels as some species naturally produce significantly more oil (per year per unit area of land) than terrestrial oil seed crops. Several marine species such as Porphyridiuim, Chlorella, and Tetraselmis are currently under investigation as sources rich in hydrocarbons suitable for biofuel production.

Biomaterials Another area of interest is the development of novel biomaterials inspired by marine organisms. Areas of particular interest are the mechanism of calcium- and silica-based structure formation which is found in many phytoplankton, formation of hard chitinous shells in many larger marine organisms such as oysters and crabs, as well as the very powerful bioadhesives produced by mussels and barnacles. Proteins form the basis of a number of naturally occurring adhesive molecules which display a number of attractive features, particularly for clinical and other specialist applications. These features include the ability to adhere strongly to both smooth and uneven surfaces with a high degree of bonding strength and the ability to form and maintain bonds in very humid and wet conditions. This bonding ability seems to be strongly linked to the presence of hydroxylated tyrosine residues (L-dopa; L-3,4-dihydroxyphenylalanine) in such proteins and it is thought that adhesion involves interactions between the hydroxyl groups and the target surface. The development of powerful adhesives which can cure rapidly under wet conditions and are nontoxic would be of particular value in the clinic. At present there is considerable interest in using bioadhesives in the field of ophthalmology where the use of alternatives to sutures in, for example, corneal grafts is desired in order to reduce the risks of irritation and scarring to the eye following surgery. Another area where the use of bioadhesives is being actively researched is in drug delivery where the ability to attach naturally occurring polymers which can slowly release a drug over time would be useful. This is particularly relevant for poorly soluble biological drugs based on antibodies and other large proteins which can be difficult to administer. The potential contribution that marine-derived biomaterials and bioadhesives could make to such fields is enormous.

573

Chitin is the second most abundant natural polysaccharide after cellulose and is found in the exoskeletons of crustaceans such as crabs and shrimp as well as in the cell walls of fungi and cuticles of insects. The deacetylation of chitin produces chitosan, a biopolymer with great potential in medicine. Chitosan and its derivatives possess numerous applications due to their properties which include reactive functional groups, gel-forming capability, low toxicity, and high adsorption capacity, as well as complete biodegradability and antibacterial and antifungal activities. These properties make chitosan particularly attractive in areas of research such as drug delivery and tissue engineering where a nontoxic, biodegradable scaffold with antimicrobial activity would be particularly attractive. Both chitin and chitosan can influence the immune system and are being studied extensively as biomaterials in the development of supports for accelerated wound healing. These chitosan-based materials are also being modified to improve adhesion to wound sites and for the incorporation of antimicrobial agents to minimize the risk of infection. Chitosan and its derivatives are also being used to develop scaffolds for applications in tissue engineering to grow cells to form complex structures which could ultimately be used to replace damaged tissues and organs. The elaborate silica-based structures (frustules) which are exhibited by many diatoms have been of interest to materials scientists for many years and recent studies have begun to reveal some of the characteristics that are present in these silica shells, including an understanding of the proteins and other molecules involved in structure formation (Figure 4). The highly precise nature of the structures has led to suggestions that the silica structures can be used directly as either templates for microfabrication or as materials for use in microprocesses such as filters in microfluidics. By understanding and manipulating the growth environment of any given diatom it may be possible to modify the precise geometry of the natural silica shells it produces and the resultant frustules could then be modified using standard microengineering techniques to create new nanostructures with potential applications in the development of medical devices. The glass-like properties of diatom frustules, the remains of which form diatomite (diatomaceous earth), have over 300 recorded commercial applications. The fine pores present in the frustules make them especially useful in filtration processes and the bulk of diatomite is used for this purpose. It has also been suggested that frustules might have applications in the development of new optical devices.

(c) 2011 Elsevier Inc. All Rights Reserved.

574

MARINE CHEMICAL AND MEDICINE RESOURCES

(a)

Future Prospects The marine environment, representing 70% of the Earth’s surface, is a vast untapped resource for new chemicals and enzymes, often with characteristics which are considerably different to anything discovered in the terrestrial environment. A major hurdle to the study and exploitation of this resource has been the inaccessibility of the oceans but advances in science and technology are now providing new approaches to isolating and characterizing the organisms present. This should lead to a considerable increase in the number of novel marine-derived chemicals and medicines available to the market in the near future.

(b)

Glossary actinomycetes Major group of bacteria commonly isolated from low-nutrient environments. Rich source of unusual small molecules (e.g., antibiotics) and enzymes. cyanobacteria Aquatic photosynthetic bacteria widely found throughout nature. macroalga(e) Seaweed(s). microalga(e) Microscopic single-cell plants. nutraceuticals Natural chemicals, usually contained in foods, with potential benefits to human health.

See also Global Marine Pollution. Inverse Modeling of Tracers and Nutrients. Marine Biotechnology. Metal Pollution. Nuclear Fuel Reprocessing and Related Discharges. Pollution: Approaches to Pollution Control. Pollution: Effects on Marine Communities. Pollution, Solids. Radioactive Wastes. Thermal Discharges and Pollution. Trace Element Nutrients.

(c)

Further Reading

Figure 4 Scanning electron microscope images of (a) the diatom Thalassiosira, (b) the dinoflagellate Gonyualax, and (c) the dinoflagellate Peridinium showing the intricacies of cell walls as inspirations to novel biomaterials.

Gullo VP, McAlpine J, Lam KS, Baker D, and Petersen F (2006) Drug discovery from natural products. Journal of Industrial Microbiology and Biotechnology 33: 523--531. Higuera-Ciapara I, Fe´lix-Valenzuela L, and Goycoolea FM (2006) Astaxanthin: A review of its chemistry and applications. Critical Reviews in Food Science and Nutrition 46: 185--196. Hussein G, Sankawa U, Goto H, Matsumoto K, and Watanabe H (2006) Astaxanthin, a carotenoid with potential in human health and nutrition. Journal of Natural Products 69: 443--449.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE CHEMICAL AND MEDICINE RESOURCES

Khan Z, Bhadouria P, and Bisen PS (2005) Nutritional and therapeutic potential of Spirulina. Current Pharmaceutical Biotechnology 6: 373--379. Ko¨nig GM, Kehraus S, Seibert SF, Abdel-Lateff A, and Mu¨ller D (2005) Natural products from marine organisms and their associated microbes. ChemBioChem 7: 229--238. Lovley DR (2006) Microbial fuel cells: Novel microbial physiologies and engineering approaches. Current Opinion in Biotechnology 17: 327--332. Marszalek JR and Lodish HF (2005) Docosahexaenoic acid, fatty acid-interacting proteins, and neuronal function: Breastmilk and fish are good for you. Annual Review of Cell and Developmental Biology 21: 633--657. Napier JA and Sayanova O (2005) The production of verylong-chain PUFA biosynthesis in transgenic plants: Towards a sustainable source of fish oils. Proceedings of the Nutritional Society 64: 387--393.

575

Newman DJ and Hill RT (2005) New drugs from marine microbes: The tide is turning. Journal of Industrial Microbiology and Biotechnology 33: 539--544. Shi C, Zhu Y, Ran X, Wang M, Su Y, and Cheng T (2006) Therapeutic potential of chitosan and its derivatives in regenerative medicine. Journal of Surgical Research 133: 185--192. Shick JM and Dunlap WC (2002) Mycosporine-like amino acids and related gadusols: Biosynthesis, accumulation, and UV-protective functions in aquatic organisms. Annual Review of Physiology 64: 223--262. Suttle CA (2005) Viruses in the sea. Nature 437: 356--361. Wilt FH (2005) Developmental biology meets materials science: Morphogenesis of biomineralized structures. Developmental Biology 280: 15--25.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE FISHERY RESOURCES, GLOBAL STATE OF J. Csirke and S. M. Garcia, Food and Agriculture Organization of the United Nations, Rome, Italy & 2009 Elsevier Ltd. All rights reserved.

short-term exceptionally high initial catches (a phenomenon known as ‘overshooting’) can be obtained when a new fishery on a previously unexploited stock develops too fast. However, the lower catch values in recent years may also be indicative of an increase in the number of resources being overfished or depleted.

Introduction

Global Levels of Exploitation

The Fisheries Department of the Food and Agriculture Organization of the United Nations (FAO) monitors the state of world marine fishery resources and produces a major review every 6–8 years, with shorter updates presented every 2 years to the FAO Committee on Fisheries (COFI) as part of a more general report The State of Fisheries and Aquaculture (SOFIA). The latest major FAO review of the state of world marine fishery resources was issued in 2005 with a shorter update in SOFIA 2006 and a further updating focusing on fishery resources that can be found partly or entirely on the high seas also published in 2006. This article draws significantly from sections of the above FAO publications and uses catch information available from 1950 to 2004 (the last year for which global catch statistics are available). With a view to offering a comprehensive description of the global state of world fish stocks, the short analysis provided below considers successively: (1) the relation between 2004 and historical production levels; (2) the state of stocks, globally and by regions according to information from the FAO reports above; and (3) the trends in state of stocks since 1974, globally and by region.

The recent FAO reviews report on 584 stock or species groups (stock items) being monitored. For 441 (or 76%) of them, there is some more-or-less recent information allowing some estimates of the state of exploitation. These ‘stock’ items are classified as underexploited (U), moderately exploited (M), fully exploited (F), overexploited (O), depleted (D), or recovering (R), depending on how far they are from ‘full exploitation’ in terms of biomass and fishing pressure. ‘Full exploitation’ is used by FAO as loosely equivalent to the level corresponding to maximum sustainable yield (MSY) or maximum long-term average yield (MLTAY).

Relative Production Levels The catch data available for the 19 FAO statistical areas (Table 1) or regions of the world’s oceans indicate that four of them are at or very close to their maximum historical level of production: the Eastern Indian Ocean and the Western Central Pacific Oceans reached maximum production in 2004 while the Western Indian Ocean and the NE Pacific produces 95% and 90% of the maximum, produced in 2003 and 1987, respectively. All other regions are presently producing less than 90% of their historical maximum, for various reasons. This may result, at least in part, from natural oscillations in productivity caused by ocean climate variability or by the fact that

576

1. U and M stocks could yield higher catches under increased fishing pressure, but this does not imply any recommendation to increase fishing pressure. 2. F stocks are considered as being exploited close to their MSY or MLTAY and could be slightly under or above this level because of natural variability or uncertainties in the data and in stock assessments. These stocks are usually in need of (and in some cases already have) effective control on fishing capacity in order to remain as fully exploited, and avoid falling in the following category. 3. O or D stocks are exploited beyond MSY or MLTAY levels and have their production levels reduced; they are in need of effective strategies for capacity reduction and stock rebuilding. 4. R stocks are usually at very low abundance levels compared to historical levels. Directed fishing pressure may have been reduced by management or because of profitability being lost, but may nevertheless still be under excessive fishing pressure. In some cases, their indirect exploitation as bycatch in another fishery might be enough to keep them in a depressed state despite reduced direct fishing pressure. According to information available in 2004 (Figure 1), 3% of the world fish stocks (all included)

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE FISHERY RESOURCES, GLOBAL STATE OF

Table 1

577

Ratio between recent (2004) total catch and maximum reached catch, by FAO statistical areas

FAO area (code and name)

Year when maximum catch was reached

Total catch in 2004 (103)

Ratio of 2004 over maximum catch (%)

18 21 27 31 34 37 41 47 51 57 61 67 71 77 81 87 48 58 88

1968 1968 1976 1984 1990 1988 1997 1978 2003 2004 1998 1987 2004 2002 1992 1994 1987 1972 1983

0 2353 9952 1652 3392 1528 1745 1726 4147 5625 21 558 3050 11 011 1701 736 15 451 125 8 3

0 52 77 66 82 77 66 53 95 100 87 90 100 87 80 76 25 4 28

– – – – – – – – – – – – – – – – – – –

Arctic Sea Atlantic Northwest Atlantic Northwest Atlantic, Western Central Atlantic, Eastern Central Mediterranean and Black Sea Atlantic, Southwest Atlantic, Southwest Indian Ocean, Western Indian Ocean, Western Pacific, Northwest Pacific, Northwest Pacific, Western Central Pacific, Eastern Central Pacific, Southwest Pacific, Southwest Atlantic, Antarctic Indian Ocean, Antarctic Pacific, Antarctic

Recovering (R)

1%

Depleted (D)

7%

Over exploited (O)

17%

Fully exploited (F)

52%

Moderately exploited (M) Under exploited (U) 0%

20% 3% 10%

20%

30%

40%

50%

60%

Figure 1 State of world fish stocks in 2004. Reproduced with permission from FAO (2005) Review of the State of World Marine Fishery Resources. FAO Fisheries Technical Paper, 457, 235pp, figure A2.1. Rome: FAO.

appeared to be underexploited, 20% moderately exploited, 52% fully exploited, 17% overfished, 7% depleted, and 1% recovering. On the one hand, this indicates that 25% of the world stocks (O þ D þ R) for which some data are available are below the level of abundance corresponding to MSY or are exploited with a fishing capacity well above this level. They require management to rebuild them at least to the level corresponding to MSY as provided by the 1992 UN Convention on the Law of the Sea (UNCLOS). As

52% of the stocks appear to be exploited around MSY and most, if not all, also require that capacity control measures be applied to avoid the negative effects of overcapacity, it appears that 77% (F þ O þ D þ R) of the world stocks for which data are available require that strict capacity and effort control be applied in order to be stabilized or be rebuilt around the MSY biomass levels, and possibly beyond. Some of the fisheries concerned may already be under such management schemes. On the other hand, Figure 1 also indicates that 23% (U þ M) of the world stocks for which some data are available are above the level of abundance corresponding to MSY, or are exploited with a fishing capacity below this level. Considering again that 52% of the stocks are exploited around MSY, this means that 75% of the stocks (U þ M þ F) are at or above MSY level of abundance, or are exploited with a fishing capacity at or below this level, and should be therefore considered as compliant with UNCLOS basic requirements. These two visions of the global situation of fishery stocks indicate that the ‘glass is half full or half empty’ and both are equally correct depending on which angle one takes. From the ‘state of stocks’ point of view, it is comforting to see that 75% of the world resources are still in a state which could produce the MSY, as provided by UNCLOS. From the management point of view, it should certainly be noted that 77% of the resources require stringent management and control of fishing capacity. As mentioned above, some of these (mainly in a few

(c) 2011 Elsevier Inc. All Rights Reserved.

578

MARINE FISHERY RESOURCES, GLOBAL STATE OF

60% World overall Tunas and tuna-like

50%

Oceanic sharks Straddling stocks

40%

30%

20%

10%

0% U

M

F O State of exploitation

D

R

Figure 2 State of exploitation of world highly migratory tuna and tuna-like species, highly migratory oceanic sharks, and straddling stocks (including high seas fish stocks). Reproduced with permission from Maguire J-J, Sissenwine M, Csirke J, Grainger R, and Garcia S (2006) The State of World Highly Migratory, Straddling and Other High Seas Fishery Resources and Associated Species. FAO Fisheries Technical Paper, 495, 84pp, figure 59. Rome: FAO.

developed countries) are already under some form of capacity management. Many, however, would require urgent action to stabilize or improve the situation. For 25% of them, energetic action is required for rebuilding. The situation appears more critical in the case of fish stocks that occur partly or entirely, and can be fished in the high seas (Figure 2). While the state of exploitation of highly migratory tuna and tuna-like species is very similar to that of the world overall, the state of exploitation of highly migratory oceanic sharks appears to be more problematic, with more than half of the stocks listed as overexploited or depleted. The state of straddling stock (including high seas stocks) is even more problematic with nearly two-thirds being classified as overexploited or depleted.

State of Stocks by Region When the available information is examined by regions (Figure 3), the percentage of stocks exploited at or beyond levels of exploitation corresponding to MSY (F þ O þ D þ R) and needing fishing capacity control ranges from 43% (for the Eastern Central Pacific) to 100% (in the Western Central Atlantic). Overall, in most regions, 70% of the stocks at least are already fully exploited or overexploited.

The percentage of stocks exploited at or below levels corresponding to MSY (U þ M þ F) ranges from 48% (in the Southeast Atlantic) to 100% (in the Eastern Central Pacific). An indication of how weak (or strong) management and development performance can be given by the proportion of stocks that are exploited beyond the MSY level of exploitation (O þ D þ R), that in the latest reviews were ranging from 0% (for the Eastern Central Pacific) to 52% (for the Southeast Atlantic).

Global Trends The trends in the proportion of stocks in the various states of exploitation as taken from the various FAO reviews since 1974 (Figure 4) shows that the percentage of stocks maintained at MSY level, or fully exploited (F) has slightly increased since 1995, reversing the previous decreasing trend since 1974. The underexploited and moderately exploited stocks (U þ M), offering some potential for expansion, continue to decrease steadily, while the proportion of stocks exploited beyond MSY levels (O þ D þ R) increased steadily until 1995, but has apparently leveled off and remained more or less stable at around 25% since. The number of ‘stocks’ for which information is available has also increased during the same period, from 120 to 454.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE FISHERY RESOURCES, GLOBAL STATE OF

579

Southeast Atlantic (47) Southeast Pacific (87) Southern Oceans (48, 58, and 88) Northeast Atlantic (27) Tuna and tuna-like species total Mediterranean and Black Sea (37) Southwest Atlantic (41) Western Central Atlantic (31) Northeast Pacific (67) Eastern Central Atlantic (34) Western Indian Ocean (51) Eastern Indian Ocean (57) Southwest Pacific (81) Northwest Atlantic (21) Northwest Pacific (61) Western Central Pacific (71) Eastern Central Pacific (77) 0%

10%

20%

30%

40%

Underexploited + moderately exploited

50%

60%

Fully exploited

70%

80%

90%

100%

Overexploited + depleted + recovering

Figure 3 Percentage by FAO fishing areas of stocks exploited at or beyond MSY levels (F þ O þ D þ R) and below MSY levels (U þ M). Reproduced with permission from FAO (2005) Review of the State of World Marine Fishery Resources. FAO Fisheries Technical Paper, 457, 235pp, figure A2.2. Rome: FAO.

60% 50% 40% 30% 20% 10% 0% 1970

1975

1980

1985

1990

1995

2000

2005

Underexploited + moderately exploited Fully exploited Overexploited + depleted + recovering Figure 4 Global trends in the state of exploitation of world stocks since 1974. Reproduced with permission from FAO (2005) Review of the State of World Marine Fishery Resources. FAO Fisheries Technical Paper, 457, 235pp, figure A2.3. Rome: FAO.

Discussion The perspective view of the state of world stocks obtained from the series of FAO fishery resources reviews indicates clearly a number of trends. Globally, between 1974 and 1995, there was a steady increase in the proportion of stocks classified as ‘exploited beyond the MSY limit’, that is, overfished, depleted, or recovering (after overexploitation and depletion). These conclusions are in line with earlier findings summarized by Garcia, de Leiva, and

Grainger in Figure 5. Since the findings by Garcia and Grainger in 1996 were based on a sample of the world stocks, severely constrained by availability of information to FAO, the conclusions are considered with some caution. A key question is: To what extent does the information available to FAO reflect reality? There are many more stocks in the world than those referred to by FAO. In addition, some of the elements of the world resources referred to by FAO as ‘stocks’ are indeed conglomerates of stocks (and often of species). One should therefore ask what validity a statement made for the conglomerate has for individual stocks (stricto sensu). There is no simple reply to this question and no research has been undertaken in this respect. However, while recognizing that the global trends observed reflect trends in the monitored stocks, it is also noted that the observations generally coincide with reports from studies conducted at a ‘lower’ level, usually based on more insight and detailed data. For instance, an analysis on Cuban fisheries using the same approach as used by Garcia and Grainger for the whole world leads to surprisingly similar conclusions, using less coarse aggregations with an even longer time series. There is of course the possibility that stocks become ‘noticed’ and appear in the FAO information base as ‘new’ stocks only when they start getting into trouble and scientists having accumulated enough data start dealing with them, generating reports that FAO can access. This could explain the increase in the percentage of stocks exploited beyond MSY since

(c) 2011 Elsevier Inc. All Rights Reserved.

580

MARINE FISHERY RESOURCES, GLOBAL STATE OF 100% 90% 80% 70%

Recovering

60% Senescent 50% Mature

40% 30%

Developing

20%

Undeveloped

10% 0% 1951− 1956− 1961− 1966− 1971− 1976− 1981− 1986− 1991− 1996− 55

60

65

70

75

80

85

90

95

2000

Figure 5 Stage of development of the 200 major marine fishery resources: 1950–2000. Reproduced with permission from FAO (2005) Review of the state of world marine fishery resources. FAO Fisheries Technical Paper, 457, 235pp, figure A2.4. Rome: FAO.

1974. This assumption, however, does not hold for at least two reasons. 1. The number of ‘stock items’ identified by FAO but for which there is not enough information has also increased significantly with time, from seven in 1974 to 149 in 1999, clearly showing that new entries in the system are not limited to welldeveloped fisheries or fish stocks in trouble. 2. From the 1980s, based on the recognition of the uncertainties behind identification of the MSY level, and recognizing also the declines due to decadal natural fluctuations, scientists have become more and more reluctant to definitely classify stocks as ‘overfished’. The apparent ‘plateauing’ of the proportion of stocks with excessive exploitation in the world oceans may in part be due to this new trend. While the trend analyses of world fisheries landings (Figure 5) tend to suggest that the proportion of senescent and recovery fisheries (taken as grossly corresponding to those being exploited beyond MSY) is still on the increase, the resources analyses summarized in Figure 4 indicate that there is a ‘flattening’ in the proportion of fish stocks that are overexploited (or exploited beyond MSY). Even if the latter may be indicative that the ‘deterioriation’ process leading to the overexploitation of marine fishery resources has slowed down and that eventually the past trend can be reversed, there is no evidence of clear improvements in the state of

exploitation of world stocks, and with 25% the proportion of stocks overexploited or depleted is still high.

See also Demersal Species Fisheries. Dynamics of Exploited Marine Fish Populations. Ecosystem Effects of Fishing. Fish Predation and Mortality. Fisheries: Multispecies Dynamics. Fisheries Overview. Fishery Management. Mariculture Overview. Molluskan Fisheries. Open Ocean Fisheries for Deep-Water Species. Open Ocean Fisheries for Large Pelagic Species. Pelagic Fishes. Salmon Fisheries, Atlantic. Salmon Fisheries, Pacific. Small Pelagic Species Fisheries. Southern Ocean Fisheries.

Further Reading Baisre JA (2000) Chronicles of Cuban Marine Fisheries (1935–1995): Trend Analysis and Fisheries Potential. FAO Fisheries Technical Paper, 394. Rome: FAO. Csirke J Vasconcellos M (2005) Fisheries and long-term climate variability. In: Review of the State of World Marine Fishery Resources. FAO Fisheries Technical Paper, 457, pp. 201–211. Rome: FAO. FAO (2005) Review of the State of World Marine Fishery Resources. FAO Fisheries Technical Paper, 457, 235pp. Rome: FAO.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE FISHERY RESOURCES, GLOBAL STATE OF

FAO (2007) The State of World Fisheries and Aquaculture 2006, 162pp. Rome: FAO. Garcia SM and De Leiva Moreno I (2000) Trends in world fisheries and their resources: 1974–1999. The State of World Fisheries and Aquaculture 2000. Rome: FAO. Garcia SM and Grainger R (1996) Chronicles of Marine Fishery Landings (1950–1994): Trend Analysis and

581

Fisheries Potential. FAO Fisheries Technical Paper, 359. Rome: FAO. Maguire J-J, Sissenwine M, Csirke J, Grainger R, and Garcia S (2006) The State of World Highly Migratory, Straddling and Other High Seas Fishery Resources and Associated Species. FAO Fisheries Technical Paper, 495, 84pp. Rome: FAO.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL DIVING PHYSIOLOGY G. L. Kooyman, University of California San Diego, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1589–1595, & 2001, Elsevier Ltd.

Introduction Marine mammals are the last major group of vertebrates to adapt widely to the marine environment. Reptiles and birds preceded them by tens if not hundreds of millions of years. The marine reptiles had their greatest success in the Mesozoic. Like the dinosaurs most had disappeared by the end of the Cretaceous, except for sea turtles and crocodiles. Marine diving birds were also present in the Mesozoic, including possibly some penguins. However, as with marine mammals, their greatest diversification occurred during the Tertiary. The lack of competition from such successful and formidable marine reptiles as the mosasaurs, ichthyosaurs, and pleisiosaurs may have enabled this adaptive radiation. Nevertheless, then as now, all three groups had similar physical obstacles to overcome in adapting to marine life. These problems stimulated the evolution of some of the most extreme and unusual physiological and morphological adaptations ever achieved by vertebrates. The ancestors of whales were the first to begin the invasion of the sea sometime during the Eocene, more than 60 million years ago (Ma). Sea cows, the only herbivorous marine mammal, originated about 50 Ma during the late Eocene, and pinnipeds followed about 30 Ma in the late Oligocene. Pelagic species wander the vast offshore regions of the world’s oceans, and dive in waters with depths up to thousands of meters. Because the greatest challenges of the physical environment are preeminent in this region, the pelagic whales and pinnipeds will be discussed in greatest detail. There are seven major physical obstacles to overcome that require extreme physiological adaptations to life in the oceans. 1. Anoxia: diving into a world that is without oxygen for an air-breathing mammal. 2. Density: just a short distance from the surface the hydrostatic pressure becomes extreme. 3. Breathing: the less time taken for respiration, the more time at depth to search for prey or to avoid being eaten.

582

4. Vision: even in the best conditions of water clarity in pelagic tropical waters this is a region of twilight to eternal darkness. 5. Acoustics: the limited field of vision underwater increases the importance of hearing over long distances compared to land mammals. 6. Cold: even the warmest tropical sea is 10–151C cooler than the internal temperature of a hotblooded marine mammal. 7. Viscosity: there is a reason animals underwater appear to move in slow motion – their movements are slowed by the viscosity of water. Selection pressure for adaptations to overcome these physical barriers is great and has resulted in some very consistent morphological and physiological adaptations that, in some cases, make it easy to recognize a marine mammal from only a small part of its anatomy. Some of the more salient anatomical features are discussed in relation to their function. Just as there are variations and gradations on the theme of adapting to the marine environment, so too there are extremes that are exemplified by the most pelagic and the deepest divers. Table 1 shows statistics from each major group regarding the simple assessment of diving ability by the maximum and routine depths and durations. It should be noted that even though the diving ability of some species is impressive, the exploitative ability of marine mammals is superficial considering that the average depth of the world’s oceans is 3.5 km and the maximum depth is 11 km. Emphasis will be placed on five of the seven adaptations mentioned.

Adaptations to Anoxia When marine mammals dive below the surface they enter an anoxic environment even though there is much dissolved oxygen in the surrounding water. Lacking gills or any means of extracting oxygen from the water they are without oxygen, except for that stored within their bodies, until they return to the surface. The brain must not be without oxygen for more than three minutes or irreversible damage occurs, and dysfunction occurs even sooner. With such a short margin of resistance, marine mammals had to develop special adaptations to protect the brain and other organs and tissues sensitive to oxygen deprivation. Some of the broad categories of adaptation are: (1) oxygen stores; (2) redistribution of blood flow by cardiovascular adjustments; (3) reduced

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL DIVING PHYSIOLOGY

583

Table 1 Routine and maximum diving characteristics of selected marine mammals from some of the major groups that hunt pelagically Species

California sea lion (Zalophus californianus) Northern elephant seal (Mirounga angustirostris) Sperm whale (Physeter catodon) Bottlenose whale (Ampullatus hyperoodon)

Family

Dive duration (min)

Dive depth (m)

Mean

Max.

Mean

Otariidae

2

10

62

274

Phocidae

22

90

520

1581

Physeteridae

35

73

792/466

2035

Ziphiidae

37

70

1060

1453

metabolism during the dive; (4) behavioral patterns that encourage oxygen conservation; (5) reliance on anaerobic metabolism in organs tolerant to hypoxia. Oxygen Stores

The oxygen consumed during a breath hold is stored in three compartments, the respiratory system, the blood, and the body musculature (Figure 1). The respiratory oxygen store is of marginal value since about 80% of the volume is nitrogen, and because there is little gas exchange between the lung and blood while the animal is at depth. The blood oxygen store is dependent on the blood volume, red cell volume, and the concentration of hemoglobin in the red blood cells. As the cell volume increases, so does viscosity, increasing the resistance to blood flow. Some marine mammals have very high blood oxygen storage capacity, whereas in others the blood oxygen storage capacity is little different from that in terrestrial mammals. In contrast, there is an increased concentration of the oxygen-binding protein myoglobin in muscle in all marine mammals which sets them apart from all other mammals. Myoglobin is 3–15 times more concentrated in the muscle of diving compared to terrestrial mammals, and there may be some relationship between the depth of dives the animal routinely makes and the level of the myoglobin in the muscle (Tables 1 and 2). The deepest divers seem to have the largest oxygen stores. In humans the total store is 20 ml O2kg 1 body mass, which is about a fifth of that in elephant seals (nearly 100 ml O2 kg 1 body mass). Using the human average as a standard of comparison for the typical terrestrial mammal, the elephant seal has a blood volume three times greater, a hemoglobin concentration 1.5 times more, and a myoglobin concentration approximately 10 times more (Figure 1).

Max.

Cardiovascular Adjustments

Specializations of the cardiovascular system varies among the different families of marine mammals. The least cardiovascular modification appears to be in the sea lion, sea otter and manatee. At the other extreme are the cetaceans with numerous variations or entirely new structures, most of which have unknown function. For example, most toothed whales have an extreme development of the thoracic retia mirabilia. This complex network of arteries is invested in the dorsal aspect of the thorax as well as embedded between the ribs. One of its main functions appears to be to provide the primary blood supply to the brain. This role has been usurped from the internal carotid, which does not reach the brain before it ends as a tapered down and occluded vessel. The reason for this complexity, and other possible functions of the thoracic retia are unknown.

Northern elephant seal

California sea lion

97 (4, 71, 25) 39 (21, 45, 34)

Bottlenose dolphin

36 (34, 27, 39) 20 (24, 57, 15) Human

Figure 1 Distribution of oxygen stores within the three major compartments of lung, blood and muscle. Groups represented are the toothed whales (bottlenose dolphin), sea lion and fur seals (California sea lion), and true seals (northern elephant seal). The first number is total body oxygen store; numbers in parentheses are for the lung, blood, and muscle (ml kg l). (Modified from Kooyman (1989).)

(c) 2011 Elsevier Inc. All Rights Reserved.

584

MARINE MAMMAL DIVING PHYSIOLOGY

Table 2 Myoglobin concentrations in the major groups of marine mammals Species

Myoglobin (g 100 g 1 wet weight)

Manatee (Trichechus manatus) Sea otter (Enhydra lutris) Walrus (Odobenus rosmarus) California sea lion (Zalophus californianus) Weddell seal (Leptonychotes weddellii ) Northern elephant seal (Mirounga angustirostris) Fraser’s dolphin (Lagenodelphis hosei ) Sperm whale (Physeter catodon)

0.4 3.1 3.0 2.8 5.4 5.7 (8 months old) 7.1 5.0

A more universal structure is the aortic bulb, an enlargement of the aortic arch, that functions as a capacious, elastic chamber. This bulb absorbs much of the flow energy developed during systole by the left ventricle. This absorbed pulse is then more evenly spread through the rest of the cardiac cycle. The maintenance of blood flow pressure is especially effective during the bradycardia that occurs during a dive. There is considerable modification of the venous circulation in seals. The intravertebral extradural vein that lies within the vertebral canal above the spinal cord is responsible for most of the brain return flow. It also drains portions of the back musculature and the pelvic area. Much of the blood volume returns via the intercoastals to the azygous vein and then to the anterior vena cava to the heart. The major blood return of the body is from the inferior vena cava, which drains into a large hepatic sinus. This vessel passes through a narrow restriction, the vena cava sphincter, at the diaphragm before entering the thoracic cavity. The sphincter is a circular muscle capable of reducing flow return to the heart. Finally, at least in pinnipeds, there is an enlarged spleen that acts as a reservoir for about 50% of the total red blood cell volume while the seal is resting or sleeping. The splenic mass in terrestrial mammals is about 0.5–2% of body mass, whereas it is 4–10% in seals. Once the seal begins to dive these cells are injected into the circulation and raise the hematocrit about 50% above the resting value. Once a diving bout is concluded much of the red blood cells are again stored in the spleen.

The most complete information on cardiovascular function is from seals in which information has been obtained while the animals are making routine foraging dives. Most data from other diving mammals have been collected during trained dives, or during resting submersions. In general, as the dive begins, a rapid onset of bradycardia ensues that may range from 20–90% of the resting heart rate. This is dependent on the duration of the dive, probably the swim speed, and whether the dive is routine or an evasive response that incurs some level of stress. The latter dives are when the most extreme declines in heart rate occur. Concurrent with the reduced heart rate there is the necessary decline in cardiac output, which is greater than would be predicted from the change in heart rate alone. This happens because there is also a reduction of 50% in stroke volume. During the longest dives, there is a reduction in splanchnic blood flow to hypoxic-tolerant organs such as the liver and kidneys, and in somatic blood flow to many of the muscles. The brain, which is not hypoxic tolerant, does not have a reduced blood flow and it becomes a major consumer of the stored oxygen of the circulatory system. What the rate of oxygen consumption is during the dive remains one of the most important unanswered questions in diving physiology. Metabolic Responses

Even a modest reduction in metabolic rate of the liver, kidney, and gastrointestinal tract will help to conserve the body’s oxygen store since they account for about 50% of the total resting oxygen consumption. A small reduction can be easily made up while the animal is ventilating at the surface. The other major consumer is the locomotor muscles. Even under conditions of foraging when the animal may be anxious to swim to the prey patch, the axiom of ‘make haste slowly’ is applicable. It must conserve the muscle oxygen as much as possible, and one option for doing this is to take advantage of its natural buoyancy. Gliding, the behavior that achieves this result, has now been measured. In four species, two seals, a bottlenose dolphin, and a blue whale, it is now known that during the descent below a depth of 20–40 m the animals glide to greater depths. One elephant seal was observed to glide down to 400 m. In this condition only the brain remains uncompromising in its need of a large oxygen requirement. It has been estimated that gliding to depth can result in an energy saving of over 50% compared to a dive in which there is swimming at all times. These are conservative estimates because the drag caused by the attached camera must have been

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL DIVING PHYSIOLOGY

30 25

Post-dive blood lactate (mM)

substantial for all species except the blue whale, and this would incur a cost that in the unfettered animal would be much less. Even at depth there is an additional reduction of metabolic rate for swimming as the streamlined animal strokes intermittently and glides as much as possible. Away from surface effects, the resistance to forward propulsion is at a minimum. The economy in swim effort results in important savings in oxygen consumption as propulsive muscle does less work. In addition, it is likely that there is a reduced blood flow to muscle as it relies more on the internal store of oxygen rather than that from the circulating blood, which is the main oxygen source for the hypoxic-intolerant brain. The high concentration of myoglobin in all marine mammals indicates that it is a key adaptation for diving. As the muscle oxygen depletes, the need for supplemental energy from anaerobic catalysis of creatine phosphate and glycogen rises. Both of these compounds produce adenosine triphosphate (ATP) that is essential for electron transfer, and the conversion of chemical energy into mechanical work. However, anaerobic glycolysis results in an inefficient use of the glycogen store because of incomplete combustion of glucose to lactate which results in a metabolic acidosis that has a limit of tolerance. Diving mammals have a broad tolerance of acidosis because of their exceptional capacity to buffer the acidity of lactate. Most divers avoid this condition, and reliance on anaerobic metabolism occurs only in exceptional cases when the dive has to be extended beyond the routine foraging durations. This threshold is when a net production of lactate results in a rise in arterial blood lactate after the dive. This has been called the aerobic diving limit (ADL). When dive durations are plotted against lactate concentration in arterial blood (Figure 2) there is a distinct inflection where lactate concentration increases sharply. Beyond this dive duration anaerobic metabolism contributes a greater amount to the energy needs of the animal, mostly in the muscle. There is a cost in addition to the inefficient use of glycogen stores. An imbalance occurs in metabolic endproducts, and there is acidification in the cells and the circulatory system which must be restored to normal acid–base balance. To do this there may be an extended surface period for recovery, or if the recovery is continued during the next dive, it is likely to be shorter because of the reduced oxygen and glycogen stores which cannot return to full capacity until the acid–base balance of the body returns to normal levels. In order to avoid the problem inherent in relying on extensive anaerobic metabolism during a dive, most dives of marine mammals are within the

585

20 15 10 5 0 _5 0

10

20

30

40

50

60

70

Dive duration (min)

Figure 2 Peak concentration of lactate in arterial blood after dives of different duration in adult Weddell seals. The inflection represents the transitions from completely aerobic dives and is considered the aerobic diving limit (ADL). Open circles reflect blood values of dives in which there is no net production of lactate, and black circles are those in which there was a net production. (Reproduced with permission from Kooyman (2000).)

ADL. Circumstances during which the ADL might be exceeded occur during foraging when a rich source of prey might be at such a depth that the dive has to be longer than the ADL, or during avoidance of a predator when diving unusually deep might aid in a successful escape. For the Weddell seal it might be to travel an exceptionally long distance under ice to a new breathing hole.

Adaptations to Pressure Once marine mammals developed the capacity to breath hold for a few minutes they were exposed to the second most dominating physical effect of the marine environment – the density of water. Just a short distance from the surface, the hydrostatic pressure becomes overwhelming for any terrestrial animal which lacks the adaptations found in marine mammals. In order to tolerate the effects of pressure, marine mammals have adapted to live with it rather than to resist it as human made submersibles do. Instead of an outer shell that protects the internal organs from the crushing pressure of depth, marine mammals give and absorb the pressure. The chest is almost infinitely compliant allowing the lungs to be compressed to a solid organ. There are no air sinuses in the skull such as the facial sinuses of humans, and the vascular lining of the middle ear can expand and reduce in volume to match the ambient pressure. There are no disparities of pressure between the inside and outside of the body. This has many

(c) 2011 Elsevier Inc. All Rights Reserved.

586

MARINE MAMMAL DIVING PHYSIOLOGY

ramifications, but the most important relates to how diving mammals avoid absorbing the lung gases that comprise 80% nitrogen while diving to great depths. Such a volume is adequate, depending on the distribution of blood, to cause blood and tissue nitrogen tensions to reach a level where gas bubbles could form during the animal’s rapid ascent. The result is decompression sickness or the ‘bends’. By yielding to the compressive effects of pressure marine mammals avoid the problem. The lungs of marine mammals are similar in all groups, but distinctive from terrestrial mammals (Figure 3). As a result of robust cartilaginous support that is absent from terrestrial mammals the most peripheral airways are less compliant. As the chest wall compresses, it allows the lungs to collapse; this takes place in a graded fashion with the gas from the alveoli being forced into the upper airways where gas exchange does not occur (Figure 4). Consequently, the gas is sequestered while the animal is at depth. For some seals this collapse may occur at depths of only 20 m. No matter how much deeper the dive is, the arterial blood nitrogen tension does not rise above 2300 mmHg, within the range where gas bubbles will not form even if the decompression rate is extremely rapid. Curiously, the most robust peripheral airways are not always in the deepest divers.

Furthermore, in some of these species the airways seem to be armored to a far greater degree than is necessary to ensure a graded collapse as they descend to depth.

Adaptations for Ventilation Because of the compliant chest wall and the robust peripheral airways the lungs of most marine mammals empty to an unusually low volume, about 5–10% of total lung capacity (TLC), compared to the 25% residual volume of terrestrial mammals. This makes possible a vital capacity of about 90% of TLC, which is often equivalent to tidal volume in species that rapidly ventilate such as dolphins. For a fast-swimming dolphin or sea lion to make effective use of such a large tidal volume it needs to be turned over rapidly during the brief time the blow hole or nostrils are near the surface (Figure 5). The bottlenose dolphin can turnover about 90% of its TLC within 0.7 s. Indeed, these dolphins and some whales aid this process by exhaling most of the tidal volume just before breaking the surface, so that most of the time at the surface can be used for inhaling. Inhalation is slower because although the chest wall is actively expanded, lung expansion is a passive

Phocidae Human

Cartilage Alveoli

Alveolar duct 0

Alveolar sac

Muscle 0

1 mm

1 mm Alveolar sac

Delphinidae Otariidae Cartilage

Sphincter muscle Muscle

0

1 mm

0

1 mm

Figure 3 Diagrams of the terminal airways and alveoli of a human and three major groups of marine mammals. The human alveolar duct is thin-walled and exchanges gas with the capillaries. The seal (Phocidae) has a short alveolar duct if present at all and the terminal bronchiole is reinforced with cartilage and muscle. The sea lion (Otariidae) has no gas exchanging surfaces except within the alveolar sac, and the cartilage is robust throughout the terminal bronchiole. The dolphin (Delphinidae) has similar robust cartilage reinforcement within the terminal bronchiole, but in addition there is a series of sphincter muscles. This muscle configuration is unique to the toothed whales. (Modified from Denison DM and Kooyman GL (1973) The structure and function of the small airways in pinniped and sea otter lungs. Respir. Physiol. 17: 1–10.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL DIVING PHYSIOLOGY

Time (min)

10

Bronchiole

8

Alveolus

Flow (VC / s)

Tursiops Capillary

Depth (m)

Phocoena

>1

0

0

587

50

6

4

2

Human

0 100

80

60

40

20

0

Volume (%VC)

100

Absorption collapse

Figure 4 Diagram of the graded collapse of the alveolar gas into the upper airways of a seal as it descends to depth. It is presumed that the same type of collapse occurs in other kinds of marine mammals. The time axis is relative, but emphasizes the short period over which this collapse occurs as the seal descends to depth. The arrows within the alveolus and across the alveolar membrane indicate the direction of gas flow. The curved arrow indicates the further closure of the alveolar space as absorption collapse occurs. The thickness of the arrow’s stem indicates the rate. The deeper the seal descends the faster the uptake of gas into the capillaries. Thus, the staircase configuration of the collapsed alveoli showing that the collapse will occur sooner the deeper the animal descends. (Modified from Kooyman GL, Schroeder JP, Denison DM et al. (1972) Blood N2 tensions of seals during simulated deep dives. American Journal of Physiology 223: 1016–1020.)

process that relies on the elastic properties of the lung for inflation.

Figure 5 The rapid expiration of gas from the excised lungs of a porpoise (Phocoena phocoena) during a forced expiration. The forced expiration curve from a trained bottlenose dolphin (Tursiops truncatus). The maximum rate of expiration of a human is shown for comparative purposes. (Modified from Kooyman and Sinnett (1979) Mechanical properties of the harbor porpoise lung. Respiratory Physiology 36: 287–300; Kooyman and Cornell (1981) Flow properties of expiration and inspiration in a trained bottlenose porpoise. Physiological Zoology 54: 55–61.)

has a fineness ratio, i.e., body length divided by the maximum body diameter, of about 4.5. This shape greatly reduces form drag, which at the speeds marine mammals swim, is over 90% of the total drag. Also the routine swim speed of all marine mammals is about 2–3 m s 1. At these speeds and through such a dense medium the rates of progress are slower than that of many land mammals, but the burst and glide swimming that are used and the support to the body mass makes their progress seem almost effortless, and accounts for the low metabolic rates discussed earlier.

Adaptations to Light Extremes Adaptations for Locomotion The similarity in shape of different marine mammals, from seals to whales, can be seen in Figure 1. There are strict rules for this, which are related to density and viscosity of water (see Fish Locomotion). The density of water is about 1000 times greater than air, and at 01C the kinematic viscosity is about seven times that of air. In order to minimize resistance to movement, marine mammals like many fish had to evolve a special shape of the body. The ideal shape

Marine mammals are creatures of the night. Even those that may be active during the day search for most of their food at depths where the light level is similar to twilight or less. Hence, they might be thought of as marine bats. Some echolocate to find their prey, whereas others hunt visually. Seals and sea lions are visual hunters and like many nocturnal mammals they have large eyes. The eye of the southern elephant seal has an internal anterior and horizontal diameter of 52 by 61 mm, compared to

(c) 2011 Elsevier Inc. All Rights Reserved.

588

MARINE MAMMAL DIVING PHYSIOLOGY

the human eye of 24 by 24 mm. These eyes have a high concentration of rods, the light-sensitive element within the retina, that is 10–50 times more dense than the human eye. At least some have visual pigments adjusted to the blue light that is the dominant ambient light source at depth. All have a tapetum lucidum layer behind the retinal layer and a large pupillary area. In comparison to humans, a diurnal primate whose pupil area ranges from 3.2 to 50 mm2, or the domestic cat, whose pupil ranges from 0.9 to 123 mm2, the northern elephant seal pupil area ranges from 0.9 to 422 mm2. Less extreme is the California sea lion, which dives to much shallower depths, and whose pupillary dimensions are only 8.4–220 mm2. At light levels where the sea lion pupil is expanded maximally, which is roughly equivalent to a clear, full-moon night, that of the elephant seal is only at 22% of its maximum diameter. Indeed the maximum pupillary expansion of the elephant seal does not occur until there is total darkness. At this level the eyes are still functional because although no light is produced by the physical environment, biological light is produced by most marine organisms. Hence, when elephant seals descend to depths beyond the limit of surface light there is still much light in the environment. As marine mammals commute with great rapidity to and from the depths it is essential that there is rapid adaptation to the dark. Whereas land mammals may take tens of minutes to adapt to darkness, elephant seals may do so within 4–6 min. The shallower-diving harbor seal takes 18 min and humans about 22 min. Elephant seals may be helped in this adaptation because the contracted pupil is so small that even at the surface the amount of light reaching the retina is not great enough to saturate the rod receptors.

Conclusions The management of oxygen stores has been a long standing topic of study in the history of marine mammal physiology since the first key works were published in the 1930s and 1940s. Progress in understanding the adaptations of marine mammals to conserve and utilize those stores efficiently has been fitful, depending on the techniques available. The research emphasis has shifted from restrained

and forced submersions in the laboratory to experiments on free-ranging animals. New and powerful tools are becoming available that will enhance progress in this field. In other areas such as determining the function of some of the vascular retia of whales, or the function of some of the peculiar lung structure in these animals, progress has ceased despite the fact that some of this poorly understood anatomy appears to play a major role in the adaptation of these animals to the marine environment. The lack of investigations relates to unavailability of the animals, and the intrusive measures that would be required to determine function.

See also Bioacoustics. Fish Locomotion. Fish Vision.

Further Reading Butler PJ and Jones DR (1997) Physiology of diving of birds and mammals. Physiological Review 77(3): 837--899. Kooyman GL (1989) Diverse Divers: Physiology and Behavior. Zoophysiology Series, vol. 23. New York: Springer-Verlag. Kooyman GL (2000) Diving physiology. In: Perrin WF, Wursig B, and Thewissen JGM (eds.) Encyclopedia of Marine Mammals. San Diego: Academic Press. Kooyman GL and Ponganis PJ (1997) The challenges of diving to depth. American Scientist 85: 530--539. LeBoeuf BJ and Laws RM (1994) Elephant Seals: Population Ecology, Behavior and Physiology. Berkeley and Los Angeles: University of California Press. Levenson DH and Schusterman RJ (1999) Dark adaptation and visual sensitivity in shallow and deep-diving pinnipeds. Marine Mammal Science 15: 1303--1313. Shadwick RE (1998) Elasticity in arteries. American Scientist 86: 535--541. Watkins WA, Daher MA, Fristrup KM, and Howald TJ (1993) Sperm whales tagged with transponders and tracked underwater by sonar. Marine Mammal Science 9: 55--67. Williams TM, Davis RW, Fuiman LA, et al. (2000) Sink or swim: strategies for cost-efficient diving by marine mammals. Science 288: 133--136.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL EVOLUTION AND TAXONOMY J. E. Heyning, The Natural History Museum of Los Angeles County, Los Angeles, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1596–1602, & 2001, Elsevier Ltd.

Since Eocene times some 57 million years ago, several groups of land mammals have independently evolved adaptations for a marine existence. With just over 120 modern species (Table 1), marine mammals are not taxonomically diverse compared to other groups of marine organisms, yet they are important components of many ecosystems. Marine mammals have evolved from only two terrestrial groups, or clades, of mammals. The first is the Order Carnivora, which includes such familiar creatures as cats, dogs, and bears. From this order arose the pinnipeds (seals, sea lions, walruses), the sea otter, and the polar bear. The other clade is the Ungulata, a group that includes all the orders of modern hoofed animals. Evolving from this rank are the cetaceans (whales, dolphins, and porpoises), the sirenians (dugongs and manatees), and the extinct hippo-like desmostylians. Table 1

Marine mammals Common name

Number of living species

Order Cetacea Suborder Mysticeti Family Balaenidae Family Neobalaenidae Family Eschrichtiidae Family Balaenopteridae Suborder Odontoceti Family Physeteridae Family Kogiidae Family Ziphiidae Family Platanistidae Family Iniidae Family Monodontidae Family Phocoenidae Family Delphinidae

Baleen whales Right whales Pygmy right whale Gray whale Rorqual whales Toothed whales Sperm whale Pygmy sperm whales Beaked whales Indian river dolphins River dolphins Beluga and narwhal Porpoises Oceanic dolphins

1 2 20 1 3 2 6 36

Order Carnivora Family Phocidae Family Otariidae Family Odobenidae

Carnivores Seals Sea lions and fur seals Walruses

19 16 1

Order Sirenia Family Dugongidae Family Trichechidae

Sea cows Dugongs Manatees

4 1 1 8

1 3

Systematics is the science of defining evolutionary relationships among organisms. A phylogeny is a hypothesis of those evolutionary relationships, and is the foundation for any evolutionary study. It is the distribution of characters on a phylogenetic tree that is used to evaluate whether similar features in two organisms are a result of inheritance from a common ancestor or evolved via convergence. No proposed phylogeny can be proven, as proof would require the unattainable knowledge of past evolutionary events. Hence, researchers today can only infer past events from phylogenetic reconstructions of evolutionary relationships. Most modern systematists use a philosophical approach called cladistics. The basic tenets of cladistics are quite simple: organisms are deemed to be related based on shared derived characters called synapomorphies. Derived characters are defined as having arisen in the common ancestor of the taxa and are subsequently passed on to their descendant taxa. Groups of related taxa and their descendants are called clades regardless of taxonomic rank. Monophyletic groups are those that include the ancestor and all of its descendants. Paraphyletic groups are those which include the ancestor taxa, but not all the descendants. Paraphyletic groups often reflect a certain level of morphological organization or ‘grade,’ and exclude some or all of the more derived descendants. For example, in classical cetacean taxonomy, the suborder Archaeoceti is unquestionably paraphyletic, as this group of fossil cetaceans includes the most basal species, but not the descendant suborders Mysticeti and Odontoceti.

Order Carnivora Several groups of carnivores have evolved to a partially aquatic existence, although all must return to land, at least to give birth. The polar bear, Ursus maritimus, appears to be a very recent evolutionary divergence of the brown bear, Ursus arctos, lineage. Analysis of molecular data suggests that polar bears are most similar to those brown bears from the islands off south-east Alaska. Fossils suggest that the two species diverged in the middle Pleistocene. There are three species of marine carnivores in the weasel/otter family (Mustelidae). Along the northeast coast of North America, the extinct sea mink, Mustela macrodon, has been found primarily from Native American midden sites. The second is the chungungo or marine otter, Lutra felina, a poorly

(c) 2011 Elsevier Inc. All Rights Reserved.

589

590

MARINE MAMMAL EVOLUTION AND TAXONOMY

known species inhabiting South American waters. Lastly, is the well-known sea otter, Enhydra lutris, of the temperate and subarctic North Pacific. These two otters are closely related to the fresh water otters. The sea otter appears to be related to the late Miocene to early Pliocene Enhydritherium, found along both coasts of North America and Europe. Pinnipeds

The pinnipeds (from the Latin meaning ‘fin-footed’) are a group of the marine mammals, which includes the seals, sea lions, and walrus (Figure 1). Pinnipeds arose from the arctoid (bear, dogs, weasels, etc) lineage of carnivores. The pinnipeds consist of three living families, the Phocidae (true seals), the Otariidae (fur seals and sea lions), and the Odobenidae (walruses). The terrestrial ancestor of the pinnipeds has been the subject of considerable debate. Two schools of thought exist. One, citing primarily biogeographical, and paleontological evidence, supports a diphyletic or dual origin, attributing the sea lions, fur seals, and walruses to an ursid (bear-like) ancestor evolving in the North Pacific and the true seals to a mustelid (weasels, otters, etc.) ancestor evolving in the North Atlantic. The other school, using molecular,

karyological, and morphological evidence supports a monophyletic origin for all three pinniped families stemming from an unresolved arctoid ancestor. There is growing consensus that pinnipeds constitute a monophyletic group. The extant true seals, Phocidae, are distinguished from the sea lions by the lack of external ear flaps, relatively short forelimbs with claws, backward directed hindlimbs that do not permit quadrupedal movement on land, and locomotion in water by sculling the hindlimbs. They are divided into two groups, the Monachinae (‘southern’ seals) and the Phocinae (‘northern’ seals). The oldest phocid fossils are from the late Oligocene of the North Atlantic. The sea lions and fur seals (or ‘eared’ seals), Otariidae, can be distinguished from the phocids by the presence of external ear flaps, elongate flipperlike forelimbs, rotatable hindlimbs that allow quadrupedal locomotion on land, and locomotion in water by flapping of the fore-flippers. They are divided into two groups, the Arctocephalinae (fur seals) and the Otariinae (sea lions). The diagnostic feature is the presence of abundant underhair in the fur seals. The oldest fossil otariids are from the early Miocene of the North Pacific. The walrus family includes mostly extinct species consisting of three groups. The most basal are

Figure 1 Skeleton of a modern southern sea lion (Otaria flavescens). Photo courtesy of the Natural History Museum of Los Angeles County.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL EVOLUTION AND TAXONOMY

archaic nontusked walruses referred to as imagotarines retaining the ancestral fish-eating diet of other pinnipeds whereas the Dusginathinae are an extinct group apparently specializing in squirt/suction feeding behavior for foraging on shellfish and/or crustaceans. Dusignathines have enlarged canines in both jaws providing a four-tusked image to the skull. These two groups are distinguished from the Odobeniinae that only develop tusks in the upper jaw and eventually gave rise to the modern walrus, Odobenus rosmarus, which feeds primarily by sucking bivalves out of their shells.

Order Cetacea Until quite recently it was commonly asserted that whales were fish, and not mammals. Today, the cetaceans (whales, dolphins, and porpoises) are readily recognized as mammals. Their relationship to the ungulates (hoofed terrestrial mammals including horses, cows, pigs, camels, elephants, etc.) is generally accepted, though their closest relation among the living ungulates remains a topic of research. The Order Cetacea consists of three suborders: the Archaeoceti (an extinct group of archaic whales), the Odontoceti (toothed whales), and the Mysticeti (baleen whales). The hypothesis that the order Cetacea is monophyletic is supported by an overwhelming amount of morphological, cytological, and molecular evidence.

591

The karyotypes of cetaceans are amazingly conservative when compared to other groups of mammals. The typical chromosome count for most cetaceans is 2N ¼ 44. The exceptions are sperm whales (Physeteridae and Kogiidae), beaked whales (Ziphiidae), and right whales (Balaenidae). For right whales and beaked whales, the lower counts (2N ¼ 42) are a result of the fusion of different chromosome pairs, respectively. The chromosomebanding pattern among cetaceans is also astonishingly conservative. In modern cetaceans (Figure 3), the body shape is fish-like – streamlined with fin-like flippers and flukes. The hind limbs are but vestiges tucked within the body wall, and the nostrils are situated high on the head and termed blowholes. It is because cetaceans differ so strikingly from their terrestrial kin that it is difficult to discern intuitively which, among the other orders of ungulate mammals, are their closest kin. There is now convincing fossil evidence that landdwelling extinct mesonychians are closely related to the ancestor of whales; either they are the sistertaxon to the whale ancestor, or whales arose from within the paraphyletic extinct family Mesonychidae (Figure 2). There are many similarities between whales and at least some mesonychids in such features as the construction of the cheekteeth, the humerus, and the venous drainage of the skull. An analysis of the postcranial skeleton of the

Gray whales Roqual whales

Pygmy right whales

BALEEN WHALES

Right whales

Mesonychid Archaeocete (ancestor of whales) (early whale)

Sperm whales

TOOTHED WHALES

Ganges river dolphins River dolphins Dolphins Porpoises White whales Beaked whales

Figure 2 Phylogeny of the modern families of cetaceans based on morphological and molecular data (modified from Heyning, 1995).

(c) 2011 Elsevier Inc. All Rights Reserved.

592

MARINE MAMMAL EVOLUTION AND TAXONOMY

Figure 3 Skeleton of a modern baleen whale (Balaena mysticetus). (From Van Beneden and Gervais, 1880.)

mesonychid Pachyaena suggests that this animal may have had a body form similar to tapirs, capybaras, or suids (pigs), many of which are excellent swimmers. Some molecular studies indicate that the Cetacea may have actually evolved from within the order Artiodactyla. Suborder Archaeoceti

Archaeocetes represent a primitive grade-level taxon that includes all cetaceans that lack the derived telescoped pattern of skull bones found in either mysticetes (baleen whales) or odontocetes (toothed whales). Archaeocetes are further characterized by an elongate snout, a narrow braincase, a skull with large temporal fossa and well-defined sagittal and lambdoidal crests, a broad supraorbital process of the frontal bone, and bony nares situated some distance posterior to tip of the snout. The archaeocete grade was ‘extinct’ by the end of the Eocene. The archaeocetes include five families: Pakicetidae, Ambulocetidae, Protocetidae, Remingtonocetidae, and Basilosaurocetidae. Most Pakicetids have been unearthed from terrestrial, freshwater, or nearshore deposits of the eastern Tethys, a massive epicontinental sea that divided Eurasia from Africa/India. The oldest and most primitive whale, Pakicetus inachus was described from Pakistan. Many of these fossils from the eastern Tethys Sea, some with known hind limbs, are significant in that they serve as morphological transitions between land-dwelling mammals and fully aquatic cetaceans. The Ambulocetidae occupied tidal and estuary habitats. The enigmatic Ambulocetus natans differs from all other known archaeocetes in that its eyes are elevated above the overall profile of the skull, and its hind feet are elongate imparting a crocodile-like appearance. These hindlimbs were probably important for aquatic locomotion. The Protocetidae is defined by details of the orbits, teeth, and sacrum. In the eastern Tethys, protocetids are

found in nearshore waters, whereas protocetids from the western Tethys died in shallow offshore regions and their North American kin have been found in shallow nearshore deposits. Remingtonocetidae have extremely long and narrow skulls with small eyes and are known from rock units from Pakistan and India. These archaeocetes are found in sediments indicating a coastal environment or nearshore shallow marine deposits. Though they are interpreted to be the most derived of the archaeocete families, the Basilosauridae were among the first archaeocete fossils to be discovered. Basilosaurids are typically divided into two subfamilies: the Durodontinae with unspecialized vertebrae and the Basilosauinae with extremely elongate vertebral bodies. It has been suggested that durodontines were ancestral to both mysticetes and odontocetes. However, further evidence suggests that the durodontines represent the sister taxa to the modern Cetacea clades. One of the most dramatic morphological changes of early cetaceans was the shift from quadruped locomotion on land to the axial undulation of swimming in the ocean. These include changes of the vertebral column, limb structure, and of the tail flukes. One striking conclusion is how quickly these transitions occurred. At the dusk of the Early Eocene, Pakicetus and its contemporaries were quadruped animals that drank fresh water. A few million years later in the Middle Eocene, Rodhocetus and its kin were swimming with the aid of tail flukes and drinking sea water. By the Late Eocene, Basilosaurines possessed such exceedingly small hind limbs that it is unlikely that they were of much use in terrestrial locomotion. Hence, one can speculate that by the Late Eocene cetaceans had severed all links to the land. One hypothesis is that archaeocetes first evolved in fluvial or estuarine environments of the eastern Tethys and subsequently dispersed as more morphologically and physiologically derived forms conquered the oceans. All of the most primitive and chronologically oldest fossil archaeocetes are found

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL EVOLUTION AND TAXONOMY

along the shores of the eastern Tethys, whereas the more morphologically derived and fully marine protocetids and basilosaurids of the Middle and Late Eocene are found in rocks from Asia, North Africa, North America, New Zealand, and Antarctica. Suborder Odontoceti

Some cetaceans possess teeth (odontocetes), whereas others have baleen for filtering out prey (mysticetes). There are numerous other synapomorphies that unite the Odontoceti such as the maxilla (upper jaw) which has telescoped back over the frontal bone of the skull. All living and all but the most basal fossil odontocetes have asymmetrical skulls, and all modern species have asymmetrical facial soft anatomy. This asymmetry results from an enlargement of the facial soft anatomy on the right side. All odontocetes have a complex series of air sacs off their nasal passages. All extant odontocetes possess an enlarged fatty melon in front of the nasal passages, which is distinctly different from the diminutive melon-like structure observed in mysticetes. Odontocetes are unique among all tetrapods in that the distal narial passages coalesce to form a single nostril or blowhole. The large melon, complex nasal anatomy, and asymmetry all appear to be correlated with the ability to echolocate. The sperm whales are represented today by the giant sperm whale (Physeteridae) and the dwarf sperm whales (Kogiidae). This clade is the first to diverge from the lineage of living odontocetes. All sperm whales are recognized by the presence of a spermaceti organ in their head and very asymmetrical skulls. All living species are known or suspected to be deep divers. The beaked whales (Ziphiidae) are a very diverse group with at least 20 known living species. This group is typified by extreme sexual dimorphism in that males of most species have a one or two enlarged pairs of teeth used primarily for fighting other males, presumably for establishing breeding dominance. The river dolphin (genus Platanista) of the Indian subcontinent is the sole living representative of the family Platanistidae. The bone pattern of the palate and the elaborate crests of bone on top of the skull define this family. The other living river dolphins and their kin belong to the family Iniidae, with living representatives found in the waters of the Yanzee and Amazon river basins, and coastal waters of eastern South America. This clade is defined by the extreme asymmetry in the nasal sacs. The remaining odontocetes are the closely related families of the Monodontidae (narwhal and beluga), Phocoenidae (porpoises) and the taxonomically diverse oceanic dolphins (Delphinidae).

593

This clade represents the vast majority of living species, and includes those most familiar to most people. Suborder Mysticeti

The mysticetes, with their edentulous mouths lined with filtering baleen, are one of the distinct clades among the mammals. The evolutionary transition from capturing single prey to filtering numerous prey items out of mouthfuls of sea water has ramifications not only in the morphology, but also the behavioral ecology of these, the largest of all animals. Mysticetes evolved from cetaceans that possessed teeth. Certain cranial features predate the loss of teeth in the mysticete clade and, therefore, the most basal mysticetes retain teeth. The oldest known mysticete is the toothed species Llanocetus denticrenatus from the Late Eocene of Seymour Island, Antarctica. The next oldest specimens are those of the Oligocene cetotheres whose wide, edentulous palates strongly imply the presence of baleen. As baleen is made of the protein keratin, it typically decomposes with the other soft tissues rarely leaving a fossil trace. These fossil discoveries now represent a moderately good morphological series from the archaeocetes to modern mysticetes. Three families of extinct mysticetes are recognized. They are the Llancetidae, the Aetiocetidae, and the Cetotheridae. The family Llanocetidae is based on one species, Llanocetus denticrenatus, with an estimated total length of perhaps 10 m. Aetiocetids are relatively small toothed mysticetes known only from the shorelines of the North Pacific. Chonecetus and some species of Aetiocetus retain the primitive eutherian mammal tooth count. Aetiocetus polydentatus with its expanded toothcount exhibits an incipient stage of the derived feature of supernumerous teeth as seen in later cetaceans. The Cetotheriidae represent a phylogenetically heterogeneous assemblage that is truly a ‘wastebasket’ taxon with over 60 named species within 30 or so genera of unknown affinities. The rostrum is typically broad and flat, not dissimilar to that found in the primitive aetiocetids and modern balaenopterids. The oldest cetotheres are Cetotheriopsis lintianus from Austria and the relatively complete skull of Mauicetus lophocephalus from New Zealand, both from the Late Oligocene. The modern mysticetes are divided into four families: the Balaenidae, Neobalaenidae, Eschrichtiidae, and Balaenopteridae. The systematics of these families are much less contentious than that for the modern odontocetes, though a few taxa remain elusive with regard to their relationship within and among other modern mysticete groups.

(c) 2011 Elsevier Inc. All Rights Reserved.

594

MARINE MAMMAL EVOLUTION AND TAXONOMY

The right whales (Balaenidae) are heavy-bodied mysticetes with large arched heads and cavernous mouths to accommodate the extremely large filtering surface formed by the extraordinarily long baleen plates. Balaenids lack throat grooves. There are two species of modern balaenids, the bowhead and the right whale. Balaenids have skulls that are narrow and highly arched. The baleen is long, narrow, with a fine fringe. The dorsal fin is absent. The oldest fossil of an extant mysticete family is the balaenid Morenocetus parvus from the earliest Miocene of Argentina. The family Neobalaenidae is represented solely by the poorly known Southern Hemisphere pygmy right whale, Caperea marginata. Some workers have considered Caperea to be a balaenid. However, there is ever-growing consensus, based on morphology and molecular data, that Caperea is not an ‘aberrant’ right whale, but is instead more likely the sister group to the rorqual/gray whale clade. The dorsal fin is small, yet distinctive. The throat grooves of Caperea are highly variable in depth, and virtually absent in some individuals. The rostrum is only somewhat arched, intermediate between the conditions seen in gray whales (Eschrichtiidae) and right whales (Balaenidae). The ribs are unique among cetaceans, living or fossil, in that they are broad and overlap each other in profile. The enigmatic gray whale, Eschrichtus robustus, is the sole member of the family Eschrichtiidae. The overall mottled gray color and small dorsal fin followed by a series of dorsal ridges characterize gray whales. The two to five throat grooves are well delineated and are confined to the gular region. The rostrum is attenuate and moderately arched. The yellowish to white baleen is relatively short and moderately wide. Gray whales have lost a digit in their flipper. The only fossil gray whale is a Late Pleistocene individual of superb preservation. However, this animal is indistinguishable from the modern species and its relative young age does not help to resolve the relationship between gray whales and other baleen whales. The Balaenopteridae, also known as the rorquals, are immediately recognized by their numerous throat grooves. These distinctive throat grooves are numerous, ranging from 14–22 grooves in the humpback whale (Megaptera novaengliae) to 56–100 grooves in the fin whale (Balaenoptera physalus). Balaenopterids are the ‘greyhounds of the sea.’ Their bodies are the sleekest among living mysticetes. The dorsal fin is always present and tends to be inversely proportional to body size. The rostrum is extremely broad and flat. Rorqual whales have a very complex interdigitating pattern of bony sutures between the

rostral bones and those of the braincase proper. Only four digits are present in the flippers. The humpback is unique among balaenopterids in that it has extremely elongate flippers. There are three major cranial character suites that have been used to ascertain the phylogenetic position of fossil baleen-bearing whales. These are the position and shape of the occipital shield, the complexity of interdigitation between the bones of the rostrum (nasals, premaxillae, and maxillae) and the braincase proper, and slope of the supraorbital process of the frontal bone. Ancestrally, the occipital shield does not extend very far anteriorly providing dorsal midline exposure of the parietal and frontal bones. The most primitive character state of this feature is seen on the various toothed mysticetes. In the most advanced character state, the occipital is close to the nasals and premaxillae on the vertex. This condition is found in modern balaenopterids, neobalaenid, and also in some cetotheres. The complex interdigitation of the bony sutures is clearly derived. Incipient interdigitation of the rostrum and braincase is seen in cetotheres. Modern balaenopterids uniquely possess a supraorbital process of the frontal that is flat and horizontal until it reaches the braincase and then abruptly turns dorsally to the skull vertex. The result is a large region over the supraorbital process for the greatly enlarged temporalis muscle required to close the mouth after engulfing tons of water. The cetothere Cetotherium has distinct crests on the temporal ridge along the contact with the frontals which suggests a condition that foreshadows the state found in modern balaenopterids. Although differing somewhat in detail, most morphologically based phylogenies of baleen whales as well as molecular-based studies suggest that the balaenids were the first clade to diverge, followed by the Neobalaenidae, then the Eschrichtiidae and Balaenopteridae as sister taxa. Several studies using molecular sequence data have implied that the ancestry of the gray whale (Eschrichtiidae) is located near or within the genus Balaenoptera (Balaenopteridae).

Order Sirenia The sirenians, manatees, dugongs, and their extinct relatives, are a fully aquatic herbivorous group of mammals. The living species are restricted to tropical and subtropical waters; however, fossil species appear to have inhabited temperate waters and the range of the recently extinct Steller’s sea cow extended into Arctic waters. Modern dugongs feed primarily on seagrasses, whereas the Steller’s sea cow fed on brown algae, and manatees feed on a variety of freshwater aquatic plants. Sirenians are part of the

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL EVOLUTION AND TAXONOMY

595

Figure 4 Skeleton of a modern manatee (Trichechus manatus). Photo courtesy of the Natural History Museum of Los Angeles County.

ungulate clade called Tethytheria, which also includes the elephants and their extinct relatives, the extinct marine desmostylians, and in some classifications, the hyraxes. All living groups have a unique form of tooth replacement in which new teeth originate at the back of the toothrow, then slowly move forward, and finally the worn down tooth drops out of the front. The oldest and most primitive sirenians are found in Eocene rocks; Prorastomus from the Early and Middle Eocene of Jamaica and Protosiren from the Middle Eocene of Pakistan, North Africa, and Europe. Ancient sirenians had four limbs and were amphibious, but as with the cetaceans, modern species have but vestiges of the pelvic girdle and are propelled entirely by tail flukes (Figure 4). There are two living groups of sirenians: the manatees (family Trichechidae) and the dugong (Dugongidae). The fossil record is rich with taxa, which is not the case for the living species. There are but two species of recent dugongids: the dugongs, Dugong dugon, of the Indo-Pacific and the recently extinct Steller’s sea cow, Hydrodamalis gigas. Modern Dugongids have pointed flukes similar to cetaceans and are exclusively marine. The upper incisors are tusk-like in males. The Steller’s sea cow was hunted to extinction in its last refugia of Commander Islands in 1768, but it occurred recently (19 000 years ago) as far south as Monterey, California. Dugongids were widespread in Miocene times and apparently spread from the Atlantic into the Pacific prior to the formation of the Isthmus of Panama. Dugongids subsequently became extinct in the Atlantic. There are three closely related species of

manatees all distributed in the Atlantic: the West Indian manatee, Trichechus manatus; the African manatee, T. senegalensis; and the exclusively freshwater Amazon manatee, T. inunguis. Modern manatees have rounded tail flukes and are primarily found in fresh water although the West Indian manatees are not uncommonly found in marine waters. The fossil record for manatees is far sparser than that of dugong.

See also Baleen Whales. Marine Mammals: Sperm Whales and Beaked Whales. Sea Otters. Seals. Sirenians.

Further Reading Berta A and Deme´re´ T (eds.) (1994) Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History. Fordyce RE and Barnes LG (1994) The evolutionary history of whales and dolphins. Annual Review of Earth and Planetary Sciences 22: 419--455. Heyning JE (1997) Sperm whale phylogeny revisited: analysis of the morphological evidence. Marine Mammal Science 13: 596--613. Rice DW (1998) Marine Mammals of the World: Systematics and Distribution. The Society for Marine Mammalogy. Special Publication No. 4. Thewissen JGM (ed.) (1998) The Emergence of Whales: Evolutionary Patterns in the Origin of Cetacea. New York: Plenum Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS P. J. Corkeron, James Cook University, Townsville, Australia S. M. Van Parijs, Norwegian Polar Institute, Tromsø, Norway Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1603–1611, & 2001, Elsevier Ltd.

Introduction Marine mammals are renowned as great travelers. The migrations of some whales and seals span ocean basins, but other species have home ranges limited to a few square kilometers. Factors that influence marine mammals’ movements strongly include how they give birth and mate, and how their food is distributed in space and time. These differ substantially across the marine mammal groups, and so they are dealt with separately in this chapter. Marine and terrestrial ecosystems differ substantially as environments for air-breathing homeotherms. The density of sea water leads to far lower energetic costs of locomotion for animals living in marine systems. Living in a dense medium also allows mammals to grow to larger sizes than is possible on land, and the energetic cost of travel decreases exponentially with an animal’s size. These factors, coupled with the connectivity of oceans, mean that marine mammals generally have far larger ranges, or longer migratory routes, than terrestrial mammals. However, marine mammals’ need for gaseous oxygen means that, unlike most marine animals, they must return regularly to the air–water interface to breathe. Also unlike most marine vertebrates, mammals give birth to live young. These young need to breathe immediately after birth, so they must either be remarkably good swimmers at birth or be born on land. Cetaceans, sea otters, and sirenians give birth at sea, but pinnipeds and polar bears give birth on land. This leads to major differences in their movements. Locomotion represents an energetic cost to animals, so all individual animals are selected to balance the cost of travel against the benefits gained by travel. Unlike air, which is distributed evenly at the water surface, the food resources of marine mammals are found in patches of differing scales, varying in both space and time. Other marine species are

596

distributed through the oceans’ depths, so marine mammals have a third dimension available for their travel. Changes in the density of sea water with depth mean that marine mammals may expend little energy when descending in deep water, but must display physiological adaptations to deal with the extreme pressure associated with these deep dives. Animals’ movements can be considered on several temporal and spatial scales. Migrations, generally on an annual cycle, involve persistent movement between two destinations. An animal’s home range is that area within which it carries out most of its normal activities throughout the year. Classically, home ranges are not considered to include migratory movements.

Cetaceans Two factors appear to be responsible for substantial differences between the movement patterns of mysticetes and odontocetes. Most mysticetes feed in polar or cold temperate waters, highly seasonal environments. Therefore, mysticetes’ prey are more heavily based on an annual cycle than the prey of most odontocetes. Also, although all cetaceans give birth aquatically, the location for giving birth appears to be particularly important to mysticetes.

Mysticetes The annual cycle of most baleen whales is characterized by migrations between polar or cold temperate summering grounds and warm temperate, subtropical or tropical wintering grounds. In general, whales feed on their summering grounds and breed on their wintering grounds. These migrations include movements of nearly 8000 km by some humpback whales, the longest known annual movements of any mammal. Generally, whales’ longitudinal (east–west) movements are relatively small when compared with their latitudinal (north–south) movements. Individual animals tend to return to the same summering and wintering areas over several years. Some species – right (Eubalaena spp.), gray (Eschrichtius robustus), and humpback whales (Megaptera novaeangliae) – migrate relatively close to coastlines. Balaenoptera spp. tend to migrate further offshore, and their movements are less well known than those of the coastal migrators. Species

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

also vary in the distance over which they migrate; for example, right whales tend to cover less latitudinal range than humpback or gray whales. Variation in Migratory Patterns

Not all species of baleen whale demonstrate this annual cycle. Bowhead whales, Balaena glacialis, undertake substantial longitudinal movements around the coasts of Alaska, Canada, and Siberia, but the southernmost extent of their movements remains close to the pack ice edge. At the other thermal extreme, some tropical Bryde’s whales, Balaenoptera brydei, may not migrate at all. A population of humpback whales in the Arabian Sea also appears not to migrate. There can be great intraspecific variability in the distances traveled by baleen whales. Humpback

597

whales that congregate in Caribbean waters to breed come from discrete feeding groups in Atlantic waters, including western Greenland, off Newfoundland/Labrador, the Gulf of St Lawrence, off the northeastern USA, Iceland, and from north of Norway (Figure 1). From Antarctic waters, some humpback whales migrate into equatorial waters in the Northern Hemisphere, but others seem not to migrate north of approximately 151S. Individuals of migratory species do not necessarily migrate every year. Some female and juvenile humpback whales, and female southern right whales do not arrive at their wintering grounds every year. Large baleen whales were sighted south of the Antarctic Convergence in winter by early expeditions, and Antarctic minke whales, Balaenoptera bonarensis, have been seen inside the pack ice during winter. There are records of blue (B. musculus) fin

?

2000 km

Figure 1 Migratory paths of humpback whales in the North Atlantic.

(c) 2011 Elsevier Inc. All Rights Reserved.

598

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

(B. physalus) and humpback whales wintering in high latitude waters of the North Atlantic. There are records of remarkable longitudinal movements by some baleen whales. Tagging programs during Antarctic whaling demonstrated that individual blue whales could move around the Antarctic continent. Individual humpback whales, apparently males, have been observed overwintering in Hawaiian and Japanese waters in different years. Individual southern right whales have moved between the coast of South America and islands in the central South Atlantic. As some populations of baleen whales recover from previous overhunting, further variability in migratory behavior is becoming evident. Some gray whales now feed in waters off British Columbia, well south of their primary feeding grounds in Arctic waters. Wintering southern right whales off the east coast of Australia are occurring in more northern waters. Further recoveries of baleen whale populations should reveal more behavioral variability at both wintering and summering grounds. Why Do Baleen Whales Migrate?

Why baleen whales use high latitude feeding grounds is clear – polar systems produce incredible quantities of baleen whales’ prey in the warmer months. The enigmatic aspect of baleen whale migration is why most whales travel to low latitude breeding grounds. Current hypotheses regarding why baleen whales migrate relate to calf survivorship: that calves are born away from polar waters so that in the first few weeks of life they avoid either cold waters, stormy waters, or predation by killer whales, Orcinus orca. Information available at present does not allow definitive conclusions on which of these competing (although not mutually exclusive) hypotheses is correct.

Odontocetes Most odontocetes are smaller than mysticetes, and do not show their annual migratory patterns, nor move over ocean basins. Sperm whales, Physeter macrocephalus, are the exception to this. The movements of most odontocete species are not well known. Sperm Whales

Baleen whales travel long distances latitudinally, but sperm whales are known more for their deep foraging dives. However, male sperm whales also

Pole

Equator

Figure 2 Diagrammatic representation of the latitudinal distribution of sperm whales in the Northern Hemisphere. (Reproduced with permission from Mann et al., 2000.)

undertake significant latitudinal movements. Females with calves and juveniles live in matrilineal groups in tropical and subtropical waters. These groups occupy home ranges with a long axis in the order of 1000 km. Young males leave their natal groups and move into higher latitude waters with conspecifics of similar age. As they mature, male sperm whales become less sociable and move into polar waters. Mature males migrate from their high latitude feeding areas to tropical waters to mate, but whether these are annual migrations is unclear (Figure 2). Despite these long-distance migrations, small areas can be important. On one well-studied (and highly productive) feeding ground of only 20  30 km, up to approximately 90 males can be seasonally resident. Females’ foraging ranges are larger, and it appears that their large ranges are a strategy to cope with interannual variability in environmental productivity. Delphinids

Bottlenose dolphins Bottlenose dolphins, Tursiops truncatus and T. aduncus, are among the best studied of the delphinids. They are found from temperate to equatorial waters and from shallow waters by the coast to the deep ocean. Throughout their range, two ecotypes occur: inshore and offshore forms, with offshore animals being more robust. The relationship between these two forms and the two species of Tursiops is unclear, confounding comparisons. Bottlenose dolphins ranging patterns vary from animals with relatively small home ranges to migratory animals. Bottlenose dolphins living in sheltered coastal waters, particularly bays, tend to have home ranges of several tens of square kilometers, varying in size with gender, age, and reproductive status. Most

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

individuals appear to remain within these ranges throughout their life. Animals living in shallow waters off open coasts (e.g. California, South Africa) can range over 4100 km of coastline. Elsewhere (e.g. eastern Australia) dolphins in shallow waters off open coasts can have small home ranges, similar to those of bay animals elsewhere. In some areas, populations of bottlenose dolphins undertake relatively long migrations. Some inshore dolphins off the US east coast migrate annually over approximately 400 km of coastline, while offshore animals in the same area move even further. Most information on bottlenose dolphins comes from studies of individually identified animals. Logistically, these studies are likely to concentrate on animals with relatively small ranges. Satellite tracking offshore for bottlenose dolphins has started to reveal the extent over which they can range. Two animals tracked off Florida traveled 2050 km in 43 days, and 4200 km in 47 days. As these animals were rehabilitated after stranding, the extent to which their movements are representative of normal movements is debatable, but they demonstrate the ranging capacities of offshore bottlenose dolphins. Killer Whales

Killer whales are found in all the world’s oceans, from polar to equatorial waters, but their densities in polar waters are substantially higher than in tropical waters. There are two genetically distinct forms of killer whales: residents, feeding mainly on fish, and transients, that are marine mammal predators. Although these forms are best described from the waters off British Columbia, similar forms have been reported from Antarctic waters. Despite the names, the differences in ranging of these two forms of killer whales are not necessarily great. Off British Columbia and Washington, the ranges of transient pods extend to approximately 140 000 km2 and those of residents to 90 000 km2. As with bottlenose dolphins, logistics limit knowledge of the extent of killer whales’ ranges. Migratory behavior of killer whales remains poorly understood. The longest movements documented to date are of three identified individual transient killer whales that moved 2660 km along the Pacific coast of North America, from 581410 N to 361480 N over 3 years. Soviet whaling data from the Antarctic suggest that some killer whales undergo annual migrations to at least temperate latitudes, but sightings of killer whales in the Antarctic pack ice in winter demonstrate that not all animals migrate. Killer whales’ movements seem tied to movements of their prey: north–south movements of some killer whales along the North American west coast appear to

599

be at least partially in response to the presence of migrating gray whales. Killer whales off coastal Norway seem to follow the herring migration. Killer whales appear to move into nearshore waters at several areas in the subantarctic – Peninsula Valdes Argentina – Marion Island, Macquarie Island, the Crozet Archipelago – coincident with seal pupping, although this may reflect the limits of shore-based observation to determine killer whales’ real movements. Other Odontocetes

Some inshore delphinids seem to demonstrate the variability in ranging behavior shown by bottlenose dolphins. Some individual humpback dolphins (Sousa spp.) in bay and estuarine environments have small home ranges (tens of square kilometers), others in more coastal waters range along hundreds of kilometers of coastline. Along 2000 km of the open west coast of South Africa there are three distinct matrilineal assemblages of Indian humpback dolphins, S. plumbea. Marine Irrawaddy dolphins, Orcaella brevirostris, appear to have small ranges (tens of square kilometers), centered around the mouths of rivers. Movement patterns of delphinids in offshore waters are less well understood, but it is clear that animals range over at least thousands of square kilometers. Surveys off the west coast of the USA, covering over 10 degrees of latitude and extending to 550 km offshore, demonstrated seasonal changes in the distribution of small cetaceans. For example, northern right whale dolphins, Lissodelphis borealis, moved into continental shelf waters of the Southern California Bight in the winter, presumably from waters further offshore. Pacific white-sided dolphins, Lagenorhynchus obliquidens, and Dall’s porpoises, Phocoenoides dalli, moved into more southern waters in winter. Belugas (Delphinapterus leucas) and narwhals (Monodon monoceros) are ice-associated odontocetes found in Northern Hemisphere waters. Both species’ annual movements through Arctic waters are closely tied to the movements of pack ice. One satellite-tagged narwhal traveled 46300 km in 89 days, moving through 461 of latitude. Satellite tagging studies of both species reveal times when they move to the vicinity of glaciers, for reasons that are unclear.

Pinnipeds Pinnipeds comprise three families: Phocidae, the true or haired seals; Otariidae, the eared or fur seals; and Odobenidae, the walrus, Odobenus rosmarus.

(c) 2011 Elsevier Inc. All Rights Reserved.

600

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

Unlike cetaceans or sirenians, pinnipeds have retained some terrestrial traits while adapting to foraging at sea. Although all pinnipeds give birth on land or ice, the importance of terrestrial sites for suckling, rest, mating, molting, predator avoidance, thermoregulation, or saving energy varies considerably across species. Within species, individuals’ movement patterns vary depending both upon local environment conditions and individuals’ age, sex, or breeding status. Breeding

Pinniped movement patterns differ from other marine mammals due to the need for females to give birth and (generally) nurse on land. Female pinnipeds exhibit three maternal strategies, ‘aquatic nursing,’ the ‘foraging cycle,’ and the ‘fasting cycle.’ To date, the walrus is the only species to exhibit aquatic nursing. Female walrus give birth, fast for a few days, and then take their calves with them while they forage at sea. Lactation is extended and lasts for up to 2 or 3 years. This strategy bears some resemblance to that of cetaceans and sirenians. Most phocid and several otariid species exhibit the foraging cycle strategy. In these species, females forage at sea during lactation, returning to nurse their pups. Lactation can be relatively prolonged in these cases, most otariids suckle for weeks or months. Lactation is generally less prolonged in phocid species. Harbor seals, Phoca vitulina, nurse their pups for 24 days, and forage at sea during the latter stage of nursing (Figure 3). Female phocids using the fasting strategy remain on land throughout the whole nursing period. This period is relatively short, around 18 days for gray seals, Halichoerus grypus, 23 days for southern elephant seals, Mirounga leonina, and only 4 days for hooded seals, Cystophora cristata. Therefore, throughout the breeding season, the movements of female seals are constrained to different degrees by the need to suckle. Although male pinnipeds are not limited by the need to nurse a pup, their movements during the breeding season are constrained by their need to mate. Males either obtain access to females on land or in the water. Approximately half of the pinniped species mate on land. However, walrus, two otariid species, and 15 species of phocid mate aquatically. Land mating pinnipeds remain onshore to defend either territorial access to females, such as in elephant seals, Mirounga spp. or access to resources used by females, as in Antarctic fur seals, Arctocephalus gazella. Males of aquatic mating pinnipeds, such as harbor seals, were previously thought not to restrict their movements during the mating season.

However, recent evidence has shown that males perform vocal and dive displays in small discrete areas and limit their movements to areas frequently used by females throughout the mating season (Figure 3). Breeding activities clearly regulate movements of both female and male pinnipeds. Nonbreeding

There is less variation between movement patterns of males and females outside of the breeding season, although differences do exist. During this period movement patterns are more strongly governed by resource (generally prey) availability. Pinnipeds exhibit marked variation in the distance they move in order to reach suitable feeding grounds or breeding habitats. Harbor seals remain faithful to a single site or small group of local sites and usually do not forage 450 km from their haul-out sites. Similarly, southern sea lions, Otaria flavescens, forage up to 45 km offshore, with occasional longer trips extending to 4150 km. In contrast, northern elephant seals, Mirounga angustirostris, carry out long-distance migrations of several thousand kilometers, travelling from California to the north-eastern Pacific Ocean twice a year (Figure 4). Pups and juvenile animals are unable to travel as extensively as adults. In harbor seals, mothers restrict their foraging range while accompanied by their pups. Juvenile northern elephant seals do not dive as deeply, move more slowly, and do not migrate as far as adults during the first few years of their lives. Haul-out Sites

Most pinnipeds remain faithful to a single or small group of haul-out sites. Groups of seals at different sites within a local haul-out area may show consistent differences in sex or age structure. Seasonal changes in haul-out site result from changes in foraging grounds, seasonal availability of prey, and characteristics of haul-out sites. Frequently, females and pups predominate at certain sites during the pupping season. Sheltered isolated areas may be chosen because of the lack of disturbance or terrestrial predators. Other sites may be used predominantly during molting. Timing of trips to sea in relation to tidal and diel cycles varies considerably both within and between areas. Seals using haul-out sites that are available throughout the tidal cycle have activity patterns dominated by the diel cycle. When seals use haul-out sites that are only available over low tide, the tidal cycle has a more dominant effect. These effects are less pronounced when pinnipeds engage in longer foraging trips.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

601

Female harbor seal distribution

km N 0

5

10

Moray Firth, Scotland

= Pupping haul-out sites = Female foraging grounds

Male harbor seal distribution

km N 0

5

10

Moray Firth, Scotland

= Male distribution

Figure 3 Ranges of harbor seals in the Moray Firth, Scotland. (From Thompson PM, Miller D, Cooper R and Hammond PS (1994) Changes in the distribution and activity of female harbour seals during the breeding season: implications for their lactation strategy and mating patterns. Journal of Animal Ecology 63: 24–30. Van Parijs SM, Hastie GD and Thompson PM (1999) Geographic variation in temporal and spatial patterns of aquatic mating male harbour seals. Animal Behavior 58: 1231–1239.)

(c) 2011 Elsevier Inc. All Rights Reserved.

602

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

Figure 4 Movements of 22 adult males (red) and 17 adult females (yellow) tracked by satellite during spring and fall migrations from An˜o Nuevo, California, during 1995, 1996, and 1997. (Reproduced with permission from Le Boeuf et al., 2000.)

Small-scale movements in the vicinity of haul-out areas can be affected by the risk of predation. For example, northern elephant seals approaching major breeding areas alter their diving behavior in a way that appears to reduce the risk of attack by the white shark, Carcharodon carcharias. Pinnipeds hauling out on sand or rocky shorelines are exposed to different constraints to those hauling out on ice. Ice breeding pinnipeds often show a partiality to hauling out on a particular type of ice. As ice changes often and suddenly, they may be more constrained in the choice of haul-out site at particular periods of the year. Bearded seals, Erignathus barbatus, haul out on ice floes, the availability of which alters seasonally, daily, and hourly. Therefore, movements of both female and male bearded seals reflect the availability of a particular type of ice within an area. The annual movements of harp seals, Phoca groenlandica, in the Barents Sea demonstrate their relationship with both ice and food availability.

3

2

500 km

1

Figure 5 Annual movements of harp seals in the Barents Sea, including movements during recent invasions of the Norwegian coast. (1) Svalbard; (2) Franz Josef Land; (3) Novaja Zemlja. Gray shading shows the molting area, hatching shows the breeding area. The dashed line indicates the extent of movements of harp seals during invasions over recent years. (Data primarily from Haug T, Nilssen KT, Øien N and Potelov V (1994) Seasonal distribution of harp seals (Phoca groenlandica) in the Barents Sea. Polar Research 13: 163–172.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

These seals breed in the White Sea in February/ March, then molt in April/May in the White Sea and southern Barents Sea. In March and April, adult females seem to move westward on a short feeding migration. Between June and September, they are found in open water or in pack ice from Novaja Zemlja to Svalbard. They tend to be more associated with pack ice edge in September and October, moving east and north east to the vicinity of Franz Josef Land. Through November they seem to migrate ahead of the advancing pack to the southern coast of Novaja Zemlja, from where they move to their breeding grounds (Figure 5). Recently, in some years this pattern has been altered, with the westward movement of immature seals along the coast of northern Norway in winter (December–March). As these seals were in poor condition, these movements appear to be related to attempted foraging.

Sirenians Being generally herbivorous, sirenians’ foraging behavior differs substantially from other marine mammals. As animals of tropical and subtropical waters, their movements and food sources also show less seasonal variation than those of most other marine mammals. Dugongs, Dugong dugon, are the only extant sirenians that are truly marine, and so they will be the focus of discussion here. Dugongs occur in shallow (generally o20 m) waters of the tropical and subtropical Indo-West Pacific. Most dugongs live in relatively small home ranges, in the order of tens to around a hundred square kilometers. Occasionally, satellite-tagged individuals undertake longer movements, up to 600 km from their home range and then, after a period of up to several weeks, return to their home range. There are no apparent age- or gender-related patterns to this, and reasons for these movements are unclear. Within their home range, dugongs’ diel movements are tidally influenced, especially if seagrass beds occur in banks that are o1 m deep or exposed at low tide. In at least one area, dugongs graze in large herds at the same site over weeks or months. This ranging and foraging pattern, termed ‘cultivation grazing’, encourages the growth of the pioneer seagrass species that are dugongs’ preferred food. During periods of extreme food shortage in their home range, dugongs are known to travel along several hundred kilometers of coastline in search of new feeding grounds.

Glossary Migration Persistent destinations.

movement

between

two

603

Home range An animal’s home range is that area within which it carries out most of its normal activities throughout the year. Home ranges are not usually considered to include migratory movements. Haul-out site Area (land or ice) where seals remove themselves from water. Foraging The process by which animals obtain food – includes searching for, capturing, and ingesting food. Matrilineal (social unit) Social system where female relatives remain associated, thereby providing the basic unit of the animals’ society.

See also Baleen Whales. Marine Mammal Diving Physiology. Marine Mammal Overview. Marine Mammal Social Organization and Communication. Marine Mammal Trophic Levels and Interactions. Marine Mammals: Sperm Whales and Beaked Whales. Sea Otters. Seals. Sirenians.

Further Reading Boness DJ and Bowen WD (1996) The evolution of maternal care in pinnipeds. Bioscience 46: 645--654. Clapham PJ (1996) The social and reproductive biology of humpback whales: an ecological perspective. Mammal Review 26: 27--49. Corkeron PJ and Connor RC (1999) Why do baleen whales migrate? Marine Mammal Science 15: 1228--1245. Gaskin DE (1982) The Ecology of Whales and Dolphins. London: Heinemann. Hammond PS, Mizroch SA and Donovan GP (1990) Individual Recognition of Cetaceans: Use of Photoidentification and Other Techniques to Estimate Population Parameters. Reports of the International Whaling Commission, Special Issue 12. Le Boeuf BJ and Laws RM (eds.) (1994) Elephant Seals: Population Ecology, Behavior and Physiology. Berkeley, CA: University of California Press. Le Boeuf BJ, Crocker DE, Costa DP, et al. (2000) Foraging ecology of northern elephant seals. Ecological Monographs 70(3): 353--382. Mann J, Connor RC, Tyack PL, and Whitehead H (eds.) (2000) Cetacean Societies. Field Studies of Dolphins and Whales. Chicago: University of Chicago Press. Marsh H, Eros C, Corkeron PJ, and Breen B (1999) A conservation strategy for dugongs: implications of Australian research. Marine and Freshwater Research 50: 979--990. Moore SE and Reeves RR (1993) Distribution and movement. In: Burns JJ, Montague JJ, and Cowles CJ

(c) 2011 Elsevier Inc. All Rights Reserved.

604

MARINE MAMMAL MIGRATIONS AND MOVEMENT PATTERNS

(eds.) The Bowhead Whale, pp. 313--386. Lawrence, KS: Society of Marine Mammalogy. Preen AR (1995) Impacts of dugong foraging on seagrass habitats: observational and experimental evidence for

cultivation grazing. Marine Ecology Progress Series 124: 201--213. Renouf D (ed.) (1991) Behavior of Pinnipeds. London: Chapman and Hall.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL OVERVIEW P. L. Tyack, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1611–1621, & 2001, Elsevier Ltd.

Introduction The term ‘marine mammals’ is an ecological grouping that lumps together a phylogenetically diverse set of mammals. The only thing linking them is that marine mammals occupy and rely upon aquatic habitats for all or much of their lives. The cetaceans and the sirenians (see the relevant Encyclopedia articles), or dugongs and manatees, live their entire lives at sea, only coming on land during perilous stranding events. Cetaceans (Table 1) evolved from ungulates whose modern members include pigs, while sirenians (Table 2) evolved from ungulates related to the modern elephant. One member of the bear family (Ursidae), the polar bear, is categorized as a marine mammal, while two members of the family Mustelidae are considered marine: the sea otter and the marine otter of Chile (Table 3). Seals and walruses also evolved from terrestrial carnivores. Most biologists lump the seals and walruses into a suborder called the Pinnepedia, and these are often referred to as pinnipeds, meaning ‘finlike feet’. Seals are divided into two families: the Otariidae, or eared seals, and the Phocidae, or true seals (Table 4). Walruses are categorized as a separate family, the Odobenidae (Table 4). The definition of marine mammals is somewhat arbitrary – river dolphins are considered marine mammals, but river otters are not. Inclusion in this category can have real consequences in the United States, since marine mammals have special protection under the US Marine Mammal Protection Act.

Taxonomy The basic evolution of marine mammals from carnivores and ungulates is well established, but the details of phylogeny and taxonomy at the species level are in flux, owing in part to recent molecular genetic data. Since there is a relatively well-established nomenclature that has been stable for the past 20 years or so, and which forms the basis for species management, this nomenclature will be used in the following tables, while it is recognized that the

developing synthesis of molecular and morphological data will probably change some of the species relations. For example, the US and international agencies responsible for protecting endangered species split right whales of the northern and southern oceans into two species; some taxonomists lump right whales into one species, but the North Pacific and North Atlantic right whales are clearly separated and recent genetic data suggests the division of right whales into three species: Southern, Northern Pacific, and North Atlantic. Where the designation of endangered species focuses on the species level, this lumping of right whales of the North Atlantic and North Pacific, which are highly endangered, with right whales of the South Atlantic, which are doing well, could have profound policy implications.

Adaptations of Marine Mammals The success of marine mammals is something of a puzzle. How did their terrestrial ancestors, adapted for life on land, compete against all of the life forms that were already so well adapted to the marine environment? Multicellular organisms arose in the oceans of the earth about 700 million years ago (MYa), and for the next 100 million years or so, there was a remarkable burst of evolutionary diversification, as most of the basic body plans of life in the sea evolved. By 350–400 MYa, multicellular animals expanded from the ocean into terrestrial environments, with another evolutionary radiation. Mammals only reentered the sea about 60 MYa, and thus have had less than one-tenth of the time available to the original marine metazoans for adaptation to this challenging environment. The ocean poses a hostile environment to mammals, yet a remarkable diversity of mammalian groups from carnivores such as bears and otters to ungulates have adapted to the marine environment. Most marine animals maintain their bodies at the temperature of the surrounding sea water and obtain any oxygen required for respiration directly from oxygen dissolved in sea water. Marine mammals need to breathe air and maintain their bodies at temperatures well above the typical temperature of sea water. Many marine animals release thousands to millions of eggs into the hostile marine environment, where the odds are that only a handful will survive. Mammals produce only a handful of offspring at a time, and must rely upon parental care to enhance the survival of their offspring.

(c) 2011 Elsevier Inc. All Rights Reserved.

605

606

MARINE MAMMAL OVERVIEW

Table 1

Scientific and common names of cetaceans along with conservation statusa

Scientific name

Common name

ESA

MMPA

IUCN

Suborder Mysticeti Family Balaenidae Balaena mysticetus Eubalaena australis Eubalaena glacialis

Baleen whales Right whales Bowhead whale Southern right whale Northern right whale

EN EN EN

DEP DEP DEP

EN

Family Neobalaenidae Caperea marginata

Pygmy right whale

Family Balaenopteridae Balaenoptera acutorostrata Balaenoptera borealis Balaenoptera edeni Balaenoptera musculus Balaenoptera physalus Megaptera novaeangliae

Rorquals Minke whale Sei whale Bryde’s whale Blue whale Fin whale Humpback whale

EN

DEP

EN

EN EN EN

DEP DEP DEP

EN EN VU

Family Eschrichtiidae Eschrichtius robustus

Grey whale

EN (NW Pacific)

DEP (NW Pacific)

Suborder Odontoceti Family Physeteridae Physeter macrocephalus

Toothed whales Sperm whales Sperm whale

EN

DEP

VU

Family Kogiidae Kogia breviceps Kogia simus

Pygmy sperm whale Dwarf sperm whale

Family Ziphiidae Berardius arnuxii Berardius bairdii Hyperoodon ampullatus Hyperoodon planifrons Mesoplodon bidens Mesoplodon bowdoini Mesoplodon carlhubbsi Mesoplodon densirostris Mesoplodon europaeus Mesoplodon ginkgodens Mesoplodon grayi Mesoplodon hectori Mesoplodon layardii Mesoplodon mirus Mesoplodon pacificus Mesoplodon peruvianus Mesoplodon stejnegeri Tasmacetus shepardi Ziphius cavirostris

Beaked whales Arnoux’s beaked whale Baird’s beaked whale Northern bottlenose whale Southern bottlenose whale Sowerby’s beaked whale Andrew’s beaked whale Hubb’s beaked whale Blainville’s beaked whale Gervais’ beaked whale Ginkgo-toothed beaked whale Gray’s beaked whale Hector’s beaked whale Strap-toothed beaked whale True’s beaked whale Longman’s beaked whale Pygmy beaked whale Stejneger’s beaked whale Tasman’s beaked whale Cuvier’s beaked whale

Family Monodontidae Delphinapterus leucas Monodon monoceros

Beluga; white whale Narwhal

DEP (Cook Inlet stock)

VU

Family Delphinidae Cephalorhynchus commersonii Cephalorhynchus eutropia Cephalorhynchus heavisidii Cephalorhynchus hectori Delphinus capensis Delphinus delphis Feresa attenuata

Oceanic dolphins Commerson’s dolphin Black dolphin Heaviside’s dolphin Hector’s dolphin Long-beaked common dolphin Common dolphin Pygmy killer whale

EN

(Continued )

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL OVERVIEW

Table 1

607

Continued

Scientific name

Common name

ESA

Globicephala macrorhynchus Globicephala melas Grampus griseus Lagenodelphis hosei Lagenorhynchus acutus Lagenorhynchus albirostris Lagenorhynchus australis Lagenorhynchus cruciger Lagenorhynchus obliquidens Lagenorhynchus obscurus Lissodelphis borealis Lissodelphis peronii Orcaella brevirostris Orcinus orca Peponocephala electra Pseudorca crassidens Sotalia fluviatilis Sousa chinensis Sousa teuszii Stenella attenuata Stenella clymene Stenella coeruleoalba Stenella frontalis Stenella longirostris Steno bredanensis Tursiops truncatus

Short-finned pilot whale Long-finned pilot whale Risso’s dolphin Fraser’s dolphin Atlantic white-side dolphin White-beaked dolphin Peale’s dolphin Hourglass dolphin Pacific white-sided dolphin Dusky dolphin Northern right whale dolphin Southern right whale dolphin Irrawaddy dolphin Killer whale Melon-headed whale False killer whale Tucuxi Indo-Pacific hump-backed dolphin Atlantic hump-backed dolphin Pantropical spotted dolphin Clymene dolphin Striped dolphin Atlantic spotted dolphin Spinner dolphin Rough-toothed dolphin Bottlenose dolphin

Family Phocoenidae Phocoena dioptrica Neophocaena phocaenoides Phocoena phocoena Phocoena sinus Phocoena spinnipinis Phocoenoides dalli

Porpoises Spectacled porpoise Finless porpoise Harbor porpoise Cochito; Vaquita Burmeister’s porpoise Dall’s porpoise

Family Platanistidae Platanista gangetica Platanista minor Family Iniidae Inia geoffrensis Lipotes vexillifer Pontoporia blainvillei

River dolphins Ganges river dolphin Indus susu River dolphins Boutu; boto Baiji Franciscana

MMPA

IUCN

DEP (ETP)

DEP (ETP) DEP (mid-Atlantic coastal)

EN

DEP

EN

DEP

EN

DEP

VU CR

a

CR ¼ Critically endangered; DEP ¼ Depleted; EN ¼ Endangered; ESA ¼ US Endangered Species Act; IUCN ¼ International Union for the Conservation of Nature; MMPA ¼ US Marine Mammal Protection Act; TH ¼ Threatened; VU ¼ Vulnerable. (Sources: Reynolds and Rommel, 1999; IUCN red book.)

Some of the mammalian adaptations that are wellsuited for terrestrial life appear to create risks and drawbacks of life in the sea, yet these adaptations opened new niches for marine mammals in four of the five trophic levels of marine ecosystems. These adaptations may be particularly important for predators. For example, the mammalian adaptation of homeothermy, or maintaining a constant body temperature, requires a metabolic rate 10–100 times that of an animal whose body stays at ambient temperature. Biochemical reactions occur at different rates at different temperatures, so animals that do not

regulate their temperature as precisely as mammals may not have sensory or motor systems operating as rapidly. All a predator needs is a small marginal advantage in ability to detect or locate prey or in locomotor ability and maneuverability in order to succeed. The high metabolic cost of homeothermy in marine mammals may thus be offset by their potential predatory advantage. This high-cost/high-gain strategy is unique in the oceans to marine mammals and marine birds, although some predatory marine fish also warm muscles or the central nervous system for similar advantage. Marine mammals provide

(c) 2011 Elsevier Inc. All Rights Reserved.

608

MARINE MAMMAL OVERVIEW

Table 2

Scientific and common names of sirenians along with conservation statusa

Scientific name

Common name

ESA

MMPA

IUCN

Family Trichechidae Trichechus inunguis Trichechus manatus Trichechus senegalensis

Amazonian manatee West Indian manatee West African manatee

EN EN TH

DEP DEP DEP

VU VU

Family Dugongidae Dugong dugon

Dugong

EN

DEP

VU

a

DEP ¼ Depleted; EN ¼ Endangered; ESA ¼ US Endangered Species Act; IUCN ¼ International Union for the Conservation of Nature; MMPA ¼ US Marine Mammal Protection Act; TH ¼ Threatened; VU ¼ Vulnerable. (Sources: Reynolds and Rommel, 1999; IUCN red book.)

Table 3 Scientific and common names of marine mustelids and ursids along with conservation statusa Scientific name

Common name

ESA

MMPA

IUCN

Enhydra lutris Lutra marina Ursus maritimus

Sea otter Marine otter Polar bear

TH EN

DEP DEP

EN EN

a

DEP ¼ Depleted; EN ¼ Endangered; ESA ¼ US Endangered Species Act; IUCN ¼ International Union for the Conservation of Nature; MMPA ¼ US Marine Mammal Protection Act; TH ¼ Threatened; VU ¼ Vulnerable. (Sources: Reynolds and Rommel, 1999; IUCN red book.)

parental care to their young, and can better afford to put more reproductive effort into fewer offspring than is typical of marine organisms. Marine mammals often feed in areas that are separated from the places where they give birth. This has led to an adaptation unknown among terrestrial mammals (except for bears) – the ability to lactate while fasting. Mammals evolved an ear specialized to analyze the frequency content of sound. Air-breathing mammals also have a vocal tract well-adapted to producing loud and complex sounds. When mammals took to the sea, these acoustic adaptations took on a special importance. Light does not penetrate more than a few hundred meters in clear open ocean and only a few meters in some coastal areas, so vision, which is a primary distance sense in air, has a limited range under water. Sound, by contrast, propagates extremely well under water. Many marine mammals have evolved specialized abilities of hearing and sound production that are used for communication and exploring the environment. Many of the toothed whales feed on elusive and patchy prey in the dark of the deep sea or at night. Many species have evolved a high-frequency biosonar that allows them to detect and chase prey. Many baleen whales migrate thousands of kilometers

annually, yet are quite social, and they produce lowfrequency sounds that can be detected at ranges of hundreds if not thousands of kilometers.

Locomotion Terrestrial animals have evolved special adaptations to support their bodies and for moving along a solid substrate. Efficient swimming puts very different selection pressures on marine compared to terrestrial locomotion. Water is close to the same density as most animal tissues. This freed aquatic mammals such as the sirenians and cetaceans completely from the need to support their bodies. However, water is also much more viscous than air. This resistance to motion allows marine animals to move by pushing against the medium, but also creates a strong selective pressure for mechanisms to reduce drag. Two different kinds of drag forces act on marine mammals. Viscous drag selects for a smooth skin and a low ratio of surface area to volume. Pressure drag relates to how fluid flows around the body given pressure differences, and this selects for a streamlined hydrodynamic shape. Most marine mammals that swim long distances in the water thus have a slick skin and a low-drag shape that makes them look quite different from their terrestrial ancestors (see Figure 1).

How Different Marine Mammals Fall on a Continuum from Aquatic to Terrestrial While all marine mammals use the marine environment, some species are amphibious and need to function in air and in water. The polar bear, for example, is a strong swimmer but retains a body form quite similar to that of terrestrial bears (see Figure 1). Polar bears live on sea ice when they can, and their primary diet is

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL OVERVIEW

Table 4

609

Scientific and common names of pinnipeds along with conservation statusa

Scientific name

Common name

Family Odobenidae Odobenus rosmarus

Walrus Walrus

Family Otariidae Subfamily Otariinae Eumetopias jubatus Neophoca cinarea Otaria byronia Phocarctos hookeri Zalophus californianus

Eared Seals Sea lions Stellar or Northern sea lion Australian sea lion Southern sea lion New Zealand sea lion California sea lion

Subfamily Arctocephalinae Arctocephalus australis Arctocephalus forsteri Arctocephalus galapagoensis Arctocephalus gazella Arctocephalus philipii Arctocephalus pusillus Arctocephalus townsendi Arctocephalus tropicalis Callorhinus ursinus

Fur seals Falkland or South American fur seal New Zealand fur seal Galapagos fur seal Antarctic fur seal Juan Fernandez fur seal Australian or S. African fur seal Guadelupe fur seal Subantarctic fur seal Northern fur seal

Family Phocidae Subfamily Phocinae Cystophora cristata Erignathus barbatus Halichoerus grypus Phoca caspica Phoca fasciata Phoca groenlandica Phoca hispida Phoca larga Phoca sibirica Phoca vitulina

Earless or True seals Northern phocids Hooded seal Bearded seal Gray seal Caspian seal Ribbon seal Harp seal Ringed seal Larga seal Baikal seal Harbor (US) or common (UK) seal

Subfamily Monachinae Hydrurga leptonyx Leptonychotes weddellii Lobodon carcinophagus Mirounga angustirostris Mirounga leonina Monachus monachus Monachus schauinslandi Monachus tropicalis Ommatophoca rossii

Southern phocids Leopard seal Weddell seal Crabeater seal Northern elephant seal Southern elephant seal Mediterranean monk seal Hawaiian monk seal Caribbean monk seal Ross seal

ESA

MMPA

IUCN

TH

DEP

EN

VU

VU VU TH

DEP

VU

DEP

VU

EN (NE Atl) VU

EN (Saimaa)

DEP

VU EN (Saimaa)

EN EN EN

DEP DEP DEP

CR EN EX

a

CR ¼ Critically endangered; DEP ¼ Depleted; EN ¼ Endangered; ESA ¼ US Endangered Species Act; EX ¼ Extinct; IUCN ¼ International Union for the Conservation of Nature; MMPA ¼ US Marine Mammal Protection Act; TH ¼ Threatened; VU ¼ Vulnerable. (Sources: Reynolds and Rommel, 1999; IUCN red book.)

seals, but they often live ashore during four months in the later summer and fall when sea ice has receded. Otariid seals can walk on land, but have a body shape transformed for swimming compared to their terrestrial ancestors (see Figure 1). The finlike limbs of seals are designed to push against the water with a large cross-sectional area. Marine mammals with limbs designed for swimming have large muscles attached to

shorter bones, giving more power and a greater mechanical advantage to the swimming motion. The sirenians and cetaceans never come onto land except during a stranding, and their bodies are the most transformed for swimming – neither group has legs or separate hind flippers, but rather a broad tail (see Figure 1). Cetaceans and sirenians use their axial musculature running along the entire vertebral column

(c) 2011 Elsevier Inc. All Rights Reserved.

610

MARINE MAMMAL OVERVIEW

Otariid Sea Lion Phocid Seal

Polar Bear

Sea Otter

Dog Sirenian

Baleen Whale

Dolphin

Figure 1 Body outlines and skeletons of selected marine mammal species. (From Reynolds and Rommel (1999), Figures 2–8.)

as they swim. Their tail flukes do not need to function for walking, only for swimming, and they are more efficient at generating thrust than are the seal flippers.

Sensory Systems Just as these groups differ in their adaptation for locomotion in air and under water, so their sensory systems fall in different places on the continuum from adaptation to terrestrial versus aquatic life. The cetaceans are so fully adapted for marine life that they have lost most olfaction, which senses airborne odors. As mentioned above, sound is a particularly useful modality in the sea, but even for acoustic communication there is variation in the importance of airborne versus underwater sound. The otariid pinnipeds, sea otter, and polar bear communicate primarily in air; some phocid seals communicate both in air and under water; while sirenians, cetaceans, some phocid seals, and the walrus use sound to communicate primarily underwater. There are differences in the relative importance of hearing in air versus water for three different pinniped species whose hearing has been tested in both environments. The Otariid California sea lion is adapted to hear best in air; the phocid harbor seal can hear equally well in air and under water; and the auditory system of the phocid northern elephant seal is adapted for underwater sensitivity at the expense of aerial hearing.

Feeding Marine mammals feed at a variety of trophic levels from herbivore to top predator. The sirenians are herbivores and eat sea grass that grows in coastal waters of the tropics. Several marine mammals specialize on benthic prey. The walrus feeds primarily on benthic mollusks; sea otters feed on benthic mollusks, echinoderms and decapod crustaceans. Sea otters are a keystone predator; their feeding on echinoderms is thought to structure kelp forest ecosystems. Grey whales feeds on benthic amphipods in the Bering Sea; their feeding turns over 9–27% of the seafloor there each feeding season, which probably enhances the abundance of colonizing benthic species. Baleen whales do not have teeth, and baleen whales other than the grey whale use their baleen to strain prey from sea water. They have been compared to grazers, because they feed on whole patches of prey, but their prey is typically large zooplankton or even schooling fish, which often require a predatorlike ability to find and pursue prey. Most seals and toothed whales chase individual prey items, often fish or squid. Very little is known about the feeding behavior of deep diving toothed whales such as sperm whales. Our ignorance of the behavior of sperm whales and their deep squid prey may cause us to underestimate their ecological importance, for it has been calculated that they must take out of the ocean about the same biomass as all human fisheries.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL OVERVIEW

Annual Feeding and Reproductive Cycles of Marine Mammals Polar Bear and Pinnipeds

Polar bears and seals have a strong annual breeding cycle, with a short birthing and mating season, and delayed implantation that allows them to mate in spring well before they settle into a winter den for a short gestation period. Polar bears feed primarily on seals in the Arctic sea ice. Feeding is particularly intense during the spring birthing season for ringed seals, when up to a quarter of ringed seal breathing holes show predation attempts by polar bears. Up to 44% of the seal pups may be taken by polar bears at this time. Spring is not only the prime feeding season but also the breeding season for polar bears. Pregnant females delay implantation of the fetus until the feeding season is over. By summer, the sea ice habitat breaks up, forcing the polar bears either to summer on pack ice or to come ashore. A female who has come to ashore cannot hunt efficiently, and she may fast. By late September or early October, she will enter a winter den. Once she enters the den, the fetus implants and the pregnancy progresses, with a gestation period of only 3–4 months. The young are small and poorly developed when born. The eyes do not open for 40 days and the infant must nurse several times an hour and rely upon the mother to keep warm. Cetaceans, sea otters, and sirenians never need to come to shore and do not have any lairs or dens that act as refuges. Seals are like the polar bear in that

they must give birth on land or ice. Most seals also have a strong annual breeding cycle, with a short birthing and mating season and delayed implantation. A primary difference between otariid and phocid seals is that otariid mothers will leave their pups on shore as they forage. By contrast, many phocid mothers stay with their pups and fast throughout lactation. Many phocids give birth on the ice. Selection appears to favor a short, intense period of lactation in this setting. For example, the hooded seal suckles her young for an average of 4 days. The milk is 460% fat and the female suckles more than twice an hour. The pup gains more than 7 kg d1 during this period. Annual Cycle of Baleen Whales

Researchers debate whether cetacean swimming is more or less energetically costly than terrestrial locomotion, but living in a buoyant medium has certainly freed cetaceans to grow to larger sizes. The blue whale is the largest animal ever to live on Earth, many times more massive than the largest dinosaurs. Aquatic animals are freed from some of the constraints on size imposed on terrestrial animals, but baleen whales are also larger than any other aquatic animals. This large size is part of a suite of adaptations driven by an annual migratory cycle of most baleen whales, feeding in the summer and mating and giving birth in the winter (see Figure 2). Most species of baleen whales have adapted to take advantage of a seasonal burst of productivity in the summer in polar waters. Summer

Breeding 25

35

35

WEANING

25

45

45

55

55

65

65

Migration

Resting

Feeding

Lactation

CONCEPTION

SECOND YEAR CALVING

CONCEPTION

Breeding Migration Feeding

Latitude (°S)

FIRST YEAR

Pregnancy

611

J J A S O N D J F M A M J J A S O N D J F M A M J Month, starting with conception Figure 2 Annual migratory feeding and reproductive cycle of baleen whales. (Adapted from Reynolds and Rommel (1999), Figures 6–19.)

(c) 2011 Elsevier Inc. All Rights Reserved.

612

MARINE MAMMAL OVERVIEW

is the feeding season for most migratory baleen whales, and they feed intensively during the summer, building up fat reserves. The primary prey are invertebrate zooplankton such as krill and copepods; some balaenopterid whales also feed on schooling fish. Once the polar pulse of summer productivity is over, there is little reason to stay in these waters. Most baleen whales migrate away from the feeding grounds to separate winter breeding and calving grounds. They seldom feed during these seasons, and live off of their fat reserves from summer feeding. Most baleen whales migrate to tropical waters for the winter breeding season. It is tempting to assume that they migrate to the tropics to avoid the winter cold and storms in polar seas, but many smaller marine mammals can thermoregulate well in polar areas in winter, so this raises questions about the reasons for migrations to tropical waters in winter. Grey whales migrate roughly 8000 km from feeding grounds in the Bering Sea to breeding grounds in the coastal lagoons of Baja California, but bowhead whales remain in the Bering Sea during winter. There is a need for better modeling of the energetic costs and benefits of migrating thousands of kilometers to warmer water. Many whale species select calm, protected waters with relatively low predation risk for calving, and this may also influence the choice of breeding area. The annual cycle of baleen whales and the way in which they forage have selected for large size. Baleen

whales engulf many prey items within a dense patch of prey, straining prey from sea water using the baleen for which they are named. The habit of feeding on as many prey within a patch as possible selects for large size in the whales. If whales feed in the summer months, and must live off of their energy stores for the rest of the year, then they must be large enough to store enough energy for the whole year. If whales feed in high latitudes and winter in low latitudes, then they must swim thousands of kilometers, and larger animals can swim more efficiently. All of these factors have acted in concert to select for large size in these largest of animals. The growth of a baleen whale calf is truly extraordinary. During the first half-year of life, a blue whale calf will grow on average 3.5 cm d1 and 80 kg/day (Figure 3). This is supported entirely by the mother’s milk, which comes from fat reserves since the mother is still migrating up to the feeding area. The annual cycle may select for such rapid growth so that the calf is ready to wean during the summer feeding season.

Prolonged Growth and Maturation in Odontocetes Odontocetes do not have as pronounced an annual feeding cycle as baleen whales. Many odontocetes do have a breeding season, but few reproduce on an annual basis. Most odontocetes have a gestation

Sperm Whale

Blue Whale 25

WEANING MALES AND FEMALES

25

15

20 SEXUAL MATURITY MALES

Body length (m)

Body length (m)

20

10

SEXUAL MATURITY FEMALES

5

15 WEANING 10

5

0

0 0

(A)

SEXUAL MATURITY

10

20

30 Age (y)

40

50

0 (B)

10

20

Age (y)

Figure 3 Growth and maturation of (A) sperm whale and (B) blue whale (data from Best, 1979; Lockyer, 1981). Baleen whales live well beyond the 20 years indicated in the figure, which is designed to show ages at which these animals reach full growth. In fact, there is some evidence that bowhead whales may survive for more than a century.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL OVERVIEW

period of about one year, but the larger odontocetes have longer gestation, up to 14–16 months in the sperm and killer whales. Most dolphins and larger toothed whales have a prolonged period of dependency that contrasts particularly strikingly with the record short periods of lactation in some phocid seals. For example, the short-finned pilot whale and sperm whale may suckle young for over 10 years in some cases. The prolonged maturation of toothed whales makes particularly stark contrast with the large baleen whales. Figure 3 shows the growth and maturation of the blue whale versus the sperm whale. The blue whale has weaned by 6 months and is sexually mature by 10 years. By contrast, a male sperm whale suckles for many years and may not reach sexual maturity until 20–25 years of age. Even though baleen whales mature more rapidly than sperm whales, they may live longer. Recent evidence suggests that bowhead whales may live more than a century. Even where lactation may not be this prolonged, there is still often a prolonged period of dependency in odontocetes. For example, bottlenose dolphins in Sarasota, Florida, are sighted with their mothers for 3–4 years, typically until the mother has another calf. There is a paradox in this prolonged dependency. On the one hand, odontocete young are extremely precocial in their senses and motor abilities for vocalizing and swimming. On the other, they appear to need many years of parental care for success. Many biologists believe that this parental care does not just involve the nutritional care of suckling, but may also involve a long period in which the young learn from the mother and other conspecifics. Among the odontocetes, porpoises and river dolphins live life in the fast lane, weaning within a year of age. These species grow and mature more rapidly than most other odontocetes, and do not rely upon as long a period of dependency.

Navigation and Migration Many marine mammals are highly mobile, and may swim over 100 km d1. Some coastal animals such as sea otters, dugongs, or coastal bottlenose dolphins may have home ranges only kilometers to tens of kilometers in scope. All species are mobile enough to require abilities to navigate. Some ice-loving seals range over kilometers, but must find holes in the ice to breathe through. Failure to find these holes could easily be fatal. When Weddell seals are caught at a breathing hole and transported several kilometers by scientists to a new man-made one, they can swim under thick ice to find the original hole. These seals

613

must be very skilled at timing their dives to be sure they can return to this original hole when they need to breathe. When sunlight is available, antarctic Weddell seals and arctic ringed seals appear to use downwelling light to navigate. It seems that holes and cracks in the ice, and under-ice features, may provide the same kind of cues for landmarks and routes as used by terrestrial animals. If vision is not available seals restrict their diving but appear to be able to use acoustic cues to locate holes. Seals use some sounds to orient, but ringed seals may avoid sounds of stepping or scraping at an airhole. This makes sense, since these sounds may come from a polar bear or Inuit waiting to kill a seal as it surfaces. At close range, even a blindfolded seal can use its vibrissae to center in an airhole. These animals thus rely upon local knowledge and a combination of sensory cues for navigation. Many marine mammals, such as the baleen whales, migrate thousands of kilometers annually. Scientists have suggested that marine mammals may use visual, acoustic, chemical, and even geomagnetic cues to orient for migration, but little is known about the sensory basis of orientation or migration. Most whale calves will stay with their mother on their initial migration from calving ground to feeding ground, so they may have an opportunity to learn about migration routes. When bowhead whales migrate, they make low-frequency calls, apparently to coordinate movements and to detect ice obstacles in their path. Most pelagic odontocetes live for several years in the group in which they were born, providing opportunities for learning about navigation. Elephant seals also have migrations of thousands of kilometers, but a young seal is left on the beach by its mother. When a weaned seal leaves the beach, it is thought to leave alone and to have to learn diving, foraging, and orientation on its own. The tracks of satellite-tagged migrating seals and whales are often remarkably well-oriented. Coupling such tags with experimental manipulations may illuminate the sensory basis of migration in marine mammals.

Conservation of Marine Mammals Many populations of marine mammals have been endangered by humans. They have traditionally been tempting targets for commercial hunting because they carry quantities of valuable meat and fat or oil, and because humans can catch them when they surface to breathe. Species that hauled out on land were particularly vulnerable. Human hunters drove the Steller’s sea cow (Hydrodamalis gigas) to extinction in about 25 years in the mid-eighteenth century. In

(c) 2011 Elsevier Inc. All Rights Reserved.

614

MARINE MAMMAL OVERVIEW

the twentieth century most of these commercial hunts were regulated to protect populations. The US Marine Mammal Protection Act was passed in large measure to reduce the unintentional killing of dolphins in a tuna fishery in the eastern tropical Pacific. Marine mammals are still killed during fishing activities or by ghost gear. Many of the great whales remain so endangered from commercial hunting that no human take is allowed. In spite of the prohibition on taking severely endangered whales such as the northern right whale, of which fewer than 300 individuals survive, right whales are regularly killed by entanglement in fishing gear or by collision with large ships. As direct mortality has been increasingly controlled, newer threats to marine mammals have also surfaced. Humans have treated waterways as dumping grounds for toxic wastes, and most contaminants in water ultimately enter the sea. Some marine mammals carry heavy loads of organochlorine compounds such as polychlorinated biphenyls (PCBs) or DDT and toxic elements such as mercury or cadmium. We know little about what pathologies may be linked to these exposures, but some evidence suggests associations between impaired reproduction and exposure to some organochlorine compounds. Whales have died after eating fish contaminated with saxitoxin from a harmful dinoflagellate bloom. Many human seafaring activities also create noise pollution. Humans use sound to explore the oceans with sonar and use intense sounds to explore geological strata below the seafloor. The motorized propulsion of ships has increased ocean noise globally, and underwater explosions have killed endangered whales and river dolphins. All of these threats can be considered forms of habitat degradation. There has also been growing concern about the effects of climate change upon some marine mammals. Many marine mammals depend upon sea ice. In years with little sea ice, harp seals may have decreased reproduction and increased mortality, and polar bears may not be able to feed as well, leading to lower reproductive rates. If global warming reduces the amount or quality of ice used by marine mammals, this may degrade or eliminate critical habitats. Most current laws to protect marine mammals were designed to prevent humans from killing animals directly. Since marine mammals sample most trophic levels of marine ecosystems and can be counted and observed at the surface by

humans, many biologists consider marine mammals to be indicator species, helping us to monitor the health of marine ecosystems. New ways of thinking will be required to protect marine mammal populations from habitat degradation.

See also Acoustic Noise. Acoustics in Marine Sediments. Anthropogenic Trace Elements in the Ocean. Baleen Whales. Bioacoustics. Chlorinated Hydrocarbons. Copepods. Inherent Optical Properties and Irradiance. Krill. Marine Mammal Diving Physiology. Marine Mammal Evolution and Taxonomy. Marine Mammal Trophic Levels and Interactions. Marine Mammals: Sperm Whales and Beaked Whales. Metal Pollution. Phytoplankton Blooms. Primary Production Distribution. Primary Production Processes. Sea Ice. Sea Ice: Overview. Sea Otters. Seals. Sirenians. Sonar Systems.

Further Reading Berta A and Sumich JL (1999) Marine Mammals: Evolutionary Biology. San Diego: Academic Press. Best PB (1979) Social organization in sperm whales, Physeter macrocephalus. In: Winn HE and Olla BL (eds.) Behavior of Marine Mammals: Current Perspectives in Research, vol. 3: Cetaceans, pp. 227--289. New York: Plenum Press. Castellini MA, Davis RW, and Kooyman GL (1992) Annual Cycles of Diving Behavior and Ecology of the Weddell Seal. Bulletin of the Scripps Institution of Oceanography. San Diego: University of California Press. Le Boeuf BJ and Laws RM (eds.) (1994) Elephant Seals: Population Ecology, Behavior and Physiology. Berkeley: University of California Press. Lockyer C (1981) Growth and energy budgets of large baleen whales from the Southern Hemisphere. Food and Agriculture Organization of the United Nations Fisheries, Series 5, Mammals in the Seas 3: 379--487. Kanwisher J and Ridgway S (1983) The physiological ecology of whales and porpoises. Scientific American 248: 110--120. Reynolds JE III and Rommel SA (eds.) (1999) Biology of Marine Mammals. Washington, DC: Smithsonian Press. Rice DW (1998) Marine Mammals of the World: Systematics and Distribution. Lawrence, KS: Society for Marine Mammalogy. Riedman M (1991) The Pinnipeds: Seals, Sea Lions and Walruses. Berkeley: University of California Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION P. L. Tyack, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1621–1628, & 2001, Elsevier Ltd.

Introduction All animals face the same basic behavioral problems – obtaining food, avoiding predators and parasites, orienting in the environment, finding and selecting a mate, maintaining contact with relatives and group members. When the ungulates and carnivores that ultimately evolved into today’s marine mammals entered the sea, the basic problems did not change, but the context in which the animals had to solve them changed radically. Different marine mammal groups have a different balance in the extent to which they use the underwater versus the in-air environments. All sirenians and cetaceans live their entire lives in the sea, and cetaceans show the most elaborate and extreme specializations for life in the sea. All marine mammals other than sirenians, the sea otter, and cetaceans spend critical parts of their lives on land or ice. These other species, including the pinnipeds and polar bear (Ursus maritimus), rely upon land refuges for giving birth, and for taking care of the young. The ocean is a hostile environment for airbreathing mammals. There is little room for error – if an animal misjudges a dive or becomes incapacitated, it may have only minutes to correct the error or there is a risk of drowning. In the days of sail, humans responded to the notion that ‘the sea is a harsh mistress’ with an apprenticeship system, whereby a young cabin boy spent years learning the ropes before being entrusted to make the decisions required of a captain. Some cetaceans have similarly long periods of dependency when the young can learn how to feed, avoid predators, dive, and orient within their natal group. Pilot whales and sperm whales may even continue to suckle up to 13–15 years of age, and pilot whale females have a postreproductive period when they switch their reproductive effort fully to parental care. By contrast, some seals that suckle on land have drastically curtailed the period of lactation, so that their young can leave the land-based refuge early. The hooded seal suckles her young twice an hour for an average of

just four days before the pup is weaned. Even seals that lactate longer may still leave the young to an early independence. Elephant seals are deep divers on a par with the sperm whale, yet they must learn to navigate the sea alone. When an elephant seal pup is weaned, the mother leaves the pup on the beach. Pups spend about 2.5 months on the rookery, learning to swim and dive. They must fend for themselves as they make their first pelagic trip, lasting about four months. This solo entry into the sea exerts a heavy cost; fewer than half of the pups survive this trip.

Feeding Behavior Behavioral ecologists divide feeding behavior into a sequence of steps: searching for prey, pursuit, capture, and handling prey. Marine mammals use many different senses to detect their prey. Walruses, sea otters, and gray whales feed on benthic prey. The walrus uses vibrissae in its mustache to sense shells in the mud; experiments with captive walrus show that they can use vibrissae to determine the shape of objects. Most seals are thought to use vision to find their prey – seals that feed at depth have eyes adapted to low light levels. Most toothed whales produce click sounds, and those species tested in captivity have sophisticated systems of echolocation. When a dolphin detects a target and closes in, it usually increases the repetition rate of its clicks into a buzz sound. If other animals intercept this sound, they may learn about prey distribution whether or not the echolocating animal wants to broadcast this information. Captive experiments show that one dolphin can detect an object by listening to echoes from the clicks of another dolphin. Both of these features of echolocation may favor coordinated social feeding. Many questions remain about how some marine mammals search for prey; we have no idea how baleen whales find patches of prey in the water column. Most marine mammals catch and process prey using their teeth, as most mammals do (Figure 1A–C), but the mysticete whales have baleen, a sievelike set of plates that descend from the lower jaw (Figure 1D), instead of teeth. Baleen whales are able to engulf many prey items in one gulp and then use this baleen to strain out the sea water. Baleen whales feed on patches of prey, and will often aggregate when they are feeding on large patches; the larger the patch, the more whales in the group. Whales feeding on

(c) 2011 Elsevier Inc. All Rights Reserved.

615

616

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

Grinding molars

Shearing molars (B)

(A)

Piercing teeth Sieving plates (C)

(D)

Figure 1 Feeding mechanisms of baleen whales and toothed marine mammals. (A) Polar bears, otters, and most seals have shearing teeth similar to those of terrestrial carnivores. (B) Sirenians have grinding molars that are replaced throughout life. (C) Toothed whales are homodonts; some beaked whales have only one tooth per row, most dolphins have dozens per row. (D) Baleen of a balaenid whale. Adapted from Figures 2–17 of Reynolds and Rommel (1999).

slow-moving prey, such as copepods, may feed in loose aggregations; those feeding on more mobile prey, such as schooling fish, may feed in a more coordinated fashion. Humpback whales may feed in stable groups in which each individual plays a distinct role. Toothed whales tend to feed on mobile prey such as fish and squid, and they must catch one prey item at a time in their mouths. When dolphins feed on schooling fish, they usually feed socially in a coordinated group. Some dolphins appear to corral the school near the surface, while others swim through capturing a few fish. Marine mammalogists have often described these groups as cooperative, but little is known about the costs and benefits of social feeding, so one must be careful to distinguish between coordinated feeding behavior and behavior following more complex models of cooperation. The normal usage of ‘cooperation’ in English is just to work together toward a common goal; in ecology, studying cooperation demands measuring the costs and benefits of different behaviors. Different evolutionary models of cooperation involve different kinds of exchanges. One simple model covers the situation in which animals feeding together on a patch may each feed more efficiently than if they were feeding alone – this may apply to some coordinated feeding in whales. One-sided cooperation may be favored when one animal provides benefits to related animals; kin selection of this sort favors

parental care and even cooperation with more distant relatives under some circumstances. More complex models of cooperation between unrelated animals involve separated interactions of reciprocation, in which one animal may provide a benefit to a partner, expecting the partner to reciprocate at some later date. When a lone dolphin discovers a patch, it may direct a feeding call to other animals, which approach to join in. Feeding calls of this sort may evolve through reciprocation. If search costs are high and if a dolphin discovers a patch too large for it to consume, it may benefit from calling to partners in the expectation that the partners may reciprocate. Sea otters use tools to process food; they dive to catch animals with hard shells, and then bring the prey to the surface, where they break it open on a stone that they carry while foraging. Some marine mammals may also have social mechanisms for processing prey. Some toothed whales catch large oceanic fish such as mahi-mahi (Coryphaena spp.), which are almost as big as they are. It may take one animal to hold the prey and another to rip off pieces in order to consume the flesh efficiently.

Defense from Predators Marine mammals have different strategies for defending themselves from predators. Many of the toothed whales are thought to use their social groups

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

to increase their probability of detecting predators and to protect vulnerable members of the group. Behavioral ecologists studying many animals have found that animals feeding in a group spend less time than lone animals breaking off from feeding to look for predators, and that they also detect the predators at larger ranges, enabling more successful escape strategies. K. Norris has urgued that schools of dolphins function in this way to integrate sensory information from each member of the group using social communication to bolster the sensory abilities of each individual. Some of the large toothed whales use a social defense from predators once the predators are detected. The best observations come from sperm whales. The most dangerous predators of sperm whales are killer whales and human whalers. When killer whales attack a sperm whale group, the adult sperm whales circle around the young with their flukes facing out, and they will attack approaching killer whales with their flukes. If an adult is injured, the sperm whales will circle around that animal to protect it as well. Human whalers knew about this, and they would often harpoon a whale in the group to wound it and then kill each adult in turn as the whales remained nearby to protect the wounded whale. Other marine mammals appear to have a strategy opposite to grouping for protection from predators. When humpback whale females have a young calf, they do not join with other females but seem to space themselves out in protected clear waters. You might think that these whales are so big they do not need help to defend themselves from predators, but sperm whales are almost as big, yet rely upon social defense. Smaller animals such as seals may also spread out rather than group in response to predator pressure and some seals appear to dive in response to predators, adding a third dimension to this response. The ocean does not have many hiding places, but is so vast that these responses may represent a strategy to spread out to avoid detection.

Finding and Selecting a Mate Evolutionary biologists assume that selection acts to maximize lifetime reproductive success, including both an animal’s own offspring and those of relatives, weighted according to the degree of genetic relationship. Mating behavior is closely related to this goal. The mating behavior of male and female mammals differs in part because of the physiology of mammalian reproduction. Female mammals gestate their young internally and provide nutrition after

617

birth through lactation. Once a female has become pregnant, she must usually wait until her young is born and often even weaned before becoming receptive again. This means that reproduction in females is limited primarily by their ability to gather energy to produce young. While male mammals can provide parental care, it is more common for them not to do so, and paternal care is not known among marine mammals. A male is capable of inseminating another female soon after a previous mating. Males often compete for the opportunity to mate with females, and reproductive success in males is often limited by the number of females they can inseminate. The ratio between the number of receptive males to the number of receptive females is called the operational sex ratio. The more receptive males per female, the higher the selection pressure for competition between males for access to females. Females have a variety of mating strategies, where ‘strategy’ is used in an evolutionary rather than purely cognitive sense. A female may have one or more estrous cycles per year, and these may be spontaneous or induced by copulation. A receptive female may search for and select a male either on the basis of specific resources the male may provide, or of judgments of male quality as a mate. A female who is not relying upon resources provided by a male may either elicit competition between the sperm of several males with which she has mated, or incite behavioral competition between males and select the winner. There are also a variety of mating strategies available to males. A male may defend, from other males, resources used by females. An example of this occurs with fur seals, where males will defend areas on the beach that females use to give birth and for thermoregulation. The goal of the male strategy is to exclude other males from their territory, in hopes that females coming to their territory for the resources will become receptive there. Alternatively, males may defend females directly. Elephant seal females may cluster on a beach. Males establish a dominance hierarchy before the breeding season, and the dominant few males defend receptive females from other males. In bottlenose dolphins, coalitions of 2 or 3 males also have been observed to defend females to prevent them from choosing another mate. Males within a coalition have a strong bond, and are typically sighted together most of the time for many years. The males may chase and herd a female away from the group in which they initially find her, and may escort her for days. Why does this involve coalitions of males? It may be impossible for a lone male to preempt female choice with such maneuverable animals in a three-dimensional medium.

(c) 2011 Elsevier Inc. All Rights Reserved.

618

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

Male strategies often depend upon the distribution of females. In the bottlenose dolphin case, females live in small fluid groups and males search for one receptive female, often guarding her when found. By contrast, sperm whale females live in stable groups of several females with their young. Adult male sperm whales join with these groups for varying amounts of time during the breeding season. Computer models suggest that males should rove between female groups if the duration of estrous is greater than the time it takes to swim between groups. This illustrates a general pattern that the distribution of females is often driven by the distribution of resources, while the distribution of males during the breeding season is often driven by the distribution of receptive females. Most of the strategies listed above can be viewed as strategies used by males to limit or preempt the ability of a female to select a mate. In situations where males have a limited ability to preempt this choice, males may evolve signals to attract mates and may display to influence female choice. If females select mates on the basis of the display, then the male displays may be under a strong selection pressure to develop whatever features are used by females to select a mate. Darwin distinguished this kind of sexual selection from natural selection. Sexual selection stems from competition between members of the sex with the least parental investment (typically males) for access to mating with the sex that provides the largest parental investment (typically females). Earlier paragraphs described selection arising from competition between males; this is called intrasexual

selection, and often leads to the development of weapons and large body size. Selection arising from competition between males to influence female choice is called intersexual selection. Intersexual selection often leads to the evolution of elaborate displays, called reproductive advertisement displays. Examples of reproductive advertisement displays include the songs of birds and whales. Songs are usually defined as acoustic displays in which sequences of discrete sounds are repeated in a predictable pattern. Songs are known from a variety of marine mammals. The songs of the humpback whale are well known and sound so musical to our ears that they have been commercial bestsellers. Male humpback whales sing for hours, usually when they are alone during the breeding season. Bowhead whales, Balaena mysticetus, produce songs that are simpler than those of humpbacks, consisting of a few sounds that repeat in the same order for many song repetitions (Figure 2). As with humpback song, bowhead songs appear to change year after year. Bowhead whales winter in the Bering Sea, and humans have seldom studied them during their winter breeding season, but their songs have been recorded during their spring migrations past Point Barrow, Alaska. The long series of 20 Hz pulses produced by finback whales may also be a reproductive advertisement display. The seasonal distribution of these 20 Hz series has been measured near Bermuda, and it matches the breeding season quite closely. Some pinnipeds also repeat acoustically complex songs during the breeding season. The bearded seal,

1980 1000 400 100

Frequency (Hz)

1985 1000 400 100 1986 1000 400 100 1988 5000 1000 50

0

10

20

30 Time (s)

40

50

60

Figure 2 Spectrogram of the song of bowhead whales, Balaena mysticetus, recorded over 4 years during their spring migration. From Figure 4.10 of Tyack and Clark (2000).

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

singer, and many acoustic features of these songs are consistent with a role in female choice.

Maintaining Contact with Recognition Calls Mother–Infant Recognition

When mammalian young are born, they need to suckle for nutrition, and many species depend upon the mother for thermoregulation and protection from parasites and predators. This dependency has created a selection pressure for a vocal recognition system to regain contact when mother and offspring are separated. These problems of recognition between mother and young are acute in colonially breeding otariid seals. A female otariid may leave her young pup on land in a colony of hundreds to thousands of animals, feed at sea for a day or more and, when she returns, must find her pup to feed it. In the Galapagos fur seal, Arctocephalus galapagoensis, both mother and pup produce and recognize distinctive contact calls, and the mother often makes a final olfactory check before allowing a pup to suckle. The young of many dolphin and other odontocete species are born into groups comprising many adult females with their young, and they rely upon a mother–young bond that is more prolonged than that of otariids. As was described in the introduction, many of these species have unusually extended parental care. Bottlenose dolphin calves typically remain with their mothers for 3–6 years. These dolphin calves are precocious in locomotory skills, and swim

kHz

No modulation

n

Si inte lent rva l

Ph ras e2 a

3 ra se Ph

Ph ra

Ph ra se

se 2

1

Types of frequency modulation superimposed on the carrier frequency drawn below

Mo a

Erignatus barbatus, produces a sirenlike warbling song that includes rapid frequency modulation superimposed upon slower modulation of the carrier frequency (Figure 3). The songs of bearded seals are produced by sexually mature adult males during the breeding season. Male walruses, Odobenus rosmarus, also perform visual and acoustic advertisement displays near herds of females during the breeding season. Males inflate modified pharyngeal pouches that can produce a metallic bell-like sound. When walruses surface during these displays, they may make loud sounds in air, including knocks, whistles, and loud breaths. They then dive, producing distinctive sounds under water, usually a series of sharp knocks followed by the gonglike or bell-like sounds. Antarctic Weddell seals also have extensive vocal repertoires and males repeat underwater trills during the breeding seasons. Males defend territories on traditional breeding colonies. These trills have been interpreted as territorial advertisement and defense calls. There is evidence that marine mammal songs play a role both in male–male competition and in female choice. Evidence for intrasexual selection includes observations that aggressive interactions between singers and other males are much more commonly observed than sexual interactions between singers and females and song appears to maintain distance between singers. However, just because the responses of males to song may be seen more frequently than those of females does not mean that the subtler responses of females to singers are not biologically significant. In many species, females will approach and join with a

619

4 3 2 1

0

10

20

30

50 40 Time (s)

60

70

80

Figure 3 Lower panel: spectrographic portrayal of songs of the bearded seal, Erignatus barbatus. Additional frequency modulation is added to this carrier frequency; this warbling modulation is indicated in the upper panel. From Figure 1 of Ray et al. (1969).

(c) 2011 Elsevier Inc. All Rights Reserved.

620

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

out of sight of the mother within the first few weeks of life. Young calves often swim with animals other than the mother during these separations. The combination of early calf mobility and prolonged dependence selects for a mother–offspring recognition system in bottlenose dolphins. Dolphin mothers and their young calves use tonal whistles as signals for individual recognition. Observations of captive dolphins suggest that whistles function to maintain contact between mothers and young. When a dolphin mother and her young calf are separated involuntarily in the wild, they whistle at higher rates; during voluntary separations in the wild, the calf often whistles as it returns to the mother. Experimental playbacks to wild dolphins show that mothers and their calves respond preferentially to each others’ signature whistles, even after the calves become independent from their mothers.

Individual Recognition in Dolphins and Signature Whistles

Dolphins use whistles not just for mother–infant recognition but also for individual recognition throughout their lives. Calves show no reduction in whistling as they wean and separate from their mother. Adult males are not thought to provide any parental care, but they whistle just as much as adult females. Bottlenose dolphins may take up to two years to develop an individually distinctive signature whistle, but once a signature whistle is developed, it can be stable for decades (Figure 4). The signature whistles of dolphins are much more distinctive than similar recognition signals produced by other mammals. These results suggest that signature whistles Mother no. 16

1976

continue to function for individual recognition in older animals. Group Recognition in Killer Whales

Many marine mammals live in kin groups, and social interactions within these groups may have a powerful effect on fitness. The different structures of these cetacean societies create different kinds of problems of social living, and there appears to be a close connection between the structure of a cetacean society and the kinds of social communication that predominate in it. For example, stable groups are found in fish-eating or Resident killer whales, Orcinus orca, in the coastal waters of the Pacific Northwest of the United States, and these whales also have stable group-specific vocal repertoires. The only way a group of these killer whales, called a pod, changes composition is by birth, death, or rare fissions of very large groups. The vocal repertoire of killer whales includes discrete calls, which are stereotyped with acoustic features that change slowly and gradually over decades. Each pod of Resident killer whales has a group-specific repertoire of discrete calls. Each whale within a pod is thought to produce the entire call repertoire typical of that pod. Different pods may share some discrete calls, but none share the entire call repertoire. Since discrete calls change gradually, pods that diverged more recently produce more similar versions of some calls (Figure 5). The entire repertoire of a pod’s discrete calls can thus be thought of as a group-specific vocal repertoire. Different pods may have ranges that overlap and pods may associate together for hours or days before diverging. These group-specific call repertoires in killer whales are thought to indicate pod Female calf no. 140 Born 1984

10

Recording date

0 0.5

1985

1.0

10

10

0

0 0.5

1987

1.0

0.5

1.0

0.5

1.0

10

10

0

0 0.5

1.0

Figure 4 Spectrograms of signature whistles from one wild adult female bottlenose dolphin, Tursiops truncatus, recorded over a period of 11 years and of one of her calves at 1 and 3 years of age. Note the stability of both signature whistles. From Figure 11.11 of Mann et al. (2000).

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL SOCIAL ORGANIZATION AND COMMUNICATION

N8i-A1, A4, A5, H

N7i-A1, A4, A5 8 6 4 2 1

1

2

2

N8ii-C, D

N7ii-A1, A4, A5, H, I1

individual social bonds; group-specific vocal repertoires have been reported for species such as killer whales with stable groups, and population-specific advertisement displays have been reported among species such as humpback whales and some seals where adults appear to have neither stable bonds nor stable groups.

8

See also

6

Baleen Whales. Bioacoustics. Marine Mammal Migrations and Movement Patterns. Marine Mammal Overview. Marine Mammal Trophic Levels and Interactions. Marine Mammals: Sperm Whales and Beaked Whales. Sea Otters. Seals. Sirenians.

4

Frequency (kHz)

621

2 1

2

3

N8iii-B, I1

N7iii-B, H 8 6

Further Reading

4 2

8

N7iv-C, D

N8iv-B, I1

6 4 2

Time

0

500 ms

Figure 5 Spectrogram of group-specific subtypes of calls N7 and N8 of resident killer whales, Orcinus orca. Not only does each group have a group specific call repertoire, but there are different subtypes for many calls, and different sets of groups produce each subtype. On the upper left of each spectrogram is the subtype of the call and following the dash is a listing of the groups that make this particular call. From Figure 1.11 of Au et al. (2000).

affiliation, to maintain pod cohesion, and to coordinate activities of pod members. Correlation of Acoustic Recognition Signals and Social Organization

Most communication signals evolve for the solution of specific problems of social life. In fact, communication and social behavior can be viewed as two different ways of looking at the same thing. There is a clear correlation between the types of social bonds and recognition signals seen in different cetacean groups. Individual-specific signals have been reported for species such as bottlenose dolphins with strong

Bradbury JW and Vehrencamp SL (1998) Principles of Animal Communication. Sunderland, MA: Sinauer Associates. Norris KS, Wu¨rsig B, Wells R, and Wu¨rsig M (1994) The Hawaiian Spinner Dolphin. Berkeley: University of California Press. Mann J, Connor R, Tyack PL, and Whitehead H (2000) Cetacean Societies: Field Studies of Dolphins and Whales. Chicago: University of Chicago Press. Ray C, Watkins WA, and Burns JJ (1969) The underwater song of Erignathus (Bearded seal). Zoologica 54: 79--83. Reynolds JE III and Rommel SA (eds.) (1999) Biology of Marine Mammals. Washington DC: Smithsonian Press. Trillmich F (1981) Mutual mother–pup recognition in Gala´pagos fur seals and sea lions: cues used and functional significance. Behaviour 78: 21--42. Tyack PL (1986) Population biology, social behavior, and communication in whales and dolphins. Trends in Ecology and Evolution 1: 144--150. Tyack PL and Sayigh LS (1997) Vocal learning in cetaceans. In: Snowdon CT and Hausberger M (eds.) Social Influences on Vocal Development, pp. 208--233. Cambridge: Cambridge University Press. Tyack PL and Clark CW (2000) Communication and acoustic behavior of dolphins and whales. In: Au WWL, Popper AS, and Fay R (eds.) Hearing by Whales and Dolphins. Springer Handbook of Auditory Research Series, pp. 156--224. New York: Springer Verlag. Whitehead H (1990) Rules for roving males. Journal of Theoretical Biology 145: 355--368. Watkins WA, Tyack P, Moore KE, and Bird JE (1987) The 20-Hz signals of finback whales (Balaenoptera physalus). Journal of the Acoustical Society of America 82: 1901--1912. Zahavi A and Zahavi A (1997) The Handicap Principle. Oxford: Oxford University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS A. W. Trites, University of British Columbia, British Columbia, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1628–1633, & 2001, Elsevier Ltd.

Introduction Trophic levels are a hierarchical way of classifying organisms according to their feeding relationships within an ecosystem. By convention, detritus and producers (such as phytoplankton and algae) are assigned a trophic level of 1. The herbivores and detritivores that feed on the plants and detritus make up trophic level 2. Higher order carnivores, such as most marine mammals, are assigned trophic levels ranging from 3 to 5. Knowing what an animal eats is all that is needed to calculate its trophic level. Marine mammals are commonly thought to be the top predator in marine ecosystems. However, many species of fish occupy trophic levels that are on par or are above those of marine mammals. Some species such as killer whales and polar bears (that feed on other marine mammals) are indeed top carnivores, but others such as manatees and dugongs feed on plants at the bottom of the food web. Thus, marine mammals span four of the five trophic levels. Marine mammals are a diverse group of species whose behaviors, physiologies, morphologies, and life history characteristics have been evolutionarily shaped by interactions with their predators and prey. It is therefore difficult to generalize about how marine mammals affect the dynamics and structure of their ecosystems. Similarly, it is difficult to generalize about how the interactions between marine mammals and their prey (or between marine mammals and their predators) affect one another, as well as how they affect the dynamics of unrelated species. Nevertheless, some insights into marine mammal trophic interactions can be gleaned from mathematical models and from field observations following the overharvesting of marine mammal populations in the nineteenth and twentieth centuries.

Trophic Levels (Diet Composition) Trophic levels depend on what a species eats. As an example, a fish consuming 50% herbivorous-

622

zooplankton (trophic level 2) and 50% zooplanktoneating fish (trophic level 3) would have a trophic level of 3.5. Trophic levels (TL) can be calculated from Pn TL ¼ 1 þ

ðTL  DCi Þ i¼1 Pn i i¼1 DCi

½1

where n is the number of species or groups of species in the diet, DCi is the proportion of the diet consisting of species i and TLi is the trophic level of species i. Thus, the trophic level of the predator is determined by adding 1.0 to the average trophic level of all the organisms that it eats. Applying eqn [1] to marine mammals shows that sirenians (dugong and manatees) have a trophic level of 2.0, whereas blue whales (which feed on large zooplankton, trophic level 2.2) are at trophic level 3.2 ( ¼ 1.0 þ 2.2). Moving higher up the food chain, Galapagos fur seals have a trophic level of 4.1. Their diet consists of approximately 40% small squids, 20% small pelagic fishes (such as clupeoids and small scombroids), 30% mesopelagic fishes (myctophids and other groups of the deep scattering layer) and 10% miscellaneous fishes (from a diverse group consisting mainly of demersal fish). Substituting these proportions into eqn [1], along with the respective mean trophic levels (TLi ) of these four types of prey (3.2, 2.7, 3.2, and 3.3, respectively), yields a trophic level of 4.11 for Galapagos fur seals. A polar bear that feeds exclusively on ringed seals (3.8) would have a trophic level of 4.8. Dugongs and manatees occupy the lowest trophic level (2.0) of all marine mammals. They are followed (see Figure 1) by baleen whales (3.35: range 3.2–3.7), sea otters (3.45: range 3.4–3.5), pinnipeds (3.97: range 3.3–4.2), and toothed whales (4.23: range 3.8– 4.5), with the highest trophic level belonging to the polar bear (4.80). Trophic interactions between marine mammals and other species can be depicted by flowcharts showing the flow of energy between species in an ecosystem. An example is shown in Figure 2 for the eastern Bering Sea. Each of the boxes in this flowchart represents a major species or group of species within this system during the 1980s. The boxes are arranged by trophic levels and are proportional in size to their biomass. Lines connecting the boxes show the relative amounts of energy flowing between the groups of species.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS

(Sub) Order or Family

No. of species

Odontoceti

(64)

Pinnipedia

(35)

Mustelidae

(2)

Mysticeti

(11)

Physeteridae and Kogiidae

(3)

Ziphiidae

(19)

Delphinidae

(32)

Platanistidae

(2)

Monodontidae

(2)

Phocoenidae

(6)

Otariidae

(15)

Phocidae

(19)

Mustelidae

(2)

Balaenopteridae

(6)

Odobenidae

(1)

Eschrichtidae

(1)

Balaenidae and Neobalaenidae

(4)

Trophic level

4.23

3.0

3.5

4.0

623

Common name

Toothed whales

3.97

Seals, sea lions and walruses

3.45

Otters

3.35

Baleen whales

4.37

Sperm whales

4.30

Beaked whales

4.21

Ocean dolphins

4.10

River dolphins

4.10

Beluga and narwhal

4.08

Porpoises

4.03

Eared seals

3.95

True seals

3.45

Otters

3.43

Rorquals

3.40

Walrus

3.30

Gray whale

3.20

Right and bowhead whales

4.5

Trophic level Figure 1 Mean trophic levels for 112 species of marine mammals grouped by families, orders and suborders. Numbers of species averaged within each grouping is shown in brackets. Species not shown are dugong and manatees (Sirenia: trophic level 2.0) and polar bears (Ursidae: trophic level 4.8).

Figure 2 shows a large number of flows in the Bering Sea emanating from three species at trophic level 3 – pollock, small flatfish and pelagic fishes. Major level 4 consumers include large flatfish, deepwater fish, other demersal fishes, marine mammals and birds. Thus, large flatfish and other species of fish share the pedestal with marine mammals as top predators of marine ecosystems. These fish are also major competitors of marine mammals. Trophic levels depicted in Figures 1 and 2 are approximate, and are based on generalized diets and the mean trophic levels of prey types. In actual fact,

trophic levels of most marine mammals probably vary from season to season, or from year to year, because diet is unlikely to remain constant. How much they might vary is not known, but is probably within 70.2 trophic levels.

Trophic Levels (Stable Isotopes) Diets have traditionally been described from stomach contents of shot or stranded animals. This has been augmented by the identification of prey from

(c) 2011 Elsevier Inc. All Rights Reserved.

624

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS

5

Flow Connector Harvest Flow to detritus Respiration

Sperm whales

Beaked whales

0.75

0.01 Toothed whales 0.09

0.19

4

0.84

Baleen whales

Trophic level

0.001

Seals

0.004

3.58

Steller sea lion

0.66

5.26

0.001 Other demersal fish

0.007

0.050

0.28 Piscivorous birds

0.128

Large flatfish 2.96

Deep-water fish 0.65

Walrus bearded seal

18.67

Jellyfish

0.0

44.25 25.00

0.212 Small flatfish 0.3 0.108

18.09

.95

27.10

Epifauna Large Zooplankton Herbivorous zooplankton

Benth.P. Feeders

0.3

12.75

2.0

Infauna 521.36 270.96

638.00

1

7.5

0.211 Pelagics

Juvenile pollock 0-1

3

2

Cephalopods

1.895 Adult pollock2+

0.04

0.0

0.3

Phytoplankton

Detritus

Figure 2 Flowchart of trophic interactions in the eastern Bering Sea during the 1980s. All flows are in t km2 y1. Minor flows are omitted as are all backflows to the detritus. Note that size of each box is roughly proportional to the biomass therein, and that each box is placed according to its trophic level in the ecosystem.

the bony remains found in feces (primarily from pinnipeds), and from fatty acid signatures of prey species that have been laid down in the blubber of marine mammals. Unfortunately, the diets of most species of marine mammals are poorly understood due to incomplete sampling across time and space. There is another way to estimate trophic levels without stomach contents or other dietary information. It is referred to as stable isotope analysis and relies on the relative concentration of two isotopes (nitrogen-14 and nitrogen-15). Marine mammals and other organisms tend to accumulate the heavier isotope (nitrogen-15) in their tissues. Thus, as matter moves from one trophic level to the next, the ratio of the two isotopes shifts by a roughly constant amount. Trophic levels can be calculated by dividing the difference between the isotopic ratio in the marine mammal tissue and the isotopic ratio of the organism at the bottom of the local food chain, by this constant difference between trophic levels. A comparison of the isotopic estimates of trophic levels for species in Prince William Sound, Alaska with estimates derived from dietary analysis (eqn [1]) suggests that the two techniques produce comparable

results. One of the strengths of the isotopic analysis is that it can be conducted from biopsy samples and does not require killing the animal to examine stomach contents. This is particularly useful for assessing the trophic levels of cetaceans. Stable isotope analysis can also be used to probe the past to learn about the trophic levels that marine mammal populations once occupied. As predators, marine mammals are better samplers of the marine environment than biologists. Thus, analyzing seasonal and annual changes in the nitrogen concentrations contained along growing whiskers and baleen can provide a time series of dietary information. Similarly, trophic levels can be calculated from nitrogen concentrations in bones and teeth archived in museums or recovered from archaeological digs. Another useful stable isotope ratio is the relative concentration of carbon-13 and carbon-12. Studies have shown that there is a very slight enrichment of carbon from one trophic level to another (0.1– 0.2%). In the marine environment, slight enrichment occurs at low trophic levels, but not among vertebrate consumers. Thus, isotopic carbon ratios are not useful for assessing trophic level, but they are

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS

useful for tracking carbon sources through a food chain and for assessing long-term changes in ocean productivity. Isotopic analyses of marine mammal tissues have shown that species inhabiting the northern oceans have higher nitrogen-isotope ratios than those from southern oceans. This indicates that southern species feed at lower trophic levels, and presumably consume larger amounts of invertebrates. Measuring the isotopic carbon ratio of baleen plates from bowhead whales further shows that primary productivity declined in the Bering Sea through the 1970s–1990s. This drop in primary productivity may reflect an overall lowering of carrying capacity and may have a bearing on the observed decline of Steller sea lions, harbor seals, and northern fur seals during this period. Thus, isotopic analysis is a useful tool for estimating trophic levels of marine mammals, and for detecting shifts in ocean productivity and diets of marine mammals.

Trophic Interactions Changes at one level of a food web can have cascading effects on others. One of the best ways to explore the direct and indirect impacts of competition and predation by marine mammals on other species is with mathematical descriptions of ecosystems (i.e., ecosystem models). Ecosystem models, such as the one developed for the Bering Sea (Figure 2), allow changes in abundance to be tracked over time, and predictions to be made about the strength and significance of predator–prey interactions on each other, and on other components of their ecosystem. Major changes have occurred in the abundance of a number of species in the Bering Sea since the mid1970s. Most notable has been the decline of Steller sea lions, harbor seals, crabs, shrimp and forage fishes (such as herring and capelin). In contrast, populations of walleye pollock and large flatfish (mostly arrowtooth flounder) increased through the 1970s and 1980s. Some have felt that commercial whaling prompted these changes by removing a major competitor of pollock – the baleen whales. Mathematically, removing whales can be shown to positively affect pollock by reducing competition for food. However, whaling alone is insufficient to explain the 400% increase in pollock that is believed to have occurred. Overall, the models developed to date suggest that changes in the biomass of marine mammals have little or no effect on changes in the biomass of other groups in the Bering Sea. Most impacts on this northern marine ecosystem appear to

625

be associated with changing the biomass of lower trophic levels (such as primary production). The conclusions drawn from the eastern Bering Sea model may be indicative of marine mammals in long-chained food webs, and may not reflect the role of marine mammals in shorter-chained food webs such as in the Antarctic. A case in point is the increase in abundance of krill-eating Antarctic fur seals, crabeater seals, leopard seals, and penguins that followed the cessation of commercial whaling. Commercial whaling removed over 84% of the baleen whales from the Antarctic and ‘freed up’ millions of tons of krill for other species to consume. Some believe that the increase in these other krilleating species is now impeding the recovery of Antarctic whales. Sea otters and sea urchins form another shortchained food web with strong trophic interactions. By the turn of the twentieth century, sea otters had been hunted to near extinction. Without predation by otters, sea urchin populations grew unchecked and overgrazed the fleshy algae along the Pacific coast of North America. The once productive kelp forests became underwater barrens. With the reintroduction of sea otters however, productivity increased three-fold as urchins were removed and kelp and other fleshy algae began to regenerate. Kelp provides habitat for fish and invertebrates, changes water motion, and can affect onshore erosion and the recruitment of fish and invertebrates. Thus, sea otters can change the state of near-shore ecosystems and the way they function. Other examples of marine mammals affecting their prey include harbor seals in freshwater lakes, and killer whales preying on sea otters in Alaska. A number of lakes in Quebec, Canada, are home to land-locked harbor seals that feed on trout. Studies have shown that the trout in these lakes are younger and spawn at younger ages than adjacent lakes without harbor seals. The trout also grow faster and attain smaller sizes in the lakes inhabited by harbor seals. Marine mammals may also significantly affect prey abundance, as in the case of killer whales eating sea otters, Steller sea lions, and other warm-blooded species. Killer whales were observed eating sea otters along the Aleutian Islands in the 1990s and may be responsible for reported declines in sea otter population abundance. Killer whales have also been implicated as a contributing factor in the decline of Steller sea lions and may be impeding their recovery. Despite the apparent effects of some species of marine mammals on their prey, there are a number of cases where mass removals of marine mammals did not appear to have a major effect on other

(c) 2011 Elsevier Inc. All Rights Reserved.

626

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS

components of their ecosystems. Examples are the overhunting of elephant seals and California sea lions along the coast of California, the overhunting of northern fur seals in the Bering Sea, and the culling of harbor seals in British Columbia. One explanation for the lack of tractable impacts in these cases is that their food webs are more complex relative to other systems (i.e., predators consuming many different species of prey, may have no noticeable impact on any single prey type). Another reason might be related to the type of marine ecosystems that these species inhabit (i.e., whether they inhabit shelf or deep-water systems, or whether they are primarily benthic or mid-water feeders). Further insights might be gained by developing ecosystem models for these systems. Quantifying the feeding relationships between marine mammals and other species provides a means for assessing competition between species at similar trophic levels. Some species may significantly compete with more than one species. In the Bering Sea, for example (Figure 2), baleen whales and pollock have high overlaps in their diets (73–86%). There is also a significant amount of competition between seals and adult pollock for prey. Toothed whales, for example, compete primarily with beaked whales and seals, whereas the largest competitors of sea lions appear to be seals, toothed whales, and large flatfish. Fish, it turns out, can be major competitors of marine mammals. Competition can affect body growth, reproduction and survival of marine mammals. In the Bering Sea and Gulf of Alaska, for example, the growth of Steller sea lions and northern fur seals (as measured by length) appears to have been stunted during the 1980s compared to the 1970s. Eastern Pacific populations of gray whales also appear to be in poorer condition (as measured by the ratio of girth to body length) in the 1990s compared to earlier decades. These changes in body size may be densitydependent responses to reduced prey availability or may be indicative of populations that have approached or attained their carrying capacities. Reductions in prey abundance have been recorded in the Antarctic (i.e., krill), and along the coasts of California and South America during El Nin˜o events. Pinniped pups born during these periods of reduced prey abundance incur high rates of mortality (typically 2–3 times normal levels) and are weaned at lower weights than normal (typically 15–20% lighter). Lactating females must also spend longer periods of time away from their pups to search for prey. Such temporal changes in prey abundance may result in the loss of an entire year class, and may be

one of the evolutionary forces that shaped the life history of marine mammals (i.e., they are long-lived, have low reproductive rates and can endure shortterm reductions in prey abundance). Although it has not yet been demonstrated for marine mammals, reductions in prey availability can theoretically delay the onset of sexual maturity, and reduce fertility (by causing a female to not ovulate, or by causing a fetus to be reabsorbed or aborted). Reduced nutrition may also compromise an organism’s resistance to disease, and may increase vulnerability to predation. Food deprivation may mean, for example, that a seal must spend increased amounts of time searching for prey and less time hauled out on shore away from predators such as killer whales and sharks.

Conclusions Calculating trophic levels is a necessary first step to quantifying and understanding trophic interactions between marine mammals and other species in marine ecosystems. This can be achieved using dietary information collected from stomachs and scats, or by measuring isotopic ratios contained in marine mammal tissues. These data indicate that marine mammals occupy a wide range of trophic levels beginning with dugong and manatees (trophic level 2.0), and followed by baleen whales (3.35), sea otters (3.45), seals (3.95), sea lions and fur seals (4.03), toothed whales (4.23), and polar bears (4.80). With the aid of ecosystem models and other quantitative analyses, the degree of competition can be quantified, and the consequences of changing predator–prey numbers can be predicted. These analyses show that many species of fish are major competitors of marine mammals. A number of field studies have also shown negative effects of reduced prey abundance on body size and survival of marine mammals. However, there are fewer examples of marine mammal populations affecting their prey due perhaps to the difficulty of monitoring such interactions, or to the complexity of most marine mammal food webs.

See also Baleen Whales. Bioacoustics. Fishery Management. Large Marine Ecosystems. Marine Mammal Diving Physiology. Marine Mammal Evolution and Taxonomy. Marine Mammal Migrations and Movement Patterns. Marine Mammal Overview. Marine Mammal Social Organization and Communication. Marine

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMAL TROPHIC LEVELS AND INTERACTIONS

Mammals, History of Exploitation. Marine Mammals: Sperm Whales and Beaked Whales. Network Analysis of Food Webs. Sea Otters. Seals. Sirenians.

Further Reading Bowen WD (1997) Role of marine mammals in aquatic ecosystems. Marine Ecology Progress Series 158: 267--274. Christenson V and Pauly D (eds) (1993) Trophic Models of Aquatic Ecosystems. ICLARM Conference Proceedings 26. Estes JA and Duggins DO (1995) Sea otters and kelp forests in Alaska: generality and variation in a community ecological paradigm. Ecological Monographs 65: 75--100. Greenstreet SPR and Tasker ML (eds.) (1996) Aquatic Predators and Their Prey. Oxford: Fishing News Books.

627

Kelly JF (2000) Stable isotopes of carbon and nitrogen in the study of avian and mammalian trophic ecology. Canadian Journal of Zoology 78: 1--27. Knox GA (1994) The Biology of the Southern Ocean. Cambridge: Cambridge University Press. Pauly D, Trites AW, Capuli E, and Christensen V (1998) Diet composition and trophic levels of marine mammals. Journal of Marine Science 55: 467--481. Trillmich F and Ono K (1991) Pinnipeds and El Nin˜o: Responses to Environmental Stress. Berlin: SpringerVerlag. Trites AW (1997) The role of pinnipeds in the ecosystem. In: Stone G, Goebel J, and Webster S (eds.) Pinniped Populations, Eastern North Pacific: Status, Trends and Issues, pp. 31--39. Boston: New England Aquarium, Conservation Department. Trites AW, Livingston PA, Mackintosh S, et al. (1999) Ecosystem Change and the Decline of Marine Mammals in the Eastern Bering Sea: Testing the Ecosystem Shift and Commercial Whaling Hypotheses. Fisheries Centre Research Reports 1999, Vol. 7(1).

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS AND OCEAN NOISE D. Wartzok, Florida International University, Miami, FL, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The oceans are much more transparent to acoustic energy than to electromagnetic energy in the frequency ranges at which animal sensory systems operate. Consequently it is not surprising that both human and animal underwater communication systems rely on acoustic transmission. Some baleen whales are presumed to hear as low as 10 Hz and some river dolphins as high as 200 kHz. Therefore virtually all human activity on or in the oceans results in the incidental addition of sound in the hearing range of one or more marine mammal species. The intentional and unintentional introduction of acoustic energy into the oceans by human activities constitutes ocean noise so far as marine mammals are concerned. The issue of marine mammals and sound has been the subject of four National Research Council reports and several special issues of journals. Human-generated sound is not the only sound that constitutes noise for marine mammals. Natural sources of sound include those generated by the interaction of wind with the sea surface, waves breaking on shorelines, precipitation, thunder, earthquakes, and ice cracking. Biological sources of sound include snapping shrimp, fish choruses, and the vocalizations of other marine mammals. Marine mammals have evolved in this cacophony of sounds and have developed mechanisms to compensate for background noise, so they can successfully perform acoustically mediated prey detection, predator avoidance, navigation, and intraspecific communication. However, when human-generated sound exceeds the evolved adaptive capacity though intensity, duration, pervasiveness, particular signal characteristics, or as an additive factor to natural sounds, marine mammals can be negatively affected. While there is a general consensus that anthropogenic ocean sound has increased over the past decades, there are few instances in which ambient noise measurements have been made at the same locations over a span of decades. A few reports suggest that at certain locations in the Northern Hemisphere ambient noise in the 20–80-Hz band has increased

628

approximately 12 dB over the past four decades. Each 3-dB increase represents a doubling of the ambient noise (see XXX for definition of dB). Shipping noise is the predominant contributor to ambient noise in the 20–80-Hz range. Shipping is expected to increase in gross tonnage and speed of transport, both of which will lead to further increases in ocean noise unless quiet ship technologies are incorporated in the design phase of these new vessels. Although naval sonar and seismic exploration activities also transmit much energy into the ocean each year, these sources are at higher frequencies (sonar) and thus do not propagate as far, or are not omnidirectional (sonar and seismic), as is shipping. Thus they make less contribution to the global ambient. It has been difficult to demonstrate specific effects of noise on marine mammals and even more difficult to partition the noise contribution to the suite of natural and anthropogenic factors potentially affecting marine mammals. Multiyear habitat abandonment has been demonstrated for killer whales (Orcinus orca) in a British Columbia archipelago where acoustic harassment devices were employed to protect aquaculture facilities and for gray whales (Eschrichtius robustus) where a Baja California lagoon was abandoned during the period of industrial activities. In both of these cases, the habitat was reoccupied with the cessation of the acoustic harassment. The other clear example of the effect of anthropogenic sound on marine mammals is the stranding of primarily beaked whales (Ziphiidae) in response to mid-frequency naval sonar. Sound has not been considered a significant factor in any of the documented major declines of marine mammal populations, for example, Steller sea lions (Eumetopias jubatus), Aleutian Islands sea otters (Enhydra lutris), Alaskan harbor seals (Phoca vitulina richardsi), and fur seals (Callorhinus ursinus). However, it is important to note that these all involved species much easier to monitor than cetaceans. It has been estimated that the probability of sighting a beaked whale directly on the trackline of the survey vessel is only 23% under ideal conditions and is closer to 2% under typical conditions. Thus population level effects of sound on beaked whales would be difficult to detect if they were occurring. Because of these uncertainties and the difficulty of obtaining definitive data, the National Research Council (2005) concluded ‘‘On the one hand, sound may represent only a second-order effect on the conservation of marine

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS AND OCEAN NOISE

mammal populations; on the other hand, what we have observed so far may be only the first early warnings or ‘tip of the iceberg’ with respect to sound and marine mammals.’’

Zones of Influence Ocean noise affects marine mammals in different ways depending on the characteristics of the sound and its intensity. The influence of particular characteristics of the sound is not well understood. For example, the stranding of beaked whales in the presence of naval mid-frequency sonar appears to depend on characteristics of the sound rather than the received intensity, although exactly what those characteristics are and how they result in stranding are unknown. Figure 1 is a diagram illustrating the range of possible effects of sound on marine mammals based on received intensity of the sound. In most cases, the limit of auditory detectability will be the threshold for any response. Humans can show a range of physiological reactions to nonaudible, lowfrequency noise. Whether marine mammals have similar responses is unknown. If a sound is above the background level and above the auditory threshold at a particular frequency, then the animal can detect that sound. Beluga whales detect the return of their echolocation signal when it is only 1 dB above the background and gray whales react to playbacks of

629

predator killer whale vocalizations when the signal is 0 dB above the ambient. Behavioral Response

Whether the animal responds to the sound and what response the animal has to the sound depend on a suite of factors, including age, gender, season, location, behavioral state, prior exposure, and individual variation. For example, beluga whales are more sensitive to ship noise when they are confined to leads, meters to hundreds of meters wide open water channels between ice sheets, than when not so confined. Migrating gray whales changed their orientation when exposed to a Low Frequency Active sonar that was in their migratory path but ignored the same signal source at even higher received sound levels when it was located seaward of their migratory path. Right whales (Eubalaena glacialis) and fin whales (Balaenoptera physalus) are more tolerant of stationary noise sources than those moving toward them, whereas bottlenose dolphins (Tursiops truncates) show aversive behaviors in response to speedboats and jet skis even when they are not approaching. Humpback whales (Megaptera novaeangliae) respond at lower received levels to stimuli with sudden onset than they do to continuous sound sources. One of the strongest recorded reactions of a marine mammal to a sound is the response of beluga

Zones of noise influence

• Injury − acoustic trauma • Hearing loss − permanent threshold shift • Temporary threshold shift • Avoidance, masking • Behavioral disturbance declining to limits of audibility

Figure 1 Close to an intense source, sound may be loud enough to cause death or serious injury. Somewhat farther away, an animal might have less serious injury, such as hearing loss. Temporary threshold shifts occur at greater distances. Animals may avoid exposures at even greater distances or they may not move from the area but still be affected through masking of important auditory cues from the environment. They may show barely observable behavioral disturbance at distances comparable with the limit of audibility. The different distances for the different effects define different areas for each zone. Reproduced from National Research Council (2005) Marine Mammal Populations and Ocean Noise: Determining When Noise Causes Biologically Significant Effects. Washington, DC: The National Academies Press, with permission from The National Academies Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

630

MARINE MAMMALS AND OCEAN NOISE

whales in the High Arctic to the first icebreaker of the season. Belugas have been recorded fleeing for more than 80 km from the icebreakers and exhibiting a range of other behavioral changes at received sound levels between 94 and 105 dB re 1 mPa, yet 1 or 2 days later showing no reaction to icebreaker sound of 120 dB re 1 mPa. Even greater habituation is shown by belugas in Bristol Bay, Alaska, where they continue to move into the Kvichak River to feed on salmon smolt even when purposely harassed by motorboats. On the other hand, an elephant seal (Mirounga angustirostris) showed sensitization rather than habituation in response to repeated presentations of broadband pulsed signals that were similar to predatory killer whale echolocation clicks. Masking

As the received sound level rises above the limits of detectability and possible behavioral response, the sound has the potential to interfere with the natural uses of sound by marine mammals in that frequency range. The noise-induced reduction of acoustic information with respect to conspecifics, predators, prey, and other environmental clues is termed masking. Mammals have dealt with masking throughout their evolutionary history and have developed a number of ways to maintain normal functions in the presence of masking sounds. Noise is only effective in masking a signal if it is within a certain critical band around the signal’s frequency. The actual degree of masking depends upon the amount of noise energy within this critical frequency band. Although critical bands have been measured in only a few marine mammals, at frequencies above 1 kHz, the critical bands for cetaceans are narrower than they are for most other mammals. In addition to narrower critical bands, marine mammals also reduce masking by directional hearing. Unless the masking sound is directly on axis with the sound of interest, directional hearing provides a significant gain in the signal-to-noise ratio. Odontocetes have good directional hearing above 1 kHz. The directivity index (DI) is a measure of the ability of a receiver to reduce the effects of omnidirectional noise and is expressed as the decibel level above the signal the omnidirectional noise must be increased in order to mask the signal. For bottlenose dolphins, the DI increases from 10.4 dB at 30 kHz to 20.6 dB at 120 kHz. Animals adapt to masking in a number of ways. For example, a beluga whale that was moved to a location with higher levels of continuous background noise increased both the average level and frequency of its vocalizations. A beluga that was required to

echolocate on an object placed in front of a source of noise reduced masking by reflecting its sonar signals off the water surface to ensonify to the object. The strongest echoes from the object returned along a path that was off axis from the noise. This animal’s ready application of such complex behavior suggests the existence of many sophisticated strategies to reduce masking effects. Masking compensation responses in noncaptive situations include: beluga whales increasing call repetition and shifting to higher peak frequencies in response to boat traffic; gray whales increasing the amplitude of their vocalizations, changing the timing of vocalizations, and using more frequency-modulated signals in noisy environments; humpback whales increasing the duration of their songs by 29% when exposed to low-frequency active sonar; and killer whales increasing call duration over time as the number of whale-watching boats increased. No hearing thresholds have been measured in baleen whales. Their vocalizations emphasize frequencies in the same range as that of the increased background noise due to shipping. This raises the possibility of significant masking of important signals, particularly conspecific communication signals for these whales. The adaptations used to enhance signaling in the presence of background noise such as increased amplitude, longer duration, and repetition are not available in the situations where the animals are passively listening for acoustic cues such as waves breaking on distant shores or noises produced by schools of fish prey. These cues can be only a few decibels above the normal ambient and are easily masked by increased noise levels. The passive sonar equation can be used to show that as noise increases by a set amount, the range at which a signal can be detected decreases by a constant proportion, termed the range reduction factor (RRF). For example, a 6-dB increase in noise has an RRF of 2 for transmission loss by spherical spreading and an RRF of 4 for transmission loss by cylindrical spreading. In many situations spatial distribution is important (e.g., breeding humpback whales), and in these situations the reduction in area monitored may be a more realistic measure of the effect of increased noise. The area will be reduced proportional to RFF2. Temporary Threshold Shift

When animals are exposed to more intense sound levels, they experience a temporary loss of hearing sensitivity, or temporary threshold shift (TTS). The amount of the TTS depends on the energy of the

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS AND OCEAN NOISE

stimulus. A reduction in sensitivity in response to an intense or prolonged stimulus is the usual response of a sensory system and within normal ranges is reversible. TTS has been studied extensively in humans and laboratory animals and results from a combination of anatomic and metabolic alterations occasioned by the intense sound. There have been few studies of TTS in marine mammals. Among the odontocetes, only the bottlenose dolphin and the beluga whale have been tested and among the pinnipeds only harbor seals (Phoca vitulina vitulina), California sea lions (Zalophus californianus), and elephant seals. For some species, only a single animal has been tested. Given the small number of animals tested and recognizing the substantial variation in TTS in more extensively studied laboratory species, any conclusions regarding TTS in marine mammals are tentative. A just measurable TTS of about 6 dB has been produced with stimuli ranging from an impulse of 1-ms duration at 226-dB peak-to-peak re 1 mPa at 1-m to 30-min exposure to octave bandwidth noise of 179-dB re 1 mPa. A doubling of exposure time has approximately the same effect as a reduction in intensity of 3 dB. This relationship between exposure time and sound intensity shows that the various stimuli producing a just measurable TTS all delivered approximately the same amount of energy to the auditory system. For a 1 s exposure, this TTS threshold is 195 dB re 1 mPa. When the time course of recovery from TTS was tested immediately following the cessation of the fatiguing stimulus, recovery was within 4–40 min depending on experimental conditions and magnitude of the TTS. Even TTS of 23 dB usually recovered within 30 min of exposure. When the time course of recovery was not measured immediately after cessation, there was full recovery when the animals were next tested 24 h later. Permanent Threshold Shift

As the name implies, an animal does not recover hearing sensitivity when it experiences a permanent threshold shift (PTS). PTSs are both functionally and anatomically different from TTSs. PTS results in losses of the cells that convert sound energy into neural signals. No experiments have been done on marine mammals that have resulted in a PTS. The TTS measured so far is well below the level that would be expected to grade into PTS. For terrestrial animals approximately 40-dB TTS is required for a PTS. As with TTS, PTS is less likely to occur if the noise bursts are shorter and the intervening periods between intense noise are longer.

631

Nonauditory Effects

Direct nonauditory effects of sound have not been demonstrated in marine mammals, but some have been observed in other animals or humans or in vitro. Humans exposed to intense sounds can experience dizziness (Tullio phenomenon), nystagmus, and neurological disturbance in the absence of hearing loss. Acoustic resonance Tissues associated with airfilled cavities can be subjected to shear forces when those cavities resonate. Two important factors for resonance are the relationship between the dimensions of the cavity and the wavelength of the sound and the tuning or amplification of the resonance. The latter is described by the Q value, with a high Q indicating greater resonance amplitude. The resonance frequency of beluga and bottlenose whale lungs has been determined to be 30 and 36 Hz, respectively, with relatively low Q values of 2.5 and 3.1, respectively. Thus, at these low frequencies, there is a modest amplification of the resonance magnitude. The resonance characteristics of other air-filled cavities have not been measured but the consensus is that the magnitude of the resonance is not great enough to cause tissue damage. Rectified diffusion and activation of microbubbles Bubble growth either through acoustically driven rectified diffusion or acoustic activation of micro-bubbles could lead to symptoms similar to decompression sickness in human divers. The basic requirement in either model is that the tissues be highly supersaturated. In some deep-diving marine mammals, such as beaked whales, supersaturation has been calculated to exceed 300%. Beaked whales that stranded subsequent to naval sonar activities have shown bubbles in a number of tissues. However, the current understanding of the exposures of the whales that stranded is that the received sound levels were too low to cause bubble growth or activation (demonstrated at 210 dB re 1 mPa in vitro). A more likely explanation is that a behavioral response to lower received levels initiated the cascade of events resulting in the stranding. Studies in right whales, a species not involved in naval sonar stranding incidents, have shown that received levels of 133 dB re 1 mPa of an unfamiliar signal will cause abrupt changes in diving behavior. Blast injury Blast injuries to marine mammals have only rarely been observed. The best-documented case is that of humpback whales that died within 3 days after detonation of 1700–5000-kg Tovex (a trinitrotoluene (TNT) clone) blasts. The mechanical traumas

(c) 2011 Elsevier Inc. All Rights Reserved.

632

MARINE MAMMALS AND OCEAN NOISE

in the whale ears were consistent with classic blast injuries in humans including round window rupture, ossicular chain disruption, bloody effusion in the ear region, and bilateral periotic fractures. These traumas result from eruptive injury during the rarefactive portion of the shock wave when inner ear fluid pressures are much greater than ambient. Stress Both high-intensity, short-duration stimuli and long-term exposure to much lower levels of noise can result in elevated levels of stress. Some studies conducted in humans have shown that exposure to chronic noise elevates neuroendocrine and cardiovascular indices of stress and results in diminished performance on cognitive tests of reading ability and long-term memory. For most studies in humans, it has been difficult to demonstrate statistically significant effects of chronic noise, although there is a consistent trend toward increased cardiovascular risk if the daytime exposure level exceeds 65 dB(A) (see XXX for a discussion of the differences between in-air and underwater decibel reference levels and measurements). Within this context, the 12-dB increase in low-frequency noise due to shipping in the past four decades could be a stress factor in addition to its role in masking communications. Another component of chronic noise, at least in the North Atlantic, is the long-range propagation of the sounds from seismic surveys. Autonomous hydrophones located near the mid-Atlantic ridge frequently, particularly in the summer, recorded sounds of seismic surveys taking place over 3000 km from the recording location. The effects of such long-term increases in anthropogenic sound on the stress response of marine mammals have not been determined. There have been two studies of the short-term effects of noise on the stress response of marine mammals. One detected no change in behavior or catecholamine levels of captive beluga whales exposed to playbacks of the operating noise from a semisubmersible drilling platform at a source level of 153 dB re 1 mPa at 1 m. In contrast, a beluga whale exposed to high-level (4100 kPa) impulsive sounds and high-intensity tones had significantly elevated norepinephrine, epinephrine, and dopamine levels. A bottlenose dolphin similarly exposed did not show elevated catecholamines, but did show an increase in aldosterone and a decrease in absolute monocyte levels after exposure to a seismic water gun. Among the range of possible mechanisms contributing to the stranding of beaked whales exposed to naval sonar is an acute stress phenomenon, that of hemorrhagic diathesis. A precondition for hemorrhagic diathesis is a depletion or lack of clotting factors or platelet dysfunction. Humans with a

hereditary deficiency in clotting factors develop subarachnoid and inner ear hemorrhages similar to those seen in the beaked whales. No studies have been conducted on the clotting ability of beaked whale blood, but in the few cetacean species studied to date, all have shown a lack of certain clotting factors. None of the species studied have stranded in association with naval sonar; so if hemorrhagic diathesis is a contributor to beaked whale strandings, the level of stress and the physiological responses to that stress are different in beaked whales.

Magnitude of the Problem Other than the strandings of primarily beaked whales in association with naval mid-frequency sonar, anthropogenic sound has not been identified as a cause of marine mammal mortality. Even in the case of the beaked whales and naval sonar, the sequence of behavioral and physiological responses leading to stranding and death is unknown. The total number of beaked whales that are known to have died as a result of anthropogenic sound or in association with alleged naval activity is over 3 orders of magnitude fewer than the number of cetaceans killed annually in direct fisheries bycatch. The preceding paragraph reflects our current understanding but is more likely a reflection of our lack of knowledge than a true assessment of the situation. In order to better gauge the extent of the problem, we need a currency in which to measure the effects of anthropogenic sound other than mortality. As the preceding discussion has shown, anthropogenic sound can affect marine mammals in ways that range from behavioral responses, to inhibited communication and sensing of the environment, to temporary and permanent auditory threshold shifts, to trauma producing morbidity or mortality. The National Research Council (2005) proposed assigning lethality equivalents to sublethal effects of sound. There was recognition that sound which interfered with foraging, displaced animals, masked communication, or caused TTS had the potential to reduce lifetime reproductive fitness and should be given a proportional weighting. For example, if anthropogenic sound interfered with prey detection by passive listening so that it caused a decrease of 0.1% in the lifetime reproductive fitness of each animal affected, then the lethal equivalent would be 0.001, and if 1000 animals were affected, that would be the equivalent of removing one animal from the population. The zones of influence in Figure 1 show that the largest number of animals will be affected at the level

(c) 2011 Elsevier Inc. All Rights Reserved.

Sound Source Level Frequency Duration Duty cycle +++

Hourly to seasonal

1 +

Response

MARINE MAMMALS AND OCEAN NOISE

633

Behavior

Behavior change Orientation Breathing Vocalizing Diving Resting Mother−infant spatial relationships Avoidance ++

Dose 2 +

Seasonal to yearly Yearly to generational

Life function immediately affected

Generational to multigenerational

3 0

Survival Migration Feeding Breeding Nurturing Response to predator +

Vital rates Stage-specific Survival Maturation Reproduction +

Time and energy budgets

4 +++

Population effect Population growth rate Population structure Transient dynamics Sensitivity Elasticity Extinction probability +

Figure 2 The conceptual ‘population consequences of acoustic disturbance’ model describes several stages required to relate acoustic disturbance to effects on a marine mammal population. Five groups of variables are of interest, and transfer functions specify the relationships between the variables listed, for example, how sounds of a given frequency affect the vocalization rate of a given species of marine mammal under specified conditions. A typical type of dose–response transfer function is illustrated. Each box lists variables with observable features (sound, behavior change, life function affected, vital rates, and population effect). In most cases, the causal mechanisms of responses are not known. For example, survival is included as one of the life functions that could be affected to account for such situations as the beaked whale strandings in response to naval mid-frequency sonar, in which it is generally agreed that exposure can result in death. The ‘ þ ’ signs at the bottom of the boxes indicate how well the variables can be measured at the present time. The indicators between boxes show how well the ‘black box’ nature of the transfer functions is understood; these indicators scale from ‘ þ þ þ ’ (well known and easily observed) to ‘0’ (unknown). Reproduced from National Research Council (2005) Marine Mammal Populations and Ocean Noise: Determining When Noise Causes Biologically Significant Effects. Washington, DC: The National Academies Press, with permission from The National Academies Press.

of a behavioral change. The National Research Council in 2005 presented a model of the population consequences of behavioral change in response to anthropogenic sound (Figure 2). This model summarizes what we need to know and indicates how well we know various components leading from behavioral response to population effects. We can measure characteristics of the sound sources well and can do a pretty good job of observing behavioral changes. There is less understanding of how a given sound causes a particular behavioral change under a certain set of conditions. Much of the uncertainty regarding the significance of an observable behavioral response is related to the difficulty in determining how that behavioral change can cause a significant change in a life function. Little is known about the functional response of a behavioral change, and the measurement of the life function altered is difficult for most marine mammal species. Integrating changes in migration, feeding, breeding, etc., over a lifetime in order to determine effects on

vital rates is currently beyond our capabilities. However, when the changes in vital rates are known, the population-level consequences are readily determined. Unfortunately, at least a decade of work will be required before this conceptual model can become a predictive model and can be used to determine the lethal equivalents of sound-induced behavioral changes and subsequent population effects. The uncertainty regarding the magnitude of the problem leaves an extensive gray area between under-regulation resulting in unacceptable harm to marine mammal populations and over-regulation unnecessarily inhibiting essential human activities on and in the ocean.

See also Acoustic Noise. Acoustics, Arctic. Acoustics, Deep Ocean. Acoustics, Shallow Water. Baleen Whales. Fish: Hearing, Lateral Lines (Mechanisms, Role in

(c) 2011 Elsevier Inc. All Rights Reserved.

634

MARINE MAMMALS AND OCEAN NOISE

Behavior, Adaptations to Life Underwater). Marine Mammals: Sperm Whales and Beaked Whales. Sonar Systems.

Further Reading Cox TM, Ragen TJ, Read AJ, et al. (2006) Understanding the impacts of anthropogenic sound on beaked whales. Journal of Cetacean Research Management 7: 177--187. D’Spain GL, D’Amico A, and Fromm DM (2006) Properties of the underwater sound fields during some well documented beaked whale mass stranding events. Journal of Cetacean Research Management 7: 223--238. Fernandez A, Edwards JF, Rodriguez F, et al. (2005) ‘Gas and fat embolic syndrome’ involving a mass stranding of beaked whales (family Ziphiidae) exposed to anthropogenic sonar signals. Veterinary Pathology 42: 446--457. Finneran JJ, Schlundt CE, Dear R, Carder DA, and Ridgway SH (2002) Temporary shift in masked hearing thresholds in odontocetes after exposure to single underwater impulses from a seismic watergun. Journal of the Acoustical Society of America 111: 2929--2940. Hildebrand J (2005) Impacts of anthropogenic sound. In: Reynolds JE, III, Perrin WF, Reeves RR, Montgomery S, and Ragen TJ (eds.) Marine Mammal Research: Conservation beyond Crisis, pp. 101--123. Baltimore, MD: Johns Hopkins Press. McDonald MA, Hildebrand JA, and Wiggins SM (2006) Increases in deep ocean ambient noise west of San Nicolas Island, California. Journal of the Acoustical Society of America 120: 1--8.

Merrill J (ed.) (2004) Human-generated ocean sound and the effects on marine life. Marine Technology Society Journal 37(4). National Research Council (1994) Low-Frequency Sound and Marine Mammals: Current Knowledge and Research Needs. Washington, DC: The National Academies Press. National Research Council (2000) Marine Mammals and Low-Frequency Sound. Washington, DC: The National Academies Press. National Research Council (2003) Ocean Noise and Marine Mammals. Washington, DC: The National Academies Press. National Research Council (2005) Marine Mammal Populations and Ocean Noise: Determining When Noise Causes Biologically Significant Effects. Washington, DC: The National Academies Press. Potter J and Tyack PL (2003) Special Issue on Marine Mammals and Noise. IEEE Journal of Oceanic Engineering 28: 163. Richardson WJ, Greene CR, Malme CI, and Thompson DH (1995) Marine Mammals and Noise. San Diego, CA: Academic Press. Wartzok D and Ketten DR (1999) Marine mammal sensory systems. In: Reynolds JE, III and Rommel S (eds.) Biology of Marine Mammals, pp. 117--175. Washington, DC: Smithsonian Institution Press. Wartzok D, Popper AN, Gordon J, and Merrill J (2004) Factors affecting the responses of marine mammals to acoustic disturbance. Marine Technology Society Journal 37: 6--15.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS, HISTORY OF EXPLOITATION R. R. Reeves, Okapi Wildlife Associates, QC, Canada Copyright& 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1633–1641, & 2001, Elsevier Ltd.

Introduction Products obtained from marine mammals – defined to include the cetaceans (whales, dolphins, and porpoises), pinnipeds (seals, sea lions, and walrus), sirenians (manatees, dugong, and sea cow), sea otter, and polar bear – have contributed in many ways to human survival and development. Maritime communities, from the tropics to the poles, have depended on these animals for food, oil, leather, ivory, bone, baleen, and other materials. Some marine mammal products have had strategic value to nations. For example, for several centuries, streets and homes in much of the western world were illuminated with sperm oil candles and whale oil lanterns. Delicate machinery and precision instruments were lubricated with the head oil of toothed whales. Whale oil was an important source of glycerine during World War I and a key ingredient in margarine during and after World War II. Other uses of marine mammal products have been more frivolous. Seal penises are sold as aphrodisiacs; narwhal (Monodon monoceros) and walrus (Odobenus rosmarus) tusks and polar bear (Ursus

Figure 1 A male narwhal with a 2 m tusk killed in the eastern Canadian Arctic, 1975. The tusk ivory of narwhals and walruses continues to provide an important incentive for hunting them, although both species are also valued as food by native people. Most narwhal tusks are sold and exported, intact, as novelties or trophies. Photo by RR Reeves.

maritimus) hides are displayed as ‘trophies’ in homes and offices (Figure 1). Spermaceti and ambergris, both obtained from sperm whales (Physeter macrocephalus), were highly valued by the perfume and cosmetics industries. Baleen used to be a stiffener for ladies’ hoop skirts and undergarments. And, of course, the pelts of fur seals and sea otters have always been in great demand in luxury fur markets. In general, the history of marine mammal exploitation is marked by overuse and abuse, with most wild populations having been severely overhunted. Some species and populations were extirpated or brought to the brink of extinction. Many others have been reduced and fragmented as a result of too much exploitation. It was not until well into the twentieth century that any serious restrictions were imposed on the sealing and whaling industries for the sake of conservation.

Cetaceans Small Cetaceans (Dolphins, Porpoises, and the Smaller Toothed Whales)

Harpoon hunting of small cetaceans has occurred virtually all around the world, but mainly in coastal and shelf waters (Figure 2). It continues most notably in Japan, where 15 000–20 000 Dall’s porpoises (Phocoenoides dalli) and at least several hundred dolphins, short-finned pilot whales (Globicephala macrorhynchus), and false killer whales (Pseudorca crassidens) are taken annually with hand harpoons, and about 150 additional pilot whales and Baird’s beaked whales (Berardius bairdii) are taken each year with mounted harpoon guns. The meat of small cetaceans is highly valued in Japan. Eskimos in Greenland, Canada, and Alaska (USA) continue their long tradition of hunting white whales (Delphinapterus leucas) and narwhals. Although they formerly used kayaks, hand harpoons, and lances, today most of the hunting involves outboard-powered boats and high-powered rifles. Only in north-western Greenland are the traditional hunting techniques still used to any extent. Altogether, several thousand white whales and narwhals are taken each year. In addition, Greenlanders kill close to 2000 harbor porpoises (Phocoena phocoena) with rifles (Figure 3). The skin of small cetaceans is a delicacy in the Arctic. When saved, the meat and viscera are either eaten by people or fed to dogs. A large commercial hunt for short-beaked common dolphins (Delphinus delphis), bottlenose

(c) 2011 Elsevier Inc. All Rights Reserved.

635

636

MARINE MAMMALS, HISTORY OF EXPLOITATION

Figure 3 Meat from harbor porpoises shot in West Greenland is sold at an open-air market in Nuuk. Harbor porpoises contribute to a diverse array of wild foods consumed by the Greenlandic people, including fish, reindeer, seal, and whale. Photo by Steve Leatherwood, September 1987.

Figure 2 Whale hunters from Barrouallie, St Vincent, Lesser Antilles, at sea in pursuit of short-finned pilot whales (foreground) in the 1960s. The harpoon mounted on the bow of the sailboat was fired with a shotgun. Killed whales were cut into manageable pieces alongside the boat, and these pieces were brought on board to be taken ashore. On the beach, the pieces were cut in strips, hung on bamboo racks to dry, and sold to buyers from Kingstown. In 1968, the average pilot whale was worth about $40 US. Photo by David K Caldwell.

dolphins (Tursiops truncatus), and harbor porpoises was conducted in the Black Sea, using rifles and purse seine nets, from the nineteenth century into the late twentieth century. In the 1930s, nearly 150 000 dolphins and porpoises were taken in a single year. Although dolphin hunting was banned in the Soviet Union in 1966 and in Turkey in 1983, large kills were still being made in the Turkish sector of the Black Sea as recently as 1991. Oil and animal feed (‘fish meal’) were the main products, but the hunting was also prosecuted as a means of predator control. Fishermen viewed the cetaceans as serious competitors. Some small cetaceans, particularly the pilot whales (Globicephala spp.), false killer whale, and melonheaded whale (Peponocephala electra), strand (i.e., come ashore) en masse in numbers ranging from tens to hundreds. This phenomenon remains unexplained but is known to occur naturally. Early coastal people would have welcomed mass strandings, as they represented windfalls of food and other useful products. It is not difficult to imagine their making the leap to a

drive fishery, in which groups of animals were ‘herded’ toward shore, forced into shallow waters, and killed with lances or long knives. The first pilot whale drives in the Faroe Islands apparently took place at least four centuries ago, and similar drive fisheries existed elsewhere in the North Atlantic. In the Faroes, the whales have been used principally as food for humans, but in the other areas oil was a major incentive. In the post-World War II Newfoundland drive fishery, most of the catch (which reached nearly 10 000 pilot whales in 1956) was used to feed ranch mink. Drive hunting of cetaceans in the North Atlantic continues only in the Faroes, where hundreds, and in some years well over a thousand, long-finned pilot whales (Globicephala melas) and Atlantic white-sided dolphins (Lagenorhynchus acutus) are taken, and in West Greenland, where white whales and occasionally pilot whales are driven. Drive fisheries for small cetaceans have also developed in the Solomon Islands and Japan. The Solomons example represents one of the more bizarre forms of exploitation of marine mammals. There, fishermen in dugout canoes fan out across a wide expanse of ocean to search for schools of dolphins and small whales. Large stones are struck together underwater to produce aversive sounds and scare the animals in the desired direction. Eventually, the school is guided into an enclosed harbor where the animals are quickly dispatched. Although some of the meat is cooked and eaten, the primary purpose of the hunt is to obtain ‘porpoise teeth.’ Porpoisetooth necklaces must be given to a woman’s parents as ‘bride price,’ an essential item in marriage transactions.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS, HISTORY OF EXPLOITATION

Dolphin drive fisheries have existed in Japan since the late fourteenth century. Initially, sail-assisted rowing boats were used, but motor vessels were introduced in the 1920s, allowing the hunters to cover much larger areas in their search for schools of small cetaceans. In recent years, high-speed motor boats have been used to find and drive ashore striped dolphins (Stenella coeruleoalba), pantropical spotted dolphins (Stenella attenuata), bottlenose dolphins, Risso’s dolphins (Grampus griseus), pilot whales, and a number of other species. Long seine nets were used to catch bottlenose dolphins along the Atlantic coast of the United States starting in the late eighteenth century and continuing at Cape Hatteras, North Carolina, until the late 1920s. A line of nets was set parallel to the shore, and when a school of dolphins moved into the area between the net line and the shore, the fishermen used nets to shut off escape, then swept the dolphins onto shore. Oil was the main prize, but a supple, durable shoe leather was also made from the hides. Commercial whalers and traders in the Arctic used large seine nets to trap schools of white whales, beginning as early as the 1750s in Hudson Bay and continuing in some areas (e.g., Svalbard) until as recently as 1960. Hides, oil, and dog food were the main products of these commercial netting operations. Large Cetaceans (Baleen Whales and the Sperm Whale)

People in the Arctic were hunting bowhead whales (Balaena mysticetus) as long ago as the middle of the first millennium AD, and western Europeans were taking right whales (Eubalaena glacialis) by the beginning of the second. The technology and culture of subsistence whaling spread eastward within the Arctic and Subarctic from the Bering Strait region. Commercial whaling originated with the Basques, who had begun hunting right whales in the Bay of Biscay by the eleventh century. Initially, small open boats were launched from shore when a whale was sighted. However, the spread of whaling was relentless as Dutch, German, Danish, and British entrepreneurs vied to dominate the rich whaling grounds in the cold latitudes of the North Atlantic. In the 1760s, with the invention of a means to boil blubber on board the ship, it became possible to make extended offshore voyages, often lasting several years. The whaling fleets from New England, Great Britain, and France grew to dominate the industry. From the late eighteenth century to the early 1900s, commercial whaling ships penetrated all of the world’s oceans except the Antarctic. The sperm

637

whale bore the brunt of this activity (Figure 4). More than 225 000 were killed by American whalers alone from 1804 to 1876. During the peak years from the early 1830s to mid-century, over 100 000 barrels of sperm oil were delivered annually by more than 700 vessels working out of American ports. The nineteenth-century whalers often hunted blackfish (their name for pilot whales) while searching for sperm whales. They also made special voyages in pursuit of right, humpback (Megaptera novaeangliae), gray (Eschrichtius robustus), and bowhead whales. Only the fast-swimming finner whales – the blue, fin, sei, Bryde’s, and minke (Balaenoptera musculus, B. physalus, B. borealis, B. edeni/brydei, and B. acutorostrata/bonaerensis, respectively) – were beyond their capabilities to capture. Modern whaling, characterized by engine-driven catcher vessels and deck-mounted harpoon cannons firing explosive grenades, began in Norway in the 1860s. These inventions made possible the routine capture of any species, including the elusive finners. They also led to exploitation of the richest whaling ground on the planet, the Antarctic. In the first threequarters of the twentieth century, factory ships from several nations, including Norway, Great Britain, Germany, Japan, the United States, and the Soviet Union, operated in the Antarctic. At its pre-War peak in 1937–38, the modern industry’s 356 catcher boats, associated with 35 shore stations and as many floating factories, killed nearly 55 000 whales, 84% of them in the Antarctic. Having exhausted the stocks of right, bowhead, gray, and humpback whales in other areas, the industry rapidly proceeded along the same path in the Antarctic, reducing the

Figure 4 A small sperm whale killed by artisanal whalers at St Vincent, Lesser Antilles, during the 1960s. These whalers hunt for a variety of small and medium-sized cetaceans; sperm whales are taken only occasionally. The meat and oil are used locally. Photo by David K. Caldwell.

(c) 2011 Elsevier Inc. All Rights Reserved.

638

MARINE MAMMALS, HISTORY OF EXPLOITATION

largest species first and then turning its attention to the next largest. Commercial whaling declined in the 1970s as a result of conservationist pressure and depletion of the whale stocks. The last whaling stations in the United States and Canada were closed in 1972, and the last station in Australia ceased operations following the 1978 season. By the end of the 1970s, only Japan, the Soviet Union, Norway, and Iceland were still engaged in commercial whaling. With the decision by the International Whaling Commission (IWC) in 1982 to implement a global moratorium on commercial whaling, Japan and the Soviet Union made their final large-scale factory-ship expeditions to the Antarctic in 1986/87, and Japan stopped its coastal hunt for sperm whales and Bryde’s whales in 1988. Iceland closed its whaling station in 1990, and shortly thereafter withdrew its membership in the IWC. Contrary to the widespread belief that commercial whaling had ended, however, Norway and Japan continued their hunting of minke whales through the 1990s and into the 2000s. By formally objecting to the IWC moratorium, Norway reserved its right to carry on whaling. Thus, Norwegian whalers have continued to kill more than 500 minke whales each year in the North Atlantic. Using a provision in the whaling treaty that allows member states to issue permits to hunt protected species for scientific research, Japan has continued taking more than 400 Antarctic minke whales and 100 North Pacific minke whales annually. The main incentive for continued commercial whaling is the demand for whale meat and blubber, particularly in Japan. Norway is eager to re-open the international trade in whale products so that stockpiles of blubber can be exported to Japan. Aboriginal hunters in Russia, the United States (Alaska), and Canada kill several tens of bowheads and 100–200 gray whales every year. This hunting is primarily for human food. However, from the 1960s to early 1990s, gray whales taken by a modern catcher boat and delivered to native settlements in north-eastern Russia were used partly to feed foxes on fur farms. In recent years, native people in Washington State (USA), British Columbia (Canada), and Tonga (a South Pacific island nation) have expressed interest in re-establishing their own hunts for large cetaceans in order to reinvigorate their cultures. In the spring of 1999, the Makah Indian tribe in Washington took their first gray whale in more than 50 years.

Pinnipeds Sealing began in the Stone Age, when people attacked hauled-out animals with clubs. Later methods

included the use of traps, nets, harpoons thrown from skin boats, and gaff-like instruments for killing pups on ice or beaches. The introduction of firearms transformed the hunting of pinnipeds and caused an alarming increase in the proportion of animals that were killed but not retrieved, especially in those hunts where the animals were shot in deep water before first being harpooned. This problem of sinking loss also applies to many of the cetacean hunts mentioned above. In addition to their meat and fat, the pelts of some seals, especially the fur seals and phocids, are of value in the garment industry. Markets for oil and sealskins fueled commercial hunting on a massive scale from the late eighteenth century through the middle of the twentieth. The ivory tusks and tough, flexible hides of walruses made these animals exceptionally valuable to both subsistence and commercial hunters. Thousands of walruses are still killed every year by the native people of north-eastern Russia, Alaska, north-eastern Canada, and Greenland. The killing is accomplished with highpowered rifles, and in some areas harpoons are still used to secure the animal. Walrus meat and blubber are eaten by people or fed to dogs, and the tusks are either used for carving or sold as curios. Native hunters in the circumpolar north also kill more than a hundred thousand seals each year, mainly ringed seals (Pusa hispida) but also bearded (Erignathus barbatus), ribbon (Histriophoca fasciata), harp (Pagophilus groenlandicus), hooded (Cystophora cristata), and spotted seals (Phoca largha) (Figure 5). Seal meat and fat remain important in the diet of many northern communities, and the skins are still used locally to make clothing, dog traces, and hunting lines. There is also a limited commercial export market for high-quality sealskins and a strong demand in Oriental communities for pinniped penises and bacula. The sale of these items, along with walrus and narwhal ivory, white whale and narwhal skin (maktak), and polar bear hides and gall bladders, has helped offset the economic losses in some native hunting communities caused by the decline in international sealskin markets (Figure 6). The scale of commercial sealing, like that of commercial whaling, has declined considerably since the 1960s. It continues, however, in several parts of the North and South Atlantic. After a period of drastically reduced killing in the 1980s, the Canadian commercial hunt for harp and hooded seals has been expanded, at least in part as a result of governmental support. An estimated 350 000 harp seals were taken by hunters in eastern Canada and West Greenland in 1998. A few tens of thousands of molting pups are clubbed to death on the sea ice, but

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS, HISTORY OF EXPLOITATION

639

Figure 6 Pelt of a hooded seal stretched to dry on the side of a house in Upernavik, Northwest Greenland, June 1987. Photo by Steve Leatherwood.

held responsible for damaged fishing gear, the removal of fish from nets and lines, and the spread of parasitic worms which infect groundfish. (A)

Sirenians

(B) Figure 5 Ringed seal killed by a Greenlander off Northwest Greenland, June 1988. Photo by Steve Leatherwood.

the vast majority of the killing is accomplished by shooting. Norwegian and Russian ships continue to visit the harp and hooded seal grounds in the Greenland Sea (‘West Ice’) and Barents Sea (‘East Ice’), taking several tens of thousands of seals annually. Also in recent years, thousands of South African and South American fur seals (Arctocephalus pusillus and A. australis, respectively) have been taken in south-western Africa and Uruguay, respectively. These hunts are centuries old, having been driven initially by markets for skins and oil, and more recently by the Oriental demand for reproductive parts. Much of the hunting for pinnipeds is motivated by the desire of fishermen to see their populations reduced. Seals and sea lions are often

Sirenians have been hunted mainly for meat and blubber, which are highly prized as food. Steller’s sea cow (Hydrodamalis gigas), a North Pacific endemic and the only modern cold-water sirenian, was hunted to extinction within about 25 years after its discovery by commercial sea otter and fur seal hunters in 1741. Sea cows were easy to catch and provided the ship crews with sustenance as they carried on the hunt for furs and oil from other marine mammals. Manatee hides were traditionally used by people in South and Central America and in West Africa to make shields, whips, and plasters for dressing wounds. For a time, these hides were also in great demand for making glue and heavy-duty leather products (e.g., machinery belts, hoses, and gaskets). The hides of more than 19 000 manatees were exported for this purpose from Manaus, Brazil, between 1938 and 1942. For a much longer time, from the 1780s to the late 1950s, the commercial exploitation of manatees in South America was driven by the market for mixira, fried manatee meat preserved in its own fat. Although no large-scale commercial hunt takes place today, local people continue to kill manatees for food. It is impossible to make a reasonable guess at how many manatees are killed by villagers in West Africa and South and Central America, but the total in recent years has probably been in the thousands (all three species, Trichechus manatus, T. inunguis, and T. senegalensis, combined). Manatees are captured in many different ways, apart from simply stalking them in quiet

(c) 2011 Elsevier Inc. All Rights Reserved.

640

MARINE MAMMALS, HISTORY OF EXPLOITATION

dugout canoes and striking them with a lance or harpoon. These involve such things as stationary hunting blinds; drop traps armed with heavy, pointed wooden posts; and fence traps or nets placed strategically in the intertidal zone or in constricted channels. Dugongs (Dugong dugon), like manatees, have long been a prized food source for seafaring people. Hunting continues throughout much of their extensive Indo-Pacific range, even in areas where the species is almost extinct. Dugong hunters in some areas have used underwater explosives to kill their prey. In Torres Strait between Australia and New Guinea, portable platforms are set up on seagrass beds, and the hunter waits there overnight for opportunities to spear unsuspecting dugongs as they graze.

Sea Otter The sea otter (Enhydra lutris) has one of the most luxuriant and thus desirable pelts of any mammal species. As a result, it was eagerly hunted by aboriginal people all round the rim of the North Pacific Ocean. Also, beginning soon after Vitus Bering discovered the Commander Islands in 1741, Russian, and later American and Japanese, expeditions were mounted for the explicit purpose of obtaining sea otter furs, which commanded high prices in the Oriental market. No statistics were kept, but at least half a million sea otters were taken (or received in trade) by commercial hunters between 1740 and 1911, when the species was given legal protection. The hunters sometimes used anchored nets to catch the otters, but more often they lanced them from small boats. Once rifles became available, these were used in preference to lances. In California, sea otters were sometimes shot by men standing on shore, and in Washington, shooting towers were erected at the surfline and Indians were employed to swim out and retrieve the carcasses. Alaskan natives are still allowed to hunt sea otters as long as the furs are used locally to make clothing or authentic handicraft items. The reported annual kill during the mid to late 1990s ranged from 600 to 1200.

Polar Bear Eskimos traditionally hunted polar bears with dog teams and hand lances. The meat was eaten and the hides used for clothing and bedding. White explorers, whalers, sealers, and traders in the Arctic often killed polar bears with high-powered rifles. They also provided a commercial outlet for hides obtained by the Eskimos. In modern times, the

Eskimos hunt polar bears with rifles and search the ice in snowmobiles rather than dogsleds. Norwegian trappers and weather station crews on Svalbard formerly used poison, foot snares, and set guns to kill polar bears. The set gun consisted of a wooden box resting on poles about 75 cm above ground level, with a rifle or shotgun mounted inside. A string connected the gun’s trigger to bait placed in front of the box. When the bear took the bait, the trigger was pulled and the gun fired. Sport hunters have taken thousands of polar bears as trophies, particularly in Alaska where guided hunting with aircraft began in the late 1940s and continued until 1972. At least several hundred polar bears are still killed each year, most of them by Eskimos for meat and the cash value of their hides and gall bladders. Hunting permits issued to native communities in Canada are often sold to sport hunters, on the understanding that a local guide will be hired to accompany the hunter and that only the head and hide will be exported.

Live-capture and other Forms of Exploitation Although the numbers of marine mammals removed from the wild for captive display and research have been small in comparison to the numbers killed for meat, oil, skins, etc., the high commercial value of some species establishes this as an important form of exploitation (Figure 7). More than 1500 bottlenose dolphins were live-captured in the United States, Mexico, and the Bahamas between 1938 and 1980. Close to 70 killer whales (Orcinus orca) were removed from inshore waters of Washington State (USA) and British Columbia (Canada) and transported to oceanaria between 1962 and 1977, and about 50 were exported from Iceland in the 1970s and 1980s (Figure 8). Live killer whales and bottlenose dolphins are presently worth about $1 million and $50 000, respectively. Captive-bred animals and ‘strandlings’ (animals that come ashore and require rehabilitation) have increasingly been used to stock oceanaria, but this trend applies mainly to North America and involves primarily bottlenose dolphins, killer whales, California sea lions (Zalophus californianus), and harbor seals (Phoca vitulina). Dolphin, whale, and sea lion displays are becoming more popular in Asia and South America, and new facilities on those continents create a continuing demand for wild-caught animals, especially dolphins (Figure 9). Most polar bears and walruses brought into captivity have been young ones, orphaned when their mothers were killed by hunters. In Florida, manatees are often brought into captivity after being

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS, HISTORY OF EXPLOITATION

641

(B)

(A)

Figure 7 Commerson’s dolphins are endemic to the coastal waters of southern South America and certain subantarctic islands. There is some demand for them in North American, European, and Japanese oceanaria. (A) Here, a hoop net at the end of a pole is used in an attempt to capture a dolphin from a bow-riding group off the coast of Chile, February 1984. (B) A Commerson’s dolphin (foreground) shares an oceanarium tank with a white whale (beluga) at a zoo in Duisberg, Germany. Photos by Steve Leatherwood.

Figure 8 Killer whales are the most valuable marine mammals in the oceanarium trade. Recent success at captive breeding and rearing has relieved some of the pressure on wild populations to stock display facilities. The movie ‘‘Free Willy’’ inspired a campaign to return the whale ‘‘Keiko’’ back to its natal waters near Iceland. Photo by Steve Leatherwood.

Figure 9 Irrawaddy dolphins have a limited coastal and freshwater distribution in southeast Asia and northern Australia. They are fairly popular in Asian oceanaria, and live captures add to the stress on populations caused by incidental mortality in gillnets. This animal was recently on display at a facility in Thailand. Photo by Steve Leatherwood.

(c) 2011 Elsevier Inc. All Rights Reserved.

642

MARINE MAMMALS, HISTORY OF EXPLOITATION

Further Reading

Figure 10 Gray whales attract many tourists each year to the nearly pristine waters of Laguna San Ignacio, Baja California, Mexico. Whale-watching in the lagoon is closely regulated by Mexican authorities. The recognized economic value of nature tourism was partly responsible for the government’s decision in 2000 to reject a proposal for a large evaporative salt factory on the shores of San Ignacio. Photo by Steve Leatherwood.

injured or orphaned as a result of boat strikes. Nearly 100 sea otters were taken from Alaskan waters for public display between 1976 and 1988. It should be mentioned that marine mammals are also ‘exploited’ as the objects of tourism. Whale-, dolphin-, seal-, and sea otter-watching supports an extensive network of tour operations around the world (Figure 10). Commercial fishermen ‘exploit’ pelagic dolphins in the tropical Pacific Ocean by using them to locate schools of tuna, and this can result in large numbers of dolphins being killed by accident.

See also Baleen Whales. Marine Mammal Overview. Marine Mammals: Sperm Whales and Beaked Whales. Sea Otters. Seals. Sirenians.

Bonner WN (1982) Seals and Man: A Study of Interactions. Seattle: University of Washington Press. Bra¨utigam A and Thomsen J (1994) Harvest and international trade in seals and their products. In: Reijnders PJH Brasseur S, et al. (eds.) Seals, Fur Seals, Sea Lions, and Walrus: Status Survey and Conservation Action Plan. Gland, Switzerland: International Union for Conservation of Nature and Natural Resources. Busch BC (1985) The War Against the Seals: A History of the North American Seal Fishery. Kingston, Ontario: McGill-Queen’s University Press. Dawbin WH (1966) Porpoises and porpoise hunting in Malaita. Australian Natural History 15: 207--211. Domning DP (1982) Commercial exploitation of manatees Trichechus in Brazil c. 1785–1973. Biological Conservation 22: 101--126. Ellis R (1991) Men and Whales. New York: Alfred A. Knopf. International Whaling Commission. Annual report and special issues. Available from IWC, The Red House, 135 Station Road, Impington, Cambridge, CB4 9NP, UK. (From April 1999, these reports appear as Supplements or Special Issues of The Journal of Cetacean Research and Management, published by the IWC.) McCartney AP (ed.) (1995) Hunting the Largest Animals: Native Whaling in the Western Arctic and Subarctic. Edmonton: Canadian Circumpolar Institute, University of Alberta. Mitchell E (1975) Porpoise, Dolphin and Small Whale Fisheries of the World. Morges, Switzerland: International Union for Conservation of Nature and Natural Resources. Taylor VJ and Dunstone N (eds.) (1996) The Exploitation of Mammal Populations. London: Chapman and Hall. Tønnessen JN and Johnsen AO (1982) The History of Modern Whaling. Berkeley: University of California Press. Twiss JR Jr and Reeves RR (eds.) (1999) Conservation and Management of Marine Mammals. Washington, DC: Smithsonian Institution Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES S. K. Hooker, University of St. Andrews, St. Andrews, UK & 2009 Elsevier Ltd. All rights reserved.

Cuvier’s beaked whales. Such studies help provide important behavioral information about these species which has not previously been available from studies of dead animals.

Taxonomy and Phylogeny

Introduction Sperm whales and beaked whales are among the largest and most enigmatic of the odontocetes (toothed whales). These species tend to live far offshore in regions of deep water, and perform long, deep dives in search of their squid prey. This has generally made the study of these animals much more difficult than that of more accessible, near-shore cetacean species. In addition, the pygmy and dwarf sperm whales, and many species of beaked whale, have superficially similar external morphology, and so are often difficult to identify to species level in the wild. The study of many of these species has therefore been based primarily on examination of stranded and beach-cast animals. As a result, we currently know comparatively little regarding many of these relatively large mammals. For example, one species of beaked whale, Longman’s beaked whale, was identified from only two skulls in Australia and Somalia and has only recently been observed in the wild. Another putative species, Mesoplodon species ‘A’, has only ever been observed at sea, and our knowledge of its morphological characteristics remains far from complete. New species of beaked whales are still being discovered. For example, the lesser beaked whale and Bahamonde’s beaked whale were only identified in the last decade from specimens collected in Peru and Chile, respectively. Perrin’s beaked whale was recently discovered by analyzing DNA sequence data from five archived Californian stranded specimens initially thought to be Hector’s beaked whales or Cuvier’s beaked whales. Likewise, the dwarf and pygmy sperm whales were only recognized as separate species in the 1960s. The sperm and beaked whale species about which we know most are the sperm whale, the northern bottlenose whale, and Baird’s beaked whale. Much of our information about these species has come from scientific research programs conducted in conjunction with historic whaling operations. Longer-term, nonlethal studies of wild populations only began in the early 1980s. These focused initially on sperm whales, and today include research on populations of northern bottlenose whales, Blainville’s beaked whales, and

There are three superfamilies within the odontocetes: the Physeteroidea (sperm whales), the Ziphioidea (beaked whales), and the Delphinoidea (river dolphins, oceanic dolphins, porpoises, and monodontids). The superfamily Physeteroidea encompasses two families: the Physeteridae which contains the sperm whale, and the Kogiidae which contains the pygmy sperm whale and the dwarf sperm whale. The Ziphioidea encompasses only the family Ziphiidae, which includes at least 21 species of beaked whales (Table 1). Although some genetic studies have challenged the relationship of the sperm whales to other toothed whales, the analytical methods used to determine this have been questioned, and there is general agreement between morphological and other molecular data that the sperm whales and beaked whales are basal odontocetes (Figure 1). Physeterids appeared in the fossil record in the early Miocene deposits of Argentina (c. 25 Ma). In the past, this family included a diverse array of genera, but today is represented only by the sperm whale. The kogiids are thought to have diverged from the physeterids in the late Miocene and early Pleiocene (c. 5–10 Ma). The earliest ziphiids have been found in deposits from the middle Miocene (10–15 Ma). Relationships among the beaked whales are not clear. The six genera in this family have previously been separated into two tribes grouping Berardius and Ziphius, and grouping Tasmacetus, Indopacetus, Hyperoodon, and Mesoplodon. However, it has also been suggested that Tasmacetus, with a full set of teeth in upper and lower jaws, may be the sister group to all other living species. Current work investigating the systematics of this group using DNA sequence data in fact suggests that Tasmacetus is more closely related to Ziphius and Hyperoodon.

Anatomy and Morphology Beaked whales are characterized by the possession of a long and slender rostrum resulting in a prominent beak in most species. An evolutionary trend in

(c) 2011 Elsevier Inc. All Rights Reserved.

643

644

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

Table 1

Sperm and beaked whale species, approximate demographic distribution and size

Species Family Physeteriidae Sperm whale Family Kogiidae Dwarf sperm whale

General location

Adult size (m)

Physeter macrocephalus

Global

12–18

Kogia simus

Tropical and temperate oceanic Tropical and temperate, continental shelf and slope

2.7–3.4

11.9–12.8 Up to 9.7 7–7.5

Notes

Pygmy sperm whale

Kogia breviceps

Family Ziphiidae Baird’s beaked whale Arnoux’s beaked whale Cuvier’s beaked whale

Berardius bairdii Berardius arnuxii Ziphius cavirostris

Northern bottlenose whale Southern bottlenose whale Shepherd’s beaked whale Longman’s beaked whale Blainville’s beaked whale Gray’s beaked whale

Hyperoodon Hyperoodon Tasmacetus Indopacetus Mesoplodon Mesoplodon

Ginkgo-toothed beaked whale Hector’s beaked whale Hubbs’ beaked whale Sowerby’s beaked whale Gervais’ beaked whale

Mesoplodon ginkgodens

True’s beaked whale

Mesoplodon mirus

Strap-toothed whale Andrew’s beaked whale

Mesoplodon layardii Mesoplodon bowdoini

Stejneger’s beaked whale Lesser beaked whale

Mesoplodon stejnegeri Mesoplodon peruvianus

North Pacific Southern Ocean Global, common in Eastern Tropical Pacific North Atlantic Southern Ocean Southern temperate Australia, Somalia Temperate global Southern temperate circumglobal Temperate/tropical Indian and Pacific Oceans Southern temperate North Pacific Northern North Atlantic Temperate/tropical Atlantic Temperate N. Atlantic and temperate Southern Ocean Southern temperate South Indian and Pacific Oceans North Pacific Eastern Tropical Pacific

Bahamonde’s beaked whale Perrin’s beaked whale Mesoplodon species ‘A’

Mesoplodon bahamondii

Peru

Estim. 5–5.5

Few strandings; tentative sightings Partial skull

Mesoplodon perrini Mesoplodon sp.

California (USA) Eastern Tropical Pacific

3.9–4.4 5.5

Few strandings Sightings only

Mesoplodon Mesoplodon Mesoplodon Mesoplodon

ampullatus planifrons shepherdii pacificus densirostris grayi

hectori carlhubbsi bidens europaeus

ziphiids has led to a reduction in the number of teeth in all genera except Tasmacetus. Most species have retained only one or two pairs of teeth, set in varying positions in the lower jaw (Figure 2). In most beaked whale species, these teeth only erupt in adult males. From observations of scarring patterns on the animals, these teeth appear to function as weapons in intraspecific combat, and have become much enlarged in some species. Other features which distinguish beaked whales from other groups include the possession of two conspicuous throat grooves or creases which form a forward pointing V-shape, and the lack of a notch in the flukes. The skull morphology of beaked whales is also unique, exhibiting elevated maxillary ridges behind the nasals.

Up to 2.7

8.7–9.8 7.2–7.8 6.6–7 Over 6 Up to 4.7 Up to 5.6

Few stranded specimens Two skulls

Up to 4.9 4.3–4.4 Up to 5.3 5.1–5.5 4.5–5.2 Up to 5

5.9–6.2 4.6–4.7 Up to 5.3 Up to 3.7

The Physeteroidea are characterized by several features of the skull, including a large supracranial basin. This basin holds the ‘spermaceti organ’, a fatfilled structure, which lies behind the melon in the forehead, and is unique to these species. This structure was named for the presence of spermaceti, an oily substance thought to resemble semen (after which it was named). It is generally thought that this organ functions in sound transmission. Sperm whales show extreme sexual dimorphism both in their size (up to 18-m length of adult males compared to up to 12-m length of adult females) and in the increased ratio of head to body size such that the adult male head is relatively much larger than the adult female head. Thus whalers valued adult males for their large

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

quantities of spermaceti much more than females. Externally, sperm whales can be differentiated from other species by their narrow lower jaw, and an upper jaw which extends well past the lower one. This group also has reduced dentition. The sperm whale has teeth (18–25 pairs) only in the lower jaw. The dwarf and pygmy sperm whales have reduced numbers of teeth in the lower jaw (generally12–16

)

les

rm

a wh

pe

les

ys

a ) e wh gm les ida le) e y d r a e p e h a te ha h a wh iid w ot se w giid and to ph ked en f hy erm i o r r e l P Z a a K w e p he Ba (s (b (d Ot s ale

Figure 1 Phylogenetic diagram showing relationship of sperm whales and beaked whales to other cetaceans. Adapted from Heyning JE (1997) Sperm whale phylogeny revisited: Analysis of the morphological evidence. Marine Mammal Science 13: 596–613.

Berardius B. bairdii

Tasmacetus T. shepherdi

B. arnuxii

Hyperoodon H. ampullatus

645

pairs in pygmy sperm whales, and 8–11 pairs in the dwarf sperm whale). Some teeth may be present in the upper jaw of both dwarf and pygmy sperm whales, although this is less common in the pygmy sperm whale. The sperm whales show highly pronounced asymmetry of the skull. Most beaked whales species also have asymmetrical skulls, although Berardius and Tasmacetus have nearly symmetrical cranial characteristics, suggesting that cranial asymmetry in beaked whales has evolved independently from that in sperm whales. The asymmetry of the sperm whale head extends to the nasal passages in which the right nasal passage has been modified as part of the sound production apparatus while the left nasal passage connects to the blowhole (on the left of the sperm whale’s head). The right nasal passage appears to be capable of disconnection from the lungs and the left nasal passage, since sperm whales can breathe and produce clicks at the same time. Thus, unlike other odontocetes which possess two bilaterally placed monkey lips/dorsal bursae complexes, sperm whales have only a single complex.

Distribution and Abundance All sperm whales and beaked whales are deep-water oceanic species. However, there is wide variation in

Ziphius Z. cavirostris

H. planifrons

Mesoplodon M. densirostris

M. grayi

M. ginkgodens

M. hectori

M. carlhubbsi

M. bidens

M. europaeus

M. mirus

M. layardii

M. bowdoini

M. stejnegeri

Figure 2 Variation in position, size, and morphology of the lower jaw teeth of adult male beaked whales shown for the majority of recorded beaked whale species. After Jefferson TA, Leatherwood S, and Webber MA (1993) Marine Mammals of the World. Rome: United Nations Environment Program, FAO. http://ip30.eti.uva.nl/bis/marine_mammals.php (accessed Mar. 2008).

(c) 2011 Elsevier Inc. All Rights Reserved.

646

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

species coverage. The sperm whale is found throughout the world’s oceans, from the Tropics to the Poles. The pygmy and dwarf sperm whales also have fairly cosmopolitan distributions, and are found in temperate and tropical waters worldwide. In contrast, many beaked whale species have quite limited distributions and only two species, the densebeaked whale and Cuvier’s beaked whale, show similar ranges to the sperm whales (Table 1). Many other beaked whale species are limited to a single ocean basin, and several species pairs show an antitropical distribution, for example, Hyperoodon (North Atlantic and Southern Oceans) and Berardius (North Pacific and Southern Oceans). There are very few estimates of abundance for sperm and beaked whales. Many of these species are difficult to detect and identify at sea, and so are likely to be more common than sighting records would suggest. The status of all sperm and beaked whales as currently listed by the IUCN (International Union for the Conservation of Nature) Red Book is ‘insufficiently known’.

Foraging Ecology The majority of sperm and beaked whales are thought to feed primarily on squid. The reduced dentition of these species is thought to be due to this dietary specialization. One exception to this is the Shepherd’s beaked whale, in which both sexes possess a full set of functional teeth, and the diet appears to consist primarily of fish. The reduced dentition of the beaked whales, together with their narrow jaws and throat grooves, have been suggested to function in suction feeding. Among the males of species such as the strap-toothed whale, the elaborate growth of the strap-like teeth may limit the aperture of the gape to a few centimeters, and it is difficult to see how prey capture techniques other than suction feeding could be successful. The same mechanism is thought to be used by sperm whales which also have a comparatively small mouth area. Anecdotal evidence suggests that the lower jaw teeth of the sperm whale are not required for feeding, as apparently healthy animals have been seen with broken and badly set lower jaws resulting from past injuries. Sperm whales and beaked whale species are known to be excellent divers. Dives of up to an hour have been recorded from several beaked whale species, although many of these records have been based on surface observations of diving whales. Similarly, dives of up to 25 min have been recorded from Kogia species. Increasingly studies are now using timedepth recorders to monitor dives in more detail from

these species. Sperm whales have been recorded to dive repeatedly to depths of 500–800 m for durations of 35–45 min, with a maximum recorded dive depth of 1860 m. Of beaked whale species, time-depth data loggers have been deployed on northern bottlenose whales, Blainville’s beaked whales, and Cuvier’s beaked whales (Figure 3). All three of these species have also been found to dive regularly to depths of over 800 m, although their dive profiles show distinct species-specific patterns to their diving behavior. Sperm and northern bottlenose whales have a greater frequency of deep dives, but the deepest dive thus far recorded was from a Cuvier’s beaked whale at 1950 m, and the longest dive was 85 min. These species are thought to forage at these depths, at times at the seafloor, in search of deep-water squid. The similarity of ecological niches among beaked and sperm whales might be expected to lead to competition between these species. The relatively discrete distributions of many beaked whale species may have resulted from this. For example, several Mesoplodon species coexist in the North Atlantic, but have separate centers of distribution, with little overlap in range: Sowerby’s beaked whale has a more northerly distribution than True’s beaked whale, which in turn is found to the north of Gervais’ beaked whale. On a much smaller spatial scale, there is some suggestion of competitive exclusion between sperm whales and northern bottlenose whales in habitat use of a submarine canyon area off eastern Canada. Prey species (mainly squid) identified from the stomachs of stranded Kogia specimens suggest that these species occur primarily along the continental shelf and slope in the epi- and mesopelagic zones. Although the diets of both species overlap, the relative contribution of prey types suggests that the dwarf sperm whale feeds on smaller squid in shallower waters and thus occurs further inshore than the pygmy sperm whale.

Social Organization The social organization of the majority of sperm and beaked whale species is only poorly known, with the exception of the sperm whale. The social system of the sperm whale appears quite unlike that known for other cetaceans. Groups of females and juveniles are found in temperate and tropical latitudes (Figure 4). Males become segregated from these female groups at or before puberty, and migrate to higher latitudes. Younger males are found in ‘bachelor schools’, which consist of animals of approximately the same age. These schools decrease in size with increasing age of the members, to the point where large mature

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

Depth (m)

Depth (m)

0

0

500

500 1000

1000 Sperm whale 1 1500

14

16

18

20

1500

22

0

500

500

1000

1000 Blainville’s beaked whale 1 12

14

16

18

1500

20

0

500

500

1000

16

18

20

14

16

18

20

1000 Cuvier’s beaked whale 1 14

16

18

20

22

14 0

500

500

1000

1000 Northern bottlenose whale 16

Cuvier’s beaked whale 2

1500

0

1500

14

Blainville’s beaked whale 2 12

0

1500

Sperm whale 2 12

0

1500

647

18 20 22 Time (hour of day)

18

20

22

Northern bottlenose whale

1500 24

16

4

6 8 10 Time (hour of day)

12

Figure 3 Dive profiles are shown for 10-h segments of each of: two sperm whales (1: Gulf of Mexico, 2003; 2: Gulf of Mexico, 2002), two Blainville’s beaked whales (1: Bahamas, 2006; 2: Hawaii, 2004), two Cuvier’s beaked whales (1: Liguria, Italy, 2006; 2: Hawaii, 2004), and two records from a single northern bottlenose whale (the Gully, eastern Canada, 1997). Data were kindly provided by Robin Baird, Cascadia Research Collective, USA; Sascha Hooker, Sea Mammal Research Unit, UK; Patrick Miller, Sea Mammal Research Unit, UK; and Mark Johnson, Woods Hole Oceanographic Institute, USA.

animals are typically solitary. Sexually mature males return to the tropical waters inhabited by females in order to breed. The sexual dimorphism, scarring patterns, and vocal behavior suggest that adult male sperm whales compete for access to groups of females. Recent mark-recapture data, relative parasite loads, and indications of synchronous estrus suggest that males rove between groups of females, and remain with any one group for only a few hours, although they may revisit groups on consecutive days. Female sperm whales are found in groups of 20 or so individuals. These groups appear to consist of two or more stable units that associate for periods of c. 10 days. Genetic evidence has suggested that these groups are composed of one or more matrilines. However, there are also suggestions of paternal relatedness between grouped matrilines, and recent photoidentification studies suggest that some animals occasionally switch groups, and thus may not be of the same maternal lineages as other group members.

Whalers observed that sperm whale groups often exhibited epimeletic behavior, with individuals supporting and staying with harpooned, injured, and even dead group members. It is thought that this may be a result of the close genetic ties between the individuals in a group. Sperm whales have also been observed to exhibit allomaternal care (babysitting behavior). Calves remain on the surface when a group is feeding, presumably since they are unable to dive to the depths at which adults forage, and adults have been observed to stagger their dives such that there is always an adult at the surface with the calf. Observations of wild Kogia suggest that they typically form small groups of one to four animals, with occasional groups of up to 10 reported. However, almost nothing is known of the composition of these groups or of the behavior of these species at sea. Many beaked whale species appear to show intraspecific aggression between adult males, presumably for access to females. The prominent and

(c) 2011 Elsevier Inc. All Rights Reserved.

648

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

Figure 4 Group of female and juvenile sperm whales off the Galapagos Islands. Photograph by Sascha K. Hooker.

elaborate teeth of many beaked whale species are thought to be used in this male–male conflict (Figure 2), resulting in the extensive scarring seen on adult males. However, in other beaked whale species, males possess only comparatively small lower jaw teeth, and these do not appear to be used for fighting. The northern bottlenose whale is an example of this. Instead, this species shows marked sexual dimorphism in skull structure and associated forehead or melon shape, which is relatively small in females but is enlarged and flattened in adult males. Recent observations have suggested that this melon morphology is also associated with male–male competition, as adult males have been observed to head-butt each other. Among beaked whales, the composition of social groups is not well known. The two beaked whales for which most data have been collected are the northern bottlenose whale and the Baird’s beaked whale. Long-term photoidentification of individual bottlenose whales has in fact suggested stronger associations between males than between females. However, the aggression observed between some associated males makes further interpretation difficult. Anatomical studies of groups of Baird’s beaked whales taken in the continuing fishery off Japan are suggestive of a different type of social structure for this species. Among this species, both males

and females possess erupted teeth, and females are slightly larger than males. Males appear to reach sexual maturity an average of four years earlier than females and may live up to 30 years longer. This has led to speculation that males may be providing parental care in this species, although further work is needed to confirm this.

Acoustics, Sound Production, and Sound Reception The acoustic behavior of sperm whales is relatively well documented. These whales produce broadband clicks with a centroid frequency of 15 kHz. These clicks are thought to function primarily in echolocation, although some repetitive patterned clicks (termed codas) also appear to be used in a social context. Neither Kogia species appears to be highly vocal, although echolocation-type signals have been recorded from the pygmy sperm whale. The social whistles characteristic of other odontocete species are absent from the physeterids, and possibly also from some beaked whale species. No whistles were documented in several hours of recordings from northern bottlenose whales, which instead also appear to produce primarily echolocation-type clicks. These were superficially similar to sperm whale clicks, although

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

often at ultrasonic frequencies (B20–30 kHz). An acoustic recording tag attached to Blainville’s and Cuvier’s beaked whales has also demonstrated short, directional, ultrasonic echolocation clicks with most of their energy between 35 and 45 kHz. These clicks are produced during dives and are frequency sweeps similar to the echolocation of a bat, with regular clicks produced every 0.2–0.4 s, ending in a rapid increase in slick rate up to 250 clicks per second as the target is approached. Only a few other records of beaked whale acoustic behavior exist and the majority of these were obtained from stranded animals. Other recordings of wild animals have been made of Baird’s beaked whale and Arnoux’ beaked whale, and these included frequency-modulated whistles, burst-pulse clicks, and discrete clicks in rapid series. The sound production mechanism used by both sperm and beaked whales for echolocation is homologous with that of other odontocetes, consisting of a sound-producing complex (the ‘monkey lips’/dorsal bursae) in the upper nasal passages. Sound is propagated in the water by the melon, a lipid-filled structure which acts as an acoustic lens to focus a directional beam ahead of the animal. The echoes of this sound are then received via the fat body in the lower jaw which connects with the bulla of the middle ear. The sperm whale head is unique in comparison to other odontocetes in that the blowhole and the sound production mechanism are situated at the front of the head rather than above the eyes. The initial click propagates backward from the front of the head through the spermaceti organ to create an intense forward-directed pulse. Reverberation within the spermaceti organ generates a decaying series of pulses, with the time interval between these pulses related to the size of the head.

Predation It was previously thought that large size was adequate defense against predators, but female sperm whales (of 10–12 m) have been observed under lethal attack by ‘transient’ (mammal-eating) killer whales. Additionally, large sharks are thought to be a threat to these species, particularly to juvenile animals. Various methods of defense may be employed. For such a deep-diving species, it is surprising that deep dives are not used as a method of escape from predators. This may be because young calves are unable to dive to the depths or for the same duration as adults. Instead, sperm whales appear to show a behavioral response to the threat of predation, with the adults forming a circle (heads innermost) around the calves. Pygmy and dwarf sperm whales evacuate reddishbrown intestinal fluid when startled, in a similar

649

manner to inking by squids. The lower intestine is expanded in both species, forming a balloon-like structure filled with up to 12 l (in large specimens) of this liquid. Additionally, these species possess a crescent-shaped light-colored mark, often called a ‘false gill’, on the side of the head behind the eye and before the flippers. Along with the underslung mouth, this can lead to the mistaken identification of these animals as sharks. However, whether this patterning functions as camouflage against predation is unknown.

Conservation The larger species of sperm and beaked whales were all targeted by whaling operations in the past. The sperm whale was the most heavily hunted, primarily due to the prized spermaceti oil that it contained. This fishery spanned the seventeenth to twentieth centuries and at its peak (in the 1960s) average annual catches reached 25 000 animals. Northern bottlenose whales were also quite severely depleted by whaling. The northern bottlenose whale fishery began in the late nineteenth century and between 1880 and 1920 approximately 60 000 bottlenose whales were caught. The other species in this group which has been taken in relatively large numbers is Baird’s beaked whale. This species has been hunted in Japan since at least the seventeenth century, but has generally been taken in relatively low numbers (a maximum annual catch of 322 in 1952, but recently averaging 40 whales per year). This fishery continues even today. Current threats faced by these species range from other factors potentially causing immediate death, such as ship-strikes, to the more insidious threats of ocean plastic, chemical, and acoustic pollution. Since many sperm and beaked whales feed primarily on squid, they are very susceptible to the ingestion of plastics, apparently mistaking it for prey. Stranded animals from several sperm and beaked whale species have been found with plastic in their stomachs and in some cases, this appears to have blocked the normal function of the stomach, causing severe emaciation and likely contributing to their death. The ecological role of odontocetes as long-lived top predators also exposes these animals to increased levels of chemical pollutants. Cetaceans store energy (and pollutants) in their blubber, and have a lower capacity to metabolize some polychlorinated biphenyl (PCB) isomers than many other mammals. Foreign and toxic substances are therefore often biomagnified in odontocete species, and even species living offshore in relatively pristine environments have been found to contain high levels of pollutants. These high pollutant levels can have two major effects: (1) inhibition of immune system capacity to

(c) 2011 Elsevier Inc. All Rights Reserved.

650

MARINE MAMMALS: SPERM WHALES AND BEAKED WHALES

respond to naturally occurring diseases, and (2) potentially causing reproductive failure. There is also increasing concern about the effect of anthropogenic noise in the marine environment. Sperm whales and beaked whales appear to be particularly susceptible to the effects of such noise. Sperm whales have been observed to react to several types of underwater noise including sonar, seismic activity, and low-frequency sound. Beaked whales also appear to be particularly susceptible to highintensity underwater sound. Several beaked whale stranding events have coincided with military naval exercises and associated high-intensity underwater sound. However, the mechanism by which such sound is hazardous to these whales is not yet known, thus limiting the likely effectiveness of mitigation efforts. There is currently a high level of concern surrounding this issue.

Glossary competitive exclusion The principle that two species cannot coexist if they have identical ecological requirements. cosmopolitan Having a broad, wide-ranging distribution. echolocation The production of high-frequency sound and reception of its echoes; used to navigate and locate prey. epimeletic Caregiving behavior. mandible Lower jaw. matriline Descendants of a single female. rostrum Anterior portion or beak region of the skull that is elongated in most cetaceans. sexual dimorphism Morphological differences between males and females of a species. ultrasonic High-frequency sounds, beyond the upper range of human hearing.

See also Bioacoustics. Dolphins and Porpoises. Marine Mammal Diving Physiology. Marine Mammal Evolution and Taxonomy. Marine Mammal

Migrations and Movement Patterns. Mammal Overview. Marine Mammal Organization and Communication. Mammal Trophic Levels and Interactions. Mammals, History of Exploitation.

Marine Social Marine Marine

Further Reading Berta A and Sumich JL (1999) Marine Mammals: Evolutionary Biology. San Diego, CA: Academic Press. Best PB (1979) Social organisation in sperm whales. In: Winn HE and Olla BL (eds.) Behaviour of Marine Mammals. Vol. 3: Cetaceans, pp. 227--289. New York: Plenum. Heyning JE (1997) Sperm whale phylogeny revisited: Analysis of the morphological evidence. Marine Mammal Science 13: 596--613. Jefferson TA, Leatherwood S, and Webber MA (1993) Marine Mammals of the World. Rome: United Nations Environment Program, FAO. http://ip30.eti.uva.nl/bis/ marine_mammals.php (accessed Mar. 2008). Moore JC (1968) Relationships among the living genera of beaked whales with classification, diagnoses and keys. Fieldiana: Zoology 53: 209--298. Perrin WF, Wursig B, and Thewissen JGM (eds.) (2002) Encyclopedia of Marine Mammals. San Diego, CA: Academic Press. Rice DW (1998) Special Publication No. 4: Marine Mammals of the World – Systematics and Distribution. Lawrence, KS: The Society for Marine Mammalogy. Ridgway SH and Harrison R (1989) Handbook of Marine Mammals, Vol. 4: River Dolphins and the Larger Toothed Whales. London: Academic Press. Whitehead H (2003) Sperm Whales: Social Evolution in the Ocean. Chicago, IL: Chicago University Press. Whitehead H and Weilgart L (2000) The sperm whale: Social females and roving males. In: Mann J, Connor RC, Tyack P, and Whitehead H (eds.) Cetacean Societies: Field Studies of Dolphins and Whales, pp. 154--173. Chicago, IL: Chicago University Press.

Relevant Websites http://www.iucnredlist.org – IUCN Red List of Threatened Species. http://www.animalbehaviorarchive.org – Macaulay Library Sound and Video Catalog. http://ip30.eti.uva.nl – National Museum of Natural History.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MATS A. E. S. Kemp, University of Southampton, Southampton Oceanography Centre, Southampton, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1641–1644, & 2001, Elsevier Ltd.

The production of organic carbon in ocean surface waters and its subsequent transport to the seafloor is often referred to as the ‘biological pump’. This biological pump is an important link in the global carbon cycle because phytoplankton use dissolved carbon dioxide (CO2) gas to produce their organic matter. The concentration of dissolved CO2 in surface waters is regulated by gas exchange with the atmosphere, so if carbon is utilized by phytoplankton which then sink, more CO2 is drawn down from the atmosphere to compensate. One of the key steps in the biological pump is, therefore, the sinking of organic material from the surface waters – the faster organic material settles, the more efficiently the biological pump operates. Recent water column observations and complementary studies of deep-sea sediments have demonstrated that diatom mats and large diatoms are important, and in some cases, dominant contributors to the flux of biogenic material to the seafloor.

Research on deep-sea sediment cores, in particular, has shown that such diatoms are locally abundant and may form thick, extensive deposits that have accumulated at rates exceeding 30 cm per 1000 years. These are exceptionally high for the pelagic realm and would normally occur only in areas of high sediment supply near continental margins or beneath coastal upwelling zones. A synergy between biological oceanography and paleoceanography has shown that these diatoms are key players in the biogeochemical cycles of carbon and silica and may be important regulators of global change.

Ecological Significance of Diatom Mats and Large Diatoms In contrast to the relatively well studied, small, and rapidly reproducing diatoms that dominate primary production in coastal settings, the ecology of oceanic mat-forming and large (450 mm) diatoms is relatively poorly understood. These larger diatoms are the most widely distributed in the open ocean but had been regarded by biological oceanographers as typically occurring in very small numbers in oligotrophic (nutrient-depleted) surface waters. Such conditions dominate most of the central gyres of the

Figure 1 Mats of the needle-shaped diatom, Thalassiothrix longissima.

(c) 2011 Elsevier Inc. All Rights Reserved.

651

652

MARINE MATS

ocean and also characterize some marginal seas such as the Mediterranean. These are the deserts of the oceans, but recent observations indicate that these deserts may bloom. An increasing body of field and experimental research on several species of large diatoms has shown that they possess a range of

adaptations that allows them to exploit a unique ecological niche within the ocean. Rhizosolenia mats, for example, are able to regulate their buoyancy, migrating vertically tens or even hundreds of meters between nutrients (nitrate, phosphate, and silica) trapped at depth within the thermocline.

depth

T

thermocline forming

(A)

T

Summer Deep Chlorophyll Maximum

depth

mat-forming and large celled diatoms

thermocline and nutricline

(B)

depth

T

"Fall Dump"

rapid deposition of DCM species

(C) Figure 2 A cartoon representation of annual cycle of diatom production and flux. The autumn or fall dump of large and mat-forming diatoms is contrasted with the spring bloom or upwelling pulse. (A) Spring bloom of diatoms (or upwelling pulse) giving a growth episode lasting several days to a few weeks terminated with rapid aggregation and sinking. (B) Growth of diatoms in well-stratified summer conditions in deep chlorophyll maximum (DCM). Species adapted to low light conditions or to regulate buoyancy and migrate between nutrient-rich depths and bright surface. (C) Fall/winter mixing and resultant breakdown of stratification produces the ‘fall dump’; the mass sedimentation of diatoms which have grown episodically during the period of summer stratification. (Modified from Kemp et al., 2000.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MATS

Other species have adapted to grow in very low light conditions characteristic of nutricline depths (100 m) where they can draw on the essential nutrients and exploit mixing events caused by passing storms. A further adaptation is the coexistence of some diatom genera (including Rhizosolenia and Hemiaulus) with intracellular nitrogen-fixing symbionts that provide them with their essential chemical building blocks unavailable from the surrounding waters. Diatom mats may be up to several centimeters across and host a whole range of other organisms including smaller epiphytic diatoms, parasitic dinoflagellates, foraminifera, and small amphipods.

large and mat-forming diatoms has recently emerged (Figure 2). Most diatom production and flux was previously thought to be generated by small, rapidly growing diatoms in a spring bloom or upwelling pulse. However, a review of sediment trap studies and evidence from laminated sediments shows that large or mat-forming diatoms that grow in stratified waters in the summer and sediment massively when autumn or winter storms disrupt the water column may contribute as much or greater flux to the seafloor than the spring bloom. Ancient examples of this process are the Mediterranean sapropels, black layers whose high organic carbon content may be explained by the contribution from diatom flux.

What Causes the Concentration and Mass Sinking of Diatom Mats and Large Diatoms?

See also

How then are diatom mats and giant diatoms significant for flux to the seafloor? The existence of concentrated deposits of these diatom species in deep-sea sediments, including the giant Ethmodiscus rex, appeared to contradict the received view that they only occurred sparsely in the water column. Yet, until recently, there was only anecdotal historical evidence for the existence of substantial concentrations of these species.

Further Reading

Ocean Frontal Systems

In 1992 a breakthrough occurred, when an oceanographic experiment (the JGOFS–Joint Global Ocean Flux Study) in the eastern equatorial Pacific monitored extensive surface concentrations of the giant diatom Rhizosolenia castrecaniae. The surface concentrations were so intense that they were visible as ‘a line in the sea’ from the NASA Space Shuttle Atlantis. Parallel JGOFS seabed observations showed that these surface concentrations settled rapidly at rates of around 200 m per day over the 4000 m to the seabed. At the same time, Ocean Drilling Program cores, also from the eastern equatorial Pacific, revealed the presence of vast ancient deposits of the diatom Thalassiothrix longissima, a needle-shaped diatom that grows up to 4 mm in length and also forms tangled masses or mats (Figure 1). These layers that extend along the equator for o2000 km record ancient surface concentrations of giant diatoms that settled to the seafloor between 4 and 15 Ma. The ‘Fall Dump’

As a result of sediment trap studies and work on ancient laminated marine sediments (the tree-rings of the oceans) a further explanation for mass flux for

653

Carbon Cycle. Primary Production Distribution. Primary Production Methods. Primary Production Processes.

Alldredge AL and Gotschalk CC (1989) Direct observations of the mass flocculation of diatom blooms: characteristics, settling velocities and formation of diatom aggregates. Deep-Sea Research 36: 159--171. Goldman JC (1993) Potential role of large oceanic diatoms in new primary production. Deep-Sea Research I 40: 159--168. Grimm KA, Lange CB, and Gill AS (1997) Selfsedimentation of phytoplankton blooms in the geologic record. Sedimentary Geology 110: 151--161. Kemp AES and Baldauf JG (1993) Vast Neogene laminated diatom mat deposits from the eastern equatorial Pacific Ocean. Nature 362: 141--143. Kemp AES, Pearce RB, Koizumi I, Pike J, and Rance SJ (1993) The role of mat forming diatoms in formation of Mediterranean sapropels. Nature 398: 57--61. Kemp AES, Pike J, Pearce R, and Lange C (2000) The ‘fall dump’: a new perspective on the role of a shade flora in the annual cycle of diatom production and export flux. Deep-Sea Research II 47: 2129--2154. Sancetta C, Villareal T, and Falkowski P (1991) Massive fluxes of Rhizosolenid diatoms: a common occurrence? Limnology and Oceanography 36: 1452--1457. Smith CR, Hoover DJ, Doan SE, et al. (1996) Phytodetritus at the abyssal seafloor across 101 of latitude in the central equatorial Pacific. Deep-Sea Research II 43: 1309--1338. Smetacek V (2000) The giant diatom dump. Nature 406: 574--575. Villareal TA, Altabet MA, and Culver-Rymsza K (1993) Nitrogen transport by vertically migrating diatom mats in the North Pacific Ocean. Nature 363: 709--712. Yoder JA, Ackleson S, Barber RT, Flamant P, and Balch WA (1994) A line in the sea. Nature 371: 689--692.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MESOCOSMS J. H. Steele, Woods Hole Oceanographic Institution, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1644–1646, & 2001, Elsevier Ltd.

Controlled experiments are the basis of the scientific method. There are obvious difficulties in using this technique when dealing with natural communities or ecosystems, given the great spatial and temporal variability of their environment. On land the standard method is to divide an area of ground, say a field, into a large number of equal plots. Then with a randomized treatment, such as nutrient addition, it is possible to replicate growth of plants and animals over a season.

It is apparent that this approach is not possible in the open sea because of continuous advection and dispersion of water and the organisms in it. Bottomliving organisms are an exception, especially these living near shore, so there have been a wide range of experiments on rocky shores, salt marshes, and sea grasses. But even there, the critical reproductive period for most animals involves dispersion of the larvae in a pelagic phase. Also these experiments require continuous exchange of sea water. For the completely pelagic plants and animals, short-term experiments – usually a few days – on single species are used to study physiological responses. There can be 24-hour experimental measurements of the rates of grazing of copepods on phytoplankton in liter bottles. But for studies of longer-term interactions, much larger volumes of

Support raft Gantry Walkway 10m

Steel frame Polystyrene block

Plankton net

5m

Counterweight Bag

0

Down rope Bottom cone Sampling hose Bottom frame 40 kg steel weight Figure 1 The design of a mesocosm used in Loch Ewe, Scotland for studies of the dynamics of plankton communities and of fish larval growth and mortality (adapted from Davies and Gamble, 1979).

654

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE MESOCOSMS

water are necessary, to contain whole communities and to minimize wall effects of the containers. To this end ‘mesocosms’ – containers much larger than can fit into the normal laboratory – have been used in a variety of designs and for a diversity of purposes. The first choice is whether to construct these on land, at the sea’s edge, or to immerse them in the sea. The former has advantages in durability, ease of access, and re-use. There are constraints on the volumes that can be contained, difficulties in temperature control, and, especially, problems in transferring representative marine communities from the sea to the tanks. This approach was used originally in tall relatively narrow tanks to study populations of copepods and fish larvae; in particular to experiment on factors such as light that control vertical migration. Another use of such large tanks is to study the effect of pollutants on communities of pelagic and benthic organisms. These shore-based tanks are limited by the weight of water, usually to volumes of 10–30 m3. Enclosures immersed in the sea do not have this constraint. Instead the problems concern the strength of the flexible materials used for the walls in relation to currents and, especially, wind-induced waves. For this reason, such enclosures are placed in sheltered semienclosed places such as fiords. Nylon-reinforced polythene or vinyl reinforced with fabric have been used for these large ‘test-tubes’ containing 300– 3000 m3 (Figure 1). A column of water containing the natural plankton is captured by drawing up the bag from the bottom and fastening it in a rigid frame. The water and plankton can then be sampled by normal oceanographic methods. It is possible to maintain at least three trophic levels – phytoplankton, copepods, and fish larvae – for 100 days or more. The only necessary treatment is addition of nutrients to replace those in the organic matter that sinks out. Such mesocosms can also be used for study of the fates and effects of pollutants. These mesocosms have the obvious advantages associated with their large volumes – numerous animals for sampling, minimal wall effects. Temperature is regulated by exchange of heat through the walls. But they have various drawbacks. Not only is

655

advection suppressed but vertical mixing decreases so that the outside physical conditions are not reproduced. The greatest disadvantage, however, is lack of adequate replication. There have been only three to six of these mesocosms available for any experiment and pairs did not often agree closely. Thus each tube represents an ecosystem on its own rather than a replicate of a larger community. The need for experimental results at the community level represents an unresolved problem in biological oceanography. There are smaller-scale experiments continuing. Open mesh containers through which water and plankton pass can be a compromise for the study of small fish and fish larvae. It is now possible to mark a body of water with very sensitive tracers and follow the effects on plankton of the addition of nutrients, specifically iron, for several weeks. The concatenation of these results may have to depend on computer simulations.

See also Copepods. Fish Larvae. Iron Population Dynamics Models.

Fertilization.

Further Reading Cowan JH and Houde ED (1990) Growth and survival of bay anchovy in mesocosm enclosures. Marine Ecology Progress Series 68: 47--57. Davies JM and Gamble JC (1979) Experiments with large enclosed ecosystems. Philosophical Transcations of the Royal Society, B. Biological Sciences 286: 523--544. Gardner RH, Kemp WM, Kennedy VS, and Peterson JS (eds.) (2001) Scaling Relations in Experimental Ecology. New York: Columbia University Press. Grice GD and Reeve MR (eds.) (1982) Marine Mesocosms. New York: Springer-Verlag. Lalli CM (ed.) (1990) Enclosed Experimental Marine Ecosystems: A Review and Recommendations. Coastal and Estuarine Studies 37. New York: Springer-Verlag. Underwood AJ (1997) Experiments in Ecology. Cambridge: Cambridge University Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PLANKTON COMMUNITIES G.-A. Paffenho¨fer, Skidaway Institute of Oceanography, Savannah, GA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction By definition, a community is an interacting population of various kinds of individuals (species) in a common location (Webster’s Collegiate Dictionary, 1977). The objective of this article is to provide general information on the composition and functioning of various marine plankton communities, which is accompanied by some characteristic details on their dynamicism. General Features of a Plankton Community

The expression ‘plankton community’ implies that such a community is located in a water column. It has a range of components (groups of organisms) that can be organized according to their size. They range in size from tiny single-celled organisms such as bacteria (0.4–1-mm diameter) to large predators like scyphomedusae of more than 1 m in diameter. A common method which has been in use for decades is to group according to size, which here is attributed to the organism’s largest dimension; thus the organisms range from picoplankton to macroplankton (Figure 1). It is, however, the smallest dimension of an organism which usually determines whether it is retained by a mesh, since in a flow, elongated particles align themselves with the flow. A plankton community is operating/functioning continuously, that is, physical, chemical, and biological variables are always at work. Interactions among its components occur all the time. As one well-known fluid dynamicist stated, ‘‘The surface of the ocean can be flat calm but below that surface there is always motion of the water at various scales.’’ Many of the particles/organisms are moving or being moved most of the time: Those without flagella or appendages can do so due to processes within or due to external forcing, for example, from water motion due to internal waves; and those with flagella/cilia or appendages or muscles move or create motion of the water in order to exist. Oriented motion is usually in the vertical which often results in distinct layers of certain organisms. However,

656

physical variables also, such as light or density differences of water masses, can result in layering of planktonic organisms. Such layers which are often horizontally extended are usually referred to as patches. As stated in the definition, the components of a plankton community interact. It is usually the case that a larger organism will ingest a smaller one or a part of it (Figure 1). However, there are exceptions. The driving force for a planktonic community originates from sun energy, that is, primary productivity’s (1) direct and (2) indirect products: (1) autotrophs (phytoplankton cells) which can range from near 2 to more than 300-mm width/diameter, or chemotrophs; and (2) dissolved organic matter, most of which is released by phytoplankton cells and protozoa as metabolic end products, and being taken up by bacteria and mixo- and heterotroph protozoa (Figure 1). These two components mainly set the microbial loop (ML; (see Bacterioplankton and Protozoa, Planktonic Foraminifera)) in motion; that is, unicellular organisms of different sizes and behaviors (auto-, mixo-, and heterotrophs) depend on each other – usually, but not always, the smaller being ingested by the larger. Most of nutrients and energy are recirculated within this subcommunity of unicellular organisms in all marine regions of our planet (see Bacterioplankton, Phytoplankton Blooms, Phytoplankton Size Structure, and Protozoa, Planktonic Foraminifera for more details, especially the ML). These processes of the ML dominate the transfer of energy in all plankton communities largely because the processes (rates of ingestion, growth, reproduction) of unicellular heterotrophs almost always outpace those of phytoplankton, and also of metazooplankton taxa at most times. The main question actually could be: ‘‘What is the composition of plankton communities, and how do they function?’’ Figure 1 reveals sizes and relationships within a plankton community including the ML. It shows the so-called ‘bottom-up’ and ‘topdown’ effects as well as indirect effects like the above-mentioned labile dissolved organic matter (labile DOM), released by auto- and also by heterotrophs, which not only drives bacterial growth but can also be taken up or used by other protozoa. There can also be reversals, called two-way processes. At times a predator eating an adult metazoan will be affected by the same metazoan which is able to eat the predator’s early juveniles (e.g., well-grown ctenophores capturing adult omnivorous copepods

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PLANKTON COMMUNITIES

Dimensions (µm)

Autotrophs

657

Heterotrophs

copepods, {Large salps, doliolids } Macrozooplankton Chaetognaths Ctenophores Coelenterates

}

2000

{

}

Large diatoms Diatom chains

Mesoplankton

}

Copepodid stages and Adults of small copepods Nauplii of large copepods Early zooids of salps and doliolids

200

Microplankton

{

Diatoms { Dinoflagellates

Dinoflagellates Ciliates Nauplii

20

Nanoplankton

{

Dinoflagellates

Diatoms Dinoflagellates Flagellates

Ciliates

}

Flagellates Lab

2

Picoplankton

ile

DO

M

Bacteria

Prokaryotes

0.2 Figure 1 Interactions within a plankton community separated into size classes of auto- and heterotrophs, including the microbial loop; the arrows point to the respective grazer, or receiver of DOM; the figure is partly related to figure 9 from Landry MR and Kirchman DL (2002) Microbial community structure and variability in the tropical Pacific. Deep-Sea Research II 49: 2669–2693.

which have the ability to capture and ingest very young ctenophores). To comprehend the functioning of a plankton community requires a quantitative assessment of the

abundances and activities of its components. First, almost all of our knowledge to date stems from in situ sampling, that is, making spot measurements of the abundance and distribution of organisms in

(c) 2011 Elsevier Inc. All Rights Reserved.

658

MARINE PLANKTON COMMUNITIES

the water column. The accurate determination of abundance and distribution requires using meshes or devices which quantitatively collect the respective organisms. Because of methodological difficulties and insufficient comprehension of organisms’ sizes and activities, quantitative sampling/quantification of a community’s main components has been often inadequate. The following serves as an example of this. Despite our knowledge that copepods consist of 11 juvenile stages aside of adults, the majority of studies of marine zooplankton hardly considered the juveniles’ significance and this manifested itself in sampling with meshes which often collected merely the adults quantitatively. Second, much knowledge on rate processes comes from quantifying the respective organisms’ activities under controlled conditions in the laboratory. Some in situ measurements (e.g., of temperature, salinity, chlorophyll concentrations, and acoustic recordings of zooplankton sizes) have been achieved ‘continuously’ over time, resulting in time series of increases and decreases of certain major community components. To date there are few, if any, direct in situ observations on the activity scales of the respective organisms, from bacteria to proto- and to metazooplankton, mainly because of methodological difficulties. In essence, our present understanding of processes within plankton communities is incomplete.

Specific Plankton Communities We will provide several examples of plankton communities of our oceans. They will include information about the main variables affecting them, their main components, partly their functioning over time, including particular specifics characterizing each of those communities. In this section, plankton communities are presented for three different types of marine environments: estuaries/inshore, continental shelves, and open ocean regions. Estuaries

Estuaries and near-shore regions, being shallow, will rapidly take up and lose heat, that is, will be strongly affected by atmospheric changes in temperature, both short- and long-term, the latter showing in the seasonal extremes ranging from 2 to 32 1C in estuaries of North Carolina. Runoff of fresh water, providing continuous nutrient input for primary production, and tides contribute to rapid changes in salinity. This implies that resident planktonic taxa ought to be eurytherm as well as – therm. Only very

few metazooplanktonic species are able to exist in such an environment (Table 1). In North Carolinian estuaries, representative of other estuaries, they are the copepod species Acartia tonsa, Oithona oculata, and Parvocalanus crassirostris. In estuaries of Rhode Island, two species of the genus Acartia occur. During colder temperatures Acartia hudsonica produces dormant eggs as temperatures increase and then is replaced by A. tonsa, which produces dormant eggs once temperatures again decrease later in the year. Such estuaries are known for high primary productivity, which is accompanied by high abundances of heterotroph protozoa preying on phytoplankton. Such high abundances of unicellular organisms imply that food is hardly limiting the growth of the abovementioned copepods which can graze on auto- as well as heterotrophs. However, such estuaries are often nursery grounds for juvenile fish like menhaden which prey heavily on late juveniles and adults of such copepods, especially Acartia, which is not only the largest of those three dominant copepod species but also moves the most, and thus can be seen most easily by those visual predators. This has resulted in diurnal migrations mostly of their adults, remaining at the seafloor during the day where they hardly eat, thus avoiding predation by such visual predators, and only entering the water column during dark hours. That then is their period of pronounced feeding. The other two species which are not heavily preyed upon by juvenile fish, however, can be affected by the co-occurring Acartia, because from early copepodid stages on this genus can be strongly carnivorous, readily preying on the nauplii of its own and of those other species. Nevertheless, the usually continuous abundance of food organisms for all stages of the three copepod species results in high concentrations of nauplii which in North Carolinian estuaries can reach 100 l 1, as can their combined copepodid stages. The former is an underestimate, because sampling was done with a 75-mm mesh, which is passed through by most of those nauplii. By comparison, in an estuary on the west coast of Japan (Yellow Sea), dominated also by the genera Acartia, Oithona, and Paracalanus and sampling with 25mm mesh, nauplius concentrations during summer surpassed 700 l 1, mostly from the genus Oithona. And copepodid stages plus adults repeatedly exceeded 100 l 1. Here sampling with such narrow mesh ensured that even the smallest copepods were collected quantitatively. In essence, estuaries are known to attain among the highest concentrations of proto- and metazooplankton. The known copepod species occur during most of the year, and are observed year after year

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved.

Cop B5–100 l

Acartia Oithona Parvocalanus

Abundance

Dominant metazooplankton taxa

b

Nauplii. Copepodids and adult copepods.

NaB10–500 l

Copepod Ranges

a

o5–50 l

High spring and summer, low winter

Seasonal variability of metazoan abundance

1

r5

No. of metazoan species

1

410

B10–30 Highly variable

High at most times

Primary Productivity

b

Maximum in spring

Intermittently high

Flagellates, diatoms

Phytoplankton composition

Oithona Paracalanus Temora Doliolida

o3–30 l

1

1

Flagellates, diatoms, dinoflagellates

Intermittently high

Neocalanus Oithona Metridia

Neocalanus

Up to 1000 m

High

3

Nanoflagellates

Always low

Seasonal

High from spring to autumn

Episodic

Continuous

Phytoplankton abundance

Steady salinity, seasonal temp. variability

Subarctic Pacific

Open ocean gyres

Nutrient supply

Intermittent and seasonal atmospheric forcing

Shelves

Wide range of temperature and salinity

Estuaries

Some characteristics of marine plankton communities

Physical variables

Table 1

3

Calanus Oithona Oncaea

C. finmarchicus

Up to 1000 m

High

420

Max. in spring and autumn

Spring: diatoms Other: mostly nanoplankton

Major spring bloom

Seasonal

Major seasonal variability of temperature

Boreal Atlantic

1 3

Oithona Clausocalanus Oncaea

300–1000 m

3–10 l

Low

4100

Always low

Mostly prokaryotes, small nano- and dinoflagellates

Always low

Occasional

Steady temperature and salinity, continuous atmospheric forcing

Atlantic/Pacific

Epipelagic subtropical

MARINE PLANKTON COMMUNITIES 659

660

MARINE PLANKTON COMMUNITIES

which implies persistence of those species beyond decades. Continental Shelves

By definition they extend to the 200-m isobath, and range from narrow (few kilometers) to wide (more than 100-km width). The latter are of interest because the former are affected almost continuously and entirely by the nearby open ocean. Shelves are affected by freshwater runoff and seasonally changing physical variables. Water masses on continental shelves are evaluated concerning their residence time, because atmospheric events sustained for more than 1 week can replace most of the water residing on a wide shelf with water offshore but less so from near shore. This implies that plankton communities on wide continental shelves, which are often near boundary currents, usually persist for limited periods of time, from weeks to months (Table 1). They include shelves like the Agulhas Bank, the Campeche Banks/Yucatan Shelf, the East China Sea Shelf, the East Australian Shelf, and the US southeastern continental shelf. There can be a continuous influx yearround of new water from adjacent boundary currents as seen for the Yucatan Peninsula and Cape Canaveral (Florida). The momentum of the boundary current (here the Yucatan Current and Florida Current) passing a protruding cape will partly displace water along downstream-positioned diverging isobaths while the majority will follow the current’s general direction. This implies that upstream-produced plankton organisms can serve as seed populations toward developing a plankton community on such wide continental shelves. Whereas estuarine plankton communities receive almost continuously nutrients for primary production from runoff and pronounced benthic-pelagic coupling, those on wide continental shelves infrequently receive new nutrients. Thus they are at most times a heterotroph community unless they obtain nutrients from the benthos due to storms, or receive episodically input of cool, nutrient-rich water from greater depths of the nearby boundary current as can be seen for the US SE shelf. Passing along the outer shelf at about weekly intervals are nutrient-rich cold-core Gulf Stream eddies which contain plankton organisms from the highly productive Gulf of Mexico. Surface winds, displacing shelf surface water offshore, lead to an advance of the deep cool water onto the shelf which can be flooded entirely by it. Pronounced irradiance and high-nutrient loads in such upwellings result in phytoplankton blooms which then serve as a food source for protozoo- and metazooplankton. Bacteria concentrations in such

cool water masses increase within several days by 1 order of magnitude. Within 2–3 weeks most of the smaller phytoplankton (c. o20-mm width) has been greatly reduced, usually due to grazing by protozoa and relatively slow-growing assemblages of planktonic copepods of various genera such as Temora, Oithona, Paracalanus, Eucalanus, and Oncaea. However, quite frequently, the Florida Current which becomes the Gulf Stream carries small numbers of Thaliacea (Tunicata), which are known for intermittent and very fast asexual reproduction. Such salps and doliolids, due to their high reproductive and growth rate, can colonize large water masses, the latter increasing from B5 to 4500 zooids per cubic meter within 2 weeks, and thus form huge patches, covering several thousands of square kilometers, as the cool bottom water is displaced over much of the shelf. The increased abundance of salps (usually in the warmer and particle-poor surface waters) and doliolids (mainly in the deeper, cooler, particle-rich waters, also observed on the outer East China shelf) can control phytoplankton growth once they achieve bloom concentrations. The development of such large and dense patches is partly due to the lack of predators. Although the mixing processes between the initially quite cool intruding bottom (13–20 1C) and the warm, upper mixed layer water (27–28 1C) are limited, interactions across the thermocline occur, thus creating a plankton community throughout the water column of previously resident and newly arriving components. The warm upper mixed layer often has an extraordinary abundance of early copepodid stages of the poecilostomatoid copepod Oncaea, thanks to their ontogenetical migration after having been released by the adult females which occur exclusively in the cold intruding water. Also, early stages of the copepod Temora turbinata are abundant in the warm upper mixed layer; while T. turbinata’s late juvenile stages prefer the cool layer because of the abundance of large, readily available phytoplankton cells. As in estuaries, the copepod genus Oithona flourishes on warm, temperate, and polar continental shelves throughout most of the euphotic zone. Such wide subtropical shelves will usually be well mixed during the cooler seasons, and then harbor, due to lower temperatures, fewer metazoplankton species which are often those tolerant of wider or lower temperature ranges. Such wide shelves are usually found in subtropical regions, which explains the rapidity of the development of their plankton communities. They, however, are also found in cooler climates, like the wide and productive Argentinian/ Brazilian continental shelf about which our knowledge

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PLANKTON COMMUNITIES

is limited. Other large shelves, like the southern North Sea, have a limited exchange of water with the open ocean but at the same time considerable influx of runoff, plus nutrient supply from the benthos due to storm events, and thus can maintain identical plankton communities over months and seasons. In essence, continental shelf plankton communities are usually relatively short-lived, which is largely due to their water’s limited residence time. Open Ocean

The open ocean, even when not including ocean margins (up to 1000-m water column), includes by far the largest regions of the marine environment. Its deep-water columns range from the polar seas to the Tropics. All these regions are under different atmospheric and seasonal regimes, which affect plankton communities. Most of these communities are seasonally driven and have evolved along the physical conditions characterizing each region. The focus here is on gyres as they represent specific ocean communities whose physical environment can be readily presented. Gyres represent huge water masses extending horizontally over hundreds to even thousands of kilometers in which the water moves cyclonically or anticyclonically (see Ocean Gyre Ecosystems). They are encountered in subpolar, temperate, and subtropical regions. The best-studied ones are (see Open Ocean Convection):

• • •

subpolar: Alaskan Gyre; temperate: Norwegian Sea Gyre, Labrador– Irminger Sea Gyre; subtropical: North Pacific Central Gyre (NPCG), North Atlantic Subtropical Gyre (NASG).

The Alaskan Gyre is part of the subarctic Pacific (Table 1) and is characterized physically by a shallow halocline (B110-m depth) which prevents convective mixing during storms. Biologically it is characterized by a persistent low-standing stock of phytoplankton despite high nutrient abundance, and several species of large copepods which have evolved to persist via a life cycle as shown for Neocalanus plumchrus. By midsummer, fifth copepodids (C5) in the upper 100 m which have accumulated large amounts of lipids begin to descend to greater depths of 250 m and beyond undergoing diapause, and eventually molt to females which soon begin to spawn. Spawning females are found in abundance from August to January. Nauplii living off their lipid reserves and moving upward begin to reach surface waters by mid- to late winter as copepodid stage 1 (C1), and start feeding on the abundant small phytoplankton cells (probably

661

passively by using their second maxillae, but mostly by feeding actively on heterotrophic protozoa which are the main consumers of the tiny phytoplankton cells). The developing copepodid stages accumulate lipids which in C5 can amount to as much as 50% of their body mass, which then serve as the energy source for metabolism of the females at depth, ovary development, and the nauplii’s metabolism plus growth. While the genus Neocalanus over much of the year provides the highest amount of zooplankton biomass, the cyclopoid Oithona is the most abundant metazooplankter; other abundant metazooplankton taxa include Euphausia pacifica, and in the latter part of the year Metridia pacifica and Calanus pacificus. In the temperate Atlantic (Table 1), the Norwegian Sea Gyre maintains a planktonic community which is characterized, like much of the temperate oceanic North Atlantic, by the following physical features. Pronounced winds during winter mix the water column to beyond 400-m depth, being followed by lesser winds and surface warming resulting in stratification and a spring bloom of mostly diatoms, and a weak autumn phytoplankton bloom. A major consumer of this phytoplankton bloom and characteristic of this environment is the copepod Calanus finmarchicus, occurring all over the cool North Atlantic. This species takes advantage of the pronounced spring bloom after emerging from diapause at 4400-m depth, by moulting to adult, and grazing of females at high clearance rates on the diatoms, right away starting to reproduce and releasing up to more than 2000 fertilized eggs during their lifetime. Its nauplii start to feed as nauplius stage 3 (N3), being able to ingest diatoms of similar size as the adult females, and can reach copepodid stage 5 (C5) within about 7 weeks in the Norwegian Sea, accumulating during that period large amounts of lipids (wax ester) which serve as the main energy source for the overwintering diapause period. Part of the success of C. finmarchicus is found in its ability of being omnivorous. C5s either descend to greater depths and begin an extended diapause period, or could moult to adult females, thus producing another generation which then initiates diapause at mostly C5. Its early to late copepodid stages constitute the main food for juvenile herring which accumulate the copepods’ lipids for subsequent overwintering and reproduction. Of the other copepods, the genus Oithona together with the poecilostomatoid Oncaea and the calanoid Pseudocalanus were the most abundant. Subtropical and tropical parts of the oceans cover more than 50% of our oceans. Of these, the NPCG, positioned between c. 101 and 451 N and moving anticyclonically, has been frequently studied. It includes a southern and northern component, the latter

(c) 2011 Elsevier Inc. All Rights Reserved.

662

MARINE PLANKTON COMMUNITIES

being affected by the Kuroshio and westerly winds, the former by the North Equatorial Current and the trade winds. Despite this, the NPCG has been considered as an ecosystem as well as a huge plankton community. The NASG, found between c. 151 and 401 N and moving anticyclonically, is of similar horizontal dimensions. There are close relations between subtropical and tropical communities; for example, the Atlantic south of Bermuda is considered close to tropical conditions. Vertical mixing in both gyres is limited. Here we focus on the epipelagic community which ranges from the surface to about 150-m depth, that is, the euphotic zone. The epipelagial is physically characterized by an upper mixed layer of c. 15–40 m of higher temperature, below which a thermocline with steadily decreasing temperatures extends to below 150-m depth. In these two gyres, the concentrations of phytoplankton hardly change throughout the year in the epipelagic (Table 1) and together with the heterotrophic protozoa provide a low and quite steady food concentration (Table 1) for higher trophic levels. Such very low particle abundances imply that almost all metazooplankton taxa depending on them are living on the edge, that is, are severely food-limited. Despite this fact, there are more than 100 copepod species registered in the epipelagial of each of the two gyres. How can that be? Almost all these copepod species are small and rather diverse in their behavior: the four most abundant genera have different strategies to obtain food particles: the intermittently moving Oithona is found in the entire epipelagial and depends on moving food particles (hydrodynamic signals); Clausocalanus is mainly found in the upper 50 m of the epipelagial and always moves at high speed, thus encountering numerous food particles, mainly via chemosensory; Oncaea copepodids and females occur in the lower part of the epipelagial and feed on aggregates; and the feeding-current producing Calocalanus perceives particles via chemosensory. This implies that any copepod species can persist in these gyres as long as it obtains sufficient food for growth and reproduction. This is possible because protozooplankton always controls the abundance of available food particles; thus, there is no competition for food among the metazooplankton. In addition, since total copepod abundance (quantitatively collected with a 63-mm mesh by three different teams) is steady and usually o1000 m 3 including copepodid stages (pronounced patchiness of metazooplankton has not yet been observed in these oligotrophic waters), the probability of encounter (only a minority of the zooplankton is carnivorous on metazooplankton) is very low, and therefore the probability of predation low

within the metazooplankton. In summary, these steady conditions make it possible that in the epipelagial more than 100 copepod species can coexist, and are in a steady state throughout much of the year.

Conclusions All epipelagic marine plankton communities are at most times directly or indirectly controlled or affected by the activity of the ML, that is, unicellular organisms. Most of the main metazooplankton species are adapted to the physical and biological conditions of the respective community, be it polar, subpolar, temperate, subtropical, or tropical. The only metazooplankton genus found in all communities mentioned above, and also all other studied marine plankton communities, is the copepod genus Oithona. This copepod has the ability to persist under adverse conditions, for example, as shown for the subarctic Pacific. This genus can withstand the physical as well as biological (predation) pressures of an estuary, the persistent very low food levels in the warm open ocean, and the varying conditions of the Antarctic Ocean. Large copepods like the genus Neocalanus in the subarctic Pacific, and C. finmarchicus in the temperate to subarctic North Atlantic are adapted with respective distinct annual cycles in their respective communities. Among the abundant components of most marine plankton communities from near shore to the open ocean are appendicularia (Tunicata) and the predatory chaetognaths. Our present knowledge of the composition and functioning of marine planktonic communities derives from (1) oceanographic sampling and time series, optimally accompanied by the quantification of physical and chemical variables; and (2) laboratory/onboard experimental observations, including some time series which provide results on small-scale interactions (microns to meters; milliseconds to hours) among components of the community. Optimally, direct in situ observations on small scales in conjunction with respective modeling would provide insights in the true functioning of a plankton community which operates continuously on scales of milliseconds and larger. Our future efforts are aimed at developing instrumentation to quantify in situ interactions of the various components of marine plankton communities. Together with traditional oceanographic methods we would go ‘from small scales to the big picture’, implying the necessity of understanding the functioning on the individual scale for a comprehensive understanding as to how communities operate.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PLANKTON COMMUNITIES

See also Bacterioplankton. Chemical Processes in Estuarine Sediments. Continuous Plankton Recorders. Copepods. Estuarine Circulation. Gas Exchange in Estuaries. Gelatinous Zooplankton. Ocean Gyre Ecosystems. Phytoplankton Blooms. Phytoplankton Size Structure. Plankton. Plankton and Climate. Protozoa, Planktonic Foraminifera. Small-Scale Physical Processes and Plankton Biology. Zooplankton Sampling with Nets and Trawls.

Further Reading Atkinson LP, Lee TN, Blanton JO, and Paffenho¨fer G-A (1987) Summer upwelling on the southeastern continental shelf of the USA during 1981: Hydrographic observations. Progress in Oceanography 19: 231--266. Fulton RS, III (1984) Distribution and community structure of estuarine copepods. Estuaries 7: 38--50. Hayward TL and McGowan JA (1979) Pattern and structure in an oceanic zooplankton community. American Zoologist 19: 1045--1055. Landry MR and Kirchman DL (2002) Microbial community structure and variability in the tropical Pacific. Deep-Sea Research II 49: 2669--2693. Longhurst AR (1998) Ecological Geography of the Sea, 398pp. San Diego, CA: Academic Press. Mackas DL and Tsuda A (1999) Mesozooplankton in the eastern and western subarctic Pacific: Community structure, seasonal life histories, and interannual variability. Progress in Oceanography 43: 335--363. Marine Zooplankton Colloquium 1 (1998) Future marine zooplankton research – a perspective. Marine Ecology Progress Series 55: 197--206.

663

Menzel DW (1993) Ocean Processes: US Southeast Continental Shelf, 112pp. Washington, DC: US Department of Energy. Miller CB (1993) Pelagic production processes in the subarctic Pacific. Progress in Oceanography 32: 1--15. Miller CB (2004) Biological Oceanography, 402pp. Boston: Blackwell. Paffenho¨fer G-A and Mazzocchi MG (2003) Vertical distribution of subtropical epiplanktonic copepods. Journal of Plankton Research 25: 1139--1156. Paffenho¨fer G-A, Sherman BK, and Lee TN (1987) Summer upwelling on the southeastern continental shelf of the USA during 1981: Abundance, distribution and patch formation of zooplankton. Progress in Oceanography 19: 403--436. Paffenho¨fer G-A, Tzeng M, Hristov R, Smith CL, and Mazzocchi MG (2003) Abundance and distribution of nanoplankton in the epipelagic subtropical/tropical open Atlantic Ocean. Journal of Plankton Research 25: 1535--1549. Smetacek V, DeBaar HJW, Bathmann UV, Lochte K, and Van Der Loeff MMR (1997) Ecology and biogeochemistry of the Antarctic Circumpolar Current during austral spring: A summary of Southern Ocean JGOFS cruise ANT X/6 of RV Polarstern. Deep-Sea Research II 44: 1--21 (and all articles in this issue). Speirs DC, Gurney WSC, Heath MR, Horbelt W, Wood SN, and de Cuevas BA (2006) Ocean-scale modeling of the distribution, abundance, and seasonal dynamics of the copepod Calanus finmarchicus. Marine Ecology Progress Series 313: 173--192. Tande KS and Miller CB (2000) Population dynamics of Calanus in the North Atlantic: Results from the TransAtlantic Study of Calanus finmarchicus. ICES Journal of Marine Science 57: 1527 (entire issue). Webber MK and Roff JC (1995) Annual structure of the copepod community and its associated pelagic environment off Discovery Bay, Jamaica. Marine Biology 123: 467--479.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE POLICY OVERVIEW P. Hoagland, Woods Hole Oceanographic Institution, Woods Hole, MA, USA P. C. Ticco, Massachusetts Maritime Academy, Buzzards Bay, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1646–1654, & 2001, Elsevier Ltd.

Introduction Marine policy is an academic field in which approaches from social science disciplines are applied to problems arising out of the human use of the oceans. Usually, human actions affecting ocean resources take place within an institutional context: laws establish a system of enforceable property rights, and goods and services are exchanged through markets. Most marine policy problems involve institutional imperfections or ‘failures.’ Governance failures include ill-defined property rights, the incomplete integration of the actions of public agencies operating under separate authorities, and wasteful ‘rent seeking’ on the part of stakeholders. Market imperfections include oil spills, nutrient runoffs leading to eutrophication in coastal seas, and overexploitation of commercial fish stocks, among others. Even in the absence of technically defined

Table 1

institutional failures, problems may arise when decisions allocating marine resources are perceived to be unfair. Most marine policy issues are subsets of broader policy areas. Some examples are presented in Table 1. Marine policy can be distinguished from these more general policy areas because legal property rights in the ocean often differ from those found on land. One reason for this difference is the relatively high cost of monitoring and enforcing private property rights in a remote and sometimes hostile environment. Other reasons include the fugitive nature of biological resources and the ease with which nutrients and pollutants are dispersed by currents and other physical processes. The existence of these characteristics argues for collective action (i.e., the exercise of public authority) as a means of optimizing human uses and managing conflicts among users. The nature of collective action covers a spectrum from a centralized system of government ‘command and control’ to the implementation of decentralized ‘market-based approaches.’ The goal of marine policy analysis is to identify alternative courses of action for addressing a problem of ocean resource use and to inform public and private decision makers about the likely consequences. Consequences include physical, ecological, economic, and distributional (equity) effects.

Some examples of overlaps of marine policy with general public policy areas

General policy area

Marine policy focuses

Environment

Ocean and climate change; macronutrient fluxes; eutrophication and hypoxia; treated and untreated sewage effluent; oil and hazardous material spills; industrial chemical and heavy metal effluents; thermal effluents from power plants Commercial and recreational fisheries management; ocean minerals exploration and management; aquaculture regulation; conservation of protected species (mammals, birds, reptiles, fish, corals); ecosystem management; marine protected areas; conservation of biological diversity Offshore oil and gas development; tidal power; ocean thermal energy conversion Coastal zone management; planning; zoning uses; barrier beach protection Solid waste disposal; sewage sludge disposal; marine debris; nuclear waste disposal; incineration at sea Shipping and ports; underwater cables and pipelines; safety of life at sea; aids to navigation; international rights of passage; salvage; admiralty law Zoned training and testing areas; atomic free zones; acoustic pollution; rights of passage for military vessels Legal geography; piracy; international trade in protected species; refugees; high seas fisheries; transboundary pollution Weather prediction; hurricanes; coastal flood insurance; tsunamis; harmful algal blooms; search and rescue Funding for oceanographic research; technology transfer; basic versus applied research; large-scale science programs

Natural resources

Energy Land use Waste management Transportation Defense Foreign policy Emergency management Science policy

664

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE POLICY OVERVIEW

In any particular situation, the universe of policy alternatives is constrained by the environmental characteristics of the ocean, the range of feasible technological responses, financial resources, and, sometimes, institutional frameworks and processes. History of the Field

The emergence of marine policy as a distinct field of research dates back only about 40 years, coincident with rapid increases in ocean uses, the maturation of oceanography as a scientific field of study, and the rise of environmentalism. A number of journals specializing in public policy topics concerning the oceans, estuaries, and the coastal zone began publication in the early 1970s. Among these journals are: Coastal Management, Marine Policy, Marine Policy Reports, Marine Resource Economics, Maritime Law and Commerce, Maritime Policy and Management, Ocean Development and International Law, and Ocean and Shoreline Management. More recent additions to this list include: the International Journal of Marine and Coastal Law and the Ocean and Coastal Law Journal. Many marine policy problems predate this period, such as those relating to national security, international boundary determinations, resource exploitation, and shipping. In earlier periods, however, marine policy was not easily distinguishable from other, more general policy areas. The negotiations on the third United Nations Convention on the Law of the Sea (UNCLOS) (1970–1982) may have spurred the development of the field, as many academic institutions in the West established programs in marine policy in the early 1970 s. For example, the Marine Policy Center (then the Marine Policy and Ocean Management program) Table 2

665

was established in 1972 at the Woods Hole Oceanographic Institution by Paul Fye, who was then director of the Institution. At that time, one main purpose of the Center was to follow and analyze the potential international regulation of marine scientific research, a focus of debate at the UNCLOS negotiations. Two institutions, the Law of the Sea Institute (1966 to present) and the Center for Ocean Law and Policy at the University of Virginia (1976 to present), have published on a continuous basis the proceedings of their annual meetings on international law of the sea issues. Since 1978, the International Ocean Institute, located jointly at the University of Malta and Dalhousie University, has published an annual Ocean Yearbook that features scholarly articles on marine policy topics, compiles descriptive statistics of ocean uses and legal geography, and summarizes the activities of marine policy research centers worldwide. Social Science Disciplines

Marine policy is often described as a multidisciplinary field. Although academic degrees are issued in the United States and Europe in the field of marine policy or ‘marine affairs,’ progress in understanding marine policy problems typically occurs within the confines of more traditional social science disciplines. Alternative points of view may arise from the application of methods from different disciplines to a specific policy problem. The social sciences are divided into a number of well-established disciplines. Some of these disciplines are listed in Table 2, along with examples of recent research topics to which they have been applied. Notably, considerable overlap may exist in the

Social science disciplines and some examples of research foci

Discipline

Some example research foci

Cultural anthropology Economics

Analysis of the effects of fisheries management on fishing communities; underwater archaeology research Development of bioeconomic models of fisheries; estimating the net benefits of fisheries regulation; valuation of the nonmarket benefits of coastal and marine recreation; measurement of damages from marine pollution; evaluation of the net benefits of alternative policy instruments for controlling marine pollution Mapping and analysis of demographic, resource, and economic data using geographic information systems History of oceanography as a science; characterization of laws, social norms, and customs from earlier societies Analysis of legal institutions governing the use of marine resources; interpretation of common and statutory law with respect to ocean resource use Identification and interpretation of the principles of environmental ethics as they apply to marine resource uses and conservation Forecasting coastal and marine resource uses; demographic trends in the coastal zone; zoning the marine environment; marine protected areas; control of land use in the coastal zone Analyzing common property institutions; characterizing the effectiveness of international environmental institutions; international regime formation Effects of fisheries management on fishing communities; importance of institutions in control of resource use

Geography History Law Philosophy Planning Political science Sociology

(c) 2011 Elsevier Inc. All Rights Reserved.

666

MARINE POLICY OVERVIEW

disciplinary coverage of certain topics, such as fisheries management. Ocean Resources and Uses

The uses of the ocean for transportation, as a source of protein, and as a sink for wastes are among its oldest. In ancient times, the supply of ocean space and fish were thought to be virtually without limit. Modern humans have demonstrated that some uses of the ocean can preclude other uses, underscoring the existence of limits to the supply of space and resources, and giving rise to the potential for conflicts across uses. Since its modern development in the 1930s, oceanographic research has made significant strides in characterizing the distribution of ocean resources, although substantial uncertainties persist. A broad range of ocean uses can be mapped into a small set of ocean resources. These resources include ocean space, living resources and their habitats, nonliving resources, and energy. Table 3 lists the most prominent uses of the ocean along with a summary of typical marine policy issues that arise as a consequence of institutional imperfections.

Institutional Frameworks Marine resources, their utilization, and ocean space are all managed through a myriad of legal instruments. These instruments exist at all levels of governance, including those policies directed at local or subnational concerns, and those designed to address issues of national, regional, or global importance. A seventeenth century laissez-faire concept of ‘freedom of the seas’ was based upon the premise that the Table 3

ocean was infinite, its resources inexhaustible, its degradation impossible. These assumptions have proved to be both unrealistic and detrimental. It is now widely acknowledged that complete freedom of the seas would lead to resource waste and exploitation, economic inefficiency, and increased conflict among users. Enclosure of Ocean Space

From a pragmatic perspective, the management of ocean space involves methods of enclosure. Theoretically, the enclosure of ocean space can be derived from both national and international management regimes. In practice, it has been accomplished through the seaward extension of national jurisdictions by establishing zones of authority and use (e.g., the territorial sea and exclusive economic zone; see Law of the Sea). The primary thrust has been toward the expansion of sovereignty over ocean space previously considered open-access. Although large-scale ocean enclosures have led to reductions in international conflicts over resource use within the proscribed enclosure, such conflicts continue to persist among domestic users and over resources (e.g., straddling fish stocks) that transgress enclosure boundaries. Global Institutions

International cooperation to address marine and coastal concerns has been codified through several formal commitments. In international affairs, this institutionalization usually takes the form of a treaty or customary practice, although certain important intergovernmental organizations also exist. On the global level, both broadly based and issue-specific

Ocean uses and some leading policy issues

Use

Some leading policy issues

Commercial fishing

Overharvesting due to inappropriate management measures; overcapacity due to government subsidization; shifts to fishing lower trophic levels; impacts on habitat, species diversity, ecological functions, protected species; loss of gear; human safety risks Overharvesting due to inappropriate management measures Macronutrient pollution; spread of disease; escaped fish; interactions with protected species; loss of gear Cabotage laws; cartelization; infrastructure investments, including harbor dredging; piracy; oil and hazardous material spills; marine debris; transport of invasive species; interactions with protected species; acoustic pollution; safety of life at sea Disposal of contaminated material; government subsidization Radioactive waste disposal; chemical waste disposal; transport of pollutants from disposal sites Oil spills; benthic disturbances; habitat impacts; acoustic pollution; commercial and recreational fishery impacts Loss of ecosystems and habitat to other uses; impacts of global climate change; impacts of recreation on protected species, coral reefs; recreational boating safety Weapons tests; acoustic tests; runoff of pollutants from military sites; oil and hazardous waste spills; marine debris Erosion; industrial runoff; habitat loss; limits to public access Macronutrient and pesticide runoffs; hypoxia; hypothesized links to harmful algal blooms

Recreational fishing Aquaculture Shipping

Channel dredging Ocean dumping Minerals Recreation Defense Coastal development Agriculture

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE POLICY OVERVIEW

treaties that affect a majority of national interests have been developed. Table 4 describes some prominent examples of these agreements. The proclivity of most States to cooperate in world affairs also extends to regional arrangements. Many coastal and ocean resources transcend political boundaries and thus do not conform to jurisdictional constraints. Therefore, several regional agreements address concerns that extend beyond national jurisdictions to the interests of neighboring states. Table 5 provides several illustrations of regional institutional governance. National Institutions

Virtually all coastal nations have enacted domestic marine policies and laws to legitimize their claims to ocean resources and space. Despite the inefficiencies of fragmented policy administrations and a general lack of public input and future planning, the resulting governance regimes have brought order to the management of various ocean uses. These legislative actions have often been taken as a reaction to real or perceived threats to the health of the ocean or the overexploitation of resources. Often, these laws are designed to work in conjunction with regional and international treaties, but sometimes they do not. Table 4

667

The US Marine Mammal Protection Act is one example of a national institution that aims to conserve specific marine resources but which has come into sharp conflict with international trade law.

Institutional Integration

In general, most laws governing the use of ocean space and resources are sectoral and issue-specific. Examples include legislation pertaining to fisheries management, offshore oil and gas development, and coastal mineral extraction. The primary concern is that as ocean space, particularly coastal waters, becomes the subject of increasingly intense and diversified uses, the activities of one user group will frequently affect the interests of others. A goal of institutional integration is to discover ways in which all uses can be optimized or, at least, coexist without rancor. The hope is that integration can reduce conflicts between uses. The integration of marine policies, sometimes through the implementation of so-called multiple-use management regimes, works to eliminate the inefficiency of single-sector regulatory schemes and is believed to mirror more closely the dynamic complexity of the ocean system. For example, some marine protected areas exhibiting high degrees of marine

Prominent global agreements and organizations for ocean management

Year

Institution

Description

1992

UNCED (United Nations Conference on Environment and Development)

1982

UNCLOS (United Nations Convention on the Law of the Sea)

1973

1972

MARPOL (International Convention for the Prevention of Pollution from Ships) London Convention

1971

Ramsar Convention

1958

IMO (International Maritime Organization)

1946

IWC (International Whaling Commission)

International ‘soft law’ that helped to set the context for several international agreements targeting the interdependence of global environmental protection, sustainable development, and social equity. Most prominent for ocean management was Chapter 17 of Agenda 21 that stresses both the importance of oceans and coasts in the global life support system and the positive opportunity for sustainable development that ocean and coastal areas represent An overarching framework convention that provides both a foundation for global ocean law, and a means for individual States to direct specific coastal and marine activities The first comprehensive global convention that prevents or limits the type and amount of vessel-source pollution including oil, garbage, noxious liquid substances, sewage and plastics. Established the first global standards to govern the dumping of wastes into the ocean, including specific mandates as to what materials may be legally dumped through a permit system. Requires national initiatives by each signatory to conserve wetlands as regulators of water regimes and as habitats of distinctive ecosystems of global importance Facilitates international cooperation on matters of safety and environmental protection in maritime navigation and shipping. Its principal environmental responsibilities are to prevent marine oil pollution, provide remedies when prevention fails, and to assist the development of jurisdictional powers to prescribe and enforce pollution control standards through intergovernmental cooperation Regulates, but does not preclude, the global sustainable taking of whales through a system of quotas designed to prevent their overexploitation and possible extinction. Various management procedures and moratoria (including stout opposition to the moratoria by some commercial whaling nations) have provided an institutional framework but not a cessation of stock depletion

(c) 2011 Elsevier Inc. All Rights Reserved.

668

MARINE POLICY OVERVIEW

Table 5

Important regional institutions for coastal and ocean management

Institution

Description

UNEP (United Nations Environment Program) Oceans and Coastal Areas Program Large Marine Ecosystems (LME)

Designed to address coastal and marine environmental problems (e.g., marine pollution, fisheries conservation and development, species protection) and socioeconomic issues such as tourism common to those nations that share a communal body of water. At present, over 100 hundred States participate in twelve regional oceans and coastal area programs This concept has been proposed but it does not presently constitute a legal institution. An LME is a large region of ocean space, generally over 200 000 km2 (77 000 square miles) and situated typically within exclusive economic zones, that have unique bathymetry, hydrology, and productivity and encompass a regional functional ecological unit. Managing this comprehensive ecosystem for both the protection of biological diversity and sustainable uses requires broad regional cooperation between States The United Nations Educational, Scientific, and Cultural Organization’s (UNESCO) Man and the Biosphere Program is an international program of concerted scientific cooperation among countries directed towards finding practical solutions to environmental problems. A major function is the establishment of protected areas (including several marine and coastal reserves) of ecological significance The Cartegena Convention addresses the myriad of environmental concerns (including marine oil spills) associated with the cultural, economic and political differences exhibited throughout the wider Caribbean. As a supplement, the protocol establishes protected areas to conserve and maintain species and ecosystems, and promotes the sustainable management and use of flora and fauna to prevent their endangerment Primary goal is the conservation of tuna-like fishes and billfishes throughout the Atlantic Ocean and adjacent seas. Member nations must conduct most research, carry out analyses, and enforce ICCAT recommendations for their own nationals

Man and the Biosphere Program

Cartagena Convention

ICCAT (International Convention for the Conservation of Atlantic Tunas) Antarctic Treaty System

The Great Lakes Program

Composed of the 1972 Convention on the Conservation of Antarctic Seals; the 1980 Convention on the Conservation of Antarctic Marine Living Resources (CCAMLR); the 1988 Convention on the Regulation of Antarctic Mineral Resource Activities (CRAMRA); and the 1991 Protocol on Environmental Protection; the Antarctic Treaty System (ATS) seeks to bring institutional order to the activities of those States claiming sovereignty over Antarctic territory, or those interested in resource exploitation A comprehensive management regime for the protection and management of the Great Lakes established through the cooperation of the federal governments of the United States and Canada, eight US states, the Canadian province of Ontario, and many local and regional organizations. Targeted issues include nonpoint source pollution, water levels, navigation, recreational activities, and fishing.

biological diversity are zoned also for human uses such as tourism within a multiple-use management system. The primary purposes of integration are to ensure that links among issues are not neglected in the creation and implementation of public policies and to internalize the external costs that normally accompany the misuse of open-access resources. Integration also emphasizes responsiveness to the legitimate needs of current users while exercising stewardship responsibility on behalf of future generations. Unfortunately, a number of obstacles to the integration of marine policies remain, including incomplete scientific information, boundary disputes, lack of political will, fractionalization of government efforts, and the existence of short-term ocean management programs that may not be optimal for solving persistent problems.

zone and its resources. Integrated coastal zone management (ICZM) is a process that attempts to resolve coastal conflicts, promote the sustainability of resources, and enhance economic benefits to coastal communities. Despite some reservations as to the practicality of the concept, ICZM is designed to overcome the traditional sectoral approach to managing coastal uses by accommodating all sectors within the context of a larger planning scheme. Management tools including zoning, special area planning, land acquisition and mitigation, easements, and coastal permitting are employed to implement an ICZM program. Evolving ICZM efforts are ongoing in such diverse nations as the United Kingdom, Thailand, South Korea, and Tanzania.

Analytical Approaches Integrated Coastal Zone Management

A prime illustration of the movement toward policy integration is found in the management of the coastal

Approaches to the analysis of marine policy issues are diverse, ranging from highly quantitative models to qualitative and descriptive techniques. Whether

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE POLICY OVERVIEW

mathematical or descriptive, these approaches are unified by the presentation of policy options and the comparison, using disciplinary criteria, of alternative courses of action. Economic and political science models tend to be more quantitative, whereas models from other social science disciplines tend to be less mathematically oriented. All social science applications to marine policy problems may test hypotheses, employing rigorous statistical methods for the analysis of empirical data. These methods include the standard regression techniques as well as modern nonparametric, time series, and limited dependent variable techniques. Different analytical approaches in marine policy can be complementary, and they are commonly informed by oceanographic research findings and theory. Economic Analysis

The economic theory focusing on the management of marine resources provides the most common example of the application of a quantitative approach. Neoclassical economics emphasizes the selection of a course of action that optimizes the welfare of society through the supply and consumption of goods and services. In the marine environment, natural resources, such as fish, marine mammals, coral reefs, or entire ecosystems, represent these goods and services, and the dynamics of the ecosystem, including its response to human exploitations, provides a natural constraint to welfare optimization. A basic model, developed in the 1950s, seeks to maximize welfare in the form of producer surplus (profits, broadly defined) in a fishing fleet of identical vessels from the harvest of a single fish stock. Numerous extensions of the basic model include the addition of other ecologically related species, the incorporation of uncertainty, the investigation of nonlinear dynamics, the consideration of a non-uniform distribution of fish stocks, the analysis of consumer surplus, the introduction of competing fleets or nations (game theory), and so forth. The model has become an important tool in the analysis of the economic and biological effects of the implementation of conservation and management measures in a fishery, such as marine reserves or individual transferable quotas. Given significant uncertainty and lags in the response of the marine ecosystem to human perturbations, fisheries economists and scientists now think in terms of managing a fishery adaptively by observing how the system responds to variations in the level of fishing pressure. The economic optimization model has been utilized most commonly in the analysis of fishery management questions. Other, related applications

669

include those in the areas of marine pollution, the environmental risks of offshore oil and gas development, shipping infrastructure, ocean dumping, and marine aquaculture. Economic analysis is also directed at estimating the willingness of members of society to pay for goods and services that are not traded on established markets. Several approaches have been used to value these so-called ‘nonmarket’ commodities, some of which have generated considerable controversy. Nonmarket goods for which demand has been estimated include beach visits, water quality, marine mammals, marine protected areas, and coral reefs, among others. The purpose of estimating nonmarket values is to allow a comparison in common units of the economic values of market and nonmarket commodities when deciding on the net benefits of alternative courses of action. Organizational Studies

Social organization and cultural norms are institutional forms that may shape the feasible set of policy alternatives for any particular marine policy issue. Researchers in disciplines such as geography, sociology, history, and cultural anthropology, among others, focus their research efforts on broad- and fine-scale characterizations and mappings of social organization. Their studies include understanding the development of resource-based communities and enclaves and the ways in which coastal and marine resources are used and conserved. Through induction, empirical studies lead to theories of the natural emergence of organizational principles for the management of marine resources, including collective choice arrangements, enclosures, property right definition and enforcement, and modes of conflict resolution. One such theory that appears in the fisheries context involves the concept of comanagement, through which management responsibilities and functions are shared, according to specified rules, between the owners of the resource, or their agents, and those who are involved in its exploitation. Legal Studies

During the last 30 years, the body of law governing the human uses of the ocean has expanded and diversified at a rapid pace. At both national and international levels, virtually all uses of the sea are now regulated in some fashion. Ocean law is a dynamic institution that responds to changing ecological parameters, economic conditions, and technological and scientific advances. Legal analysts track the changing nature of the law, interpret the

(c) 2011 Elsevier Inc. All Rights Reserved.

670

MARINE POLICY OVERVIEW

way in which legal institutions affect the allocation of marine resources, and characterize the actual and potential impacts of these institutions on human behavior. One could easily argue that the courts, legislative bodies, and executive agencies with responsibility for ocean management rely on legal analysis to a much larger extent than other types of marine policy analysis. Methods of legal analysis can be characterized generally as descriptive and interpretive, relying upon: the practice of nations; the content of treaties, statutes, and rules; the interpretations of courts; and uncodified societal norms. Legal analysis may be further characterized as subjective, in the nature of advocating a particular policy to benefit the interests of one or more agencies or stakeholders. Institutional Effectiveness

In the field of international political relations and in domestic policy reviews, analysts attempt to understand the extent to which an institution is effective at attaining agreed-upon goals. For example, the degree to which an institution, such as an international agreement to control land-based marine pollution, is effective at improving the quality of the marine environment would be based upon observed changes in environmental quality measures over time. In contrast with economic analysis, studies of institutional effectiveness put forth no normative standards, such as the optimization of social welfare. The goals are determined by the participants (stakeholders, national legislatures, legations) who establish the institution. If the goal is attained, then, holding constant other motivations, such as political power, changes in economic conditions, or external influences, it is assumed that the institution has been effective in motivating its participants to take action. Lesson-drawing

Another useful analytical approach is known as ‘lesson-drawing.’ As a form of comparative political analysis, lesson-drawing focuses on the set of circumstances through which marine policies observed to be effective in one jurisdiction are potentially transferable to another. Confronted with a common problem or consistent behaviors, policy makers may be able to learn from how their counterparts elsewhere respond, and conclude that the implementation of policies in other places may be of use in their own circumstances. Lesson-drawing is particularly useful in nations that share some commonalities such as resource availability or cultural norms. The methodology involves an initial search for similar contexts and policies in other jurisdictions, the

development of a conceptual model of the application of the policy, a comparison of practices across jurisdictions, and a prediction or forecast of success after the lesson has been drawn and the policy approach adopted. Notably, the search for and discovery of lessons does not imply that there must be a common application. Realistically, one cannot expect that policies can be successfully transferred without considering the idiosyncratic characteristics of jurisdictions that may allow the policy to be effective in one place but not in another. For example, in the case of preserving marine biodiversity by zoning, there is no generic type of marine protected area that is capable of meeting every situation. The nature of a reserve, its design, and its regulatory framework all depend on the primary objectives it seeks to achieve. These identified objectives will influence the size, shape, and other design constraints of the protected area, and its implementation.

Future Prospects Marine policy will continue to grow in importance as human populations place increasing pressure on coastal space, ocean resources, and marine ecosystems. These pressures, driven by such forces as population growth, human migration to coastal areas, and expanding demand for both living and nonliving resources, will disrupt ecosystems, lead to genetic losses, and exacerbate user conflicts. As many of these problems involve institutional failures, in the future, historical customs and institutions will need to be re-examined. Solutions involving the establishment of new (or clarification of existing) property rights and their enforcement, utilizing technologies that lower the costs of monitoring and enforcing such rights, will undoubtedly come to the fore. Policy choices affecting the allocation of ocean resources lead to questions of effectiveness, or the ability of institutions to meet agreed-upon goals. Despite the steady advance of marine science and technology, policy makers must face choices across options with a high degree of uncertainty. In the face of uncertainty, policy analyses can be neither comprehensive nor fully conclusive, leading policy makers to turn increasingly toward precautionary approaches. Substantial alterations to the current institutional framework supporting coastal and ocean activities are necessitated by the shift to a precautionary approach, including a movement away from sectoral management and toward the greater integration of policies.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE POLICY OVERVIEW

See also Coastal Topography, Human Impact on. Diversity of Marine Species. Fishery Management. Fishery Management, Human Dimension. International Organizations. Large Marine Ecosystems. Law of the Sea. Mariculture, Economic and Social Impacts. Marine Protected Areas. Oil Pollution. Tidal Energy. Wave Energy.

Further Reading Anderson LG (1986) The Economics of Fisheries Management. Baltimore: Johns Hopkins University Press. Armstrong JM and Ryner PC (1981) Ocean Management: A New Perspective. Ann Arbor, MI: Ann Arbor Science. Broadus JM and Vartonov RV (eds.) (1994) The Oceans and Environmental Security. Washington: Island Press. Caldwell LK (1996) International Environmental Policy. Durham, NC: Duke University Press. Calvert P (1993) An Introduction to Comparative Politics. Hertfordshire, UK: Harvester Wheatsheaf. Cicin-Sain B and Knecht RW (1998) Integrated Coastal and Ocean Management. Washington: Island Press. Clark CW (1985) Bioeconomic Modeling and Fisheries Management. New York: Wiley-Interscience. Eckert RD (1979) The Enclosure of Ocean Resources: Economics and the Law of the Sea. Stanford, CA: Hoover Institution Press, Stanford University. Ellen E (1989) Piracy at Sea. Paris: ICC International Maritime Bureau. Farrow RS (1990) Managing the Outer Continental Shelf Lands: Oceans of Controversy. New York: Taylor Francis. Freeman AM (1996) The benefits of water quality improvements for marine recreation: a review of the empirical evidence. Marine Resource Economics 10: 385--406. Friedheim RL (1993) Negotiating the New Ocean Regime. Columbia, SC: University of South Carolina Press. Gould RA (2000) Archaeology and the Social History of Ships. Cambridge: Cambridge University Press.

671

Hsu¨ KJ and Thiede J (eds.) (1992) Use and Misuse of the Seafloor. New York: John Wiley. International Maritime (1991) The London Dumping Convention: The First Decade and Beyond. London: International Maritime Organization. Ketchum BH (1972) The Water’s Edge: Critical Problems of the Coastal Zone. Cambridge, MA: MIT Press. Mahan AT (1890) The Influence of Sea Power Upon History: 1660–1783. Newport, RI: US Naval War College. Mangone GJ (1988) Marine Policy for America, 2nd edn. New York: Taylor Francis Mead WJ, Moseidjord A, Muraoka DD, and Sorenson PE (1985) Offshore Lands: Oil and Gas Leasing and Conservation on the Outer Continental Shelf. San Francisco: Pacific Institute for Public Policy Research. Miles EL (ed.) (1989) Management of World Fisheries: Implications of Extended Coastal State Jurisdiction. Seattle: University of Washington Press. Mitchell RB (1994) International Oil Pollution at Sea: Environmental Policy and Treaty Compliance. Cambridge, MA: MIT Press. Norse EA (ed.) (1993) Global Marine Biological Diversity. Washington: Island Press. Ostrom E (1990) Governing the Commons: The Evolution of Institutions for Collective Action. New York: Cambridge University Press. Richardson JQ (1985) Managing the Ocean: Resources, Research, Law. Mt Airy, MD: Lomond Publications. Rose R (1991) What is Lesson-Drawing? Journal of Public Policy, 3--30. Sherman K, Alexander LM, and Gold BD (eds.) (1990) Large Marine Ecosystems: Patterns, Processes, and Yields. Washington: American Association for the Advancement of Science. Underdal A (1980) Integrated Marine Policy – What? Why? How? Marine Policy 4(3): 159--169. Vallega A (1992) Sea Management: A Theoretical Approach. New York: Elsevier Applied Science. Wenk E Jr (1972) The Politics of the Ocean. Seattle, WA: University of Washington Press.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PROTECTED AREAS P. Hoagland, Woods Hole Oceanographic Institution, Woods Hole, MA, USA U. R. Sumaila, University of British Columbia, Vancouver, BC, Canada S. Farrow, Carnegie Mellon University, Pittsburgh, PA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1654–1659, & 2001, Elsevier Ltd.

Introduction Marine protected areas (MPAs) are a regulatory tool for conserving the natural or cultural resources of the ocean and for managing human uses through zoning. MPAs may also be referred to as marine parks, sanctuaries, reserves, or closures; the latter two terms are used most commonly in the context of fisheries management.

Size

Although there is no discernible size limitation, the issue of geographic scale may be another defining characteristic of MPAs. On the tidelands of US coastal states, for example, the ‘public trust doctrine’ gives preference in the common law to transitory public uses, typically navigation, fishing, and hunting, over permanent private uses, such as constructing a dock. Yet the tidelands, which are quite extensive, are not referred to as an MPA. Some fishery closures can be quite large, and we would classify these as one type of MPA. The Great Barrier Reef Marine Park in Australia is the largest MPA in the world, measuring 344 million km2. Most of the world’s existing MPAs are much smaller, however, and focused on unique ocean features or sites, such as coral reefs or underwater banks. The World Bank estimates the median size of a sample of about one thousand of the world’s MPAs to be 15 840 km2 (Figure 1).

Definition

672

Number

Worldwide, MPAs have become a popular form of ocean management, and their use has expanded exponentially since they were first introduced in the late nineteenth century (Figure 2). The trend in the establishment of MPAs follows on the heels of a more general trend in the regulation of ocean uses, as an MPA represents merely a form of governance distinguishable geographically by type or severity of regulation. Regulation of the ocean has become necessary as human uses of the ocean have increased in scale and variety and as conflicts among mutually exclusive uses and users have arisen. 250 200

No. of MPAs

At a conceptual level, zoning in the ocean involves the spatial segregation of a marine area in which certain uses are regulated or prohibited. This general definition might apply to any marine area in which a set of human uses are given preference over others. For example, by law the US President may set aside hydrocarbon deposits on the US outer Continental Shelf as ‘petroleum reserves.’ However, the typical use of the term ‘protected’ implies that a primary focus of an MPA is on the conservation of either individual species and their habitats or ecological systems and functions through the regulation of ‘extractive’ or potentially polluting commercial uses, such as fishery harvests, waste disposal, and mineral development, among others. MPAs are frequently considered to be a fishery management measure, but they may be used for other purposes as well. For instance, in 1975, the first US national marine sanctuary was created around the wreck of the U.S.S. Monitor, a civil war vessel, located off the coast of North Carolina. The sanctuary was established to prevent commercial ‘treasure’ salvage and looting of the shipwreck, to regulate recreational diving, and to promote archaeological studies. In the discussion below, we focus on the use of MPAs in the field of fishery management because this use represents one of the most relevant and interesting examples.

150 100 50 0

1_ 2

2_ 3

3_ 4

4_ 5 Size

5_ 6

6 _7

>7

Figure 1 Worldwide size distribution of marine protected areas (n ¼ 991). Sizes are grouped by km2 to the powers of ten.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PROTECTED AREAS 1400

older, larger fish) will ‘diffuse’ across the boundary into the fishery. Where the behavior patterns of fish stocks are well understood, the careful placement of an MPA may be effective from a biological standpoint. One excellent example is the establishment of an MPA around a spawning aggregation in tropical fisheries.

1200

No. of MPAs

673

1000 800 600 400 200 0 1890

1910

1930

1950

1970

1990

Year Figure 2 Cumulative worldwide growth in the number of marine protected areas and estimated logarithmic trend 1898–1995.

With the expansion in the establishment of MPAs, marine scientists and policy experts have begun to take a closer look at the likely benefits and opportunity costs associated with zoning the ocean, and several recent studies have emerged. In particular, as ecological models of the marine environment become more realistic, marine policy analysts can begin to make more sophisticated examinations of how to choose among competing human uses, given the constraints presented by the natural system.

Management Objectives The extent to which an MPA may be considered an effective management tool depends on its management objectives. Here, objectives are classified under the following general categories: Biological (ecological); economic; and distributive (equity). Often, the establishment of an MPA involves objectives from more than one of these categories. To make the discussion in the next three sections more focused, we ignore the complexities of the subject, and return to them in a later section. Biology

Consider a single species fishery as a starting point, and assume that a biological objective is to increase stock size or biomass. Restrictions on fishing in certain areas are expected to lead to positive ‘refuge’ and ‘stock’ effects. The refuge effect is a static concept implying that some portion of the target stock cannot be harvested because it remains within the MPA. As a consequence, the entire stock is not exploited to the same degree as it would be in the unregulated case. The stock effect is a dynamic concept implying that fish within the MPA will grow and reproduce and that either their larva will drift out beyond the boundary of the MPA and eventually recruit to the fishery or new recruits (or possibly

Economics

The economic implications of an MPA depend critically upon the nature of the institutional framework for managing the fishery. Suppose that a fishery supplies only a small part of a large market and that, initially, it is unregulated. The first assumption implies that seafood consumers are not much affected by changes in the supply of fish from the fishery of concern. Assuming that an equilibrium is reached where harvests balance stock growth, theory suggests, and empirical investigations confirm, that the economic value of the fishery is near zero. In an unregulated fishery, fish are an unpriced factor in the production of seafood, and this implicit subsidy encourages too much fishing effort and, consequently, excessive exploitation. In the jargon of economics, ‘resource rents’ are dissipated. Depending on the scale of the variable costs of fishing, yields may fall below levels considered to be the maximum sustainable. Now suppose that an MPA is established. The refuge effect implies that the exploitable stock is smaller for any given level of fishing effort. In the absence of any complementary regulation, fishermen will exit the fishery until an open-access equilibrium sets up for the residual exploitable stock. As before, rents are dissipated at this new equilibrium, and no economic value is created through the establishment of the MPA. Over time, the stock effect might lead to an expansion of the exploitable biomass. Again, the existence of economic rents associated with an expanding biomass will attract fishermen until rents dissipate. It is conceivable that the exploitable biomass could expand to a level exceeding that in the fishery prior to the establishment of the MPA. This might happen where increasing returns exist in the production of eggs as female fish grow older and larger. A common example is the red snapper (Lutjanus campechanus), a reef fish native to the Gulf of Mexico. It has been claimed that a 10 kg red snapper produces in a single spawn more than 200 times the eggs of a fish weighing only 1 kg. Only in cases in which the stock effect more than compensates for the refuge effect and surpluses accrue to consumers due to the absence of close seafood substitutes, can a case be made that

(c) 2011 Elsevier Inc. All Rights Reserved.

674

MARINE PROTECTED AREAS

the establishment of an MPA in an otherwise unregulated fishery is valuable in an economic sense. And this result is due solely to the expansion of value to the consumer, not to the fisherman. The establishment of an MPA might be complemented with other forms of regulation. Assuming that the costs of administering fishery regulations are minor, resource rents can be realized through the implementation of management measures in conjunction with an MPA, such as taxes on either fish landings or fishing effort or the introduction of an individual tradeable quota system. However, in theory, the implementation of these alternative management measures by themselves can lead to the capture of resource rents, implying no need for an accompanying MPA. Recent research suggests, however, that, where the stock effect overcomes the refuge effect, the establishment of an MPA can lead to increases in economic value in an otherwise optimally managed fishery. Distribution of Economic Impacts

The third general category of objectives concerns the distribution of economic benefits and costs across human users. In attempting to achieve either biological or economic objectives, the effects of the establishment of an MPA on individual fishermen are not considered explicitly. For example, an economic decision rule would argue for the creation of an MPA as long as economic benefits exceed economic costs, assuming all relevant sources of benefits and costs are accounted for, without regard for the identities of the recipients of the surplus. Moreover, even if the creation of an MPA results in net benefits, the historical pattern of the distribution of gains may be shifted. One example is the creation of an MPA in the vicinity of a fishing port, forcing fishermen from that port to travel longer distances to fish. In some circumstances, such as a small fishery in a developed economy, the distributional effects may be minor, as fishermen are able to switch at low cost to other stocks or to other occupations. On the other hand, the distributive effects of an MPA may be more serious for a community that is heavily reliant on a stock for income or as a source of protein. In such cases, an objective of fairness to users may necessitate foregoing potential biological or economic gains through, say, the relocation or reduction in size of an MPA. The political economy of the management regime may dictate such a result, if users are capable of influencing the adoption of an MPA through voting, negotiation, or other means. In circumstances where some form of regulation must be imposed, it is possible that, on the basis of

equity, MPAs may be the preferred choice of fishermen, relative to alternative measures. The reason for this preference is that the establishment of an MPA does not single fishermen out on the basis of gear type or other distinguishing characteristics. Further, it may be difficult to discern ex ante which specific fishermen eventually will bear the costs or be forced to exit.

Complexities There are a number of important issues that increase the complexity of the simple scenarios described above. A few of these issues are touched on here, and the interested reader should refer to the reading list for further detail. Dynamic Responses

In the discussion above, we have ignored the potential for lags in the response of the system, including the behavior of both fish stocks and fishermen, to the implementation of an MPA. Importantly, it may take more than one fishing season for the stock effect to contribute significantly to recruitment. Further, the refuge effect does not always result in the immediate exit of fishermen from the fishery. When few opportunities exist for redeploying boats and hands elsewhere, fishermen may continue to fish in the short run, as long as they can cover their variable costs (wages, fuel, ice, etc.). In certain circumstances, fishermen might rationally delay exit, expecting the stock effect to lead to a future expansion of the fishery. If fishermen delay exit, the expected stock rebuilding may be prolonged. When environmental conditions and ecological interactions are added in, it is not hard to imagine a scenario in which an MPA appears to have no effect, at least in the short run. The lack of results may lead to political action to remove the MPA. Both fish stocks and fishing effort may be distributed nonuniformly across the fishery. This spatial distribution can be affected through the establishment of an MPA. As a consequence, location becomes an important consideration when planning an MPA. For example, recent models of plaice fisheries show that a properly located MPA can protect undersized fish when fishing effort becomes redistributed around the borders of the MPA. Ecological Relationships

MPAs have also been established to protect aggregations of species or components of ecosystems. Even where the management of a single species is of primary concern, a characterization of the biological

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PROTECTED AREAS

relationships between the species of focus and other species in the ecosystem is crucial to understanding the biological, economic, and distributional impacts of the establishment of an MPA. Where fishing technologies are nonselective, MPAs may prove beneficial in reducing the by-catch of nontarget species and minimizing the impacts of trawl gear on seafloor habitat. Biological Diversity

Recent developments in international and domestic law have emphasized the conservation of biological diversity as a biological objective, and MPAs have been suggested as one means of achieving that objective. Although the conservation of biological diversity is an appealing concept at a superficial level, basic definitional questions persist. For example, does biological diversity refer to species richness (i.e., the number of species) or to some other measure, such as the average genetic distance among a set of species? Assuming that an appropriate measure can be agreed upon, economic research has focused on the problem of maximizing a chosen diversity measure subject to limits on financial resources. When coupled with information on species distributions and ecological relationships, this research may be useful in optimizing locations and scaling the size of MPAs. Insurance and Precaution

The ocean is an uncertain environment. Substantial gaps exist in our understanding of ecological relationships among species, the linkages between environmental conditions and ecosystem states, especially given uncertainties about long-term environmental changes, and the impacts of fishing activity on habitat quality and on ecological relationships. For reasons of tractability, bioeconomic models of fisheries are often based on equilibrium assumptions, when it is not clear that, even if their existence is plausible, steady states can ever materialize. In the context of this uncertainty, MPAs have been touted as a hedge for insuring against stock depletion or collapse. Although it seems reasonable to conclude that MPAs might be useful as a hedge against uncertainty, we should heed the message of economic theory that, in the long run, some MPAs may not remedy the problem of rent dissipation, especially if they are used as the only means by which to manage fisheries. Furthermore, the presence of ineluctable uncertainty raises the question of the extent of the practical contribution that fisheries scientists and marine ecologists can make to specifying the size and

675

location of MPAs. This issue has led some observers to suggest that ‘picking’ MPAs is akin to picking securities in the stock market. They conclude that, in the long term, it may be sensible to randomly select a portfolio of MPAs that cover some agreed-upon percentage of the geography in a particular ocean region. Making estimates of the proportion of ocean area to be included in an MPA may also be problematic, as models suggest a wide range, 20–90% of the relevant area. Irreversibilities

Human uses of the ocean can result in ecological impacts that are costly or impossible to reverse. Examples include the extinction of marine fish or protected species, such as mammals or reptiles, and biomass ‘flips,’ in which the collapse of commercially important stock groupings are replaced by others. Concerns about these irreversibilities reflect the notion that there may be preferred states for marine ecosystems. Changed ecosystems could result in smaller potential economic surpluses and a different set of options for the use of the system in the future. The latter may include ‘nonmarket’ damages when protected species or unique ecosystems, such as coral reefs or underwater banks, are affected adversely. In the presence of uncertainty about human uses or ecosystem states, it may be worthwhile to delay decisions to proceed with human uses, such as fishing, that result in irreversible effects, where the development of new information reduces the uncertainty. The existence of this ‘quasi-option value’ may be a formal justification for taking the socalled ‘precautionary approach’ to fisheries management. The precautionary approach, which has now become embodied in international soft law, argues for the maintenance of commercial fish stocks at relatively high levels because, when accounting for uncertainty, the expected losses due to overexploitation exceed those due to underexploitation. Some analysts have pointed to MPAs as an essential element of a precautionary approach. The value of MPAs in this context may be most apparent when they are employed as a control in a scientific experiment designed to test hypotheses about the impacts of fishing. The partial closure of the US portion of Georges Bank to sea scallop dredging, for example, provided valuable information on the ability of that stock to rebuild in a discrete area. The designation of an MPA can be conceptualized as a kind of ‘administrative’ irreversibility, where it may be difficult to modify the designation through political processes. To many observers and interests, this kind of policy inflexibility may be the whole

(c) 2011 Elsevier Inc. All Rights Reserved.

676

MARINE PROTECTED AREAS

point to designating an MPA. Nevertheless, as environmental conditions change, ecosystems adjust, and it is sensible to have in place a management tool that can also be adjusted. The boundaries of the Canadian ‘Endeavor Hot Vents’ MPA off the coast of Vancouver, which has been proposed at the site of a deep-sea hydrothermal system, is designed to be adjusted as vents turn on and off and their associated microbial and faunal assemblages appear and disappear. Administrative Costs

MPAs have been promoted as a management tool that is less costly than alternative measures. Recent research suggests that management costs may decline with the size of an MPA as fixed costs of monitoring and enforcement are spread over larger areas (Figure 3). The degree to which MPAs are less costly to manage may depend, however, on the form of management. If MPAs are complemented with other management measures, it may be difficult to argue that the entire management regime is less costly. Many MPAs have been criticized as being ‘paper parks’ because monitoring and enforcement are minimal. In such cases, the apparent ‘savings’ in administrative costs relative to other management measures are illusory. Although some users may be dissuaded from breaking the rules inside the boundaries of an MPA, others weigh the product of the probability of apprehension and the penalty, concluding from this calculation that it is rational to ignore the rules. Even in well-monitored and enforced areas, poaching occurs, as enforcement actions in fishery closures in the US Gulf of Maine demonstrate on a regular basis. Limits on government

$ millions (2000)

1.0

budgets may imply that some portions of very large MPAs are paper parks.

Summary MPAs clearly hold promise as a rational way of managing ocean resources, but this promise should not be overstated. In particular, MPAs should not be seen as a panacea to all the problems of fisheries management. Indeed, the best way to see MPAs is probably as part of a collection of management tools and measures. As the marine counterpart to systems of national and international parks, they are conceptually easy to understand and naturally appealing to the public. Yet MPAs differ in important ways from land parks because of their relative inaccessibility, the fugitive nature of fish stocks and the physical transport of pollutants and plankton, the legal characteristics of property rights in the ocean, and the costs of monitoring human activities. As we learn more about the ocean and the workings of its environmental and ecological systems, and as demand for the special characteristics of these systems expands with growing coastal populations, we can expect the use of MPAs to grow as well.

See also Coral Reef and Other Tropical Fisheries. Deep-Sea Fishes. Dynamics of Exploited Marine Fish Populations. Ecosystem Effects of Fishing. Fisheries: Multispecies Dynamics. Fishery Management. Fishery Management, Human Dimension. Fishery Manipulation through Stock Enhancement or Restoration.

Further Reading

0.8 0.6 0.4 0.2 0.0 0.0

0.5

1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

5.0

2

Size (km ) Figure 3 An estimate for small marine protected areas (MPA) of the relationship between size and the average costs of establishing and managing an MPA. The relationship demonstrates economies of scale. Costs include the acquisition of coastal land, demolition of existing structures, development, and operating costs (capitalized at 5%). Average costs are estimated from data pertaining to size alternatives for the proposed Salt River Bay MPA in St Croix, US Virgin Islands.

Arnason R (2000) Marine protected areas: Is there an economic justification. Proceedings, Economics of Marine Protected Areas. Vancouver: Fisheries Centre, University of British Colombia. Bohnsack JA (1996) Maintenance and recovery of reef fishery productivity. In: Polunin VC and Roberts CM (eds.) Reef Fisheries, pp. 283--313. London: Chapman and Hall. Crosby M, Geenen KS, and Bohne R (2000) Alternative Access Managment Strategies for Marine and Coastal Protected Areas: a Reference Manual for Their Development and Assessment. Washington: Man and the Biosphere Program, US Department of State. Farrow S (1996) Marine protected areas: Emerging economics. Marine Policy 20(6): 439--446. Great Barrier Reef Marine Park Authority, World Bank, and World Conservation Union (1995) A Global

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE PROTECTED AREAS

Representative System of Marine Protected Areas, Vols I–IV. Washington: Environment Department, The World Bank. Gubbay S (1996) Marine Protected Areas: Principles and Techniques for Management. New York: Chapman Hall. Gue´nette S, Lauck T, and Clark C (1998) Marine reserves: from Beverton and Holt to the present. Review of Fish Biology and Fisheries 8: 251--272. Hoagland P, Broadus JM, and Kaoru Y (1995) A Methodological Review of Net Benefit Evaluation for Marine Reserves. Environment Dept. Paper No. 027. Washington: Environment Department, The World Bank. Holland DS and Brazee RJ (1996) Marine reserves for fisheries management. Marine Resource Economics 11: 157--171. Ocean Studies Board (2000) Marine Protected Areas: Tools for Sustaining Ocean Ecosystems. Washington: National Academy Press.

677

Roberts CM and Polunin NVC (1993) Marine reserves: simple solutions to managing complex fisheries? Ambio 22: 363--368. Rowley RJ (1994) Marine reserves in fisheries management. Aquatic Conservation: Marine and Freshwater Ecosystems 4: 233--254. Sanchirico JN (2000) Marine Protected Areas as Fishery Policy: an Analysis of No-Take Zones. Discussion Paper 00-23. Washington: Resources for the Future. Sumaila UR (1998) Protected marine reserves as fisheries management tools: a bioeconomic analysis. Fisheries Research 37: 287--296. Sumaila UR, Gue´nette S, Alder J, and Chuenpagdee R (2000) Addressing the ecosystem effects of fishing using marine protected areas. ICES Journal of Marine Science 57(3): 752--760.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SILICA CYCLE D. J. DeMaster, North Carolina State University, Raleigh, NC, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1659–1667, & 2001, Elsevier Ltd.

Basic Concepts In understanding the cycling of silicate in the oceans, the concept of mean oceanic residence time is commonly used. Mean oceanic residence time is defined as in eqn. [1]. ðamount of dissolved material in a reservoirÞ ðsteady-state flux into or out of the reservoirÞ

Introduction Silicate, or silicic acid (H4SiO4), is a very important nutrient in the ocean. Unlike the other major nutrients such as phosphate, nitrate, or ammonium, which are needed by almost all marine plankton, silicate is an essential chemical requirement only for certain biota such as diatoms, radiolaria, silicoflagellates, and siliceous sponges. The dissolved silicate in the ocean is converted by these various plants and animals into particulate silica (SiO2), which serves primarily as structural material (i.e., the biota’s hard parts). The reason silicate cycling has received significant scientific attention is that some researchers believe that diatoms (one of the silicasecreting biota) are one of the dominant phytoplankton responsible for export production from the surface ocean (Dugdale et al., 1995). Export production (sometimes called new production) is the transport of particulate material from the euphotic zone (where photosynthesis occurs) down into the deep ocean. The relevance of this process can be appreciated because it takes dissolved inorganic carbon from surface ocean waters, where it is exchanging with carbon dioxide in the atmosphere, turns it into particulate organic matter, and then transports it to depth, where most of it is regenerated back into the dissolved form. This process, known as the ‘biological pump’, along with deep-ocean circulation is responsible for the transfer of inorganic carbon into the deep ocean, where it is unable to exchange with the atmosphere for hundreds or even thousands of years. Consequently, silicate and silica play an important role in the global carbon cycle, which affects the world’s climate through greenhouse feedback mechanisms. In addition, the accumulation of biogenic silica on the ocean floor can tell us where in the ocean export production has occurred on timescales ranging from hundreds to millions of years, which in turn reveals important information concerning ocean circulation and nutrient distributions.

678

½1

For silicate there are approximately 7  1016 moles of dissolved silicon in ocean water. (One mole is equal to 6  1023 molecules of a substance, which in the case of silicic acid has a mass of approximately 96 g.) As described later, the various sources of silicate to the ocean supply approximately 7  1012 mol y1, which is approximately equal to our best estimates of the removal rate. Most scientists believe that there has been a reasonably good balance between supply and removal of silicate from the oceans on thousand-year timescales because there is little evidence in the oceanic sedimentary record of massive abiological precipitation of silica (indicating enhanced silicate concentrations relative to today), nor is there any evidence in the fossil record over the past several hundred million years that siliceous biota have been absent for any extended period (indicating extremely low silicate levels). Dividing the amount of dissolved silicate in the ocean by the supply/removal rate yields a mean oceanic residence time of approximately 10 000 years. Basically, what this means is that an atom of dissolved silicon supplied to the ocean will remain on average in the water column or surface seabed (being transformed between dissolved and particulate material as part of the silicate cycle) for approximately 10 000 years before it is permanently removed from the oceanic system via long-term burial in the seabed.

Distribution of Silicate in the Marine Environment Because of biological activity, surface waters throughout most of the marine realm are depleted in dissolved silicate, reaching values as low as a few micromoles per liter (mmol l1). When the siliceous biota die, their skeletons settle through the water column, where more than 90% of the silica is regenerated via inorganic dissolution. This process enriches the deep water in silicate, causing oceanic bottom waters to have as much as 10–100 times

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SILICA CYCLE

more silicate than surface waters in tropical and temperate regions. The magnitude of the deep-ocean silicate concentration depends on the location within the deep thermohaline circulation system. In general, deep water originates in North Atlantic and Antarctic surface waters. The deep water forming in the North Atlantic moves southward, where it joins with Antarctic water, on its way to feeding the deep Indian Ocean basin and then flowing from south to north in the Pacific basin. All along this ‘conveyor belt’ of deep-ocean water, siliceous biota are continually settling out from surface waters and dissolving at depth, which further increases the silicate concentration of the deep water downstream. Consequently, deep-ocean water in the Atlantic (fairly near the surface ocean source) is not very enriched in silicate (only 60 mmol l1), whereas the Indian Ocean deep water exhibits moderate enrichment (B100 mmol l1), and the north Pacific deep water is the most enriched (B180 mmol l1). This trend of increasing concentration is observed as well in the other nutrients such as nitrate and phosphate. Generalized vertical profiles of silicate are shown for the Atlantic and Pacific basins in Figure 1. The depth of the silicate maximum in these basins (typically 2000–3000 m depth) is deeper than the nutrient maxima for phosphate or nitrate, primarily because organic matter (the source of the phosphate and nitrate) is generally regenerated at shallower depths in the ocean than is silica. The nutrient concentrations in

_

0

Silicate concentration (µmol l 1) 100 150 50

200

0

oceanic deep waters can affect the chemical composition of particles settling through the water column because the vertical transport of nutrients from depth via upwelling and turbulence drives the biological production in surface waters. For example, the ratio of biogenic silica to organic carbon in particles settling between 1000 and 4000 m depth in the North Pacific Ocean (typically about 2–3) is substantially higher than that observed in the Arabian Sea (B0.7) and much higher than typical values in the Atlantic Ocean (o0.3). This chemical trend in particle flux, which is caused in part by changes in planktonic species assemblage, is consistent with the systematic increase in silicate and other nutrients along the thermohaline-driven conveyor belt of deepocean circulation. The change in the biogenic silica to organic carbon ratio throughout the ocean basins of the world turns out to be one of the most important parameters controlling the nature of biogenic sedimentation in the world (see the global ocean sediment model of Heinze and colleagues, listed in Further Reading). Silicate concentrations also can be used to distinguish different water masses. The most obvious example is at the Southern Ocean Polar Front (see Figure 2), which separates Antarctic Surface Water from the Subantarctic system. The silicate and nitrate concentration gradients across these Southern Ocean waters occur in different locations (in a manner similar to the distinct maxima in their vertical profiles). The high concentrations of silicate (50–100 mmol l1) south of the Polar Front result from windinduced upwelling bringing silicate to the surface faster than the local biota can turn it into particulate silica. Turnover times between surface waters and 100

1000 NWPacific Ocean (47°N, 162°E)

_

Depth (m)

Concentration (μmol l 1) ( , ) Temperature × 10 (˚C) ( )

80

2000

3000

4000

679

NWAtlantic Ocean (36°N, 68°W )

5000

60

40

20

0 _ 20

6000

52

Figure 1 Vertical distribution of dissolved silicate in the Atlantic Ocean and the Pacific Ocean. The Atlantic data come from Spencer (1972), whereas the Pacific data are from Nozaki et al. (1997).

55

58 Latitude (˚S)

61

64

Figure 2 Distribution of silicate (’), nitrate (~), and temperature (m) across the Drake Passage illustrating different water masses and frontal features during November, 1999.

(c) 2011 Elsevier Inc. All Rights Reserved.

680

MARINE SILICA CYCLE

deep waters in the Southern Ocean are on the order of 100 years as compared to values of 500–1000 years in the Atlantic, Indian, and Pacific Oceans. In terms of chemical equilibrium, biogenic silica is undersaturated by a factor of 10–1000-fold in surface waters and by at least a factor of 5 in deep waters. Therefore, siliceous biota must expend a good deal of energy concentrating silicate in their cells and bodies before precipitation can take place. This is quite different from the case of calcium carbonate (a material used by another type of plankton to form hard parts), which is supersaturated severalfold in most tropical and temperate surface waters. Deep waters everywhere are undersaturated with respect to biogenic silica (although to different extents). Therefore, inorganic dissolution of silica takes place in the water column as soon as the plankton’s protective organic matter is removed from the biota (typically by microbial or grazing activities). It is not until the siliceous skeletal material is buried in the seabed that the water surrounding the silica even approaches saturation levels (see later discussion on sedimentary recycling), which diminishes the rate of dissolution and enhances preservation and burial.

The Marine Silica Cycle Sources of Dissolved Silicate to the Ocean

Figure 3 shows the main features of the marine silica cycle as portrayed in a STELLATM model. The main source of silicate to the oceans as a whole is rivers, which commonly contain B150 mmol l1 silicate but depending on location, climate, and local rock type, can range from 30 to 250 mmol l1. The silicate in rivers results directly from chemical weathering of rocks on land, which is most intense in areas that are warm and wet and exhibit major changes in relief (i.e., elevation). The best estimate of the riverine flux of dissolved silicate to the oceans is B6  1012 mol Si y1. Other sources include hydrothermal fluxes (B0.3  1012 mol Si y1), dissolution of eolian particles (0.5  1012 mol Si y1) submarine volcanic activity (negligible), and submarine weathering of volcanic rocks (B0.4  1012 mol Si y1). Ground waters may contribute additional silicate to the marine realm, but the magnitude of this flux is difficult to quantify and it is believed to be small relative to the riverine flux. A more detailed discussion of the various sources of silicate to the marine environment can be

0.5 Atmospheric supply

Si concentration in surface waters

Estuaries\ shelves\ margins Surface Ocean Upwelling ~3 Riverine supply

197

Open-ocean supply Particle flux

5.7

Si concentration in deep waters

~200 Burial Estuary \ shelf \ margin 2.4 _ 4.1 Downwelling

Hydrothermal flux/submarine weathering Deep Ocean 0.7

Burial deep ocean 2.4 _ 3.2

Figure 3 STELLATM model of the global marine silica cycle showing internal and external sources of silicate to the system, internal recycling, and burial of biogenic silica in the seabed. The various reservoirs are shown as rectangles, whereas the fluxes in and out of the reservoirs are shown as arrows with regulating valves (indicating relationships and functional equations). The flux values (indicated by numbers inside the boxes) have units of 1012 mol y1.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SILICA CYCLE

found in the works by Treguer et al. and DeMaster (listed in Further Reading). All of these sources of silicate to the oceanic water column are considered to be external. As shown in Figure 3, there are internal sources supplying silicate to oceanic surface waters and they are oceanic upwelling and turbulence. Because of the strong gradient in silicate with depth, upwelling of subsurface water (100–200 m depth) by wind-induced processes and turbulence can bring substantial amounts of this nutrient (and others) to surface waters that typically would be depleted in these valuable chemical resources as a result of biological activity. This upwelling flux (B100  300  1012 mol Si y1), in fact, is 20–50 times greater than the riverine flux. The extraction of silicate from surface waters by siliceous biota is so efficient that nearly 100% of the nutrient reaching the surface is converted into biogenic silica. Therefore, the production of biogenic silica in oceanic surface waters is comparable to the flux from upwelling and turbulent transport. Riverine sources may be the dominant external source of silicate to the oceans, but they sustain only a negligible amount (only a few percent) of the overall marine silica production. Internal recycling, upwelling, and turbulence provide nearly all of the silicate necessary to sustain the gross silica production in marine surface waters. Therefore, changes in oceanic stratification and wind intensity may significantly affect the flux of nutrients to the surface and the overall efficiency of the biological pump. Silicate dynamics in the water column have been simulated using a general circulation model. The results of this study by Gnanadesikan suggest that the model distributions of silicate in the ocean are very sensitive to the parametrization of the turbulent flux. In addition, according to the model, the Southern Ocean and the North Pacific were the two major open-ocean sites where net silica production occurred, accounting for nearly 80% of the biogenic silica leaving the photic zone. If the entire ocean (surface waters, deep waters, and near-surface sediments) is considered as a single box, the external fluxes of silicate to the ocean (mentioned above) must be balanced by removal terms in order to maintain the silicate levels in the ocean at more or less a constant value over geological time. From this point of view the flux of silicate from oceanic upwelling and turbulence can be treated as part of an internal cycle. The dominant mechanism removing silica from this system is burial of biogenic silica. There is some controversy about where some of this burial takes place, but most scientists believe that burial of biogenic silica (or some chemically altered by-product thereof) is the primary way that silicate is removed from the ocean.

681

Removal of Silica from the Ocean

The sediments with the highest rate of silica accumulation (on an areal basis) occur beneath the coastal upwelling zones, where strong winds bring extensive amounts of nutrients to the surface. Examples of these upwelling areas include the west coast of Peru, the Gulf of California, and Walvis Bay (off the west coast of South Africa). Diatom skeletons are the dominant form of biogenic silica in these deposits. The sediments in these upwelling areas accumulate at rates of 0.1–1.0 cm y1 and they contain as much as 40% biogenic silica by weight. The burial rate for silica in these areas can be as high as 1.7 mol cm2 y1. In calculating the total contribution of these areas to the overall marine silica budget, the accumulation rates (calculated on an areal basis) must be multiplied by the area covered by the particular sedimentary regime. Because these upwelling regimes are confined to such small areas, the overall contribution of coastal upwelling sites to the marine silica budget is quite small (o10%, see Table 1), despite the fact that (for a given area) they bury biogenic silica more rapidly than anywhere else in the marine realm. The sediments containing the highest fraction of biogenic silica in the world occur in a 1000 km-wide Table 1

The marine silica budget

Source/Sink

Flux (1012mol Si y  1)

Sources of silicate to the ocean Rivers Hydrothermal emanations Eolian flux (soluble fraction) Submarine weathering

5.7 0.3 0.5 0.4

Total supply of silicate

B7  1012

Sites of biogenic silica burial Deep-Sea Sediments Antarctic Polar Front Non Polar Front Bering Sea North Pacific Sea of Okhotsk Equatorial Pacific Poorly siliceous sediments Continental margin sediments Estuaries Coastal upwelling areas (e.g., Peru, Walvis Bay, Gulf of California) Antarctic margin Other continental margins Total burial of biogenic silica

(c) 2011 Elsevier Inc. All Rights Reserved.

2.4–3.2 0.7–0.9 0.7–1.1 0.5 0.3 0.2 0.02 o0.2 2.4–4.1 o0.6 0.4–0.5 0.2 1.8–2.8 5–7  1012

682

MARINE SILICA CYCLE

belt surrounding Antarctica. These sediments typically contain B60% biogenic silica by weight (the majority of which are the skeletons of diatoms). Most of these sediments, however, accumulate at rates of only a few centimeters per thousand years, so their silica burial rate is quite small (o0.008 mol Si cm2 y1), accounting for 1  1012 mol y1 of silica burial or less than 20% of the total burial in the marine environment. Beneath the Polar Front (corresponding to the northern 200–300 km of the Southern Ocean siliceous belt), however, the sediment accumulation rates increase dramatically to values as high as 50  103 cm y1. Regional averages can be as high as 19  103 cm y1, yielding a silica burial rate of 0.08 mol Si cm2 y1. Unfortunately, many of the siliceous deposits beneath the Polar Front occur in areas of very rugged bottom topography (because of the submarine Antarctic Ridge), where sediments are focused into the deeper basins from the flanks of the oceanic ridge crests. This distribution of accumulation rates would not create a bias if all of the sedimentary environments are sampled equally. However, it is more likely to collect sediment cores in the deep basins, where the deposits are thicker and accumulating more rapidly, than it is on the flanks where the sediment coverage is thinner. The effects of this sediment focusing can be assessed by measuring the amount in the seabed of a naturally occurring, particle-reactive radioisotope, thorium-230 (230Th). If there were no sediment focusing, the amount of excess 230Th in the sediments would equal the production from its parent, uranium-234, in the overlying water column. In some Polar Front Antarctic cores there is 12 times more excess 230Th in the sediment column than produced in the waters above, indicating that sediment focusing is active. Initial estimates of the biogenic silica accumulation beneath the Polar Front were as high as 3  1012 mol Si y1, but tracer-corrected values are on the order of 1  1012 mol Si y1. There are other high-latitude areas accumulating substantial amounts of biogenic silica, including the Bering Sea, the Sea of Okhotsk, and much of the North Pacific Ocean; however, the accumulation rates are not as high as in the Southern Ocean (see Table 1). The high rate of silica burial in the highlatitude sediments may be attributed in part to the facts that cold waters occurring at the surface and at depth retard the rate of silica dissolution and that many of the diatom species in high latitudes have more robust skeletons than do their counterparts in lower latitudes. Moderately high silica production rates and elevated silica preservation efficiencies (approximately double the world average) combine to yield high-latitude siliceous deposits accounting

for approximately one-third of the world’s biogenic silica burial. If the focusing-corrected biogenic silica accumulation rates are correct for the Polar Front, then a large sink for biogenic silica (B1–2  1012 mol Si y1) needs to be identified in order to maintain agreement between the sources and sinks in the marine silica budget. Continental margin sediments are a likely regime because these environments have fairly high surface productivity (much of which is diatomaceous), a relatively shallow water column (resulting in reduced water column regeneration as compared to the deep sea), rapid sediment accumulation rates (10– 100  103 cm y1) and abundant aluminosilicate minerals (see Biogenic Silica Preservation below). The amount of marine organic matter buried in shelf and upper slope deposits is on the order of 3  1012 mol C y1. When this flux is multiplied by the silica/organic carbon mole ratio (Si/Corg) of sediments in productive continental margin settings (Si/Corg ¼ 0.6), the result suggests that these nearshore depositional environments can account for sufficient biogenic silica burial (1.8–2.8  1012 mol Si y1) to bring the silica budget into near balance (i.e., within the errors of calculation).

Biogenic Silica Preservation As mentioned earlier, all ocean waters are undersaturated with respect to biogenic silica. Surface waters may be more than two orders of magnitude undersaturated, whereas bottom waters are 5–15fold undersaturated. The solubility of biogenic silica is greater in warm surface waters than in colder deep waters, which, coupled with the increasing silicate concentration with depth in most ocean basins, diminishes the silicate/silica disequilibrium (or corrosiveness of the water) as particles sink into the deep sea. This disequilibrium drives silica regeneration in oceanic waters along with other factors and processes such as particle residence time in the water column, organic and inorganic surface coatings, particle chemistry, particle aggregation, fecal pellet formation, as well as particle surface area. Recycling of biogenic silica occurs via inorganic dissolution; however, the organic coating that siliceous biota use to cover their skeletons (inhibiting dissolution) must be removed by microbial or zooplankton grazing prior to dissolution. This association is highlighted by the fact that bacterial assemblages can accelerate the dissolution of biogenic silica in the water column. An important aspect of biogenic silica dissolution pertains to surface chemistry and clay-mineral

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SILICA CYCLE

formation on the surface of siliceous tests. Incorporation of aluminum in the initial skeleton as well as aluminosilicate formation on skeletal surfaces during settling and burial greatly decrease the solubility of biogenic silica, in some cases by as much as a factor of 5–10. It appears that some of this ‘armoring’ of siliceous skeletons occurs up in the water column (possibly in aggregates or fecal pellets), although some aluminosilicate formation may occur in flocs just above the seabed as well as deeper in the sediment column. The occurrence of clay minerals on skeletal surfaces has been documented using a variety of instruments (e.g., the scanning electron microscope). The nature of the settling particles also affects dissolution rates in the water column. If siliceous skeletons settle individually, they settle so slowly (a timescale of years to decades) that most particles dissolve before reaching the seabed. However, if the siliceous skeletons aggregate or are packaged into a fecal pellet by zooplankton, sinking velocities can be enhanced by several orders of magnitude, favoring preservation during passage through the water column. Siliceous tests that have high surface areas (lots of protruding spines and ornate surface structures) also are prone to high dissolution rates and low preservation in the water column relative to species that have more robust skeletons and more compact structures. Very few studies have documented silica production rates in surface waters, established the vertical fluxes of silica in the water column, and then also examined regeneration and burial rates in the seabed. One place that all of these measurements have been made is in the Ross Sea, Antarctica. In this high-latitude environment, approximately one-third of the biogenic silica produced in surface waters is exported from the euphotic zone, with most of this material (27% of production) making it to the seabed some 500–900 m below. Seabed preservation efficiencies (silica burial rate divided by silica rain rate to the seafloor) vary from 1% to 86%, depending primarily on sediment accumulation rate, but average 22% for the shelf as a whole. Consequently, the overall preservation rate (water column and seabed) is estimated to be B6% in the Ross Sea. On a global basis, approximately 3% of the biogenic silica produced in surface waters is buried in the seabed. The total preservation efficiencies for different ocean basins vary, with the Atlantic and Indian Oceans having values on the order of 0.4– 0.8% and the Pacific and Southern Oceans having values of approximately 5–10%. Sediment accumulation rate can make a large difference in seabed preservation efficiency. In the Ross Sea, for example, increasing the sediment

683

accumulation rate from 1–2 to 16  103 cm y1, increases the seabed preservation efficiency from 1– 5% up to 50–60%. In most slowly accumulating deep-sea sediments (rates of 2  103 cm y1 or less), nearly all of the biogenic silica deposited on the seafloor dissolves prior to long-term burial. Increasing the sedimentation rate decreases the time that siliceous particles are exposed to the corrosive oceanic bottom waters, by burying them in the seabed where silicate, aluminum, and cation concentrations are high, favoring aluminosilicate formation and preservation. Consequently, continental margin sediments with accumulation rates of 10–100  103 cm y1 are deposits expected to have high preservation efficiencies for biogenic silica and are believed to be an important burial site for this biogenic phase. Estuaries extend across the river–ocean boundary and are generally regions of high nutrient flux and rapid sediment burial (0.1–10 cm y1). They commonly exhibit extensive diatom production in surface waters, but may not account for substantial biogenic silica burial because of extensive dissolution in the water column. For example, on the Amazon shelf approximately 20% of the world’s river water mixes with ocean water and silicate dynamics have been studied in detail. Although nutrient concentrations are highest in the low-salinity regions of the Amazon mixing zone, biological nutrient uptake is limited because light cannot penetrate more than a few centimeters into the water column as a result of the high turbidity in the river (primarily from natural weathering of the Andes Mountains). After the terrigenous particles have flocculated in the river–ocean mixing zone, light is able to penetrate the warm surface waters, leading to some of the highest biogenic silica production rates in the world. However, resuspension on the shelf, zooplankton grazing, and high water temperatures lead to fairly efficient recycling in the water column and nearly all of the dissolved silicate coming down the river makes it out to the open ocean. The Amazon shelf seabed does appear to exhibit clay-mineral formation (primarily through replacement of dissolving diatoms), but the burial fluxes are expected to be small relative to the offshore transport of silicate and biogenic silica.

Biogenic Silica in Marine Sediments As mentioned above, the primary biota that construct siliceous skeletons are diatoms, radiolaria, silicoflagellates, and siliceous sponges. Diatoms are marine algae. These phytoplankton account for 20– 40% of the primary production in the ocean and an even greater percentage of the export production

(c) 2011 Elsevier Inc. All Rights Reserved.

684

MARINE SILICA CYCLE

from the photic zone. Diatom skeletons are the primary form of biogenic silica in deposits associated with coastal upwelling areas, high-latitude oceans (predominantly in the Pacific and the Southern Oceans), and the continental margins (Figure 4). In equatorial upwelling areas radiolarian skeletons commonly occur in marine sediments along with the diatom frustules. Radiolaria are zooplankton that live in the upper few hundred meters of the water column. Their skeletons are larger and more robust than many diatoms; consequently their preservation in marine sediments is greater than that of most diatoms. Silicoflagellates account for a very small fraction of the biogenic silica in marine sediments because most of them dissolve up in the water column or in surface sediments. They have been used in some continental margin sediments as a paleo-indicator of upwelling intensity. Siliceous sponge spicules can make up a significant fraction of the near-interface sediments in areas in which the sediment accumulation rate is low (o5  103 cm y1). For example, on the Ross Sea continental shelf, fine sediments accumulate in the basins, whereas the topographic highs (o400 m water depth) have minimal fine-grained material (because of strong currents and turbulence). As a result, mats of siliceous sponge spicules occur in high abundance on some of these banks. To measure the biogenic silica content of marine sediments, hot (851C) alkaline solutions are used to dissolve biogenic silica over a period of 5–6 hours. The silicate concentration in the leaching solution is measured colorimetrically on a spectrophotometer

and related to the dry weight of the original sedimentary material. In many sediments, coexisting clay minerals also may yield silicate during this leaching process; however, this contribution to the leaching solution can be assessed by measuring the silicate concentration in the leaching solution hourly over the course of the dissolution. Most biogenic silica dissolves within 2 hours, whereas clay minerals release silicate at a fairly constant rate over the entire leaching period. Consequently, the contributions of biogenic silica and clay-mineral silica can be resolved using a graphical approach (see Figure 5).

Measuring Rates of Processes in the Marine Silica Cycle There are several useful chemical tracers for assessing rates of silicate uptake, silica dissolution in the water column, and particle transport in the seabed. Most of these techniques are based on various isotopes of silicon, some of which are stable and some of which are radioactive. Most of the stable silicon occurring naturally in the ocean and crust is 28Si (92.2%) with minor amounts of 29Si (4.7%) and 30Si (3.1%). By adding known quantities of dissolved 29Si or 30Si to surface ocean waters, the natural abundance ratios of Si can be altered, allowing resolution of existing biogenic silica from silica produced after spiking in incubation studies. Similarly, if the silicate content of ocean water is spiked with either dissolved 29 Si or 30Si, then, as biogenic silica dissolves, the ratio of the silicon isotopes will change in proportion

Weight % silica from clay minerals

Weight % silica extracted

14 12 10 8

Weight % biogenic silica

6 4 70 _ 80 cm Carmen Basin

2

0

1

2

3

4

5

Duration of extraction (h)

Figure 4 Micrograph of diatoms (genus Corethron) collected from an Antarctic plankton tow near Palmer Station.

Figure 5 Graphical approach to resolving silicate originating via biogenic silica dissolution from that generated via clay-mineral dissolution during the alkaline leach technique used to quantify biogenic silica abundance. This sample was from the Gulf of California, Carmen Basin.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SILICA CYCLE

to the amount of silica dissolved (enabling characterization of dissolution rates). In addition, the measurement of natural silicon isotopes in sea water and in siliceous sediments has been suggested as a means of assessing the extent of silicate utilization in surface waters on timescales ranging from years to millennia. Addition of radioactive 32Si (half-life 160 y) to incubation solutions recently has been used to simplify the measurement of silica production rates in surface ocean waters. In the past, 32Si has been difficult to obtain, but recent advances in production and isolation protocols have made it possible to produce this radioisotope for oceanographic studies. Distributions of naturally occurring 32Si in the water column and seabed can be used to determine deepocean upwelling rates as well as the intensity of eddy diffusion (or turbulence) in the deep ocean. This same radioactive isotope can be used to evaluate rates of bioturbation (biological particle mixing) in the seabed on timescales of hundreds of years.

See also Carbon Cycle. Current Systems in the Atlantic Ocean. Current Systems in the Southern Ocean. Pelagic Fishes.

Further Reading Craig H, Somayajulu BLK, and Turekian KK (2000) Paradox lost, silicon-32 and the global ocean silica cycle. Earth and Planetary Science Letters 175: 297--308. DeMaster DJ (1981) The supply and removal of silica from the marine environment. Geochimica et Cosmochimica Acta 45: 1715--1732.

685

Dugdale RC, Wilkerson FP, and Minas HJ (1995) The role of a silicate pump in driving new production. Deep-Sea Research 42: 697--719. Gnanadesikan A (1999) A global model of silicon cycling: sensitivity to eddy parameterization and dissolution. Global Biogeochemical Cycles 13: 199--220. Heinze C, Maier-Reimer E, Winguth AME, and Archer D (1999) A global oceanic sediment model for long-term climate studies. Global Biogeochemical Cycles 13: 221--250. Nelson DM, DeMaster DJ, Dunbar RB, and Smith WO Jr (1996) Cycling of organic carbon and biogenic silica in the southern Ocean: estimates of water-column and sedimentary fluxes on the Ross Sea continental shelf. Journal of Geophysical Research 101: 18519--18532. Nelson DM, Treguer P, Brzezinski MA, Leynaert A, and Queguiner B (1995) Production and dissolution of biogenic silica in the ocean: revised global estimates, comparison with regional data and relationship to biogenic sedimentation. Global Biogeochemical Cycles 9: 359--372. Nozaki Y, Zhang J, and Takeda A (1997) 210Pb and 210Po in the equatorial Pacific and the Bering Sea: the effects of biological productivity and boundary scavenging. Deep-Sea Research II 44: 2203--2220. Ragueneau O, Treguer P, Leynaert A, et al. (2000) A review of the Si cycle in the modern ocean: recent progress and missing gaps in the application of biogenic opal as a paleoproductivity proxy. Global and Planetary Change 26: 317--365. Spencer D (1972) GEOSECS II. The 1970 North Atlantic Station: Hydrographic features, oxygen, and nutrients. Earth and Planetary Science Letters 16: 91--102. Treguer P, Nelson DM, Van Bennekom AJ, et al. (1995) The silica balance in the world ocean: A re-estimate. Science 268: 375--379.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SNOW R. S. Lampitt, University of Southampton, Southampton, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1667–1675, & 2001, Elsevier Ltd.

Historical Developments The debate about how material is transported to the deep seafloor has been going on for over a century but at the beginning of the last century some surprisingly modern calculations by Hans Lohmann

Introduction Marine snow is loosely defined as inanimate particles with a diameter greater than 0.5 mm. These particles sink at high rates and are thought to be the principal vehicles by which material sinks in the oceans. In addition to this high sinking rate they have characteristic properties in terms of the microenvironments within them, their chemical composition, the rates of bacterial activity and the fauna associated with them. These properties make such particles important elements in influencing the structure of marine food webs and biogeochemical cycles throughout the world’s oceans. Such an apparently simple definition, however, belies the varied and complex processes which exist in order to produce and destroy these particles (Figure 1). Similarly it gives no clue as to the wide variety of particulate material which, when aggregated together, falls into this category and to the significance of the material in the biogeochemistry of the oceans. Marine snow particles are found throughout the world’s oceans from the surface to the great depths although with a wide range in concentration reaching the highest levels in the sunlit euphotic zone where production is fastest. The principal features which render this particular class of material so important are: 1. high sinking rates such that they are probably the principal vehicles by which material is transported to depth; 2. a microenvironment which differs markedly from the surrounding water such that they provide a specialized niche for a wide variety of faunal groups and a chemical environment which is different from the surrounding water; 3. Elevated biogeochemical rates within the particles over that in the surrounding water; 4. Provision of a food source for organisms swimming freely outside the snow particles.

686

1 cm

(A)

1 mm

(B) Figure 1 Examples of marine snow particles. (A) Aggregate comprising living chain-forming diatoms. Scale bar ¼ 1 cm. (B) Aggregate containing a variety of types of material including phytoplankton cells rich in plant pigment (brown colored).

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SNOW

came to the conclusion that large particles must be capable of transporting material from the sunlit surface zone to the abyssal depths. This was based on his observations that near bottom water above the abyssal seabed sometimes contained a surprising range of thin-shelled phytoplankton species, some still in chains and with their fine spines well preserved. He thought that the fecal pellets from some larger members of the plankton (doliolids, salps, and pteropods) were the likely vehicles and in many cases he was entirely correct. The process of aggregation is now thought to involve a variety of different mechanisms of which fecal production is only one (see Particle Aggregation Dynamics). In 1951 Rachael Carson described the sediments of the oceans as the material from the most stupendous snowfall on earth and this prompted a group of Japanese oceanographers to describe as ‘marine snow’ the large particles they could see from the submersible observation chamber Kuroshio (Figure 2). This submersible was a cumbersome device and did not permit anything but the simplest of observations to be made. The scientists did, however, manage to collect some of the material and reported that its main components were the remains of diatoms, although with terrestrial material appearing to provide nuclei for formation. In spite of these observations and the outstanding questions surrounding material cycles in the oceans, it was, until the late 1970s, a widely held belief that the deep-sea environment received material as a fine ‘rain’ of small particles. These, it was assumed, would take many months or even years to reach their ultimate destination on the seafloor. The separation of a few kilometers between the top and bottom of the ocean was thought sufficient to decouple the two ecosystems in a substantial way such that any seasonal variation in particle production at the surface would be lost by the time the settling particles reached the seabed. This now seems to have been a fundamental misconception. Part of the reason for this has been lack of understanding of the role of marine snow aggregates.

Methods of Examination As stated above, the first use of the term was from submersible observations and in situ visual observation still serves a valuable role (Figure 2). This may be from manned submersibles or the rapidly evolving class of remotely operated vehicles (see Remotely Operated Vehicles (ROVs))or by subaqua divers. Photographic techniques have developed fast over the past decade and the standard photography

687

(A)

Water bottles

CTD Camera

Flash light Fresnel lenses

(B) Figure 2 Devices for observing snow particles. (A) The submersible Kuroshio as used in the 1950s. (B) A photographic system incorporating a variety of other sensors and water bottles.

systems which provided much of the currently available data on distribution are being replaced by high definition video systems linked to fast computers which can categorize particles in near real time. Holographic techniques are also being developed for in situ use and these provide very high resolution images along with the three-dimension coordinates of the particles and organisms surrounding them. An important goal in any of these studies is to be able to obtain undamaged samples of marine snow particles so that they can be examined under the microscope, chemically analysed, and used for experimentation. Some types of marine snow are,

(c) 2011 Elsevier Inc. All Rights Reserved.

688

MARINE SNOW

however, very fragile and the devices which are used to collect water often destroy their structure unless special precautions are taken in the design such that the water intake is very large and turbulence around it is reduced (Figure 3A). In situ pumping systems (Figure 3B) have been used for several years to collect material but in this case separation of the large marine snow particles which have distinctive characteristics of sinking rate, chemistry and biology from the far more

(B)

(A)

(C)

abundant smaller particles is very difficult and may lead to erroneous conclusions as a result of this forced aggregation of different types of particle. Sediment traps are the principal means by which direct measurement can be made of the downward flux of material in the oceans (see Temporal Variability of Particle Flux) and it may, therefore, be supposed that this provides a means to collect undisturbed marine snow particles (Figure 3C).

(D)

Figure 3 Methods of collecting marine snow particles. (A) A 100 l water bottle ‘The Snatcher’; (B) a large volume filtration system; (C) a sediment trap; (D) a subaqua diver.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SNOW

Motor Pulley

Tank

Foam-covered steel rods

Camera Strobe

689

the presence of transparent exopolymer particles in snow particles dominated by some groups of phytoplankton and these are thought to be crucial components for the creation of some snow particles. Bacteria are invariably present in large numbers in snow particles. Sinking Rate, Density, and Porosity

Pillow block bearing

Figure 4 Device for producing marine snow particles in the laboratory.

Although the material collected in sediment traps may be considered as that which is sinking fast and is certainly the material which mediates downward flux, the characteristics of individual marine snow particles or even particles with features in common can not be determined as the boundaries between particles are not retained in the sediment trap sample jars. Furthermore the preservatives and poisons which are usually employed in such devices (formaldehyde, mercuric chloride, etc.) prevent experimentation on the recovered material and some types of chemical analysis. Laboratory production of marine snow particles was first achieved in the 1970s (Figure 4) enabling researchers to produce a plentiful supply of material under controlled and hence repeatable laboratory conditions. This is done by enclosing water samples which have high concentrations of phytoplankton in rotating water bottles for upwards of several hours. After a while the sheer forces encourage aggregation of the solid material. Much progress has been made with these particles and insights have been gained into the ways in which they are formed and the factors which cause their destruction. We should now ask the most basic of questions about this important class of material: what is marine snow? how is it distributed in time and space? and why is it of such significance?

Characteristics of Marine Snow Microscopic Composition

The microscopic composition of marine snow particles reflects to a major degree the processes responsible for their creation, and at any one location, the composition varies rather little; in some places it is dominated by diatoms or dinoflagelates and in other places by the mucus webs from larvaceans. Various staining techniques have been used to reveal

As illustrated in Figure 5, sinking rates increase with particle size and this extends throughout the range of particle sizes considered to be marine snow. The rates measured are consistent with other observations on the time delay between phytoplankton blooms at the surface of the ocean and the arrival of fresh material on the abyssal seafloor (see Floc Layers). Although the dry weight of individual particles increases with increasing size, the density decreases and porosity increases (Figure 5). The relationships between size and these physical characteristics is invariably a poor one with much scatter in the data points, a factor which introduces considerable uncertainty in trying to deduce important rates such as downward particulate flux from the size distribution of particles. This variability is due in large part to the variety of sources of material which comprise the final aggregated snow particle, but also to the different processes which contribute to the aggregation mechanism. The processes occurring within the snow particles, such as microbial degradation and photosynthesis, will also have an effect on the physical properties. Chemical Composition

From the perspective of the biogeochemical processes occurring in the ocean, a knowledge of the chemical composition of snow particles and its variation with region, depth, and time are essential. The basic elemental composition in terms of carbon and nitrogen shows an increase, as expected, with particle size, but the proportion of the dry weight which is carbon or nitrogen tends to show a slight decrease reflecting the incorporation of lithogenic material with time as the aggregate becomes larger (Figure 5). In general, the composition of snow particles is not very different from that of the smaller particles found in the same body of water and from which the snow particles have been formed and to which they contribute when they are fragmented. This reflects the frequent and rapid transformations between the different size classes in the sea. Biological Processes Within the Marine Snow Particles

Biological processes all occur at higher rates within marine snow particles than in comparable volumes

(c) 2011 Elsevier Inc. All Rights Reserved.

690

MARINE SNOW 3

not, however, imply that formation of marine snow encourages nutrient regeneration at shallower depths. However, the high sinking rates of the marine snow particles tend to increase the depth of remineralization and the balance between these two processes can not be described with confidence at the present time.

10

_1

Sinking rate (m day )

y = 50 x

0.26

2

10

Microenvironments 1

10

_1

10

0

10

10

2

1

10

Diameter (mm)

(A) 3

10

y = 0.99 x

Larvacean snow Diatom snow Fecal snow Misc. snow

_1

POC (μg agg )

10

2

0.25

1

10

0

10

0

10 (B)

1

2

10 10 _1 3 Aggregate size ESV (mm agg )

10

3

3

Larvacean snow Diatom snow Fecal snow Misc. snow

_1

Total chlorophyll (ng agg )

10

2

y = 2.72 x

0.53

Adsorption and Scavenging

10

1

10

0

10

0

10 (C)

1

Diffusion of solutes within any particulate entity will depend on the physical composition of the particle to a large extent and the effect of such diffusion will depend on the distances involved. With increasing size of particle, the chances of generating distinctive microenvironments within the particle increase. There are several examples in the literature where anoxic conditions have been found within snow particles as a result of enhanced oxygen consumption within the body of the particle and diffusion rates being insufficiently high to restore the concentrations outside the particle. These experiments have been done using laboratory-made snow particles or naturally collected specimens, but in both cases exclusion of metazoans which may fragment snow particles introduces some doubts about the frequency of anoxia in the natural environment. If proved to be common in ‘the wild’, the effect is to create microenvironments within the particle which will have a profound effect on the chemical and biological processes which take place within it.

2

10 10 _1 3 Aggregate size ESV (mm agg )

3

10

Figure 5 Relationships between marine snow particle size and (A) sinking rate, (B) organic carbon content and (C) chlorophyll content.

of water outwith the snow. Furthermore, in the case of microbial activity, the rates of growth are often substantially higher than for comparable weights of smaller particles. Flagellate and ciliate populations also tend to be enhanced, presumably taking advantage of the elevated levels of bacterial activity. The effect of this is for snow particles to be sites where nutrient regeneration is enhanced. This does

During descent through the water column, sinking particles accumulate smaller particles in their path and this scavenging is probably an important means by which small particles with very low sinking rates are transported to depth. Layers of fine particles commonly referred to as nepheloid layers (see Nepheloid Layers) may be reduced in intensity by the descent through them of sticky marine snow particles. Similarly, as with any solid surface, dissolved chemicals may be adsorbed onto marine snow particles and hence be drawn down in the water column at a much faster rate than would be possible by diffusion or downwelling. Food Source for Grazers, Enrichment Factors

Not only do marine snow particles provide an attractive habitat in which smaller fauna and flora can live, but they are also a food source for a variety of planktonic organisms and fish. The species most commonly found associated with snow particles tend to be copepods, particularly cyclopoids, but there

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SNOW

1000

_

Zooplankters (no. aggregate 1)

100

10

1

0.1

0.01 0.1

10

1 Equivalent radius (cm)

(A) 5

_1

Normalized abundance (ml particle )

10

4

10

Distribution in Time and Space

3

2

10

1

10

0

10

_1

0.1

1 Equivalent radius (cm)

(B) 10 000

Enrichment factor

All invertebrates 1000

100

10

Bacteria

1 0.01 (C)

are also examples of heterotrophic dinoflagellates, polychaete larvae, euphausiids, and amphipods. In the case of the last of these groups one species, Themisto compressa, which was previously thought to be an obligate carnivore, was found to feed voraciously on marine snow particles. Recent experimental work has demonstrated the tracking behavior of zooplankton whereby they follow the odor trail left in the water by a sinking snow particle. Such sophisticated behavior suggests that snow particles are important food sources for some species of zooplankton. Enhancement of the concentration of bacteria or zooplankton associated with snow particles can be expressed as an ‘enrichment factor’ and, as shown in Figure 6, this factor appears to decrease with increasing size but at a different rate for the different faunal groups. The effect of this is that with increasing size, the metazoan zooplankton appear to become more important.

a = 3.25 b = 2.27 R 2=0.96

10

10

691

0.1 1 Equivalent radius (cm)

10

Figure 6 (A) Abundance of zooplankters associated with marine snow particles. (B) Abundance of all invertebrates after normalization to the size of the individual particles. (C) Enrichment factors for bacteria (), ciliates (J), heterotrophic flagellates (&), and all invertebrates.

Marine snow particles are found throughout the world’s oceans in all parts of the water column. They are not uniformly distributed either in space or time but are usually found in higher concentrations in the upper water column and in the more productive regions of the oceans. Although it had been suspected since the early observations that their concentration decreases with increasing depth, this has been confirmed only recently. The profiles now becoming available do not however suggest a simple decrease. There is considerable structure, undoubtedly related to the processes of production, destruction, and sinking, which are related to the physics, biology, and chemistry of the water column and of the particles themselves. Figure 7 shows an example of a profile from the north-east Atlantic. Bearing in mind the strong seasonal variation which can occur even well below the upper mixed layer and the different techniques employed by different researchers to obtain profiles, a common story seems to be emerging. Apart from profiles near to the continental slope where snow concentrations tend to increase near the seabed due to resuspension, there is generally a rapid fall in concentration over the top 100 m. Peak concentrations are not, however, found throughout the upper mixed layer but are located at its base, a feature which is directly related to the rates of production and loss of the marine snow particles in this highly dynamic part of the water column. Sinking rates may well decrease significantly in this part of the water column as the particles sink into water of higher density.

(c) 2011 Elsevier Inc. All Rights Reserved.

692

MARINE SNOW

Abundance (no. /litre) (in size range 0.6 _ 9.8 mm) 0

100

200

300

400

500

0

50

snow at 270 m depth in the open ocean of the northeast Atlantic. It can be seen that the period of highest biological productivity in the spring elicits strong peaks in the marine snow concentration particularly in the largest size categories. There are several examples of diel changes in marine snow concentration within the upper few hundred meters and this is probably related to the activity of the zooplankton.

Production and Destruction 150

Sigma t

Depth (m)

100

200

250

300 Seabed at 4700 m Figure 7 Vertical profile of marine snow concentrations (m) in the north-east Atlantic (481N 171W).

As might be expected for material which is so intimately related to the biological cycles of the oceans, a strong seasonal cycle in marine snow concentration is usually found even at depths below the euphotic zone. Figure 8 shows the concentration of marine

Production of marine snow particles is, almost by definition, by a process of aggregation (see Particle Aggregation Dynamics). It can be divided into processes related to the sticking together of smaller biogenic particles such as individual phytoplankton cells, the discharge of mucous feeding webs from organisms such as larvaceans and pteropods or the ejection of fecal material from any organisms containing a gut. Although marine snow particles, whether they are amorphous aggregates or fecal pellets, retain their physical identity for many days or weeks if stored at ambient temperature, when in their natural environment, it is likely that their residence time is only a matter of hours or a few days. With apparent sinking rates of tens to hundreds of meters per day, it is in fact essential from the perspective of the economy of the upper ocean that the sinking rate of this material is retarded. The reason for this rapid destruction is likely to be that the zooplankton which swim between one snow particle

Figure 8 Seasonal change in the concentration of marine snow particles at 270 m depth in the north-east Atlantic (481N 201W). This is expressed as a volume concentration to emphasize the importance of the largest size category which are rare but contribute significantly.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARINE SNOW

693

(A)

(B)

(C)

Figure 9 Zooplankton interactions with large particles: (A) feeding on a fecal pellet; (B) feeding on marine snow particle; (C) fragmenting marine snow particle.

and the next are able to fragment them and ingest some parts of them in the process (Figure 9). Even fecal pellets which one might think of as being an unattractive food resource, are readily broken up (Figure 9). Fragmentation by wave action seemed at one stage to be another likely mechanism but even for the more fragile particles, this now seems to be of minor importance as the sheer rates are not usually adequate to break a significant number of particles.

heterotrophic organisms obtain a living. They are a food source for a variety of free-living organisms some of which have already been shown to adopt complex behaviour in order to locate them.

See also Floc Layers. Nepheloid Layers. Particle Aggregation Dynamics. Remotely Operated Vehicles (ROVs). Temporal Variability of Particle Flux.

Further Reading

Conclusion Marine snow particles play a crucial role in regulating the supply of material to the deep sea due to their high sinking rates. They provide a microenvironment in which most rates of biogeochemical processes are enhanced and in which many

Alldredge A (1998) The carbon, nitrogen and mass content of marine snow as a function of aggregate size. DeepSea Research(Part 1) 45(4–5): 529--541. Alldredge AL and Gotschalk C (1988) In situ settling behaviour of marine snow. Limnology and Oceanography 33: 339--351.

(c) 2011 Elsevier Inc. All Rights Reserved.

694

MARINE SNOW

Alldredge AL, Granata TC, Gotschalk CC, and Dickey TD (1990) The physical strength of marine snow and its implications for particle disaggregation in the ocean. Limnology and Oceanography 35: 1415--1428. Dilling L and Alldredge AL (2000) Fragmentation of marine snow by swimming macrozooplankton: a new process impacting carbon cycling in the sea. Deep-Sea Research Part 1 47(7): 1227--1245. Dilling L, Wilson J, Steinberg D, and Alldredge A (1998) Feeding by the euphausiid Euphausia pacifica and the copepod Calanus pacificus on marine snow. Marine Ecology Progress Series 170: 189--201. Gotschalk CC and Alldredge AL (1989) Enhanced primary production and nutrient regeneration within aggregated marine diatoms. Marine Biology 103: 119--129. Kiorboe T (1997) Small-scale turbulence, marine snow formation, and planktivorous feeding. Scientia Marina (Barcelona) 61(suppl. 1): 141--158.

Lampitt RS, Hillier WR, and Challenor PG (1993) Seasonal and diel variation in the open ocean concentration of marine snow aggregates. Nature 362: 737--739. Lampitt RS, Wishner KF, Turley CM, and Angel MV (1993) Marine snow studies in the northeast Atlantic: distribution, composition and role as a food source for migrating plankton. Marine Biology 116: 689--702. Lohmann H (1908) On the relationship between pelagic deposits and marine plankton. Int. Rev. Ges. Hydrobiol. Hydrogr 1(3): 309--323 (in German). Simon M, Alldredge AL, and Azam F (1990) Bacterial carbon dynamics on marine snow. Marine Ecology Progress Series 65(3): 205--211. Suzuki N and Kato K (1953) Studies on suspended materials. marine snow in the sea. Part 1, sources of marine snow. Bulletin of the Faculty of Fisheries of Hokkaido University 4: 132--135.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARITIME ARCHAEOLOGY R. D. Ballard, Institute for Exploration, Mystic, CT, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1675–1681, & 2001, Elsevier Ltd.

Introduction For thousands of years, ancient mariners have traversed the waters of our planet. During this long period of time, many of their ships have been lost along the way, carrying their precious cargo and the history it represents to the bottom of the sea. Although it is difficult to know with any degree of precision, some estimate that there are hundreds of thousands of undiscovered sunken ships littering the floor of the world’s oceans. For hundreds of years, attempts have been made to recover their contents. In Architettura Militare by Francesco de Marchi (1490–1574), for example, a device best described as a diving bell was used in a series of attempts to raise a fleet of ‘pleasure galleys’ from the floor of Lake Nemi, Italy in 1531. In Treatise on Artillery by Diego Ufano in the mid1600s, a diver wearing a crude hood and air-hose of cowhide was shown lifting a cannon from the ocean floor. Clearly, these early attempts at recovering lost cargo were done for economic, not archaeological, reasons and were very crude and destructive. The field of maritime archaeology, on the other hand, is a relatively young discipline, emerging as a recognizable study in the later 1800s. Not to be confused with nautical archaeology which deals solely with the study of maritime technology, maritime archaeology is much broader in scope, concerning itself with all aspects of marine-related culture including social, religious, political, and economic elements of ancient societies.

Early History Sunken ships offer an excellent opportunity to learn about those ancient civilizations. Archaeological sites on land can commonly span hundreds to even thousands of years with successive structures being built upon the ruins of older ones. Correlating a find in one area to a similar stratigraphic find in another can introduce errors that potentially represent long periods of time. A shipwreck, on the other hand, represents a ‘time capsule’, the result of a momentary

event where the totality of the artifact assemblage comes from one distinct point in time. It is important to point out that maritime archaeology’s first major field efforts were not conducted underwater but, in fact, were the excavation of boats that are now located in land-locked sites. In 1867, the owners of a farm began to cart away the soil from a large mound some 86 m in length only to discover the timbers of a large Viking ship, the Tune ship, complete with the charred bones of a man and a horse revealing it to be a burial chamber. In 1880, the Gokstad burial ship was discovered in a flat plain on the west side of the Oslo fiord. It was buried in blue clay which resulted in a high state of preservation. Contained within the grave was a Viking king, his weapons, twelve horses, six dogs and various artifacts. Since the late 1890s, the excavation of boats and harbor installations in terrestrial settings continues to this day, following more or less traditional land excavation protocol. One of the most famous discoveries took place in 1954 near the Great Pyramid of Egypt. While building a new road around the Pyramid, a series of large limestone blocks were encountered beneath which was found an open pit containing the oldest intact ship ever discovered. Dating to 2600 BC, the ship measures 43 m in length, weighs 40 tons, and stands 7.6 m from the keel line. Although ships found in terrestrial settings provide valuable insight into the culture of their period, many have a more religious significance than one reflecting the economics of the period. Burial ships were commonly modified for this unique purpose and were not engaged in maritime activity at the time of their burial. Prior to the advent of SCUBA diving technology invented by Captain Jacques-Yves Cousteau and Emile Gagnan in 1942, archaeology conducted underwater by trained archaeologists was extremely rare. In fact, owing to the dangerous aspects of preSCUBA diving techniques which included the use of ‘hard hats’ employed by commercial sponge divers, archaeologists relied upon these commercial divers instead of doing the underwater work themselves. Even after SCUBA technology became available to the archaeological community it was the commercial or recreational divers by their numbers who tended to discover an ancient shipwreck site. As a result, it was not uncommon for a site to be stripped of its small, unique and easily recovered artifacts before the ‘authorities’ were notified of a wreck’s location.

(c) 2011 Elsevier Inc. All Rights Reserved.

695

696

MARITIME ARCHAEOLOGY

Even after learning of these sites, many of the early efforts in underwater archaeology were conducted by divers lacking archaeological training. The first was the famous Antikythera wreck off Greece which contained a cargo of bronze and marble sculptures. The site was initially raided in 1900 by sponge divers wearing copper helmets, lead weights, and steel-soled shoes with air supplied to them from the surface by a hand-cranked compressor. Fortunately, the location of the wreck was soon discovered by the Greek government which with the help of their navy, mounted a follow-up expedition under supervision from the surface by Professor George Byzantinos, their Director of Antiquities, which resulted in the recovery of more of its valuable cargo. The wreck turned out to be a first century Roman ship carrying its Greek treasures back to Rome including the famous bronze statue Youth, thought to be done by the last great classical Greek sculptor, Lysippos. A few years later, another Roman argosy was discovered off Mahdia, Tunisia which was also initially plundered for its Greek statuary. Again, the local government and their Department of Antiquities took control of the salvage operation which continued from 1907 to 1913 enriching the world with its bronze artwork. Unfortunately, not all of these early shipwreck discoveries were reported to the local government. Their fate was far less fortunate than those just mentioned including a first century BC wreck lost off Albenga, Italy which was torn apart by the Italian salvage ship Artiglio II using bucket grabs to penetrate its interior holds. One of the first ancient shipwrecks to be excavated under some semblance of archaeological control was carried out by Captain Cousteau in 1951 along with Professor Fernand Benoit, Director of Antiquities for Provence off Grand Congloue near Marseilles, France. It, too, was initially discovered by a salvage diver. Although no site plan was ever published after three seasons of diving, the ship was thought to be Roman from 230 BC, 31 m long, at a depth of 35 m, and carrying 10 000 amphora and 15 000 pieces of Campanian pottery bowls, pots, and 40 types of dishes with an estimated 20 tons of lead aboard. At the same time, a colleague of Cousteau, Commander Philippe Taillez of the French Navy organized a similar excavation of a first century BC wreck on the Titan reef off the French coast but in the absence of an archaeologist failed to actually document the site. Without the presence of trained archaeologists working underwater, there was little hope

that acceptable techniques would be developed that met archaeological standards. In 1958, Peter Throckmorton who was the Assistant Curator of the San Francisco Maritime Museum, went to Turkey hoping to locate an ancient shipwreck site. Like many before him, he learned from local sponge divers the location of a wreck off Cape Gelidonya on the southern coast of Turkey from which a series of broken objects had been recovered. During the following year, Throckmorton’s team made a number of dives on the site and recovered a series of bronze tools and ingots revealing the ship to be a late Bronze Age wreck from around 1200 BC but no major effort was mounted. Throckmorton and veteran diver Frederick Dumas, however, returned in 1960 along with a young archaeologist eager to make his first openocean dives, George Bass. With Bass in charge of the archaeological aspects of the diving program, they began to establish for the first time true archaeological mapping and sampling protocol. Although primitive by today’s standard, they established a traditional grid system using anchored lines followed by the use of an airlift system to carefully remove the overburden covering the wreck. Working in 28 m of water, the ship proved to be 11 m in length, covered with a coralline concretion some 20 cm thick. Its more than one ton of cargo consisted of four-handled ingots of copper and ‘bun’ ingots of bronze as well as a large quantity of broken and unfinished tools, including both commercial and personal goods. These assemblages of artifacts suggested the captain was a Syrian scrap-metal dealer who was also a tinker making his way from Cyprus along the coast of Turkey to the Aegean Sea. Maritime archaeology truly came of age with the excavation of the Yassi Adav wreck from 1961 to 1964 off the south coast of Turkey. Running offshore near the small barren island of Yassi Adav is a reef which is as shallow as 2 m that has claimed countless ships over the years. Its surface is strewn with Ottoman cannon balls from various wrecks of that period. Several of the ships that have run aground on its unforgiving coral outcrops slide off its rampart coming to rest on a sandy bottom. One such ship, lying in 31 m of water was the focus of an intense effort carried out by the University Museum of Pennsylvania under the direction of George Bass. Prior to this effort, no ancient ship had ever been recovered in its entirety; this became the objective of this project. With the help of fifteen specialists and thousands of hours underwater, the team carefully mapped the site. The techniques developed during this excavation effort became the new emerging

(c) 2011 Elsevier Inc. All Rights Reserved.

MARITIME ARCHAEOLOGY

standards for this young field of research and continue to be followed today by research teams around the world. Bass’ team first cleared the upper surface of cargo of its encrustation of weed. Its 900 amphora were then mapped, cataloged, and removed. One hundred were taken to the surface for subsequent conservation and preservation whereas the remaining 800 were stored on the bottom at an off-site location. Simple triangulation was first used to initially delineate the wreck site followed by the establishment of a complex series of wire grids. Each object was given a plastic tag and artists hovered over the grid system making numerous sketches of the site before any object was recovered. Copper and gold coins recovered from the site revealed the ship to be of Byzantine age sinking during the reign of Emperor Heraclius from AD. 610 to 641. Following the recovery of the ship’s contents, the now-exposed hull provided marine archaeologists with their first opportunity to develop techniques needed to document and recover the ship’s timbers. This effort was extremely time consuming but the resulting insight proved worth the effort, providing the archaeological community with a transitional method of ship construction between the classical ‘shell’ technique to the later evolved ‘skeleton’ technique. Its length to width ratio of 3.6 to 1 further supported earlier suggestions that ships of this period would have to be built with more streamlined hulls to outrun and outmaneuver hostile or piratical adversaries.

The Growth of Maritime Archaeology Following the excavation of the Yassi Adav wreck, members of Bass’ team then conducted the extensive excavation of a fourth century BC wreck near Kyrenia in Cyprus directed by Michael Katzev. This effort mirrored the Yassi Adav project and resulted in the raising and conservation of the ship’s preserved hull structure. Since these early pioneering efforts, numerous maritime archaeology programs have emerged around the world. Off Western Europe and in the Mediterranean maritime archaeology remains a strong focus of activities. Bass and his Institute for Nautical Archaeology (Texas A&M University in College Station, Texas) continue to carry out a growing number of underwater research projects off the southern coast of Turkey centered at their research facility in Bodrum. His excavation of the late Bronze Age Ulu Burun wreck off the southern Turkish coast led to the

697

recovery of thousands of artifacts that provided valuable insight in a period of time marked by the reign of Egypt’s Tutankhamun and the fall of Troy. His Byzantine ‘Glass Wreck’ found north of the island of Rhodes and dating from the twelfth to thirteenth century AD, continues to generate important information about this particular period of maritime trade. In addition to the well-known work with the Wasa, Swedish archaeologists have conducted excellent work in the worm-free waters of the Baltic. In Holland, Dutch archaeologists have drained large sections of shallow water areas which contain a rich history of maritime trade dating from the twelfth to nineteenth centuries. Another major program directed by Margaret Rule took place in England with the recovery and preservation of the Mary Rose, a large warship lost in July 1545 during the reign of King Henry VIII. From this project, a great deal was learned about the long-term preservation of wooden timbers which is being incorporated in other similar conservation programs. Research efforts in America span the length of its human habitation. Recent archaeological research suggests that humans arrived in North America more than 12 000 years ago when a southern route was first thought to have opened in the glacial icesheet covering the continent. Some scientists now suggest that early humans may have circumvented this barrier by way of water or overland surfaces now submerged on the continental shelf. New research programs are now being designed to work on the continental shelf looking for early evidence of Paleoindian settlements. For years Indian canoes, rafts, dugouts, and reed boats have been discovered in freshwater lakes, and sinkholes in the limestone terrains of North and Central America have attracted researchers for many years in search of human sacrifices and other religious artifacts associated with native American cultures. Ships associated with early explorers, including Columbus, French explorer Rene La Salle, the British, and countless Spanish explorers, have been the focus of research efforts in the Gulf of Mexico, the Northwest Passage, and the Caribbean while shipwrecks from the Revolutionary War and War of 1812 have been discovered in the Great Lakes and Lake Champlain. Warships associated with the American Civil War have received renewed interest including the Monitor lost off Cape Hatteras, the submarine Huntley and numerous other recent finds in the coastal waters of the US east coast.

(c) 2011 Elsevier Inc. All Rights Reserved.

698

MARITIME ARCHAEOLOGY

Within the last two decades, deep water search systems developed by the oceanographic community have been used to successfully locate the remains of the RMS Titanic, the German Battleship Bismarck, fourteen warships lost during the Battle of Guadalcanal, and the US aircraft carrier Yorktown lost during the Battle of Midway. Ships associated with World War II have been carefully documented including the Arizona in Pearl Harbor and numerous ships sunk during nuclear bomb testing in the atolls of the Pacific. Maritime archaeology is not limited to European or American investigators. A large number of underwater sites too numerous to mention have been investigated off the coasts of Africa, the Philippines, the Persian Gulf, South America, China, Japan, and elsewhere around the world as this young field begins to experience an explosive growth.

Marine Methodologies As was previously noted, early underwater archaeological sites were not discovered by professional archaeologists; they were found, instead, by commercial or recreational divers. It wasn’t until the

mid-1960s that archaeologists, notably George Bass, began to devise their own search strategies. Being divers, their early attempts tended to favor visual techniques from towing divers behind their boats, to towed camera systems, and finally small manned submersibles (Figure 1). It was not until the introduction of side-scan sonars that major new wreck sites were found. Operating at a frequency of 100 kHz, such sonar systems are able to search a swath-width of ocean floor 400 m wide, moving through the water at a typical speed of 3–5 knots (5.5–9 km h 1). Today, numerous companies build side-scan sonars each offering a variety of options ranging from higher frequencies (i.e. 500 kHz) to improved signal processing, recording, and display. Various magnetic sensors have been used effectively over the years in locating sunken shipwreck sites having a ferrous signature. This is particularly true for warships with large cannons aboard. Magnetic sensors have also proved effective in locating buried objects in extremely shallow water, on beaches, and beneath coastal dunes. Over the years, a variety of changes have taken place with regard to the actual documentation of a

Figure 1 Archaeological mapping techniques pioneered by Dr George Bass of Texas A&M University for shallow water archaeology. These techniques were heavily dependent on the use of divers and were limited to less than 50 m water depths.

(c) 2011 Elsevier Inc. All Rights Reserved.

MARITIME ARCHAEOLOGY

wreck site. Beginning in the early 1960s, various stereophotogrammetry techniques were used. More recently the SHARPS (sonic high accuracy ranging and positioning system) acoustic positioning system has proved extremely rapid and cost effective in accurately mapping submerged sites. This tracking technique coupled with electronic imaging sensors, has produced spectacular photomosaics. More recently, remotely operated vehicles have begun to enter this field of research. In 1990, the JASON vehicle from the Woods Hole Oceanographic Institution was used to map the Hamilton and Scourge, two ships lost during the War of 1812 in Lake Ontario. Using a SHARPS tracking system, the vehicle was placed in closed-loop control and made a series of closely spaced survey lines across and along the starboard and port sides of the ships. Mounted on the remotely operated vehicle (ROV) was a pencil-beam sonar and electronic still camera which resulted in volumetric models of the ships as well as electronic mosaics.

Deep-water Archaeology The shallow waters of the world’s oceans where the vast majority of maritime archaeology has been done represent less than 5% of its total surface area. For years, archaeologists have argued that the remaining 95% is unimportant since the ancient mariner stayed close to land and it was there that their ships sank. This premise was challenged in 1988 when an ancient deep-water trade route was first discovered between the Roman seaport of Ostia and ancient Carthage. The discovery site was situated more than 100 nautical miles (185 km) off Carthage in approximately 1000 m of water. Over a nine-year period from 1988 to 1997, a series of expeditions resulted in the discovery of the largest concentration of ancient ships ever found in the deep sea. In all, eight ships were located in an area of 210 km2, including five of the Roman era spanning a period of time from 100 BC to AD 400. The project involved the use of highly sophisticated deep submergence technologies including towed acoustic and visual search vehicles, a nuclear research submarine, and an advanced remotely operated vehicle. Precision navigation and control, similar to that first used in Lake Ontario in 1990, permitted rapid yet careful visual and acoustic mapping of each site with a degree of precision never attained before utilizing advanced robotics, the archaeological team recovered selected objects from each site for subsequent analysis ashore without intrusive excavation. Deep-water wreck sites offer numerous advantages over shallow water sites. Ships lost in shallow

699

water commonly strike an underwater obstacle such as rocks or reefs severely damaging themselves in the process. Each winter more storms continue to damage the site as encrustation begins to form. Commonly, the site is located by nonarchaeologists who frequently retrieve artifacts before reporting the wreck’s location. Ships that sink in deep water, however, tend to be swamped. As a result, they sink intact, falling at a slow speed toward the bottom where they come to rest standing upright in the soft bottom ooze. When they are located, they have not been looted. Sedimentation rates in the deep sea are extremely slow, commonly averaging 1 cm per 1000 years. That coupled with cold bottom temperatures, total darkness, and high pressures result in conditions favoring preservation. Although wood-boring organisms remove exposed wooden surfaces, deep sea muds encase the lower portions of the wreck in an oxygen free environment. When deep-sea excavation techniques are developed in the near future, these wrecks may provide highly preserved organic material normally lost in shallow-water sites. The Roman shipwrecks located off Carthage were found within a much larger area of isolated artifacts spanning a longer period of time. The isolated artifacts appear to have been discarded from passing ships overhead. Given the slow sedimentation rates in the deep sea, it might be possible to easily delineate ancient trade routes by looking for these debris trails, thus learning a great deal about ancient maritime trading practices. Since this new field of deep-water archaeology has grown out of the oceanographic community, it brings with it a strong expertise in deep submergence technology. The newly developed ROVs possess the latest in advanced imaging, robotics, and control technologies. Using this technology, archaeologists are able to map underwater sites far faster than their shallow water counterparts (Figure 2). Most recently, a second deep-water archaeological expedition resulted in the discovery of two Phoenician ships lost some 2700 years ago. Located in 450 m of water about 30 nautical miles (55 km) off the coast of Egypt, these two ships are lying upright. Due to local bottom currents, both ships are completely exposed resting in two-meter deep elongated depressions.

Ethics As pointed out earlier, the salvaging of cargo from lost ships goes back much farther in time than marine archaeological research. As a result, this long

(c) 2011 Elsevier Inc. All Rights Reserved.

700

MARITIME ARCHAEOLOGY

Figure 2 Archaeological techniques pioneered at the Woods Hole Oceanographic Institution and the Institute for Exploration rely exclusively on remotely operated vehicle systems with operating depths down to 6000 m.

history of maritime salvage, rooted in international law, has led to a quasipublic acceptance of salvage operations, making it difficult for the archaeological community to garner moral and legal public support to protect and preserve truly important underwater archaeological sites. ‘Finders keepers’ remains rooted in the public’s mind as a logical policy governing lost ships. Further blurring the boundary between these two extremes in the early years was the fact that marine archaeologists relied upon the very community that was removing artifacts from underwater sites to tell them where they were located. This uneasy marriage between the diving community and the archaeological community has, in many ways, stifled the growth and acceptance of the field. Its lack of development of systematization which arises from its immaturity and lack of a large database has further hindered its acceptance into mainstream archaeology. Today’s marine salvagers commonly employ individuals with archaeological experience to participate in their operations. In some cases, this results in

important documentation of the site as was the case with the salvage of the Central America. In other cases, however, they are being used to create a false impression that archaeological standards are being followed when they are not. American salvagers have, in large part, concentrated their attention on lost ships of the Spanish Main beginning with search efforts off the coast of Florida where a large number of silver- and goldbearing ships were lost in hurricanes between 1715 and 1733. A famous shipwreck in this area, the Atocha, was exploited by salvager Mel Fisher. On the Silver Bank off the Dominican Republic in the Caribbean the richly laden Nuestra Senora de la Pura y Limpia Concepcion sank in October 1641. Salvage efforts seeking to retrieve its valuable cargo began almost immediately including one by the British in 1687. Lost from memory, the Concepcion was relocated in 1978 by American treasure hunters who continue their recovery efforts to this day. Fortunately, more and more countries are beginning to enact laws to protect offshore cultural sites,

(c) 2011 Elsevier Inc. All Rights Reserved.

MARITIME ARCHAEOLOGY

but with the emergence of deep-water archaeology which is conducted on the high seas, the majority of the world’s oceans and the human history contained within them are not protected. One logical step is to add human history to the present Law of the Sea Convention that governs the exploitation of natural resources. Although this would not protect all future underwater sites, it would serve as an important first step.

See also Remotely Systems.

Operated

Vehicles

(ROVs).

Sonar

Further Reading Babits LE and Tilburg HV (1998) Maritime Archaeology. New York: Plenum Press. Ballard RD (1993) The MEDEA/JASON remotely operated vehicle system. Deep-Sea Research 40(8): 1673--1687. Ballard RD, McCann AM, Yoerger D, Whitcomb L, Mindell D, Oleson J, Singh H, Foley B, Adams J,

701

Piechota D, and Giangrande C (2000) The discovery of ancient history in the deep sea using advanced deep submergence technology. Deep Sea Research, Part I 47: 1591--1620. Bass GF (1972) A History of Seafaring Based on Underwater Archaeology. New York: Walker. Bass GF (1975) Archaeology Beneath the Sea. New York: Walker. Bass GF (1988) Ships and Shipwrecks of the Americas. New York: Thames and Hudson. Cockrell WA (1981) Some moral, ethical, and legal considerations in archaeology. In: Cockrell WA (ed.) Realms of Gold. Proceedings of the Tenth Conference on Underwater Archaeology, Fathom Eight San Marino, California. pp. 215–220. Dean M and Ferrari B (1992) Archaeology Underwater: The NAS Guide to Principles and Practice. London: Nautical Archaeology Society. Greene J (1990) Maritime Archaeology. London: Academic Press. Muckelroy K (1978) Maritime Archaeology. New York: Cambridge University Press. NESCO (1972) Underwater Archaeology – A Nascent Discipline. Paris: UNESCO.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDDIES AND SUB-SURFACE EDDIES H. T. Rossby, University of Rhode Island, Kingston, RI, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1682–1689, & 2001, Elsevier Ltd.

Introduction Meddies and related types of circular motion in the ocean belong to a class of eddy activity characterized by a highly coherent, axisymmetric circulation in the horizontal plane. These stable features have a lifetime measured in years, during which time they may drift thousands of kilometers, carrying with them waters from where they were formed. They play an important, but as yet inadequately defined and quantified, role in the transport and exchange of waters between different regions. Understanding these processes is of fundamental importance for a correct characterization of subsurface eddy processes in the ocean and their representation or parametrization in ocean circulation models. These subsurface eddies, shaped as very thin disks or lenses with an aspect ratio of B1 : 50 to B1 : 100, have a core body that rotates virtually as a solid disk, surrounded by a perimeter region of strong radial shear. Among the largest and most conspicuous of this type of eddy motion, the meddy (for Mediterranean eddy), can have diameters exceeding 100 km and life spans measured in years. We now know that this type of eddy motion, discovered in 1976, occurs in many regions of the world ocean, at shallow depths and deep, in tropical, subpolar, and arctic waters. Some eddies, such as the meddies, rotate anticyclonically, but, evidently just as likely, lenses may rotate in the other direction. From the growing observational database it now appears that the meddies do not merely drift with, but can in fact move through the surrounding waters. Their ubiquity, longevity, and mobility render them of potentially great importance in the transport, exchange, and mixing of waters between different water masses in the ocean. But there is much about these enigmatic features we have yet to understand.

Definitions On scales of tens of kilometers and larger in the ocean, fluid motion is in geostrophic balance, meaning that the pressure gradient is balanced by the

702

Coriolis force. These forces act at right angles to the direction of motion and thus do not tend to accelerate or alter the pattern of flow. For the circular motion of eddies discussed in this article, particle motion is curved rather than straight. This adds a radial acceleration, v2 =r, also perpendicular to the direction of motion, into the momentum balance, giving rise to a cyclogeostrophic balance or flow: fv þ

v2 1 qp ¼ r r qr

½1

where r is the radius of the curved motion, v is the azimuthal velocity, p is pressure, f ð¼ 2O sin ðlatitudeÞÞ is the Coriolis parameter, O is the angular velocity of the earth, and r is the density of the fluid. Again, because the forces act at right angles to the direction of motion, they do not alter the pattern of flow, i.e., the pattern is self-preserving. We call the clockwise motion of the (northern hemisphere) meddies anticyclonic, because they have a pressure maximum in the center. Low-pressure cyclonic eddies rotate in the opposite direction. Cyclones and anticyclones rotate in the opposite direction in the southern hemisphere. Potential vorticity expresses the circulation per unit volume of fluid and for our purposes can be written as ðf þ zÞ=h ¼ constant, where f is the Coriolis parameter and z represents the relative vorticity of a layer of fluid with thickness h. The conservation of this quantity, in the absence of forcing or dissipation, severely constrains the movement of fluids. For steady axisymmetric motion, a fluid parcel’s potential vorticity is automatically conserved. A measure of the intensity of rotation is given by the ratio of the relative vorticity of the core of the lens to the planetary vorticity, the Rossby number: R ¼ z=f . Typical R values for meddies range between  0.1 and  0.6, with the most extreme value reported ¼  0.85. Another number, the Burger number, expresses the ratio of strength of relative vorticity to the vortex stretching terms in the potential vorticity equation and is normally written as N 2 H 2 =ðf 2 L2 Þ. The Burger number can also be defined as the ratio of the available potential energy to kinetic energy of the lens. SOFAR (sound fixing and ranging) and the related RAFOS floats reveal how fluid parcels drift, disperse, and mix in the ocean from their trajectories, which are determined by acoustic triangulation. SOFAR floats transmit signals to stationary hydrophones;

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDDIES AND SUB-SURFACE EDDIES

RAFOS floats listen to moored acoustic sound sources. The travel times multiplied by the speed of sound in the ocean give the distances to within a few kilometers accuracy. Isopycnal RAFOS floats can also drift with the waters in the vertical. Because isopycnals move up or down or as fluid slides up or down along isopycnals such as across the Gulf Stream, isopycnal floats will accurately follow that motion. These acoustically tracked floats have been major contributors to our knowledge of the subsurface eddy field.

History During an oceanographic cruise in the Fall of 1976 to study ocean currents east of the Bahamas, a large body of very warm and salty water was observed at 1000 m depth. Shaped as a thin lens, it had a core diameter of nearly 150 km and a thickness of 500 m. Nothing like this had been observed before. Several SOFAR floats deployed in it to study the currents revealed a clockwise circular motion with a 10-day rotation period for the innermost float at 10 km radius. The proximity of this warm saline lens to the well-known Mediterranean Salt Tongue that stretches west across the ocean (centered near 301N) clearly suggested a Mediterranean origin. This discovery stimulated a search for similar temperature– salinity anomalies in the eastern Atlantic, and before long it became clear that these lenses, coined meddies for Mediterranean eddies, have a widespread distribution in the eastern Atlantic west of Spain and Portugal. We now know that these lenses belong to a class of coherent motion most characterized as thin spinning disks with a thickness to diameter (or aspect) ratio of about 1 : 100. While meddies rotate only anticyclonically, other subsurface eddies or lenses may rotate in the other direction. Their overall diameters range from perhaps less than 10 km to as large as the 150 km of the original meddy, which remains one of the largest ever found. During lifetimes of months to years, they may travel thousands of kilometers. The focus here is first on the wellstudied meddies, their structure, origin, patterns of drift, and decay. Other observations that illustrate the ubiquity of this class of eddy motion are then discussed.

The Meddy Structure

The typical meddy has a very distinctive density and velocity field. A vertical cross-section will reveal a spreading of the isopycnals such that the core or

703

center of the eddy has weak stratification bounded by layers above and below with very high stratification (Figure 1). This leads to a pressure field that is higher inside the lens than outside, which requires for equilibrium an anticyclonic rotation (clockwise in the northern hemisphere). The top panels show typical profiles of temperature, salinity, and density of meddy ‘Sharon’ in 1984 and in 1985, one year later. Averaged over the volume of the meddy the contributions of high temperature and salinity to density must cancel and equal the average density of the displaced waters (Archimedes’ principle). Hence, the core waters at 121C and 36.2 PSU, have the same density as the surrounding waters at 81C and 35.6 PSU at 1000 m where the density profiles can be seen to cross. But the vertical spread of the isopycnals results in radial density gradients such that for the lower half of the meddy the density inside is less, and for the upper half is higher, than that of the surrounding waters. If we imagine that at great depth the pressure is the same everywhere, then as we ascend into the meddy, the hydrostatic pressure will decrease less rapidly than outside so that pressure in the center of the core exceeds that of the surrounding waters by about 500 Pa (5 mbar). It is this excess pressure that maintains the orbital or rotary motion of the eddy in cyclogeostrophic balance. If we continue up through the meddy, the greater density of the core waters leads to a more rapid pressure drop than outside such that topside of the meddy at the surface the radial pressure all but vanishes. (This does not always apply, some meddies, especially recently formed ones, may have a surface signature.) Meddy ‘Sharon,’ by far the best-documented meddy, was visited four times over a 2-year period during which detailed surveys of the density, velocity, and microstructure were conducted. Orbiting SOFAR floats trapped in the meddy made it easy to relocate for subsequent visits. Vertical profiles of horizontal velocity show a maximum at about 1000 m depth and increasing linearly outward. Beyond a certain radius the velocity field decreases rapidly. Figure 1D shows the azimuthal velocity as a function of radius. The sharp transition from a linear increase to radial decay indicates the radial limit of solid body rotation. This rotation rate has a maximum at mid-depth (near 1000 m) and decreases slightly above and below. But at each depth the rotation appears to be that of a solid body; that is, the meddy can also be characterized as a stack of disks, each rotating at its own rate with the one in the center having the highest rotation frequency (Figure 1E). This is consistent with a density field that is nearly uniform in the horizontal yet stratified in the vertical.

(c) 2011 Elsevier Inc. All Rights Reserved.

704

MEDDIES AND SUB-SURFACE EDDIES

0

0

0

200

200

200

400

400

400

600

600

800

800

Depth (m)

1984

1000

800

Bg

1000

1984 Bg

600

1984

Bg

1000

1985

1200

1200

1400

1400

1400

1600

1600

1600

1800

1800

1800

1200

1985

1985

2000

4

12

Temperature (˚C)

16

2000

35.4

(B)

36.0

Salinity (PSU)

0

2000 (C)

31.6

32.0

32.4

_

σ1 (kg m 3 )

Period (days) 12

6

400

_10 _20

1984

600

_30 0

10

20

30

40 3 10 m

50

60

70 800 Depth (m)

(D)

36.6

24

_

Velocity (cm s 1)

(A)

8

1000

1200

1984

1400

1985

1600 0.0 (E)

0.4 0.8 _ _ Rotation frequency (10 5 s 1)

1.2

Figure 1 Temperature (A), salinity (B) and density (C) in the core of meddy ‘Sharon’ in October 1984 (solid), October 1985 (dotted) and background (dashed). Panel (D) shows azimuthal velocity in cm s1 as a function of radius in 1984 and panel (E) shows angular velocity (bottom) and period of rotation (top) as a function of depth.

Between the visits to the meddy in 1984 and 1985, the strong core of undiluted salt had shrunk substantially as intrusions of fresher waters from the outside reached toward the center. This was not the

case for the velocity field. Although it had shrunk in diameter, the core still evinced solid body rotation with a sharp transition to a region of radial decay beyond. Thus the dynamical structure is preserved

(c) 2011 Elsevier Inc. All Rights Reserved.

705

MEDDIES AND SUB-SURFACE EDDIES

even as the meddy loses its water mass anomaly. Figure 2 shows cross-sections of the salinity for the two surveys. Note the substantial loss of core waters but retention of a well-defined velocity structure. Evidently small-scale leakage and diffusion along isopycnals can ‘tunnel’ through the larger-scale dynamics. But what maintains the sharp velocity transition at the velocity maximum (panel (D) in Figure 1)? Fluid parcels do not remain at exactly the same radius as the meddy spins, but slowly wander in toward the center and out in an apparently random fashion. As SOFAR floats orbit around the meddy, they gradually drift in and out relative to the center, indicating a radial exchange of fluid within the core as it rotates. The very weak departures from solid body rotation within the core can be viewed alternatively as facilitating radial exchange and mixing or as the result thereof. Either way, this indicates a continuing process of homogenization of the core. This homogenization applies only to regions where the potential vorticity itself remains sufficiently uniform so as to allow the process to continue, i.e., within the core out to the radius of maximum velocity. The sharpness of the potential vorticity front for both years can be seen in Figure 3. Note how the radial boundary remains sharp despite the reduction in size. This stands in contrast to the continuing loss of salt inside the core of the meddy (Figure 2). Meddy Formation

At the south-west corner of Portugal, Cape St. Vincent, the continental slope makes a sharp turn

to the north. The Mediterranean outflow, a warm saline flow in geostrophic balance along the slope at about 1000 m depth, tends to follow the slope north. But sometimes, especially when the flow is strong, the current appears to overshoot and continues as an unbalanced flow to the west and curving to the north owing to the Coriolis force. If the curved motion is strong enough, it can fold back on itself, forming a closed loop and resulting in the genesis of a meddy. The negative relative vorticity of the meddy comes from the curvature of the flow and the negative lateral shear between the undercurrent and the bottom. The formation process was demonstrated by a series of RAFOS floats released into the Mediterranean Outflow over the period of a year. Some of these turned north along the bathymetry, but others exhibited the orbital motion we associate with meddies. These immediately broke away from the continent and started their drift to the south west. Other meddies were spawned farther north near the Estremadura Promontory, where the bathymetry also makes a sharp turn to the right. Figure 4 shows an example of such a trajectory. The timescale for eddy formation is estimated at 3–7 days with, in all, about 15–20 meddies spawned per year. Decay of the Meddy

It seems rather remarkable that these slender lenses with a core rotation period measured in days can last for hundreds of revolutions. Those that drift west to south-westward from Portugal toward the midAtlantic Ridge may reach 4–5 years of age. Meddy ‘Sharon,’ which was visited four times over a 2-year

Radius (km)

Radius (km) 31.35

0

25

50

31.35

0

25 35.7

1 (kg m 3 )

31.85 36.0

35.6

_

35.6

_

1 (kg m 3 )

35.6

35.8

36.2

31.85

36.0 35.4 36.2

35.6 32.35

32.35

5.0

5.0

2.5 2.5 2.0 2.0

(B)

1.0

2.0 2.0 2.0 2.0 2.5 2.5

10.0

5.0

5.0

2.5 2.5 2.5 2.5

3.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.5 2.5 2.5

(A)

Figure 2 Radial s1 distribution of salinity (PSU) in meddy ‘Sharon’ for 1984 (A) and 1985 (B). s1 is density relative to a pressure of 1000 dbar.

(c) 2011 Elsevier Inc. All Rights Reserved.

706

MEDDIES AND SUB-SURFACE EDDIES

400

0

10

20

Radius (km) 30 40

8 10

9 10

800 9

8 1000

38.0˚

7 6

3

1200

Latitude (N)

Depth (m)

9

60

9

8 600

39.0˚

50

5

4

4

1400

6

5

3 37.0˚

1600 (A) 400

0

10

20

30

40

Depth (m)

600

9

280/94

10 9

36.0˚

8 3

11.0˚

7

10.0˚

9.0˚

8.0˚

Longitude (W)

6

1000

1400

60

8

800

1200

50

9

10

5 3

Figure 4 Formation of a meddy off Cape St. Vincent as indicated by the trajectory of a RAFOS float. (Reproduced from Bower et al. (1997) with permission of the American Meteorological Society.)

4

3

1600 (B) Figure 3 Potential vorticity distribution in meddy ‘Sharon’ for 1984 (A) and 1985 (B).

period during its 1100 km drift south, was probably at least a year old at the time of discovery. These repeat visits documented in detail both a radial and a vertical erosion. The organized orbital motion leads to substantial and sustained vertical shear across the large surfaces at the top and bottom of the meddy. However, the increased stability associated with the crowding of the isopycnals more than suffices to suppress shear flow instabilities. Double-diffusive processes, which depend upon the fact that heat diffuses more rapidly than salt, appear to play a more important role. Vertical erosion occurred both above and below, with the greater losses along the lower perimeter due to salt-fingering. Radial erosion, as indicated by the loss of salt in the core, appears to take place by means of intrusions along isopycnals. Between the first and last visit two years apart, the vertical and horizontal scales had been more than halved such that the identifiable volume had been reduced by a factor ð8 km=30 kmÞ2  ð300=700 mÞ to about 3% of its original size. The greater radial

than vertical reduction points to some form of ablation at the perimeter rather than erosion at the top and bottom surfaces. The decrease in radial scale relative to the vertical might indicate an increase in Burger number (i.e., kinetic to potential energy ratio), but the uncertainties associated with this estimation process have left the matter unresolved. In any event, the faster radial than vertical erosion is testimony to the efficacy with which vertical stratification suppresses diapycnal exchange processes, even in the presence of enhanced vertical shear due to the rotation of the lens. Energetics

Two measures define the energetics of the meddy, the available potential energy (APE) and its kinetic energy (KE). The former is defined as the energy that would be released by restoring all the density surfaces to a reference state at which all pressure gradients and hence motion associated with the meddy will vanish. It is defined as the integral ð ½2 APE ¼ 1=2rN 2 p2 dV v

While easy to state, the accuracy of the integration depends upon a determination of the background

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDDIES AND SUB-SURFACE EDDIES

v

can be estimated fairly accurately. For most subsurface eddy studies the ratio of KE/APE, the energy Burger number, tends to be somewhat greater than unity, particularly as they age. This reflects the tendency for the aspect ratio of the lens to increase with age owing to the greater erosion around the perimeter than from above and below. For young meddies, APE and KE are of order 1014 J, which equals the output of a 1 GW electric utility plant for one day.

Sudden Death

Most meddies do not have the privilege of reaching a great age. Instead, there is a high probability of collision with one of the large number of seamounts in the eastern North Atlantic west of the Iberian Peninsula. Sometimes they fragment into smaller lenses that can continue for months to years. It has been estimated that perhaps 90% of all meddies eventually collide with a seamount, with an estimated average age at collision of 1.7 years. Those that do survive might live up to 5 years. Apparently spontaneous breakup of meddies into smaller units has been observed, but the extent to which these occur owing to internal instabilities or result from interactions with other currents or eddies that shear them apart needs further study. Given the great age that meddies can reach, it would appear that the probability of spontaneous fission is quite small. Sharp lateral shear in the ambient flow could also wear at the meddy, but the large and organized relative vorticity of the meddy, typically 0.2–0.6 times the Coriolis parameter, renders it immune to the surrounding eddy field. Examples also exist of coalescence of smaller eddies into larger ones. Significantly, almost all information about eddy interactions comes from the trajectories of SOFAR and RAFOS floats, which give us considerable spatial information as they drift about. Given the tight structure of the meddy, a single float will suffice to tell us the trajectory of the meddy, its collision with seamounts, and its possible demise. On the other hand, almost all our information on the mechanisms of aging comes from ‘Sharon,’ which, as noted above, shrank in volume by two orders of magnitude during the 2-year study. Curiously, meddies farther

to the west appear to be much larger and vigorous at a comparable or greater age. Thus, it remains unclear how well ‘Sharon’ represents the meddy population as a whole. This is of more than passing interest because meddies have been suggested as a mechanism for maintaining the Mediterranean Salt Tongue (MST) that extends across the ocean near 301N. If meddies are common in the eastern Atlantic, and it has been estimated that there might be of the order of 30 meddies at any given time, it seems remarkable that not a single meddy has been found west of the mid-Atlantic Ridge since the original meddy observation in the fall of 1976. Figure 5 shows a summary of all meddy sightings in relation to the salinity anomaly of the Mediterranean Salt Tongue. Whereas many meddies drift south, note the conspicuous absence of meddies west of B301W along the axis of the salinity anomaly, the maintenance of which also remains an enigma. Indeed, there is strong circumstantial evidence that the original meddy did not have a Mediterranean origin but came from quite far north in the northwest Atlantic. Near 501N where the North Atlantic Current abruptly turns east as the Subpolar Front, a strong anticyclonic rotation is maintained by the current. It has been observed that this semidetached circulation can subduct and move south across the Newfoundland Basin as a subsurface eddy and continue west and south across the Sargasso Sea. With the help of the Gulf Stream recirculation system, the transit time may only be 2–3 years instead of 5 years, despite the large distance involved. Other subsurface warm-core lens sightings in the western Atlantic lend further support to this alternative origin.

50˚

Longitude (N)

rest state, i.e., the accuracy with which the vertical displacement in the integral can be estimated. The N2 term represents the vertical stratification. The corresponding KE integral ð ½3 KE ¼ 1=2rv2 dV

707

40˚

30˚

60˚

50˚

30˚ 40˚ Latitude (W)

20˚

10˚

Figure 5 Summary of historical meddy observations. The diameter of the dots in the figure is about 50 km, somewhat smaller than the typical 100 km diameter of meddies. The contours show salinity anomaly relative to 35.01 PSU near a depth of 1100 m. (Reproduced from Richardson et al. (2000) with permission of Pergamon Press.)

(c) 2011 Elsevier Inc. All Rights Reserved.

708

MEDDIES AND SUB-SURFACE EDDIES

Other Subsurface Lenses Numerous other lenses have been observed and described. The structure of ‘Sharon’ seems to apply to others after appropriate scaling for size. Thus, an eddy about 50 km in diameter between 1000 and 2000 m depth located about 400 km west of Bermuda had a very similar structure. The isopycnals bend up and down forming a core region with weak stratification. The cold fresh waters in the core clearly point to a Labrador Sea origin, but where the lens itself was formed remains uncertain (analogous to the formation of meddies off Portugal containing waters originating in the Mediterranean Sea). One of the smallest yet very energetic subsurface eddies ever observed was tracked for 2 months in real time with a SOFAR float at 700 m depth west of Bermuda. A detailed hydrographic survey when the float was picked up indicated a diameter of B20 km and 300 m thickness. The core exhibited a distinct water mass anomaly with temperature, salinity, oxygen, and nitrate characteristics suggesting that the waters came from a low latitude, but where the eddy itself originated remains unclear; perhaps it was advected by the Gulf Stream north, perhaps it was formed by the meandering of the current. In any event, this lens, with a 1.5 : 100 aspect ratio, had a very fast rotation rate, about 3.5 days. An even faster rotation rate, 2 days, was observed for a small lens in the Gulf of Cadiz. This is very close to the theoretical limit where the relative vorticity of the lens exactly cancels the planetary vorticity. While the number of detailed lens studies remains limited, smaller lenses seem to have a higher rotation rate than larger ones. This suggests that, as the lenses age, they do so by decreasing their radius more rapidly than their height, so that their aspect ratio increases. The effect of this is to increase the Burger number of the lens. For a given pressure anomaly in the center, a smaller radius means a higher azimuthal velocity. The high angular velocity of these two small eddies compared to that of larger ones suggests a possible end fate in which the core remains intact as the ablation around the perimeter proceeds. Curiously, almost all reports have focused on anticyclonic lenses despite the fact that we know from float observations that cyclonic lenses occur with near equal probability. Cyclonic lenses have received much less attention. For these, the density surfaces must bow in rather than out, inviting the description ‘concave lenses.’ The best-documented examples of these have been observed in the West European Basin. Interestingly, these also carry a positive salt anomaly, apparently to the north west toward the mid-Atlantic Ridge. Indeed, there is

growing evidence that cyclonic eddies tend to drift poleward, whereas anticyclonic eddies drift equatorward. The classical argument for this is that as they age and lose their relative vorticity, they compensate for this by changing their latitude. In the case of meddy Sharon, the peak angular velocity actually increased, but the vertically averaged rotation rate clearly decreased (Figure 1E).

Discussion and Summary The discovery of the meddy has a curious history. One of the first lenslike subsurface eddies to be identified as such was found just north-east of the Dominican Republic in the fall of 1976. It was nearly 150 km in diameter and 500 m thick, and the temperature–salinity characteristics of the core of the eddy and its proximity to the axis of the Mediterranean Salt Tongue suggested a Mediterranean origin. The report of this finding stimulated the search for similar lenses in the eastern Atlantic, and soon enough many others had been found. But what makes the original discovery all the more remarkable is that no other meddy has since been sighted west of the mid-Atlantic ridge. In addition, the probability of meddies getting that far decreases rapidly owing to the high risk of collision with seamounts. Even if a meddy did cross the midAtlantic ridge, the additional 2000 km distance to the original sighting makes that observation seem all the more extraordinary if not implausible. Instead, it now appears that the original ‘meddy’ actually originated in the north-west Atlantic where the North Atlantic Current turns east at 501N. While the distance from that location almost matches that from Cape St. Vincent, Portugal, a lens originating in the North Atlantic Current (NAC) can be carried or advected rapidly to the south and west by the recirculating waters east of the NAC and south of the Gulf Stream, reducing the transit time to 2–3 years instead of 5 years. The decrease in latitude favors the anticyclonic eddy, but the nearly 50% reduction in Coriolis parameter suggests that the lens must undergo considerable adjustment. Perhaps the extraordinary width and flatness of the original meddy has its explanation here: As the Coriolis parameter f decreases, a decrease in thickness h would indicate a tendency to conserve its potential vorticity ðf þ zÞ=h: On a more speculative note, if the lens did indeed flatten and widen, this could help explain the extraordinary diameter of the original ‘meddy’ and simultaneously give it an additional lease of life against radial erosion. The fact that subsurface eddies larger than B100 km in diameter have not been observed suggests an upper limit at formation time set by inertia.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDDIES AND SUB-SURFACE EDDIES

Meddies form at Cape St. Vincent at the south-west corner of Portugal where the Mediterranean Undercurrent must make a sharp turn to the north along the continental slope. Owing to inertia, the current may overshoot, becoming geostropically unbalanced where the bottom turns north. This causes the current to curve to the right owing to the Coriolis force. For faster than normal flow, this curving flow can almost fold back on itself, resulting in a closed loop leading to the genesis of a meddy. Given the frequent rate of formation of meddies at Cape St. Vincent site, this would be an excellent place to study the formation process in greater detail. Other sharp topographic features have been identified as sites for the formation of anticyclonic lenses. In contrast, remarkably little is known about how cyclonic lenses get spun up. Perhaps they result from instabilities of fronts and/or fission from larger eddies. No lower limit to the size of subsurface lenses has been established, but, at some limit, viscosity and double-diffusive processes will dissipate what is left. Before that limit is reached, however, the lenses can still remain remarkably energetic. But the very small pressure gradients needed to balance the cyclogeostrophic motion all but guarantees that they can only be detected and identified as such by Lagrangian means.

709

Further Reading Bower AS, Armi L, and Ambar I (1997) Lagrangian observations of Meddy formation during a Mediterranean Undercurrent seeding experiment. Journal of Physical Oceanography 27: 2545--2575. Hebert D, Oakey N, and Ruddick B (1990) Evolution of a Mediterranean salt lenss: calar properties. Journal of Physical Oceanography 20: 1468--1483. Journal of Physical Oceanography March 1985. Special issue with numerous articles devoted to studies and observations of subsurface eddies. Prater MD and Rossby T (1999) An alternative hypothesis for the origin of the ‘Mediterranean’ salt lens observed off the Bahamas in the fall of 1976. Journal of Physical Oceanography 29: 2103--2109. Richardson PL, Bower AS, and Zenk W (2000) A census of Meddies tracked by floats. Progress in Oceanography 45: 209--250. Robinson AR (ed.) (1983) Eddies in Marine Science. New York: Springer Verlag. Schauer U (1989) A deep saline cyclonic eddy in the West European Basin. Deep-Sea Research 36: 1549--1565. Schultz Tokos K and Rossby T (1991) Kinematics and dynamics of a Mediterranean salt lens. Journal of Physical Oceanography 21: 879--892.

See also Double-Diffusive Convection. Drifters and Floats. Intrusions. Rossby Waves.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION A. R. Robinson and W. G. Leslie, Harvard University, Cambridge, MA, USA A. Theocharis, National Centre for Marine Research (NCMR), Hellinikon, Athens, Greece A. Lascaratos, University of Athens, Athens, Greece Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1689–1705, & 2001, Elsevier Ltd.

Introduction The Mediterranean Sea is a mid-latitude semienclosed sea, or almost isolated oceanic system. Many processes which are fundamental to the general circulation of the world ocean also occur within the Mediterranean, either identically or analogously. The Mediterranean Sea exchanges water, salt, heat, and other properties with the North Atlantic Ocean. The North Atlantic is known to play an important role in the global thermohaline circulation, as the major site of deep- and bottom-water formation for the global thermohaline cell (conveyor belt) which encompasses the Atlantic, Southern, Indian, and Pacific Oceans. The salty water of Mediterranean origin may affect water formation processes and variabilities and even the stability of the global thermohaline equilibrium state. The geography of the esntire Mediterranean is shown in Figure 1A and the distribution of deep-sea topography and the complex arrangement of coasts and islands in Figure 1B. The Mediterranean Sea is composed of two nearly equal size basins, connected by the Strait of Sicily. The Adriatic extends northward between Italy and the Balkans, communicating with the eastern Mediterranean basin through the Strait of Otranto. The Aegean lies between Greece and Turkey, connected to the eastern basin through the several straits of the Grecian Island Arc. The Mediterranean circulation is forced by water exchange through the various straits, by wind stress, and by buoyancy flux at the surface due to freshwater and heat fluxes. Evaporation 1.27 m/year, Precipitation 0.59 m/year, Mediterranean outflow (through the Gibraltar) B1.0 Sv, the inflow exceeds outflow by 5% (0.05 Sv) to compensate the water deficit of the Mediterranean, fresh water input 0.67 m/year, which comprises precipitation, river runoff and the Black Sea input, Net salt flux towards the Atlantic B2  106 kg/s.

710

Research on Mediterranean Sea general circulation and thermohaline circulations and their variabilities, and the identification and quantification of critical processes relevant to ocean and climate dynamics involves several issues. Conceptual, methodological, technical, and scientific issues include, for example, the formulation of multiscale (e.g., basin, sub-basin, mesoscale) interactive nonlinear dynamical models; the parametrization of air–sea interactions and fluxes; the determination of specific regional processes of water formation and transformations; the representation of convection and boundary conditions in general circulation models. A three-component nonlinear ocean system is involved whose components are: (1) air–sea interactions, (2) water mass formations and transformations, and (3) circulation elements and structures. The focus here is on the circulation elements and their variabilities. However, in order to describe the circulation, water masses must be identified and described.

Multiscale Circulation and Variabilities The new picture of the general circulation in the Mediterranean Sea which is emerging is complex, and composed of three predominant and interacting spatial scales: basin scale (including the thermohaline (vertical) circulation), sub-basin scale, and mesoscale. Complexity and scales arise from the multiple driving forces, from strong topographic and coastal influences, and from internal dynamical processes. There exist: free and boundary currents and jets which bifurcate, meander and grow and shed ring vortices; permanent and recurrent sub-basin scale cyclonic and anticyclonic gyres; and small but energetic mesoscale eddies. As the scales are interacting, aspects of all are necessarily discussed when discussing any individual scale. The path for spreading of Levantine Intermediate Water (LIW) from the region of formation to adjacent seas together with the thermohaline circulations are shown in Figure 2; where the entire Mediterranean is schematically shown as two connected basins (western and eastern). The internal thermohaline cells existing in the western and eastern Mediterranean have interesting analogies and differences to each other and to the global thermohaline circulation. In the western basin (Figure 3A) the basin-scale thermohaline cell is driven by deep water formed in the Gulf of Lions and

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

(A)

Western Mediterranian

711

Eastern Mediterranian Trieste

Venice Genoa

45 Marseille Lions Gulf

Barcelona

Ad

ria

Ligurian

tic

Corsica

Rome

Istanbul

Naples Sardinia Tyrrhenian

n

Albora

Tunis

Algiers

Strait Sic ily of Sicily

Athens

Rhodes Ionian

Tangier

35

n

Gibraltar Strait

ea

Majorca

Lisbon

g Ae

40

of to ait Str Otran

Balearic

Crete Cyprus Cretan Passage Levantine

30 10 (B)

Haifa

Alexandria

5

5

0

10°W

10

20

15

20°E

10°E

0°E

25

35

30

40°E

30°E

?

WMDW 20

0

ADW

1000

40°N

0

100

20

0

GDW

200

200

2000

00

0

200

20

10

00

100

0

30°N 0°E

30°E

20°E

10°E

Figure 1 (A) The Mediterranean Sea geography and nomenclature of the major sub-basins and straits. (B) The bottom topography of the Mediterranean Sea (contour interval is 1000 m) and the locations of the different water mass formations.

Aegean

Adriatic MAW

LIW

Sicily

Gibraltar

Balearic

Ionian

Levantine

Figure 2 Schematic of thermohaline cells and path of Levantine Intermediate Water (LIW) in the entire Mediterranean.

(c) 2011 Elsevier Inc. All Rights Reserved.

712

MEDITERRANEAN SEA CIRCULATION

(A) Sub-basin scale Mesoscale

(B) Sub-basin scale

Mesoscale

Basin scale

Figure 3 Schematic of the scales of circulation variabilities and interactions in (A) western Mediterranean, (B) eastern Mediterranean.

spreading from there. Important sub-basin scale gyres in the main thermocline in the Alboran and Balearic Seas have been identified. Intense mesoscale activity exists and is shown by instabilities along the coastal current, mid-sea eddies and along the outer rim swirl flow of a sub-basin scale gyre. The basin scale thermohaline cell of the eastern basin is depicted generically in Figure 3B and discussed in more detail in the next section. The basin scale general circulation of the main thermocline is composed of dominantly energetic sub-basin scale gyres linked by sub-basin scale jets. The active mesoscale is shown by a field of internal eddies, meanders along the border swirl flow of a sub-basin scale gyre, and as meandering jet segments. The Atlantic Water jet with

its instabilities, bifurcations, and multiple pathways, which travels from Gibraltar to the Levantine is a basin scale feature not depicted in Figure 3; this also pertains to the intermediate water return flow.

Large-scale Circulation Processes of global relevance for ocean climate dynamics include thermohaline circulation, water mass formation and transformation, dispersion, and mixing. These processes are schematically shown in Figure 4A and B for the western and the eastern basins. The Mediterranean basins are evaporation basins (lagoons), with freshwater flux from the

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

(A)

E

713

Q

τ

MAW LIW

LIW

?

AW

EMDW? Sicily

? MW Gibraltar

(B)

Q

EMDW Sicily LIW

Q

E Aegean

Adriatic

τ

AGDW MAW EMDW ?

LIW

?

? ?

LDW

Levantine Ionian

(C ) LIW

CIW MAW

LIW EMDW Adriatic origin

Old mid-depths uplifted waters

Old EMDW Adriatic origin

EMDW Aegean origin

Figure 4 Processes of air–sea interaction, water mass formation, dispersion, and transformation. (A) western Mediterranean, (B) eastern Mediterranean, (C) eastern Mediterranean (post-eastern Mediterranean Transient).

(c) 2011 Elsevier Inc. All Rights Reserved.

714

MEDITERRANEAN SEA CIRCULATION

Atlantic through the Gibraltar Straits and into the eastern Mediterranean through the Sicily Straits. Relatively fresh waters of Atlantic origin circulating in the Mediterranean increase in density because evaporation (E) exceeds precipitation (advective salinity preconditioning), and then form new water masses via convection events driven by intense local cooling (Q) from winter storms. Bottom water is produced: for the western basin (WMDW) in the Gulf of Lions (Figure 4A) and for the eastern basin (Figure 4B) in the southern Adriatic (EMDW, which plunges down through the Otranto Straits). Recent observations also indicate deep water (LDW) formation in the north-eastern Levantine basin during exceptionally cold winters, where intermediate water (LIW) is regularly formed seasonally. Evidence now shows that LIW formation occurs over much of the Levantine basin, but preferentially in the north, probably due to meteorological factors. The LIW is an important water mass which circulates through both the eastern and western basins and contributes predominantly to the efflux from Gibraltar to the Atlantic, mixed with some EMDW and together with WMDW. Additionally, intermediate and deep (but not bottom) waters formed in the Aegean (AGDW) are provided to the eastern basin through its straits. As will be seen below, that water formerly known as AGDW, is now identified as Cretan Intermediate Water (CIW) and Cretan Deep Water (CDW). Important research questions relate to the preconditioning, formation, spreading, dispersion, and mixing of these water masses. These include: sources of forced and internal variabilities; the spectrum and relative amounts of water types formed, recirculating within the Mediterranean basins, and fluxing through the straits, and the actual locations of upwelling. A basin-wide qualitative description of the thermohaline circulation in the western basin of the Mediterranean Sea has recently been provided by Millot (see Further Reading). Results based on cruises in December 1988 and August 1989 indicated that the deep layer in the western Mediterranean was 0.121C warmer and about 0.33 PSU more saline than in 1959. Analysis of these data together with those from earlier cruises has shown a trend of continuously increasing temperatures in recent decades. Based on the consideration of the heat and water budget in the Mediterranean, the deep-water temperature trend was originally speculated to be the result of greenhouse gas-included local warming. A more recent argument considers the anthropomorphic reduction of river water flux into the eastern basin to be the main cause of this warming trend. During the 18th and 19th centuries a number of observations of temperature and salinity were made

mostly in the surface down to intermediate layers, in certain areas of the Mediterranean. Progressively, the investigators extended their measurements into deeper layers to understand the distribution of the parameters, both horizontally and vertically. Sometimes the values were surprisingly close to the correct values but in other cases they presented significant differences due mainly to the primitive instrumentation and methods used by this time. Initially, some believed that the deep waters are motionless. Later the researchers noted the variations of the parameters occurring on a daily, monthly and seasonal basis in the upper layers and began to think about the movements and renewal of the deep waters. Regarding the distribution of salinity extremely confused and erroneous views prevailed up to 1870. However, it was already known by this time that the fresh water input is much lower and does not compensate evaporation from the sea surface. Furthermore, they noted that the source of the salt of the deep water is the surface layer. Several theories on the mechanisms governing the renewal and oxygenation of the deep layers were formulated. Moreover, they succeeded to measure currents and structure primitive maps showing prevailing circulation patterns, as the Atlantic Water inflow and the Black Sea outflow towards the Northeast Aegean. Since the beginning of the twentieth century, when the first investigations in the Mediterranean Sea took place (1908), up to the mid-1980s, both the intermediate and deep conveyor belts of the eastern basin presented rather constant characteristics. The Adriatic has been historically considered as the main contributor to the deep and bottom waters of the Ionian and Levantine basins, thus indicating an almost perfectly repeating cycle in both water mass characteristics and formation rates during this long period. Roether and Schlitzer found in 1991 that the thermohaline circulation in the eastern basin consists of a single coherent convective cell which connects the Levantine and Ionian basins and has a turnover time of 125 years below 1200 m. Their results indicated that the water formed in the Adriatic is a mixture of surface water (AW) and intermediate Levantine water (LIW) from the Mediterranean. The Aegean has also been reported as a possible secondary source, providing dense waters to the lower intermediate and/or deep layers, namely Cretan Intermediate Water (CIW), that affected mainly the adjacent to the Cretan Arc region of the eastern Mediterranean. Since 1946 increased densities were observed in the southern Aegean Sea in 1959–65 and 1970–73. These events occurred under extreme meteorological conditions. However, the quantities of the dense water produced were never enough to

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

affect the whole eastern Mediterranean. The traditional historical picture of water properties is illustrated in Figure 5A by a west–east vertical section of salinity through the eastern Mediterranean. After 1987, the most important changes in the thermohaline circulation and water properties basin-wide ever detected in the Mediterranean 0

(A) 0

78

9 8 7 3 1 0 8 77 77 77 77 76 76 75

occurred. The Aegean, which had only been a minor contributor to the deep waters, became more effective than the Adriatic as a new source of deep and bottom waters of the eastern Mediterranean. This source gradually provided a warmer, more saline, and denser deep-water mass than the previously existing Eastern Mediterranean Deep (and bottom) 6 1 3 6 7 8 9 4 9 0 1 2 75 75 72 72 72 72 72 73 73 74 74 74

_ 500

39.05 39.00 38.95 38.90 38.85 38.80 38.75 38.70 38.67 38.66 38.50 38.00 37.50

_ 1000

Pressure (dbar)

_ 1500 _ 2000 _ 2500 _ 3000 _ 3500 _ 4000 _ 4500 0

200

400

600

800

1000 1200 1400 1600 1800

77

61

60 74 72 76

62

58 57 67

54

50

30

10

12 13

9

34

Distance (km)

(B)

_ 500 39.05 39.00 38.95 38.90 38.85 38.80 38.75 38.70 38.68 38.65 38.50 38.00 37.50

_ 1000

Pressure (dbar)

_ 1500 _ 2000 _ 2500 _ 3000 _ 3500 _ 4000

0

200

400

600

715

800 1000 1200 1400 1600 1800 2000 Distance (km)

Figure 5 West–east vertical sections of salinity through the eastern Mediterranean: (A) 1987, (B) 1995, (C) 1999.

(c) 2011 Elsevier Inc. All Rights Reserved.

716

MEDITERRANEAN SEA CIRCULATION

78 7

(C)

6

5

4

34

3

1

2

_ 500 39.05 39.00 38.95 38.90 38.85 38.82 38.80 38.75 38.70 38.65 38.50 38.00 37.50

Pressure (dbar)

_ 1000 _ 1500 _ 2000 _ 2500 _ 3000 _ 3500 _ 4000 0

200

400

600

800 1000 1200 1400 1600 1800 2000 Distance (km)

Figure 5 Continued

Water (EMDW) of Adriatic origin. Its overall production was estimated for the period 1989–95 at more than 7 Sv, which is three times higher than that of the Adriatic. After 1990, CIW appeared to be formed in the southern Aegean with modified characteristics. This warmer and more saline CIW (less dense than the older one) exits the Aegean mainly through the western Cretan Arc Straits and spreads in the intermediate layers, the so-called LIW horizons, in the major part of the Ionian Sea, blocking the westward route of the LIW. These changes have altered the deep/internal and upper/open conveyor belts of the eastern Mediterranean. This abrupt shift in the Mediterranean ‘ocean climate’ has been named the Eastern Mediterranean Transient (EMT). Several hypotheses have been proposed concerning possible causes of this unique thermohaline event, including: (1) internal redistribution of salt, (2) changes in the local atmospheric forcing combined with long term salinity change, (3) changes in circulation patterns leading to blocking situations concerning the Modified Atlantic Water (MAW) and the LIW, and (4) variations in the fresher water of Black Sea origin input through the Strait of Dardanelles. The production of denser than usual local deep water started in winter 1987, in the Kiklades plateau of the southern Aegean. The combination of continuous salinity increase in the southern Aegean during the period 1987–92, followed by significant temperature drop in 1992 and 1993 caused massive

dense water formation. The overall salinity increase in the Cretan Sea was about 0.1 PSU, due to a persistent period of reduced precipitation over the Aegean and the eastern Mediterranean. This meteorological event might be attributed to larger scale atmospheric variability as the North Atlantic Oscillation. Moreover, the net upper layer (0–200 m) salt transport into the Aegean from the Levantine was increased one to four times within the period 1987–94 due not only to the dry period but also to significant changes of the characteristic water mass pathways. This was a secondary source of salt for the south Aegean that has further preconditioned dense water formation. The second period is characterized by cooling of the deep waters by about 0.351C, related to the exceptionally cold winters of 1992 and 1993. The strongest winter heat loss since 1985 in the Adriatic and since 1979 in the Aegean was observed in 1992. During this winter an almost complete overturning of the water column occurred in the Cretan Sea. The density of the newly formed water, namely Cretan Deep Water (CDW), reached its maximum value in 1994–95 in the Cretan Sea of the southern Aegean. The massive dense water production caused a strong deep outflow through the Cretan Arc Straits towards the Ionian and Levantine basins. Interestingly, the peak of the production rate, about 3 Sv, occurred in 1991–92 when the 29.2 sT isopycnal was raised up to the surface layer. While its deep-water production in the Aegean is becoming more effective with time, that in the Adriatic stopped

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

after 1992. Conditions in 1995 are illustrated in Figure 5B by a west–east vertical section of salinity through the eastern Mediterranean. The period 1995–98 is characterized by continuous decrease of CDW production, from 1 to 0.3 Sv. The level of the CDW at the area of the deep Cretan Arc Straits (i.e. Antikithira in the West and Kassos in the East) is found approximately at the sill depths (800–1000 m). The deep outflow has also been weakened, especially from the western Cretan Straits. Moreover, the density of the outflowing water is no longer sufficient to sink to the bottom and therefore the water coming from the recent Aegean outflow has settled above the old Aegean bottom-water mass, in layers between 1500 and 2500 m. On the other hand, the Aegean continues to contribute the CIW to the intermediate layers of the eastern Mediterranean. Salinity in the eastern Mediterranean in 1999 is shown in Figure 5C. The intrusion of the dense Aegean waters has initiated a series of modifications not only in the hydrology and the dynamics of the entire basin, but also in the chemical structure and some biological parameters of the ecosystem. The dense, highly oxygenated CDW has filled the deep and bottom parts of the eastern Mediterranean, replacing the old EMDW of Adriatic origin, which has been uplifted several hundred meters. This process brought the oxygen-poor, nutrient-rich waters closer to the surface, so that in some regions winter mixing might bring extra nutrients to the euphotic zone, enhancing the biological production. Since 1991, the above mentioned uplifted old EMDW of Adriatic origin has reached shallow enough depths outside the Aegean and especially in the vicinity of the Straits of the Cretan Arc, to intrude the Aegean (Cretan Sea) and compensate its deep outflow (CDW outflow). These waters, namely Transitional Mediterranean Water (TMW), gradually formed a distinct intermediate layer (150–500 m) in the south Aegean, characterized by temperature, salinity and oxygen minima, and nutrient maxima. This has enhanced the previously weak stratification and enriched with nutrients one of the most oligotrophic seas in the world. This new structure prevents winter convection deeper than 250 m. Finally, in 1998–99, the presence of the TMW was much reduced, mainly as a result of mixing. The simultaneous changes in both the upper and deep conveyor belts of the eastern Mediterranean may affect the processes and the water characteristics of the neighboring seas. The contribution of the Aegean to the intermediate and deep layers is still active. The variability in the intermediate waters can alter the preconditioning of dense water formation in the Adriatic as well as in the western Mediterranean.

717

On the other hand, the changes in the deep waters can affect the LIW formation characteristics. Whether the present thermohaline regime will eventually return to its previous state or arrive at a new equilibrium is still an open question.

Sub-basin Scale Circulation Figure 6 shows the patterns of circulation in the western Mediterranean for the various water types. The Atlantic Water in the Alboran Sea flows anticyclonically in the western portion of the western basin, while a more variable pattern occurs in the eastern portion. The vein flowing from Spain to Algeria is named the Almeria-Oran Jet. Further east, the MAW is transported by the Algerian Current, which is relatively narrow (30–50 km) and deep (200–400 m) in the west, but it becomes wider and thinner while progressing eastwards along the Algerian slope till the Channel of Sardinia. Meanders of few tens of kilometers, often ‘coastal eddies’ (Figure 7), are generated due to the unstable character of the current. The cyclonic eddies are relatively superficial and short-lived, while the anticyclones last for weeks or months. The current and its associated mesoscale phenomena can be disturbed by the ‘open sea eddies.’ The buffer zone that is formed by the MAW reservoir in the Algerian Basin disconnects the inflow from the outflow at relatively short timescales typically. Large mesoscale variability characterizes the Channel of Sicily. In the Tyrrhenian Sea both the flow along Sicily and the Italian peninsula and the mesoscale activity in the open sea are the dominant features. The flows of MAW west and east of Corsica join and form the so-called Liguro-Provenco-Catalan Current, which is the ‘Northern Current’ of the Basin along the south-west European coasts. Mesoscale activity is more intense in winter, when this current becomes thicker and narrower than in summer. There is also strong seasonal variability in the mesoscale in the Balearic Sea. Intense barotropic mesoscale eddy activity propagates seaward from the coastline around the sea from winter to spring, and induces a seasonal variability in the open sea. There is evidence that the Winter Intermediate Water (WIW) formed in the Ligurian Sea and the Gulf of Lions can be in larger amounts than that of Western Mediterranean Deep Water (WMDW). Because of their appropriate or shallower depths, these WIW can flow out at Gibraltar with LIW more easily than the WMDW. Furthermore, apart from the LIW there are also other intermediate waters of eastern Mediterranean origin that circulate and participate in the processes of the western

(c) 2011 Elsevier Inc. All Rights Reserved.

718

MEDITERRANEAN SEA CIRCULATION

(A) MAW-WIW : more or less steady paths : mesoscale currents throughout the year : wintertime mesoscale currents : wind-induced mesoscale eddies : the North Balearic Front 40

: 0 m isobath

35

_5

0

5

10

15

(B) LIW-TDW : more or less steady paths : mesoscale currents throughout the year : wintertime mesoscale currents : 0 m and 200 m (thick) isobaths 40

35

_5

0

5

10

15

(C) TDW-WMDW : more or less steady paths : mesoscale currents throughout the year : wintertime mesoscale currents : 0 m and 1000 m (thick) isobaths 40

_5

35 0

5

10

15

Figure 6 Schematics of the circulation of water masses in the western Mediterranean. (A) MAW–WIW; (B) LIW–TDW; (C) TDW– WMDW (Reproduced with permission from Millot, 1999).

topography. After filling the Algero–Provencal Basin up to depths B2000 m, the WMDW intrudes into the deep Tyrrhenian (B3900 m). The amount of unmixed WMDW in the western Mediterranean and especially in the south Tyrrhenian Sea is automatically controlled by the density of the cascading flow from the Channel of Sicily and thus from the dense water formation processes in the eastern Mediterranean. The south Tyrrhenian is a key place for the mixing and transformation of the water masses; the processes within the eastern

Mediterranean play a dominant role in the entire Mediterranean Basin. In the eastern basin energetic sub-basin scale features (jets and gyres) are linked to construct the basin-wide circulation. Important variabilities exist and include: (1) shape, position, and strength of permanent sub-basin gyres and their unstable lobes, multi-centers, mesoscale meanders, and swirls; (2) meander pattern, bifurcation structure, and strength of permanent jets; and (3) occurrence of transient eddies and aperiodic eddies, jets, and filaments.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

719

November 1998

15

17.5

16.25

20

18.75

22.5

21.25

Figure 7 Satellite imagery of sea surface temperature during November 1998 in the Western Mediterranean.

12˚ E

24˚ E

20˚ E

16˚ E

28˚ E

32˚ E

36˚ E

ASW 40˚

38˚ MAW

AIS

ISW 36˚

IA

MIJ

LSW

C

AM

PA

es

od

LS

MAW 34˚

CC

MMJ 32˚

AIS = Atlantic-Ionian Stream MIJ = Mid Ionian Jet MAW = Modified Atlantic Water ISW = Ionian Surface Water

Rh

W

Iérapetra

CC

W s pru Cy

J

MM

Shikmona

MersaMatruh

IA = Ionian Anticyclones MMJ = Mid-Mediterranean Jet ASW = Adriatic Surface Water LSW = Levantine Surface Water

PA = Pelops Anticyclone CC = Cretan Cyclone AMC = Asia Minor Current

Figure 8 Sub-basin scale and mesoscale circulation features in the eastern Mediterranean (Reproduced with permission from Malanotte-Rizzoli et al., 1997 after Robinson and Golnaraghi, 1994).

(c) 2011 Elsevier Inc. All Rights Reserved.

720

MEDITERRANEAN SEA CIRCULATION

Table 1

Upper thermocline circulation features

Feature

Type

ON85

MA86

MA87

AS87

SO91

JA95

S97

ON98

AIS MMC AMC CC Se Lev. Jets Rhodes C West Cyprus C MMA Cretan C Shikmona AC Latakia C Antalya AC Pelops AC Ionian eddies AC Cretan Sea eddies Ierapetra

P P P R T P P P P R R R P T T R

– Y Y Y Y Y Y Y Y Y Y ? – – Y Y

– Y Y N Y Y Y Y ? Y N Y Y – Y N

Y – – Y – – – – – – N – Y – – Y

Y Y Y N Y Y Y Y Y Y Y N Y Y Y Y

Y Y Y – – Y Y Y Y Y – – Y Y – Y

Y Y Y – – Y Y Y – – – – – – Y Y

Y – – – – – – – Y – – – Y Y – –

N Y Y – – Y – Y Y – – – Y N Y Y

Figure 8 shows a conceptual model in which a jet of Atlantic Water enters the eastern basin through the Straits of Sicily, meanders through the interior of the Ionian Sea, which is believed to feed the MidMediterranean Jet, and continues to flow through the central Levantine all the way to the shores of Israel. In the Levantine basin, this Mid-Mediterranean Jet bifurcates, one branch flows towards Cyprus and then northward to feed the Asia Minor Current, and a second branch separates, flows eastward, and then turns southward. Important sub-basin features include: the Rhodes cyclonic gyre, the Mersa Matruh anticyclonic gyre, and the south-eastern Levantine system of anticyclonic eddies, among which is the recurrent Shikmona eddy south of Cyprus. The diameter of the gyres is generally between 200 and 350 km. Flow in the upper thermocline is in the order of 10–20 cm s 1. A tabulation of circulation features in the eastern Mediterranean and their characteristics is presented in Table 1. Figure 7 shows the upper-thermocline main circulation features and surface waters’ pathways. Figure 4 presents the thermohaline (intermediate and deep) circulation, which has a significant vertical component. Finally, Figure 5 presents the vertical structure of the water masses in three different periods in order to follow/show the continuous transformation of the water mass structure and characteristics in the recent 13 years, that is the period of the Eastern Mediterranean Transient. During the period 1991–95, a large three-lobe anticyclonic feature developed in the south-western Levantine (from the eastern end of the Cretan Passage, 261E up to 311E), blocking the free westward LIW flow, from the Levantine to the Ionian, and causing a recirculation of the LIW within the west

Levantine Basin. Although, multiple, coherent anticyclonic eddies were also quite common in the area before 1991 (as the Ierapetra and Mersa-Matruh), the 1991–95 pattern differs significantly, with three anticyclones of relatively larger size covering the entire area. This feature seems to comprise the Mersa-Matruh and the Ierapetra Anticyclone. Moreover, the 1998–99 infrared SST images (Figure 9) indicated that the area was still occupied by large anticyclonic structures. The data sets collected in late 1998 and early 1999 indicated that this circulation pattern had been reversed to cyclonic, confirming the transient nature of these eddies. Consequently, the Atlantic Ionian Stream (AIS) was not flowing from Sicily towards the northern Ionian, but directly eastwards crossing the central Ionian towards the Cretan Passage (Table 2). The seasonal variability of the circulation of the late 1980s in the south Aegean Sea has been replaced by a rather constant pattern in the period of the EMT (1991–98). Therefore, the Cretan Sea eddies were in a seasonal evolution in the 1980s (always present), while in the 1990s there was a constant succession of three main eddies (one cyclone in the west, one anticyclone in the central region and again one cyclone in the east) that presented spatial variability.

Mesoscale Circulation The horizontal scale of mesoscale eddies is generally related to, but somewhat larger than, the Rossby radius of deformation. In the Mediterranean the internal radius is O(10–14) km or four times smaller than the typical values for much of the world ocean. The study of mesoscale instabilities, meandering, and eddying thus requires a very fine resolution sampling. For this

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

721

November 1998

18.75

20

21.25

22.5

23.75

25

Figure 9 Satellite imagery of sea surface temperature during November 1998 in the Eastern Mediterranean.

Table 2

Mediterranean water masses

Water mass name

Acronym

Aegean Deep Water Adriatic Water Cretan Deep Water Cretan Intermediate Water Eastern Mediterranean Deep Water Eastern Mediterranean Transient Levantine Deep Water Levantine Intermediate Water Modified Atlantic Water Transitional Mediterranean Water Winter Intermediate Water Western Mediterranean Deep Water

AGDW ASW CDW CIW EMDW EMT LDW LIW MAW TMW WIW WMDW

reason, only recently different mesoscale features were found in both the western and eastern basins including the mesoscale variabilities associated with the coastal currents in the western basin and open sea mesoscale energetic eddies in the Levantine basin. In the western basin, intense mesoscale phenomena (Figure 6) have been detected using satellite information and current measurements. Mesoscale activity occurs as instabilities along the coastal currents (i.e., the Algerian Current) leading to the formation of mesoscale eddies which can eventually move across the basin or interact with the current itself. Along the Algerian Current cyclonic and anticyclonic eddies develop and evolve over several months as they slowly drift eastward (a few kilometers per day). The anticyclonic eddies generally

increase in size and detach from the coast. Some may drift near the continental slope of Sardinia, where a well-defined flow of LIW exists. Here they are able to pull fragments of LIW seaward. Old offshore eddies extend deep in the water column and last from several months to as much as a year. They sometimes enter the coastal regions and interact with the Algerian Current. In the coastal zones the mesoscale currents appear to be strongly sheared in the vertical. This clearly indicates that eddies can modify the circulation over a relatively wide area and for relatively long periods of time. The coastal eddies along the Algerian coast can be especially vigorous, inducing currents of 20–30 cm s 1 strength for periods of a few weeks. More complicated variations of the currents have also been measured at 300 m and sometimes at 1000 m. Mesoscale activity has been observed in the northern basin (i.e., along the western and northern Corsican Currents). Coastal Corsican eddies are typically anticyclonic and located either offshore or along the coast of Corsica. A number of experiments were conducted to investigate the mesoscale phenomena in the Ligurian Sea. The results of a 1-year current meter array are shown for the southern coastal zone in the Corsican Channel (Figure 10D). Mesoscale currents are characterized by permanent occurrence and by a baroclinic structure with relatively large amplitude at the surface, moderate at the intermediate level and still noticeable at depth, thus indicating large vertical shear of the horizontal currents.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

(A)

(B) B

A

C

15.5

15.0

15.5

150

225

375 14.0

450

15.0 14.5

15.5

14.5

14.0

300

.5

14.5

14

_ 15.0 14.5

14.5

15.0

375

225 14.5

300

16.0

0 16. 5.5 1

150

16.5

75

0.91

75

D

14. 0

722

450 (D)

(C)

B 0

0.0

C

D

0.06

A

250 200 150 100 50 0 _ 50 _ 100 _ 150 150 100 50 0 _ 50 _ 100 50 0 _ 50

250 200 150 100 50 0 _ 50 _ 100 _ 150 150 100 50 0 _ 50 _ 100 50 0 _ 50

12 1 1 1 1 1 1 1 1 1 1 1 1 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul 81 81 81 81 81 81 82 82 82 82 82 82 82

Figure 10 (A) Mesoscale temperature cross-section from XBTs in AS87 POEM cruise along section ABCD. (B) Filtered temperature cross-section using a pyramid filter with 50 km influential radius. (C) Location of the cross-section superimposed on the dynamic height anomaly from AS87 survey (excluding XBTs). (D) Velocities from a current meter array in the Corsican Channel.

Dedicated high-resolution sampling in the Levantine basin led to the discovery of open ocean mesoscale energetic eddies, as well as jets and filaments. This was confirmed by a mesoscale experiment in August–September 1987 in the eastern basin. Mesoscale eddies dynamically interacting with the general circulation occur with diameters in the order of 40–80 km. From this analysis a notable energetic sub-basin/mesoscale interaction in the Levantine basin and a remarkable thermostad in the Ionian have been revealed. Figure 10A shows a temperature cross-section from XBT profiles collected in summer 1987 across

the Mid-Mediterranean Jet, West Cyprus Gyre, MMJ, and the northern border of the Shikmona eddy (section ABCD shown in Figure 10C). Figure 10B shows the identical XBT section after a pyramid filter, with horizontal influential distance of 50 km. The filter has removed very small scale features while maintaining the mesoscale structure.

Modeling Vigorous research in the 1980s and the developing picture of the multiscale Mediterranean circulation

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

723

50.0 Figure 11 Velocity field for the eastern Mediterranean at 10 m depth from an eddy-resolving primitive equation dynamical model.

were accompanied by a new era of numerical modeling on all scales. Modeling efforts included: water mass models, general circulation models, and data assimilative models. Dynamics in the models include: primitive equations, non-hydrostatic formulations, and quasi-geostrophy. The assimilation of the cooperative eastern Mediterranean surveys of the 1980s and 1990s into dynamical models played a significant role in the identification of sub-basin scale features. The numerical model results shown in Figure 11 depict the existence of numerous sub-basin scale features, as schematized in Figure 7. In recent years numerical modeling of the general circulation of the Mediterranean Sea has advanced greatly. Increased computer power has allowed the design of eddy resolving models with grid spacing of one-eighth and one-sixteenth of a degree for the whole basin and higher for parts of it. An example output from such a model, forced with perpetual year atmospheric forcing, which includes the seasonal cycle but not interannual variability, is shown in Figure 12. Many of these models incorporate sophisticated atmospheric forcing parameterizations (e.g., interactive schemes) which successfully mimic existing feedback mechanisms between the atmosphere and the ocean. Studies have been carried out using perpetual year atmospheric forcing, mostly aimed at studying the seasonal cycle, as well as interannual atmospheric forcing. Studies have mainly focused on reproducing and understanding the

seasonal cycle, the deep- and intermediate-water formation processes and the interannual variability of the Mediterranean. They have shown the existence of a strong response of the Mediterranean Sea to seasonal and interannual atmospheric forcing. Both seasonal and interannual variability of the Mediterranean seems to occur on the sub-basin gyre scale. The Ionian and eastern Levantine areas are found to be more prone to interannual changes than the rest of the Mediterranean. Sensitivity experiments to atmospheric forcing show that large anomalies in winter wind events can shift the time of occurrence of the seasonal cycle. This introduces the concept of a ‘memory’ of the system which ‘preconditions’ the sea at timescales of the order of one season to 1 year. In the deep- and intermediate-water formation studies, the use of high frequency (6 h) atmospheric forcing (in contrast to previously used monthly forcing) in correctly reproducing the observed convection depths and formation rates was found to be crucial. This shows the intermittent and often violent nature of the convention process, which is linked to a series of specific storm events that occur during each winter rather than to a gradual and continuous cooling over winter. The use of high-resolution numerical models in both the western and the eastern Mediterranean allowed the study of the role of baroclinic eddies, which are formed at the periphery of the chimney within the cyclonic gyre by instabilities

(c) 2011 Elsevier Inc. All Rights Reserved.

724

MEDITERRANEAN SEA CIRCULATION

46 44 42 40 38 34 32 30

_5

16

0 5 30 10 15 20 25 Temperature at depth = 10 m, Year = 2, Month = 6, Day = 15, Min = 15.9703, Max = 23.2415

17

18

19

20

21

22

35

23

Figure 12 Temperature and superimposed velocity vectors at 10 m depth from a numerical simulation of the entire Mediterranean Sea.

of the meandering rim current, in open ocean convection. These eddies were shown to advect buoyancy horizontally towards the center of the chimney, thus reducing the effectiveness of the atmospheric cooling in producing a deep convected mixed layer. These results are in agreement with previous theoretical and laboratory work. The LIW layer which extends over the whole Mediterranean was found to play an important role both in the western (Gulf of Lions) and the eastern (Adriatic) deep-water formation sites and more specifically in ‘preconditioning’ the formation process. It was shown that the existence of this layer greatly influences the depth of the winter convection penetration in these areas. This is related to the fact that the LIW layer with its high salt content decreases the density contrast at intermediate layers, thus allowing convection to penetrate deeper. This result shows the existence of teleconnections and inter-dependencies between sub-basins of the Mediterranean. A number of numerical models have been developed to simulate and understand the origins and the evolution of the Eastern Mediterranean Transient. These models indicate that the observed changes can be at least partially explained as a response to variability in atmospheric forcing. Sensitivity experiments, in which the observed precipitation anomaly of 1989–90 and 1992–93 was not included, did not reproduce properly the EMT. This confirms that this factor was a significant contributor to the occurrence and evolution of the Eastern Mediterranean Transient, since it acted as a

‘preconditioner’ to the latter by importantly increasing the salinity in the area. The enhanced deep water production in the Aegean has implied a deposition of salt in the deep and bottom layers with a simultaneous decrease higher up. As the turnover rate for waters below 1200 m has been estimated to exceed 100 years, this extra salt will take many decades to return into the upper waters. Its return, however, might well induce changes in the thermohaline circulation, considering the dependence of the two potential sources of deep water on the salinity preconditioning. It will therefore take many decades before the eastern Mediterranean returns to a new quasi-steady state. An interesting question in this connection is whether the system will recover its previous mode of operation with a single source of deep-water production in the Adriatic or evolve into an entirely different, perhaps even an unanticipated direction.

Conclusion The Mediterranean Sea is now known to have a complex thermohaline, wind, and water flux-driven multi-scale circulation with interactive variabilities. Recent vigorous research, both experimental and modeling, has led to this interesting and complex picture. However, the complete story has not yet been told. We must wait to see the story unfold and see how many states of the circulation exist, what changes occur and whether or not conditions repeat.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEDITERRANEAN SEA CIRCULATION

See also Coastal Circulation Models. Current Systems in the Mediterranean Sea. Data Assimilation in Models. Deep Convection. Elemental Distribution: Overview. Forward Problem in Numerical Models. Heat and Momentum Fluxes at the Sea Surface. Meddies and Sub-Surface Eddies. Mesoscale Eddies. Ocean Circulation. Open Ocean Convection. Regional and Shelf Sea Models. Upper Ocean Time and Space Variability. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Angel MV and Smith R (eds.) (1999) Insights into the hydrodynamics and biogeochemistry of the South Aegean Sea, Eastern Mediterranean: The PELAGOS (EU) PROJECT. Progress in Oceanography 44 (special issue): 1--699. Briand F (ed.) (2000) CIESM Workshop Series no.10, The Eastern Mediterranean Climatic Transient: its Origin, Evolution and Impact on the Ecosystem. Monaco: CIESM. Chu PC and Gascard JC (eds.) (1991) Elsevier Oceanography Series, Deep Convection and Deep Water Formation in the Oceans. Elsevier. Lascaratos A, Roether W, Nittis K, and Klein B (1999) Recent changes in deep water formation and spreading in the Eastern Mediterranean Sea. Progress in Oceanography 44: 5--36. Malanotte-Rizzoli P (ed.) (1996) Elsevier Oceanography Series, Modern Approaches to Data Assimilation in Ocean Modeling. Malanotte-Rizzoli P and Eremeev VN (eds.) (1999) The Eastern Mediterranean as a Laboratory Basin for the Assessment of Contrasting Ecosystems, NATO Science Series – Environmental Sercurity vol. 51, Dordrecht: Kluwer Academic.

725

Malanotte-Rizzoli P, Manca BB, Ribera d’Acala M, et al. (1999) The Eastern Mediterranean in the 80s and in the 90s: The big transition in the intermediate and deep circulations. Dynamics of Atmospheres and Oceans 29: 365–395 Millot C (1999) Circulation in the Western Mediterranean Sea. Journal of Marine Systems 20: 423--442. Nielsen JN (1912) Hydrography of the Mediterranean and Adjacent Waters. In: Report of the Danish Oceanographic Expedition 1908–1910 to the Mediterranean and Adjacent Waters, 1, Copenhagen, pp. 72–191. Pinardi N and Roether W (ed.) Mediterranean Eddy Resolving Modelling and InterDisciplinary Studies (MERMAIDS). Journal of Marine Systems 18: 1–3. POEM group (1992) General circulation of the Eastern Mediterranean. Earth Sciences Review 32: 285–308. Robinson AR and Brink KH (eds.) (1998) The Sea: The Global Coastal Ocean, Regional Studies and Syntheses, vol. 11. New York: John Wiley and Sons. Robinson AR and Golnaraghi M (1994) The physical and dynamical oceanography of the Mediterranean Sea. In: Malanotee-Rizzoli P and Robinson AR (eds.) Proceedings of a NATO-ASI, Ocean Processes in Climate Dynamics: Global and Mediterranean Examples, pp. 255--306. Dordrecht: Kluwer Academic. Robinson AR and Malanott-Rizzoli P (eds.) Physical Oceanography of the Eastern Mediterranean Sea, Deep Sea Research, vol. 40(6) (Special Issue), Oxford: Pergamon Press. Roether W, Manca B, and Klein B, et al. (1996) Recent changes in the Eastern Mediterranean deep waters. Science 271: 333--335. Theocharis A and Kontoyiannis H (1999) Interannual variability of the circulation and hydrography in the eastern Mediterranean (1986–1995). In: MalanotteRizzoli P and Eremeev VN (eds.) NATO Science Series – Environmental Security vol. 51, The Eastern Mediterranean as a Laboratory Basin for the Assessment of Contrasting Ecosystems, pp. 453--464. Dordrecht: Kluwer Academic.

(c) 2011 Elsevier Inc. All Rights Reserved.

MEIOBENTHOS B. C. Coull, University of South Carolina, Columbia, SC, USA G. T. Chandler, University of South Carolina, Columbia, SC, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1705–1711, & 2001, Elsevier Ltd.

Introduction Meiobenthos live in all aquatic environments. They are important for the remineralization of organic matter, and they are crucial members of marine food chains. These small (less than 1 mm) invertebrates have representatives from 20 metazoan (multicellular) phyla and three protistan (unicellular) phyla. With their ubiquitous distribution in nature, high abundances (millions per square meter), intimate association with sediments, rapid reproduction and rapid life histories, the meiobenthos have also emerged as valuable sentinels of pollution.

Definitions and Included Taxa Meio (Greek, pronounced ‘myo’) means smaller, thus meiobenthos are the smaller benthos. They are smaller than the more visually obvious macrobenthos (e.g., segmented worms, echinoderms, clams, snails, etc.). Conversely, they are larger than the microbenthos – a term restricted primarily to Protista, unicellular algae, and bacteria. Meiofauna are small invertebrate animals that live in or on sediments, or on structures attached to substrates in aquatic environments. Meiobenthos (benthos ¼ bottom living) refers specifically to those meiofauna that live on or in sediments. Meiofauna is the more encompassing word. By size, meiofauna are traditionally defined as invertebrates less than 1 mm in size and able to be retained on sieve meshes of 31–64 mm. Nineteen of the 34 multicellular animal phyla (Table 1) and three protistan (unicellular) phyla, i.e., Foraminifera, Rhizopoda, and Ciliophora, have meiofaunal representatives. Of these multicellular (metazoan) phyla, some are always meiofaunal insize (permanent meiofauna), whereas others are meiofaunal in size only during the early part of their life (temporary meiofauna) (Table 2). These are the larvae and/or juveniles of macrobenthic species (e.g., Annelida, Mollusca, Echinodermata). Members of the phylum Nematoda are the most abundant

726

meiofaunal organisms, and copepods (Arthropoda, Crustacea) or Foraminifera are typically second in abundance worldwide. Representative meiofauna taxa are illustrated schematically in Figure 1. The books listed under Further Reading by Higgins and Theil and by Giere, and any invertebrate zoology text, should allow one to identify field-collected meiofauna to phylum. Identification to the family, genus, and species level requires specialized literature. Good places to start are chapters on specific phyla in the two texts listed, and also the International Association of Meiobenthologists website: http://www.mtsu.edu/meio

Where Do Meiofauna Live? Meiofauna occupy a variety of habitats from highaltitude lakes to the deepest ocean depths. In fresh water they occur in beaches, wetlands, streams, rivers, and even the bottoms of our deepest lakes. In marine habitats they occur from the intertidal splash zone to the deepest trenches. Wherever one looks in the aquatic environment, meiofauna are likely to be found. This holds true even in heavily polluted or anoxic sediments where the only living multicellular species are often a few meiofaunal taxa. Sediment Habitats

Sediments, from the softest muds to the coarsest shell gravels and cobbles, harbor abundant meiofauna. Meiofauna associated with sediments live ‘on’ or ‘in’ the sediment. Those living on top of the sediment are epifaunal (or epibenthic) and are adapted to moving over sediment surfaces. Those living ‘in’ the sediment may burrow into the sediment (burrowing meiofauna), displacing sediment particles as they move, or they may move in the interstices between sediment grains and be called interstitial meiofauna (see Table 2). The interstitial fauna are restricted to sediments where there is sufficient space to move between the particles; typically sands and gravels. Sediments where the median particle diameter is below 125 mm provide little room for meiofauna to move between particles, and thus are inhabited by burrowing and epibenthic taxa. In those taxa having both interstitial and burrowing representatives (e.g., Nematoda, Copepoda, Turbellaria), there are often stark differences in the morphologies of the mud dwellers and sand dwellers. The sand fauna tend to be slender, since they must maneuver through narrow

(c) 2011 Elsevier Inc. All Rights Reserved.

MEIOBENTHOS

727

Table 1 A list of meiobenthic taxa (Phyla of the Kingdom Animalia). Currently, 19 phyla (bold) from the 34 recognized phyla of the Kingdom Animalia have meiofaunal representatives. Of these 19 phyla, only five are exclusively meiofaunal (bold italics) Phyla

Porifera Placozoa Cnidaria Ctenophora Plathelminthes Orthonectida Rhombozoa Cycliophora Acanthocephala Nemertea Nematomorpha Gnathostomulida Kinorhyncha Loricifera Nematoda Rotifera Gastrotricha Entoprocta Priapulida Pogonophora Echiura Sipuncula Annelida Arthropoda Tardigrada Onychophora Mollusca Phoronida Bryozoa Echinodermata Echinodermata Chaetognatha Hemichordata Chordata

Free-living

Symbiotic

Marine

Freshwater

Terrestrial

Yes Endemic Yes Endemic Yes No No No No Yes No Endemic Endemic Endemic Yes Yes Yes Yes Endemic Endemic Endemic Yes Yes Yes Yes No Yes Endemic Yes Endemic Endemic Endemic Endemic Yes

Yes No Yes No Yes No No No No Yes No No No No Yes Yes Yes Yes No No No No Yes Yes Yes No Yes No Yes No No No No Yes

No No No No Yes No No No No Yes No No No No Yes Yes No No No No No Yes Yes Yes Yes Endemic Yes No No No No No No Yes

No No Yes No Yes Endemic (marine) Endemic (marine) Endemic (marine) Endemic Yes Endemic No No No Yes Yes No Yes No No No No Yes Yes No No Yes No No No No No No Yes

(Modified from RP Higgins, unpublished, with permission.)

Table 2

Types of the meiofauna

Permanent meiofauna: always meiofaunal size Interstitial: moves between sediment particles Burrowing: displaces sediment particles Epibenthic On sediment surfaces On plants or animals Temporary meiofauna: meiofaunal size in early life only Larvae or juveniles of macrofauna: mostly bivalve molluscs and polychaete worms

interstitial openings, whereas the mud fauna are not restricted to a particular morphology and are generally larger. Since sandy habitats often occur in areas with high wave and tidal action, most

interstitial fauna have adhesive glands for attaching to sand grains so that they will not be washed away. They also tend to have a low number of eggs because their reduced body size cannot support large egg masses. Other Habitats

Meiofauna also occupy several ‘above sediment’ habitats, including rooted aquatic vegetation, moss, algae, sea ice, and various animal structures such as coral crevices, worm tubes, and echinoderm spines. Still other meiofauna are symbionts living commensally in animal tubes. Those meiofaunal assemblages living above the bottom, for example, in or on fouling communities, or on various animal structures, differ from sediment dwellers by having

(c) 2011 Elsevier Inc. All Rights Reserved.

728

MEIOBENTHOS

(D)

(C)

(B) (A)

(F)

(G)

(E)

(H) (I) (J)

(K)

(L)

(M)

(O) (N) Figure 1 Schematic diagrams of representative meiofaunal animals: (A) Annelida, Polychaeta; (B) Foraminifera; (C) Crustacea, Ostracoda; (D) Priapulida; (E) Crustacea, Copepoda; (F) Loricifera; (G) Nematoda; (H) Rotifera; (I) Kinorhyncha; (J) Plathelminthes, Turbellaria; (K) Mollusca, Gastropoda; (L) Gastrotricha; (M) Annelida, Oligochaeta; (N) Arthropoda, Halacaroidea; (O) Tardigrada. The animals are not drawn to scale. (Modified from Higgins and Thiel (1988).)

species composition and adaptive morphologies specific to particular epibenthic habitats.

Collection and Extraction of Meiofauna Qualitative sampling of meiofauna will not allow estimation of abundance per unit area, but it is useful for a general assessment of faunal richness or to accumulate one or several species for experimental work. Qualitative samples of sediment are taken by scooping sediment arbitrarily with some device (shovel, hands, grab sampler, dredge), whereas qualitative samples of meiofauna living on structures

are taken by collecting the structure itself. Such samples can be sieved live at the collection site, be taken to a laboratory for extraction of the fauna by physical or chemical means, or be preserved in their entirety for future examination. Quantitative meiofauna sampling requires that the sampling area be accurately known. For sediments, this typically involves pushing a core tube of known diameter into the sediment to a preselected depth, collecting all the sediment within the core, and ultimately counting all the fauna in the known area or volume. Quantitative samples are typically preserved in formaldehyde or alcohol and subsequently counted and identified under a microscope. They are often stained with a protein stain (e.g., Rose Bengal) to

(c) 2011 Elsevier Inc. All Rights Reserved.

MEIOBENTHOS

help distinguish the animals from surrounding sediment and organic debris. Meiofaunal abundance values are preferably expressed as number per 10 cm2, but also as number per m2. There are multiple ways of extracting meiofauna from sediments and surfaces. For live qualitative sediment samples, many species will be attracted to a focused directional light source (preferably cold fiberoptic light so as not to heat the sediments unduly) if sieved sediments are spread in a thin layer with a centimeter or so of overlying water. Sieved sediment can also be put into funnels, where established salinity and/or heat gradients will drive the fauna down the funnel and into a collecting dish. For animals clinging to surfaces, chemical relaxants – or fresh water for marine samples – will cause some fauna to release their purchase and be washed into overlying water where they can be collected onto sieves of appropriate size. For preserved quantitative samples, meiofauna can be separated from the sediments by decantation (swirling the sediment in a container and pouring off the less dense animals after the mineral particles have settled), elutriation (where water is passed through a sample continuously so that sediment is kept in suspension and the lighter animals come off with the flow), or by centrifugation in a density gradient solution so that the sediment (or debris) remains in one layer and the animals in another. All the products of extractions are sieved through a fine mesh (32–100 mm depending on the objective) and the portion retained on the mesh is observed, counted and identified under a microscope.

Distribution of Meiofauna

729

of sailing vessels, and dispersal by suspension in the water column. On a local scale, meiofaunal dispersal is either a passive process of mechanical removal due to current scour or one in which the animals actively migrate to the water column. Animals occupying the sediment surface are obviously scoured much more easily than those living deeper in the sediment. The abundance of eroded species in the water column at any given time is a function of the magnitude of local current velocity and sediment erodability. Large-scale Spatial Distribution

Meiofauna are rarely evenly distributed on, or in, a substrate. On the large scale (meters to kilometers) gradients in physical factors (e.g., salinity, tidal exposure, sediment grain size, oxygen concentrations) are primarily responsible for variances in abundance, whereas on smaller (centimeter) scales both physical and biological factors have been reported as important. Large-scale gradients lead to zonation of the fauna. For example, certain meiofauna species are confined to specific areas along salinity gradients in estuaries, across intertidal sandy and muddy habitats, and across the water depth gradient in lakes and in the ocean. With water depth, faunal changes are primarily a function of food availability (e.g., organic content of sediment), sediment type, temperature, and oxygen availability. Interestingly, the meiofauna at similar ocean depths are usually similar to each other all over the world. The same families and/or genera comprise a significant portion of the fauna at similar depths except in the Mediterranean and the Arctic, where many of the ‘deep sea’ genera also occur into shallower (o500 m) depths.

Geographic Distribution

Small-scale Spatial Distribution

Meiofauna inhabit some of the most dynamic environments imaginable (such as exposed high-energy shores) and these animals have traditionally been considered sedentary. Emphasis has centered on adaptations for remaining in close proximity to the substratum, particularly because pelagic larvae are almost nonexistent in the permanent meiofauna. Development, morphology, and biology all seem designed to ensure that the organism remains in or on the substratum. On the basis of such observations, one would expect limited worldwide distribution patterns for species. However, numerous species (identified by morphology, not by molecular genetic technologies) appear to be cosmopolitan. Plate tectonics has been invoked as a potential mechanism to describe pan-oceanic and worldwide meiofaunal distributions, as have dispersal via birds, rafting on drifting materials, transport in the ballast

Meiobenthos also exhibit spatial variation (patchiness) on a small (millimeter to centimeter) scale. A variety of factors have been suggested for the observed small-scale patchiness including: (1) microspatial variation in physical factors (oxygen, grain size); (2) food distribution; (3) physical structures in the habitat (worm tubes, algae, mud balls, etc.); (4) predation/disturbance, where a predator eats one patch of animals but not another; (5) interspecific competition, where species segregate themselves spatially to avoid competition for a resource; and (6) aggregations, where individuals come together for mating. While we presently lack a framework for experimentally testing how these factors effect microspatial distribution in the field, we know that species are aggregated more often than not. Smallscale zonation also takes place vertically in sediment. Here the vertical distribution of the fauna is

(c) 2011 Elsevier Inc. All Rights Reserved.

730

MEIOBENTHOS

controlled primarily by the level of oxygen in the sediment layers. Most meiofauna require oxygen to survive, but certain adapted species can tolerate low oxygen or no oxygen. Species living in such sediments that can tolerate hydrogen sulfide, a known animal toxin, are called the ‘sulfide fauna’ or the thiobios. Copepods are typically the meiobenthic taxon most sensitive to decreased oxygen, and generally are confined to oxic sediments. Gnathostomulida primarily live in mild sulfidic and low oxic sediments, as do some Nematoda, Turbellaria, Ciliophora, Gastrotricha and Oligochaeta (see Table 1). While oxygen content is the ultimate factor controlling most meiofaunal vertical distribution, desiccation can also be important, particularly in intertidal marine beaches. As sand dries at low tide, the fauna face desiccation stress regardless of the oxygen content. Meiofauna therefore migrate downward on an ebbing tide and upward on a flooding tide, and this happens more at midday in the summer, when drying is greatest, than at midnight in the winter.

Abundance and Diversity of Meiofauna On the average there are a million meiofaunal organisms per square meter of sediment surface, with a dry weight biomass of 0.75–2 g m2 in shallow (o100 m) waters. Highest abundance values come from intertidal muddy estuarine habitats (6–12 million per m2), lowest values from the deep sea (hundreds to thousands per m2). In general, sediment grain size is the primary factor affecting the abundance and species composition of meiofaunal organisms within a given depth range. Different species occur in muddy versus sandy versus phytal habitats. In areas where temperature varies seasonally, meiobenthos abundance and species composition also vary seasonally. Typically, maximum abundances occur in the warmer months of the year, but individual species may reach maximum abundance at other times. Year-to-year variability in abundance also can be greater than within-year seasonal variability. The highest known species diversity for a meiofaunal assemblage has been recorded for copepods from algal holdfast communities. Shallow-water algal frond assemblages and deep-sea sediments also yield high species diversities. Even though meiofaunal abundance in the deep sea is greatly reduced compared to shallow sediments, there are many different and exotic species. In shallow-water sedimentary habitats, meiofaunal diversity appears

similar worldwide, with ecologically equivalent species in different geographic regions. These communities usually have four to ten predominant species. While the database is limited and there are always difficulties interpreting diversity data, there appears to be a standard diversity range for most shallow-water meiofaunal assemblages. There is no evidence that meiofaunal species diversity increases toward the tropics. Pollution or other disturbances, such as hypoxia/anoxia, tend to decrease diversity.

Functional Role of Meiofauna Meiofauna appear to have two major functional roles in aquatic ecosystems: to serve as food for organisms higher in the food web, and to facilitate mineralization of organic material and enhance nutrient regeneration. In addition, because they exhibit high sensitivity and rapid response to anthropogenic disturbance, they are excellent sentinels of pollution. Food for Higher Trophic Levels

Meiofauna are very important nutritionally to a variety of animals that could not survive without them. Many predators go though an obligatory meiofaunal feeding stage, and copepods appear to be the major meiofauna prey item for most of these predators. These copepods primarily live in muddy sediment or on plants. Thus most predation on meiofauna takes place in muddy substrates or in areas with substantial sea grass or macroalgae. In muds, the meiofauna prey are restricted to the upper few millimeters or centimeters of oxidized sediment. Thus bottom-feeding predators only need to take a shallow bite to obtain abundant food. On aquatic plants, fish predation on meiofauna is analogous to birds eating insects on a tree. Over 90 species of juvenile fish are known to eat meiofauna, making them the major meiofaunal predators. Other predators are shrimp (prawns) and some bottom-feeding birds. Mineralization and Nutrient Regeneration

Meiofauna are important in stimulating bacterial growth, which then enhances remineralization (the conversion of organic nitrogen, phosphorus, and carbon to their inorganic forms). Meiofauna package organic molecules and, because of their relatively short life span (months) and high metabolic rate, this packaged material is returned to the system rapidly (compare, for example, the carbon tied up in a clam that lives for 2–5 years). Meiofaunal nutrients then become part of the well-known microbial loop in which they are utilized by bacteria and can be converted into dissolved organic carbon for use by

(c) 2011 Elsevier Inc. All Rights Reserved.

MEIOBENTHOS

higher trophic levels and/or remineralized for primary producers. Meiofauna typically have less than 20% of the standing biomass of the larger, more visible, macrofauna, but they turn over as much or more carbon per year. These processes are important in all kinds of habitats, but they are probably most important in those sediments with high amounts of organic matter, i.e., muds.

731

effectively assessed than macrobenthos for toxicantinduced effects since meiofauna spend their entire life cycle in sediments and are not reliant on recruitment of a planktonic larval stage. After years of neglect, meiofauna are becoming more popular subjects of pollution studies.

See also Meiofauna and Pollution Sediments are the ultimate repository for most of the persistent pollutants released to the ecosphere. Upon entering aquatic environments, most toxicants associate with dissolved organics, suspended silts, clays, and organic particulates and eventually accumulate in sediments. Meiofauna, of course, are intimately associated with this muddy-sediment geochemical soup, as they spend their entire life cycle there and have limited ability to leave. Because meiofauna reproduce very rapidly (often in 2–4 weeks), pollution effects on meiofaunal populations can be detected quickly and early in the history of contamination of a site. There have been three general approaches to using meiofauna to assess pollution: field studies, laboratory studies, and studies using replicas of the controlled natural environment (microcosm/mesocosm studies). In field studies, samples are typically collected from a polluted site and from a reference site, and differences in community (or genetic) structure between the sites are assessed. Laboratory studies usually examine the lethal effects (e.g., how many individuals die after exposure to specific dose levels of a contaminant) or sublethal effects (e.g., changes in egg production, embryonic development time, hatching success, or genetic diversity of contaminants singly or in mixture. Meiobenthic community responses to pollutants in micro/mesocosms are measurable and reproducible over reasonable time and spatial scales (owing to small organism size and rapid production/turnover), and are more

Benthic Foraminifera. Benthic Organisms Overview. Carbon Sequestration via Direct Injection into the Ocean. Deep-Sea Fauna. Fish Feeding and Foraging. Macrobenthos. Microbial Loops. Microphytobenthos. Pollution: Effects on Marine Communities. Salt Marshes and Mud Flats. Sandy Beaches, Biology of. Sea Ice. Sea Ice: Overview.

Further Reading Coull BC and Chandler GT (1992) Pollution and meiofauna: field, laboratory and mesocosm studies. Oceanography and Marine Biology Annual Reviews 30: 191--271. Giere O (1993) Meiobenthology: The Microscopic Fauna in Aquatic Sediments. Berlin: Springer-Verlag. Heip C, Vincx M, and Vranken G (1985) The ecology of marine nematodes. Oceanography and Marine Biology Annual Reviews 23: 399--489. Hicks GRF and Coull BC (1983) The ecology of marine meiobenthic harpacticoid copepods. Oceanography and Marine Biology Annual Reviews 21: 67--175. Higgins RP and Thiel H (eds.) (1988) Introduction to the Study of Meiofauna. Washington, DC: Smithsonian Institution Press. International Association of Meiobenthologists web site: http://www.mtsu.edu/meio McIntyre AD (1969) Ecology of the marine meiobenthos. Biological Reviews of the Cambridge Philosophical Society 44: 245--290. Swedmark B (1964) The interstitial fauna of marine sand. Biological Reviews of the Cambridge Philosophical Society 39: 1--42.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE J. E. Petersen, Oberlin College, Oberlin, OH, USA W. M. Kemp, University of Maryland Center for Environmental Science, Cambridge, MD, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: Experimental Ecosystems as Tools for Aquatic Research Within the last few decades there has been a clear trend within ecological science of growing reliance on manipulative experiments as a means of testing ecological theory. Many approaches are available for experimentation. An important distinction can be drawn between field and laboratory-based experiments. In field experiments, either parts of nature or whole, naturally bounded ecosystems are manipulated in place while similar areas are left as controls. In laboratory experiments, organisms, communities, and the physical substrate are transported to controlled facilities. A second distinction can be drawn between experiments in which organisms and materials freely exchange between the experiment and surrounding environment and those in which organisms and materials are enclosed and isolated either in a laboratory setting or with physical boundaries imposed in the field. The term ‘enclosed experimental ecosystem’ is used when the goal of an enclosure experiment, conducted in either laboratory or field conditions, is to explore interactions among organisms or between organisms and their chemical and physical environment. Because enclosed experimental ecosystems are intended to serve as miniaturized worlds for studying ecological processes, they are often called ‘microcosms’ or ‘mesocosms’. Enclosed experimental ecosystems have become widely used research tools in oceanographic and freshwater sciences because they allow for a relatively high degree of experimental control and replication necessary for hypothesis testing while still capturing dynamics that emerge from ecosystemlevel interactions between organisms and their physical and chemical environments. They provide a bridge between observational field studies and process-oriented lab research. Mesocosms have been

732

used to conduct experiments on a broad range of aquatic habitats. Over the last 30 years, enclosed experimental ecosystems have become important tools in both coastal and open ocean contexts to address critical research questions in the fields of chemical and physical oceanography, ecotoxicology, fisheries science, and basic and applied ecology (Figure 1). Two fundamental objectives of ecological experiments are to achieve high levels of control and realism. Control refers to the ability to relate cause and effect, to manipulate, to replicate, and to repeat experiments; realism is a measure of the degree to which results accurately mimic the dynamics of particular natural ecosystems. Trade-offs between control and realism are inherent in different experimental approaches; experiments conducted within nature tend to maximize realism, whereas physiological experiments in the laboratory allow for the highest degree of experimental control. In theory, mesocosms provide intermediate levels of both control and realism (Figure 2).

Scale Is a Crucial Issue in Mesocosm Research Scale is a crucial issue for all ocean scientists and has particular implications for researchers using enclosed experimental ecosystems. How can largescale processes be simulated and incorporated into enclosed experimental ecosystems so as to maximize realism? How can research findings be quantitatively extrapolated from small, often simplified experimental ecosystems up to whole natural ecosystems? For that matter, how can information gleaned from research in one type of ecosystem be extrapolated to other natural ecosystems that differ in scale? Recent research indicates that scale effects can be parsed into ‘fundamental effects’, that are evident in both natural and experimental ecosystems, and ‘artifacts of enclosure’, that are solely attributable to the artificial environment in mesocosms. A key objective of this contribution to the encyclopedia is to review the ways in which mesocosm experiments have been used to study the marine environment and to suggest ways in which scaling considerations can be used to improve the use of mesocosm’s research tools.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

733

Figure 1 Enclosed experimental ecosystems provide a means of conducting controlled, replicated experiments to reveal processes and interactions that occur within different marine habitats. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

History and Applications There is a rich history in the use of enclosed experimental ecosystems. The initial concept of microcosms, as hierarchically nested miniature worlds contained successively within larger worlds, has been credited to early Greek philosophers including Aristotle. Although it is difficult to date the initial scientific uses of enclosed experimental ecosystems, small glass jars and other containers were routinely used as experimental

ecosystems by the middle of the twentieth century. H.T. Odum and his colleagues were pioneers and proponents of the use of mesocosms to study aquatic ecosystems. They constructed a wide variety of experimental ecosystems including laboratory streams, containers with planktonic and vascular plant communities, and shallow outdoor ponds containing oysters and/or seagrasses. Although the word ‘microcosm’ was used initially to describe virtually all experimental ecosystems, the term ‘mesocosm’ was later adopted to

(c) 2011 Elsevier Inc. All Rights Reserved.

734

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Whole ecosystem Open plots

m

is al

Time

Re

Caged plots

Laboratory mesocosms

n

io

at

ul

s

el

od

Field mesocosms

m

m

Si

l

tro

n Co

Laboratory experiments Space

Figure 2 As the scale of experiments increases from simple laboratory experiments to complex whole-ecosystem manipulations, greater realism is possible, but control over experimental conditions declines. Simulation models can be used to synthesize and integrate results from all types of studies. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

distinguish larger experimental units from smaller bench-top laboratory systems. Some have suggested that experimental manipulations of whole aquatic ecosystems in nature are always preferable to mesocosm studies. However, the characteristically steep spatial gradients, threedimensional water exchanges, lack of boundaries, and natural variability make such whole ecosystem manipulations extremely difficult to accomplish in coastal and open ocean environments, leaving mesocosms as critical tools for controlled experimentation. A series of books devoted to aspects of experimental aquatic ecosystems mark recent progress with this research approach. There are diverse styles and applications of enclosed experimental ecosystems (Figure 3). During the last four decades, experimental microcosms and mesocosms have been developed in a diversity of sizes, shapes, and habitats to address a broad range of research questions. Small (B0.5 l) laboratory chemostat flasks have been widely used by R. Margalef and others to study plankton community dynamics, while large (30–1300 m3) plastic bag enclosures have been deployed in situ by J. Gamble, G. Grice, D. Menzel, T. Parsons, M. Reeve, J. Steele, J. Strickland, and others to study pelagic (in some cases including benthic) coastal ecosystems in Europe and North America. Similarly, mesocosm shapes vary from the tall and relatively narrow (23 m high  9.5 m deep) in situ plankton bags used by Gamble, Steele, and their colleagues to broad (350 m2 surface), shallow

(1 m deep) estuarine ponds used by R. Twilley and others. Mesocosms have been constructed to study diverse marine habitats, including planktonic regions of oceans and estuaries, deep benthos, shallow tidal ponds, coral reefs, salt marshes, and seagrasses. Composition and organization of experimental ecological communities range broadly and include: simple ‘gnotobiotic’ ecosystems where all species are selected and identified; interconnected microcosms, each containing a different trophic level; intact ‘undisturbed’ columns of sediment and overlying water extracted and contained; and tidal ponds with ‘selforganizing’ communities developed by seeding with diverse inoculant communities taken from different natural ecosystems. Marine mesocosms have been used effectively to address a range of theoretical and applied scientific questions. Early studies using in situ bag enclosures (e.g., Controlled Ecosystem Pollution Experiment (CEPEX), Loch Ewe Enclosures, Kiel Plankton Towers) examined planktonic food web responses to nutrient enrichment and introduction of toxic contaminants (e.g., copper, mercury). These experiments were designed to assess the effects of both ‘bottomup’ (resource-limited) and ‘top-down’ (herbivoreand predator-determined) controls (e.g., Figures 3(a) and 4). Although these studies were very instructive, difficulties in controlling mixing regimes and lack of replication of treatments tended to limit interpretation of results. Later studies, notably the land-based Marine Ecosystem Research Laboratory (MERL), employed mechanical mixing, added intact sediments, and increased replication (Figure 3(b)). MERL systems were used by S. Nixon, C. Oviatt, and their colleagues to investigate trophic and biogeochemical responses to similar treatments including N, P, and Si enrichment, crude oil contamination, filter-feeding, and storm mixing events. The versatile and permanent MERL facility allowed investigators to explore interactions between pelagic and benthic communities that are critical in the dynamics of shallow coastal ecosystems (e.g., Figure 5).

The Challenges and Opportunities of Scale in Mesocosm Research Two parallel trends in ecology during the last 20 years have been an increased use of mesocosms as research tools (Figure 6(a)) and an increased recognition of the importance of scale as a determinant of the patterns and processes observed in natural ecosystems (Figure 6(b)). As we have discussed, mesocosms have become widely used and accepted tools in ocean science because they provide a means of conducting

(c) 2011 Elsevier Inc. All Rights Reserved.

Figure 3 Marine mesocosm facilities have taken diverse forms including (a) Controlled Ecosystem Pollution Experiment (CEPEX, 1300 m3, 17 m deep, 3 m diameter) system in Saanich Inlet, British Columbia, 1978; (b) Marine Ecosystem Research Laboratory (MERL, 13 m3, 5 m deep, 1.8 m diameter) experimental ecosystems established in 1980; (c) rocky littoral mesocosms (23 m3, 4.7 m long, 3.6 m wide, 1.3 m deep) at Solbergstrand, Norway; (d) Plankton community mesocosms (55 l, 0.77 m deep, 0.30 m diameter) with Neuse Estuary water from University of North Carolina. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

(c) 2011 Elsevier Inc. All Rights Reserved.

735

736

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

(a) 7 Phytoplankton chlorophyll a 6

Concentration (µg l–1)

5

4

Hg + nutrients Nutrient additions

3

2 Control 1

Mercury added

Nutrients

0 (b)

70 Zooplankton (calanoid copepods) 60 Nutrients

Abundance (103 m–3)

50

40 Control

30

20 Hg + nutrients 10

0 0

10

20

30

40 50 Sampling day

60

70

80

Figure 4 Example results of in situ mesocosm experiments (CEPEX) designed to investigate ‘top-down’ (predator) and ‘bottom-up’ (nutrient) controls on phytoplankton. Inorganic nutrients were added (on days 25, 37, and 53) to two of three mesocosms to stimulate primary productivity (a). Mercury was added to one of these mesocosms (on day 9) to reduce zooplankton abundance (b). Although the experiments incorporated no replication, the findings contributed to our understanding of the importance of top-down control. Redrawn from Grice GD, Reeve MR, Koeller P, and Menzel DW (1977) The use of large volume, transparent, enclosed sea-surface water columns in the study of stress on plankton ecosystems. Helgolander Wissenschaftliche Meeresuntersuchungen 30: 118–133.

ecosystem-level experiments under replicated, controlled, and repeatable conditions. The focus on scale can be traced to a number of factors including: theoretical and technological advances that increase our understanding of causal linkages between local, regional, and global phenomena; a growing awareness of human impact at all scales; and the formalization of scale as a legitimate topic of inquiry within the emerging field of landscape ecology. This emphasis on scale is evidenced by the steady increase in the

number of journal articles listing ‘scale’ as a keyword (Figure 6(b)) and in the publication of a number of new books devoted to scaling theory. It has long been recognized that scale is an inherent design problem that may confound the interpretation of results from experimental ecosystem studies. Since their use first became prevalent in the 1970s, researchers have expressed concerns regarding scaling problems associated with mesocosms including the effects of: reduced system size and short timescale of

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Ecosystem metabolism (gO2 m–2)

800

737

Ecosystem production and respiration

600

400

Respiration

200 Production

0 0.5

1

5

10

50

10

104

105

103 Polychaetes

102

104 Total biomass

103 0.1

0.5

1

5

10

Macro-infauna biomass (m mol N m–2)

Polychaetes (# m–2)

0.1 6

10 50

NH4 input (m mol m–3 d–1) Figure 5 Example results of land-based mesocosm experiments (MERL) examining plankton–benthic responses to different levels of nutrient enrichment. Total productivity and total system respiration both respond positively to enrichment. However, the relative importance of polychaetes worms and macro-infauna changes across the gradient. Redrawn from Nixon SW, Pilson MEQ, and Oviatt CA (1984) Eutrophication of a coastal marine ecosystem – an experimental study using the MERL microcosms. In: Fasham MJR (eds.) Flows of Energy and Materials in Marine Ecosystems: Theory and Practice, pp. 105–135. New York: Plenum.

experiments, reduced ecological complexity, wall growth, limitations on animal movements, distorted mixing regimes, and unrealistic water exchange rates. A few investigators have used a simple idea of mesocosm calibration, where key properties are adjusted in experimental systems to mimic conditions in the natural environment. However, the majority of early mesocosm studies skirted the question of scaling and the problem of extrapolation altogether. By the end of the 1980s, it was clear that further progress in the application of experimental ecosystem methods to aquatic science would require focused quantitative study of how scale affects behavior in natural and experimental ecosystems and how

experimental ecosystems might be better designed to account for scale. The development of systematic techniques for extrapolating results from small experimental ecosystem studies to conditions in nature at large remains an active area of research. Recent research (e.g., at the Multiscale Experimental Ecosystem Research Center (MEERC)) has focused on developing quantitative and systematic approaches for the design and interpretation of experimental ecosystem research with a particular focus on the problem of scale. Several scaling concerns must be addressed when using mesocosm results to predict effects in natural aquatic ecosystems. The first and most obvious is

(c) 2011 Elsevier Inc. All Rights Reserved.

738

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

(a) David Kimmel

Experimental studies

3

2

Field

1

Mesocosm 0 1975

1980

1985

1990

1995

2000

2005

Tracey Saxby

Published papers (%)

4

Year (b) 14

Scale-related studies Ecotron

10 8 6 4 Terrestrial 2 0 1975

Aquatic 1980

1985

1990 Year

1995

2000

2005

Katharina Engelhardt

Published papers (%)

12

Figure 6 (a) Trends in use of field experiments and mesocosms in ecological studies as revealed from keyword searches in ecological journals. Note that field experiments and mesocosm studies are not mutually exclusive categories because the latter can be used in the field. The patterns suggest an increasing reliance on both categories of experimentation. (b) Trends in scale studies in ecology based on separate searches conducted by year for the term ‘scale’ in keywords and abstracts of journals emphasizing terrestrial research (Ecology, Oikos, Oecologia) and journals publishing only aquatic research (Limnology and Oceanography, Marine Ecology Progress Series). The number of papers identified in each year was then standardized to the total number of papers published for that year in those journals and expressed in the graph as a percent. Adapted from Petersen JE, Cornwell, JC, and Kemp WM (1999) Implicit scaling in the design of experimental aquatic ecosystems. Oikos 85: 3–18.

that experimental systems are constrained in size and duration. An extensive literature review revealed a median experimental duration of 49 days and median volume of 1.7 m3; aquatic mesocosm experiments are brief and small relative to the natural scales that characterize many important ecological processes of interest. A second problem is the presence of walls, which restrict biological, material, and energy exchange with the outside world and provide a substrate for growth of undesirable but potentially influential organisms on this artificial edge habitat. A third problem is that a host of experimental design decisions – such as how many replicates to include per treatment and whether to control light, mixing, and other properties – tend to vary together with choices of size, duration, and ecological complexity

(Figures 7 and 8). Finally, the relative importance of the air–water area, sediment–water area, and wall area, in relation to each other and to water and sediment volume, changes with physical dimensions. Unfortunately, parallel scaling problems also exist for field experiments. For example, replication tends to decrease with increasing plot size and experimental lakes and field plots tend to be orders of magnitude smaller than the natural systems for which inferences are drawn. An analysis of aquatic studies conducted in cylindrical planktonic–benthic mesocosms reveals that in designing experimental ecosystems, researchers gravitate toward a depth/radius ratio of approximately 4.5 (Figure 9). As a consequence of this bias, in general, larger mesocosms are simultaneously less

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

739

60

Frequency observed (%)

50 40 30 20 10

Small (10 m3)

Figure 7 Relationship between mesocosm size and the presence of various design characteristics in a quantitative review of the mesocosm literature. Size categories (small, medium, or large, in cubic meters) are indicated in the legend. The y-axis represents the percentage of articles in a given size class for which the design characteristic indicated is present. The overall percentage of experiments for which a given characteristic is present is indicated in parentheses within the key. ‘Defined community’ indicates that individual populations were selectively added to create the mesocosm community. Adapted from Petersen JE, Cornwell JC, and Kemp WM (1999) Implicit scaling in the design of experimental aquatic ecosystems. Oikos 85: 3–18.

104 *

Volume (m3)

102

100

10–2

*

10–4 1 2 3 4 5 Number of replicates per treatment Figure 8 Plot of mesocosm volume vs. number of replicates per treatment. Median values are represented by the bar within a box, and the 75th and 25th percentiles (i.e., the interquartile range) by the top and bottom of the box. The ends of the ‘whiskers’ represent the farthest data point within a span that extends 1.5 times the interquartile range from the 75th and 25th percentiles. Data outside this span are graphed with asterisks. Adapted from Petersen JE, Cornwell JC, and Kemp WM (1999) Implicit scaling in the design of experimental aquatic ecosystems. Oikos 85: 3–18.

influenced by wall artifacts, have less sediment area per unit volume, and have less surface area available for gas and light exchange per unit volume than do smaller systems (Figures 9(b) and 9(c)). Collectively, these scaling attributes can potentially confound interpretation, comparison, and extrapolation of findings from mesocosm experiments. One might conclude from the preceding figures and discussion that reductions, artifacts, co-variation, and distortions in scale pose an almost insurmountable obstacle to designing mesocosm studies to examine oceanic processes. Alternatively, these problems can be viewed as interesting research opportunities to advance our theoretical and practical understanding of the ‘science of scale’. A variety of mesocosm scaling experiments have been designed to shed light on two classes of effects: ‘fundamental effects of scale’ evident in both natural and experimental ecosystems (e.g., the effects of water depth), and ‘artifacts of enclosure’ attributable to the artificial environment in experimental ecosystems (e.g., the effects of wall growth). In these experiments, ecological responses are measured in relation to manipulations in experimental scales

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

(a)

(c)

103

Wall area / volume (m–1)

740

102

Scaling options Constant radius (r = c2) Constant depth (z = c1) Constant shape (r /z = C3)

10

100 10–1 10–2 10–4

(b)

100 102 Volume (m3)

104

10–2

100 102 3 Volume (m )

104

(d) 103

102

Horizontal area / v olume (m–1)

103

Depth (m)

10–2

R 2 = 0.59 p < 0.001 z /r = 4.48

10

100 10–1 10–2 10–3

10–2

10–1 100 Radius (m)

10

102

102 10

100 10–1 10–2 10–4

Figure 9 (a) Available options for conserving characteristic length relationships as the size of a cylindrical mesocosm is increased. (b) Relations between depth and radius for the cylindrical mesocosms in the ecological literature. Dots are physical dimension data from a comprehensive literature review of experiments conducted in mesocosms. (c) Surface areas of the vertical walls vs. volume. (d) Surface area of bottom and top vs. mesocosm volume. Dotted (green) lines represent scaling for constant depth and are placed at values corresponding with median depth. Dashed (red) lines that represent scaling for constant radius are placed at median radius. The solid (blue) lines represent scaling for constant shape and are derived from linear regression of radius (r) vs. depth (z), with statistics provided in (b). A clear implicit bias is evident toward scaling for constant shape. Adapted from Petersen JE, Cornwell JC, and Kemp WM (1999) Implicit scaling in the design of experimental aquatic ecosystems. Oikos 85: 3–18.

(i.e., time, space, and complexity) for a variety of coastal habitat types. Such studies suggest that it is possible to improve substantially the design of experimental ecosystems. Selected examples are discussed in the sections that follow.

Effective Design of Enclosed Experimental Ecosystems There are many issues and questions that must be considered in the design of enclosed experimental ecosystems. Design decisions are important because they affect how results can be interpreted and extrapolated to nature. Optimal design is determined by the research question under consideration. The processes, organisms, and habitats associated with

this question determine the appropriate size, shape, duration, and complexity for the experimental ecosystem. Even within a given ecosystem type, there is no single best design that will suit all research goals. Typically, the choices made will reflect a balance between three competing objectives: (1) control (the ability to relate cause and effect, to manipulate, to replicate, and to repeat experiments); (2) realism (the degree to which results accurately mimic nature); and (3) generality (the breadth of different systems to which results are applicable). There are, however, specific tools and guidelines available to aid in the experiment design process for enhancing the probability of research success. The sections below provide guidance on critical issues that must be considered.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Design Choices: Degree of Abstraction

Experimental ecosystems are a type of model. Models are, by definition, simplifications and abstractions of the reality that we hope to represent with them. As modelers, researchers select a level of abstraction that is appropriate to their research question, and the choices made have direct bearing on trade-offs between control, realism, and scale. One can distinguish between ‘generic’ and ‘ecosystem-specific’ models, which represent the two extremes in this trade-off. Generic mesocosms are used to test broad theories that potentially apply to many different kinds of ecosystems. These systems tend to be small, highly artificial, have minimal physical and biological complexity, and are not designed to represent particular natural ecosystems. This is the ecological analog of using a worm as a model for studying human physiology or behavior. In ecology, generic mesocosms have been successfully applied to address research questions pertaining to ecosystem development, predator–prey relations, stress, system stability, and species diversity. Because precise correspondence with particular ecosystems is not an objective, the researcher has considerable design flexibility in constructing generic models. The downside is that extrapolation from simplified, abstract systems to particular natural ecosystems is challenging. Ecosystem-specific mesocosms are used to test hypotheses linked to particular types of ecosystems. This is the ecological analog of using chimpanzees to study human physiology and behavior. To achieve the higher degree of realism required, these systems must incorporate the essential physical and biological features that control the dynamics in the systems that they represent. Various ecosystemspecific models have been constructed, ranging from coral reefs to coastal plain estuaries. As the desired degree of specificity and desired level of realism increase, so does the complexity of engineering necessary to achieve realistic ecological conditions (e.g., Figure 10). Design Choices: Physical Characteristics

In addition to questions related to appropriate degree of abstraction and ecosystem type, researchers face crucial design questions regarding the physical characteristics of the experimental ecosystem. For example, what are the minimum system size, experimental duration, and ecological complexity necessary to answer the research question? How will the experimental system address each of the following design decisions: light source, mixing, temperature, exchange of water and constituents, inclusion of

Experimental food chain

Natural food chain

741

Natural food web

Increasing trophic complexity

Figure 10 Experimental ecosystems are typically simplified relative to nature in terms of biodiversity and trophic (feeding) complexity. Inclusion of higher trophic levels (increased trophic depth) or more species diversity at each trophic level (increased trophic breadth) is not always feasible or desirable. Predators at high trophic levels are often large and may not exhibit normal behavior in small enclosures. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

sediments, and organism source and introduction mode? We provide a list (Table 1) of some of the key variables associated with these questions that must be considered, the design decisions associated with these variables, and the ecological properties that are potentially affected by these design decisions. Choices related to physical characteristics are obviously also dependent on resources available in terms of funds, time, equipment, and support personnel. Design Choices: Mixing and Exchange

Mixing and exchange of water and associated constituents are particularly important factors to consider in the design of enclosed experimental ecosystems. A core objective of mesocosm experiments is to isolate biological, chemical, and physical conditions to facilitate controlled manipulative experiments. This act of isolation can, however, create conditions within the mesocosm that are very different from those in nature, thereby distorting the dynamics observed in these experiments. Exchange can be defined as the net transport of water and its constituents through a system. Mixing can be defined as the physical movement of the water and its constituents within the system, generating turbulence within the fluid and homogenization of the constituents. Mixing and exchange are important aspects of natural marine ecosystems from the largest to the smallest of scales. Depending on how the system of interest is defined, mixing at one scale can sometimes be considered exchange at another scale.

(c) 2011 Elsevier Inc. All Rights Reserved.

742

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Table 1

Key variables to consider in the design of experimental ecosystems

Variable

Design decisions

Properties affected

Size

Volume, depth, radius, surface area

Time

Experimental duration, timing of perturbation, sampling frequency

Mixing

Vertical and horizontal mixing environment, mechanical mixing apparatus employed

Materials exchange Light

Frequency, magnitude, variability chemical composition, biological composition Natural or artificial, intensity, spectral properties

Walls

Construction materials, whether to clean, cleaning frequency Whether to control, how to control

Relative dominance of pelagic, benthic, and emergent producer communities, wall growth, temperature oscillations Ecological dynamics and life cycle of organism included in experiment, ability to detect seasonal and long-term effects, influence of experimental artifacts Pelagic–benthic interactions, feeding rates and behavior, access to nutrients, artifacts, and potential mortality associated with mechanical devices Recolonization rates, flushing of planktonic organisms, selection for particular organisms and communities Primary productivity, producer community composition, water temperature Relative dominance of wall growth, light environment

Temperature Ecological complexity Sediments

Species and functional group diversity, number of habitats and biogeochemical environments included From nature or synthesized, intact or homogenized, particle size, organic matter content, organisms included

Rate of biogeochemical activity, selection for particular organisms Primary productivity, trophic dynamics, biogeochemical pathways Pelagic–benthic interactions, vascular plant growth, primary productivity

Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

For example, mixing of surface and bottom waters can be thought of as an exchange that delivers nutrient-rich bottom water to the surface. Mesocosms need to be designed to either include or simulate the variety and magnitude of exchange and mixing that occur in the natural ecosystems that they are designed to represent. At intermediate (meso-)scales, mixing and exchange are crucial in estuaries and coastal waters where fresh and saltwater interact. Exchange and mixing of water are intricately linked processes that determine the estuary’s flushing rate, and in so doing they play a major role in its biological productivity and its susceptibility to pollution effects. At very small scales, microscopic organisms are influenced by relative motion of the fluid (shear) that is directly related to mixing intensity. Small-scale mixing renews nutrient and food supplies, affects contact between predators and prey, and may be a source of physical stress at high levels (Table 2). Mesocosm experiments indicate that mixing intensity can have a negative effect on copepod abundance and a highly negative effect on gelatinous zooplankton (Figure 11). At the interfaces between water and fixed solid surfaces, boundary layers (regions of reduced mixing) are formed due to effects of friction. Experimental

ecosystems will generally require special mixing mechanisms to minimize boundary layers at their walls and mimic natural boundary layers near the sediment surface (benthic boundary layers). A variety of engineering approaches can be taken to mix water in mesocosms. Spinning paddles and discs, mechanical plungers, bubbling, and water pumping have all been used as approaches to generating mixing in the water column (Figure 12). A range of techniques can be used to characterize the mesocosm mixing environment, including current meters and acoustic Doppler current profilers, as well as measurements of dye dispersion and gypsum dissolution. Scale models can be developed and used to explore mixing characteristics before full-scale experimental ecosystems are built. A range of investigations in various mesocosm systems (e.g., CEPEX, Loch Ewe, MERL, MEERC) have demonstrated the physical and ecological effects of alternative mixing regimes. The goal of these studies is to characterize the mixing environment within the water column and the mixing and flow environment across the bottom so that key mixing parameters (e.g., turbulent energy dissipation, vertical mixing rate) can be matched to natural conditions. Mesocosm researchers should familiarize themselves with the mixing literature as it relates to the design of these systems.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Table 2 Empirically determined effects of mixing phytoplankton, zooplankton, and ecosystem processes

Phytoplankton Settling rate Cell size Cell abundance Chlorophyll a Cell growth Diatom/flagellate Species composition Nutrient uptake Timing of bloom Microzooplankton (protozoa) Predation/grazing rate Growth rate (numbers) Cell size Macrozooplankton (copepods) Abundance/biomass Metabolic rate Excretion rate Predation/grazing rate Growth rate Development rate Age structure Sex ratio Ecosystem Community productivity Ecosystem productivity Ecosystem R Nutrient dynamics

on

Relationshipa

() (þ) ( þ ) or (0) (þ) ( þ ) or (  ) (þ) (O) ( þ ) or (  ) (O)

Zooplankton abundance

Variable

743

Acartia tonsa

( þ ), (  ), or (0) (þ) () (  ) or ( þ ) (þ) (þ) ( þ ) or (  ) (þ) (þ) (O) (O) ( þ ), (  ), or (0) (þ) (þ) (O)

Moerisia lyonsi

Turbulent energy Figure 11 Relationships between the abundance of Moerisia lyonsi and Acartia tonsa, and the turbulent energy dissipation rate (E) in the three mixing treatments. Turbulent energy dissipation is one of a number of important parameters that can be used to match conditions in nature and mesocosms. Data from Petersen JE, Sanford LP, and Kemp WM (1998) Coastal plankton responses to turbulent mixing in experimental ecosystems. Marine Ecology Progress Series 171: 23–41.

Liquid

surface

H = liquid height a

( þ ) symbol indicates a positive relationship between the variable and turbulence, (  ) indicates a negative relationship, (O) indicates the presence of a relationship, (0) indicates no relationship. Because mixing levels used in individual experiments included in this analysis ranged from no mixing to unrealistically high levels atypical of nature, this table can only be considered a rough summary of findings. Citations to studies in this analysis are included in Petersen JE, Sanford LP, and Kemp WM (1998) Coastal plankton responses to turbulent mixing in experimental ecosystems. Marine Ecology Progress Series 171: 23–41.

D = impeller diameter W = impeller blade width = 1/4 to 1/6 D

D = 1/4 to 1/2 T

C = impeller clearance = 1/6 to 1/2 T T = tank diameter

The rate at which water is exchanged with surrounding ecosystems is a physical feature that controls many important processes in marine systems. Indeed, the relatively high rate of primary and secondary productivity typical of coastal ecosystems is often attributed to large material exchange resulting from their position at the interface between the watershed and open ocean. Although exchange incorporates both temporal and spatial scale, it is often convenient to express water exchange in terms of ‘residence time’ (i.e., time required for incoming water to replace the entire volume of the basin or container), or alternatively as ‘exchange rate’ (i.e., (residence time)  1). Residence time is an important scaling factor to consider in natural and experimental ecosystems

Figure 12 Typical water flow patterns generated in a mesocosm provided with a single rotating axial impeller. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

because it determines whether a system is dominated by internal or external processes. The actual time a substance or organism resides in the system depends on the combination of flow rate and the rate of reaction, growth, or death inside the system. Flowthrough ‘chemostat’ experiments are commonly used to study phytoplankton growth; however, few ecosystem-level studies have attempted to simulate exchange rates that characterize specific natural

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Planktonic ecosystem

Submersed plant system

Gross primary productivity

Epiphyte biomass Epiphyte biomass (g m–2)

6

4

2

0 40

3 2 1 0

Zooplankton biomass

30

Plant growth (cm m–2 d–1)

Relative change

Gross primary productivity (g O2 m–3 d–1)

744

20 10

25

Submersed aquatic plant growth

20

Low exchange High exchange

15 10 5 0

0 Low nutrient

High nutrient

Low nutrient

High nutrient

Figure 13 Effects of water exchange rate and nutrient concentration of inflowing waters on gross primary productivity and zooplankton biomass in planktonic experimental ecosystems (left panels) and on competition between aquatic grasses and epiphytes growing on plant leaves (right panels). Values presented are experimental means 7SE. Adapted from Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer.

ecosystems, and fewer still have explicitly assessed the effects of different exchange rates on ecological dynamics. The studies that have been conducted indicate that variations in water exchange rate can have substantial effects on ecological dynamics observed in both planktonic and seagrass mesocosms (Figure 13). The specific impacts of exchange rates are regulated by the nature of the constituents being exchanged with the water, and the organisms present within the system. Depending on the actual conditions and the organisms involved, variations in water exchange sometimes have counteracting effects. For example, exchange can deliver nutrients or other resources to a system and at the same time flush out mobile organisms that might utilize those resources. The effects of exchange are distinct for systems dominated by planktonic primary producers from those that are dominated by stationary producers. It is also important to recognize that variability in exchange rates can be as important in controlling ecological dynamics as the mean rates of exchange. The various effects of exchange must be taken into careful consideration in the design of experimental ecosystems. Scaling Considerations in Design and Extrapolation

Even in the case of ‘ecosystem-specific’ mesocosms that are designed to match precisely certain natural

habitats (see section above on abstraction), experimental systems will generally be far smaller than the natural ecosystem that they are intended to represent. Scaling theory suggests that certain patterns and processes only become evident as system size and duration are increased beyond thresholds. Furthermore, scaling responses are often nonlinear and unique for specific variables. Thus, for example, patterns determined to be scale-dependent in mesocosm experiments may become scale-independent at the larger scales of natural systems (solid line in Figure 14). Likewise, relationships seen as scale-independent in mesocosms may change with scale in larger natural ecosystems (dashed line in Figure 14). Finally, it is possible that thresholds exist over which small changes in scale result in dramatic and possibly discontinuous changes in ecological dynamics. Given these possibilities, special attention is necessary to account for the potential scale dependence of observations made in mesocosms. Spatial scaling relationships such as those established between water depth and both phytoplankton primary productivity and zooplankton biomass (Figure 15) provide a basis for quantitative extrapolation. Although less information is available, it is clear that temporal as well as spatial dynamics can also profoundly affect experimental outcomes (Figure 16). In most cases, experimental interpretations and conclusions must be qualified with the acknowledgement that precise effects of scale are yet to be known.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

745

Possible scaling functions D

A

Mesocosms

B C

0 order

1st Small

Large Spatial or temporal scale

Figure 14 Hypothetical responses of two distinct ecological properties to changes in the scales over which they are observed. Mesocosms scales (shaded region of graph) are inherently smaller than the scales of most natural systems. Trajectories shown indicate how different properties may be affected differently by changes in scale. From Kemp WM, Petersen JE, and Gardner RH (2001) Scale-dependence and the problem of extrapolation: Implications for experimental and natural coastal ecosystems. In: Gardner RH, Kemp WM, Kennedy VS, and Petersen JE (eds.) Scaling Relations in Experimental Ecology, pp. 3–57. New York: Columbia University Press.

The evidence that we have presented thus far implies that mesocosms are inherently distorted representations of nature. A key question then is, can we somehow compensate for these distortions in the design and interpretation of experiments? The term ‘dimensional analysis’ encompasses a variety of techniques that are based on the proposition that universal relationships should apply regardless of the dimensions of a particular system under investigation. In general, the technique involves developing dimensionless relationships that capture the balance between processes or forces governing the dynamics of a particular system. Dimensional analysis provides a potentially valuable tool for designing experimental ecosystems so that they retain key features of nature. For example, spatially patchy distributions of resources and predators in natural ecosystems may be simulated in mesocosms by creating an exchange regime that is pulsed over time. Similarly, the effects of patchy schools of plankton-eating fish on plankton community dynamics can be simulated experimentally with periodic additions and then removal of fish from the tank. In these cases, temporal variability is substituted for spatial heterogeneity, and the dimensional properties conserved in the mesocosm study are both the duration and frequency of contact between organisms, resources, and predators. Simulation models provide an additional tool that can be used to improve both the design and interpretation of mesocosm research. Given the importance of spatial heterogeneity in controlling ecological

Gross primary productivity (g O2 m−3 d−1)

1st order

E

Mesocosms B 6 E C A

4

Chesapeake tributary

2 D

Chesapeake Mainstem

0 Mesocosms Zooplankton biomass (g C m−3)

Ecological property

0

B

2

E C

1

A D

Chesapeake tributary

Chesapeake mainstem

0 0

2

4 6 Depth (m)

8

Figure 15 Variations in primary productivity and depth with changes in water column depth for five experimental and two natural estuarine ecosystems with similar salinity. Experimental ecosystems have five different sizes or shapes and the estuarine sites are in the mainstream and a tributary of Chesapeake Bay. Data for gross primary productivity (GPP per unit water volume) are mean values measured from changes in dissolved oxygen concentration. Data are from Petersen J, Chen C-C, and Kemp WM (1997) Scaling aquatic primary productivity: Experiments under nutrient- and light-limited conditions. Ecology 78: 2326–2338. From Kemp WM, Petersen JE, and Gardner RH (2001) Scale-dependence and the problem of extrapolation: Implications for experimental and natural coastal ecosystems. In: Gardner RH, Kemp WM, Kennedy VS, and Petersen JE (eds.) Scaling Relations in Experimental Ecology, pp. 3–57. New York: Columbia University Press.

dynamics, coupling mesocosms with spatially explicit dynamic simulation models may become an increasingly valuable approach to ecological research. In this approach, mesocosms can be thought of as individual cells (grain) within a heterogeneous matrix of different habitats that cover broad spatial extent (Figure 1). Likewise, models can be used to explore effects of temporal variability that are difficult to incorporate in the design of mesocosm studies. Numerical models offer an excellent tool for exploring nonlinear feedback effects at scales that are larger than individual mesocosms.

(c) 2011 Elsevier Inc. All Rights Reserved.

746

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

Anchovy growth (mm d–1)

0.4 Chesapeake Bay populations 43 d experiment 0.2

90 d experiment 0 0

0.5

1.0 1.5 Container radius (m)

2.0

2.5

Figure 16 Variations in mean growth of bay anchovy, Anchoa mitchelli, with size (radius) of cylindrical mesocosms and with duration of experiment. In smaller containers and in longer experiments, fish exhibit lower growth rate. Shaded area indicates the range of growth rates measured in natural coastal waters. Only in the larger containers and shorter experiments were bay anchovy growth rates comparable to those reported for the estuarine waters that serve as natural habitat for these fish. Adapted from Mowitt WP, Houde ED, Hinkle D, and Sanford A (2006) Growth of planktivorouos bay anchovy Anchoa mitchelli, top-down control, and scale-dependence in estuarine mesocosms. Marine Ecology Progress Series 308: 255–269.

Conclusions Enclosed experimental ecosystems have become crucial tools for conducting controlled and repeatable studies of the ocean environment. Those who use mesocosms as research tools and those who use the results of mesocosm experiments need to understand that experimental design choices have important implications for interpretation. Mesocosms are model ecosystems and as such they represent imperfect representations of nature. A great deal is now known about how to design these experimental ecosystems, so that they capture the essential features of nature. Much remains to be learned. The information presented in this article is intended to provide the reader with an introduction to some of the key issues in mesocosm research. The interested reader is encouraged to explore the more detailed information provided in the ‘Further reading section’.

See also Fluid Dynamics, Introduction, and Laboratory Experiments. Laboratory Studies of Turbulent Mixing. Ocean Biogeochemistry and Ecology, Modeling of. One-Dimensional Models. Patch Dynamics. Three-Dimensional (3D) Turbulence. Turbulence in the Benthic Boundary Layer.

Further Reading Adey WH and Loveland K (1991) Dynamic Aquaria: Building Living Ecosystems. San Diego, CA: Academic Press.

Beyers RJ and Odum HT (1993) Ecological Microcosms. New York: Springer. Gardner RH, Kemp WM, Kennedy VS, and Petersen JE (eds.) (2001) Scaling Relations in Experimental Ecology. New York: Columbia University Press. Giesy JPJ (ed.) (1980) Microcosms in Ecological Research. Springfield, VA: National Technical Information Service. Graney RL, Kennedy JH, and Rodgers JH, Jr. (eds.) (1994) Aquatic Mesocosm Studies in Ecological Risk Assessment. Boca Raton, FL: CRC Press. Grice GD and Reeve MR (eds.) (1982) Marine Mesocosms: Biological and Chemical Research in Experimental Ecosystems. New York: Springer. Grice GD, Reeve MR, Koeller P, and Menzel DW (1977) The use of large volume, transparent, enclosed seasurface water columns in the study of stress on plankton ecosystems. Helgolander Wissenschaftliche Meeresuntersuchungen 30: 118--133. Kemp WM, Lewis MR, Cunningham FF, Stevenson JC, and Boynton W (1980) Microcosms, macrophytes, and hierarchies: Environmental research in the Chesapeake Bay. In: Giesy JPJ (ed.) Microcosms in Ecological Research, pp. 911--936. Springfield, VA: National Technical Information Service. Kemp WM, Petersen JE, and Gardner RH (2001) Scaledependence and the problem of extrapolation: Implications for experimental and natural coastal ecosystems. In: Gardner RH, Kemp WM, Kennedy VS, and Petersen JE (eds.) Scaling Relations in Experimental Ecology, pp. 3--57. New York: Columbia University Press. Lalli CM (ed.) (1990) Enclosed Experimental Marine Ecosystems: A Review and Recommendations. New York: Springer. Mowitt WP, Houde ED, Hinkle D, and Sanford A (2006) Growth of planktivorouos bay anchovy Anchoa mitchelli, top-down control, and scale-dependence in estuarine mesocosms. Marine Ecology Progress Series 308: 255--269. Nixon SW, Pilson MEQ, and Oviatt CA (1984) Eutrophication of a coastal marine ecosystem – an experimental study using the MERL microcosms. In: Fasham MJR (ed.) Flows of Energy and Materials in Marine Ecosystems: Theory and Practice, pp. 105--135. New York: Plenum. Odum EP (1984) The mesocosm. BioScience 34: 558--562. Oviatt C (1994) Biological considerations in marine enclosure experiments: Challenges and revelations. Oceanography 7: 45--51. Petersen J, Chen C-C, and Kemp WM (1997) Scaling aquatic primary productivity: Experiments under nutrient- and light-limited conditions. Ecology 78: 2326--2338. Petersen JE, Cornwell JC, and Kemp WM (1999) Implicit scaling in the design of experimental aquatic ecosystems. Oikos 85: 3--18. Petersen JE and Hastings A (2001) Dimensional approaches to scaling experimental ecosystems: Designing

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOCOSMS: ENCLOSED EXPERIMENTAL ECOSYSTEMS IN OCEAN SCIENCE

mousetraps to catch elephants. American Naturalist 157: 324--333. Petersen JE, Kemp WM, Bartleson R, et al. (2003) Multiscale experiments in coastal ecology: Improving realism and advancing theory. BioScience 53: 1181--1197. Petersen JE, Kennedy VS, Dennison WC, and Kemp WM (eds.) (2008) Enclosed Experimental Ecosystems and

747

Scale: Tools for Understanding and Managing Coastal Ecosystems. New York: Springer. Petersen JE, Sanford LP, and Kemp WM (1998) Coastal plankton responses to turbulent mixing in experimental ecosystems. Marine Ecology Progress Series 171: 23--41. Sanford LP (1997) Turbulent mixing in experimental ecosystem studies. Marine Ecology Progress Series 161: 265--293.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOPELAGIC FISHES A. G. V. Salvanes and J. B. Kristoffersen, University of Bergen, Bergen, Norway Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1711–1717, & 2001, Elsevier Ltd.

Introduction ‘Meso’ meaning intermediate and mesopelagic (or midwater) fish refers to fish that live in the intermediate pelagic water masses between the euphotic zone at 100 m depth and the deep bathypelagic zone where no light is visible at 1000 m. Most mesopelagic species make extensive vertical migrations into the epipelagic zone at night, where they prey on plankton and each other, and thereafter migrate down several hundred meters to their daytime depths. Some species are distributed worldwide, and many are circumpolar, especially in the Southern Hemisphere. Much research on distribution and natural history of mesopelagic fish was conducted in the 1970s, when FAO (Food and Agriculture Organization) searched for new unexplored commercial resources. The total biomass was at that time estimated to be around one billion tonnes with highest abundance in the Indian Ocean (about 300 million tonnes) approximately 10 times the biomass of the world’s total fish catch. No large fisheries were, however, developed on mesopelagic fish resources, perhaps due Table 1

to the combination of technology limitations and a high proportion of wax-esters, of limited nutritional value, in many species. From 1990 there was renewed interest in these species in connection with interdisciplinary ecosystem studies, when vertically and diel migrating sound-scattering layers (SSLs) turned out to be high densities of mesopelagic fish. These findings formed the basis for studies of the life history and adaptations of mesopelagic fish in the context of general ecological theory. The thirty identified families of mesopelagic fish are listed in Table 1 and typical morphologies are shown on Figure 1. The taxonomic arrangements of the families differ between various classification systems. In terms of the number of genera per family, the families Gonostomatidae, Melanostomiatidae, Myctophidae, and Gempylidae are the most diverse. Mesopelagic fish are abundant along the continental shelf in the Atlantic, Pacific, and Indian Oceans and in deep fiords, but have lower abundance offshore and in Arctic and sub-Arctic waters. Most populations have their daytime depths somewhere between 200 and 1000 m. They show several adaptations to a life under low light intensity: sensitive eyes, dark backs, silvery sides, ventral light organs that emit light of a spectrum similar to ambient light, and reduced metabolic rates for deeper-living fish. Vertically migrating species have muscular bodies, well-ossified skeletons, scales, well-developed central nervous systems, well-developed gills, large hearts, large kidneys, and usually a swim bladder. The

Families of mesopelagic fish with corresponding number of genera

Family

Number of genera

Family

Argentinidae Bathylagidae Opisthoproctidae Gonostomatidae Sternoptychidae Stomiatidae Chauliodontidae Astronesthidae Melanostomiatidae Malacosteidae Idiacanthidae Myctophidae Paralepididae Omosudidae Anotopteridae

2 2 4 20 3 2 1 6 ca.15 4 1 ca.30 5 1 1

Alepisauridae Scopelarchidae Evermannellidae Giganturidae Nemichthyidae Trachypteridae Regalecidae Lophotidae Melamphaeidae Anoplogasteridae Chiasmodontidae Gempylidae Trichiuridae Centrolophidae Tetragonuridae

Adapted from Gjøsæter and Kawaguchi (1980).

748

(c) 2011 Elsevier Inc. All Rights Reserved.

Number of genera 1 5 3 2 ca.5 3 2 2 2 2 5 20 8 1 1

MESOPELAGIC FISHES

749

8 cm

11 cm

(A)

(B) 70 cm

7.6 cm

(D)

(C) 30 cm

(E) 3 cm

18 cm

(G) (F) Figure 1 Mesopelagic fish. (A) Benthosema glaciale (Myctophidae). (B) Argyropelecus olfersii (Sternoptychidae). (C) Argentina silus (Argentinidae). (D) Maurolicus muelleri (Sternoptychidae). (E) Notolepis rissoi kroyeri (Paralepidae). (F) Astronesthes cyclophotus (Astronesthidae). (G) Bathophilus vaillanti (Melanostomiidae).

ventral light organs are species specific in some families, such as the Myctophidae. The deeper-living species have reduced skeletons, a higher water content in their muscles, lower oxygen consumption, and probably reduced swimming activity compared with species that live at shallower depths.

Life Histories Most mesopelagic fish species are small, usually 2–15 cm long, and have short life spans covering one or a few years. Some species, especially at higher latitudes, become larger and older. A few larger species such as the blue whiting Micromesistius poutassou also live in the mesopelagic habitat, but have the characteristics of epipelagic species. Because of a generally small size, mesopelagic fish have low fecundity, ranging from hundreds to a few thousand eggs. This implies a low mortality in the early life stages, whereas adult mortality is high compared with many epipelagic species. Despite their low fecundity mesopelagic fish have a higher reproductive rate than long-lived epipelagic species which have higher fecundity but a much longer generation time. Neither eggs nor larvae from mesopelagic fish appear to have fundamentally different morphology from those of epipelagic fish, and the larvae all inhabit the epipelagic zone and have growth rates comparable with larvae of epipelagic fish. The higher survival

among the early life stages of mesopelagic fish than of epipelagic species has not yet been quantified. One possible explanation could be different advective loss. The early life stages of large epipelagic populations are passively transported long-distances which means high advective loss. No particular longdistance drift pattern is yet known for mesopelagic fish and this may reflect lower advective loss and lower mortality. Generally, mesopelagic species that live at high latitudes or at shallow depths have more defined spawning seasons than those that live deeper or at lower latitudes. Some species (e.g., Maurolicus muelleri, Gonostoma ebelingi, Cyclothone pseudopallida) exhibit batch spawning, with repeated spawning throughout a prolonged season of several months. Egg diameters do not differ from those of other fish with pelagic eggs and range between 0.5 and 1.65 mm. The eggs are released either in the daytime in deep water, or epipelagically at night. Eggs and larvae have a dilute internal milieu which makes them buoyant. In some species these buoyancy chambers are later replaced by a swim bladder. Other species, especially among the deepest-living forms, do not have a swim bladder. Those with a gas-filled swim bladder often deposit increased amounts of fat in the swim bladder as the fish become older. Before metamorphosis the larvae inhabit the productive epipelagic zone. During metamorphosis the skin becomes pigmented, light organs develop, and the

(c) 2011 Elsevier Inc. All Rights Reserved.

750

MESOPELAGIC FISHES

young start to move down towards the adult habitat. Among some myctophids this ontogenetic shift in habitat is believed to be recorded in the otoliths as accessory primordia, that is, structures that appear as extra nuclei outside the true nucleus of the otoliths. Growth and age composition of some species have been studied by counting presumed annuli or daily increments in the otoliths. In cold and temperate waters both annual and daily increments may be found. In tropical waters only daily increments can usually be detected, partly because of a shorter longevity in these waters and partly because of the lack of seasonality that fish from temperate regions experience. When there are seasonal changes in the environment this is usually registered as annuli in the otoliths. Only seldom has the periodicity of the increments been validated for mesopelagic fish, Nevertheless, studies have verified the daily basis of microincrements in, for example, Maurolicus muelleri, Benthosema suborbitale, B. pterotum, B. fibulatum, Lepidophanes guentheri, Diaphus dumerilii, D. diademophilus, Lampanyctus sp., and Myctophum spinosum. Annual increments have been partially validated for M. muelleri, B. glaciale, Notoscopelus kroyeri, and Stenobrachius leucopsaurus. The usual pattern of growth towards an asymptotic size (usually expressed by fitting the von Bertalanffy growth equation to empirical data of length versus age), which is common in fish, may not occur in all mesopelagic species. Some show almost linear length increase with age and tend not to reach any asymptotic length in their lifetime. Others slow down their length increase as they become older but do reach an asymptotic length. Among widely distributed mesopelagic species, geographical variation has been found in morphology, life history or genetics. Based on morphology, 15 subspecies of M. muelleri have been identified worldwide. Meristic characters of B. glaciale and Notoscopelus elongatus differ between the Mediterranean and the North Atlantic, which suggest genetic heterogeneity. Furthermore, populations of B. glaciale in west Norwegian fiords are genetically different from each other and from the Norwegian Sea population, and their life histories also vary, with a faster growth towards a lower maximum length in the fiord populations. Genetic isolation is probably possible because of the generally deep distribution of B. glaciale combined with relatively shallow sills at the mouth of the fiords. Maurolicus muelleri in Norwegian fiords have lower mortality than those in oceanic water masses. The estimated light level at the depth occupied by M. muelleri is also lower in the fiords than off the shelf, and this may give the fiord fish improved protection from visually oriented

predators. The growth rate, reproductive strategy and predation risk also tend to differ between fiords. Sexual size dimorphism is observed in many mesopelagic species as well as in numerous other fish species. In such dimorphic species the average size of females is larger than for males. Possible explanations for such differences are lower mortality and/or higher growth among females. In some species (e.g., Cyclothone microdon, Gonostoma gracile) sex change occurs; they change from male to female as they grow older. That females are larger than males indicates that a large body size is of greater benefit for females than males, possibly because large females are more fecund than small females. Secondary sexual characters are also found in some species. Among myctophids, males have a supracaudal light organ, whereas females have an infracaudal light organ. These light organs are perhaps structures that could be associated with courtship behavior.

Behavior The behavior of mesopelagic fish has mostly been studied indirectly through monitoring of soundscattering layers (SSLs) by echosounder and by pelagic trawling to obtain samples with some in situ sightings from submersibles. These show that mesopelagic fish are often oriented obliquely or vertically in the water column and it is thought that they may be in a dormant state during daytime. Fish with extensive vertical migrations are not good animals for laboratory experiments. Attempts to keep such lightsensitive mesopelagic fish in aquaria have failed because the fish attempted to migrate downwards, or battered themselves against the walls of the container until they became lifeless. In specially designed spherical containers with water jets, captured myctophid fish have survived a maximum of 72 h. Although no long-distance horizontal spawning or feeding migrations are known for small mesopelagic fish, many species (particularly the myctophids and some stomiatoids) undertake nightly vertical feeding migrations into the productive surface layer. Species with gas-filled swim bladders are most prominent on the records of echosounders, and populations may appear as distinct layers. Such sound-scattering layers move upward after sunset and downward before dawn to their daytime depths. Vertical migration speeds up to 90 m h 1 have been measured. The entire population does not necessarily migrate to the surface every night. For instance, a considerable proportion of the adult population of Benthosema glaciale is present at daytime depths during the night, whereas juveniles are most numerous in the surface

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOPELAGIC FISHES

layers. Ontogenetic differences in daytime levels have been observed for Maurolicus muelleri. In winter and spring, juveniles are found in a separate scattering layer above the adults. Some evidence for depth segregation between the sexes is also reported; female M. muelleri tend to stay deeper than males at daytime during the spring in temperate regions. Depending on the season, females of Lampanyctodes hectoris have been reported to stay either shallower or deeper than males. During daytime mesopelagic fish can adjust their vertical position to accommodate fluctuating light intensities caused by changes in cloudiness and precipitation. The adjustment of the daytime depth levels of the scattering layer thus suggests that vertically migrating mesopelagic fish tend to follow isolumes, at least over short time periods. However, during a 24 h cycle in the summer the estimated light intensity at the depth of M. muelleri has been observed to change by three orders of magnitude. Light is a common stimulus for the vertical displacements and acts as a controlling, initiating and orientation cue during migration. It has been suggested that the ratio between mortality risk and feeding rate in fish, which locate their prey and predators by sight, tend to be at minimum at intermediate light levels. Thus, migration during dawn and dusk may extend the time available for visual feeding while minimizing the predation risk (so-called ‘antipredator-window’). At high latitudes in summer the nights become less dark, and the optimal vertical distribution for catching prey and avoiding predators is altered. For example, Maurolicus muelleri in the northern Norwegian Sea changes between winter to summer months from a daily vertical migration behavior to schooling. Schooling serves as an alternative antipredator behavior during feeding bouts in the upper illuminated productive water masses. Adaptations

Mesopelagic fish experience vertical gradients in light intensity, temperature, pressure, rate of circulation, oxygen content, food availability and predation risk. Species of mesopelagic fish have adapted morphologically and physiologically to a midwater life in various ways. Mouth morphologies are generally large horizontal mouths with numerous small teeth, typical of fish that feed on large prey, combined with fine gill rakers, typical of fish that feed on small prey. This arrangement may partially explain their success, since it enables the fish to feed on whatever prey comes along, regardless of size. Considered broadly, three main groups of mesopelagic fish can be identified based on the

751

morphology: (1) small-jawed plankton eaters, mostly equipped with swim bladders; (2) large-jawed piscivorous predators with a swim bladder; (3) largejawed piscivorous predators without a swim bladder. Table 2 lists most of the traits that are typical for these groups. The physiological and morphological adaptations in mesopelagic fish can be regarded as indirect or direct responses to light stimuli. For example, except for Omosudidae, all mesopelagic fish have pure-rod retinas which are characterized with a high density of the photosensitive pigment, rhodopsin. Eyes of mesopelagic fish tend to be large. The larger the absolute size of the eye and the greater the relative size of its pupils and lens, the better it is for gathering and registering the light from small bioluminescent flashes emitted by photophores. At times, such flashes may be frequent enough to merge into a nearly continuous background of light. Some mesopelagic fish (members of the families Gonostomatidea, Sternoptychidae, Argentinidae, Opistoproctidae, Scopelarchidae, Evermannellidae, Myctophidae, and Giganturidae) have even evolved tubular eyes with large lenses and a larger field of binocular vision, which improve resolution and the ability to judge distances of nearby objects. Coupled with short snouts such eyes enable the individuals to pick out small planktonic organisms in dim light. Tubular eye design for improved binocular vision is achieved at the cost of lateral vision. Many species have modified the eyes further with an accessory retina or even accessory lenses that also allow lateral vision. The possibility of protection for mesopelagic fish lies in camouflage coloration which matches the light conditions in their habitat. Most of them lack spines or other protrusions that may serve as a defence against predators. In the deep ocean they find protection in twilight and darkness, where dark-skinned predatory fish are also well camouflaged. In shallower waters good camouflage is provided by transparency, by reflecting light to match the background perceived by a visual predator, or in certain surroundings by having a very low reflectance. The shallow-living larvae of mesopelagic fish are generally transparent to light. During and after metamorphosis, when the mature coloration is developing, the young fish move down to the dim or dark depths of their adult habitat. The adult coloration of a large proportion of the mesopelagic fish consists of silvery sides, a silvery iris, and a dark back. Most kinds of silvery-sided fish live at the upper mesopelagic levels, where, to the eyes of a visual predator, uncamouflaged prey will stand out against the background of light, except when viewed from above.

(c) 2011 Elsevier Inc. All Rights Reserved.

752

MESOPELAGIC FISHES

Table 2 Organization of the major groups of mesopelagic fish. The comparisons of the predators are relative to the plankton consuming group

Silvery-sided fish are very vulnerable to attacks from below, particularly from black-skinned visual predators. When a visual predator looks upwards it will see its prey in silhouette. It has been argued that the ventral light organs in mesopelagic fish are an adaptation to emit light that matches the background of downwelling ambient light, in order to break up its silhouette and so to make attack from below more difficult. The ability of mesopelagic fish, which are nearly neutrally buoyant, to undertake vertical migration is

related to the structure of their myotomes. They have a muscular organization for sustained efforts with a large proportion of red muscle fibers. These are rich in fat, contain lots of glycogen, myoglobin, and many mitochondria and are richly supplied with blood and thus oxygen. White fibers that dominate the muscles of epipelagic fish, hold little or no fat, little glycogen, no myoglobin, few mitochondria and are more sparsely supplied with blood. White muscles work anaerobically in short bursts, such as rapid escape responses towards predators. The

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOPELAGIC FISHES

Figure 2 Transverse sections through the tails of three mesopelagic fish showing the extent of the red muscles (stippled). (A) Notolepis coasti, an Antarctic paralepidid. (B) Electrona antarctica, a myctophid. (C) Maurolicus, a Sternoptychidae. (D) Astronesthes lucifer. From Marshall (1977).

metabolic cost is related mostly to the requirements of the red muscle in moving the fish upward at a cruising speed in order to search for food. Little energy would be needed during descent when mesopelagic fish, whether they have a swim bladder or not, are likely to be negatively buoyant. The comparative development of red muscle in the tail of selected species is shown in Figure 2. Those with highest proportions of red muscle fibers undertake the most pronounced vertical migrations. Originally the adaptive value of daily vertical migration was related to factors such as reduced competition among species through resource partitioning; minimizing horizontal displacement through advection; and bioenergetic benefits by foraging in warm surface waters and digesting in cooler deep waters. It was also suggested that mesopelagic

753

fish could use vertical current gradients as a way of being transported to new feeding areas. Subsequently there has been more focus on the balance between predation risk and food demand and how this affects vertical distribution patterns. Emphasis has been laid on how vertical migration during dawn and dusk extends the time available for visual feeding, while minimizing the visibility towards predators. The earlier view relied on research in tropical and subtropical regions of the ocean where the fish always experience a change in temperature of about 101C between the daytime depth and surface, and where also the daily light changes are similar all year round. Although similar temperature differences also exist during the summer in temperate regions, there is hardly any temperature difference between shallow and deep water in winter, and occasionally shallow water may be colder than the deeper water. The observations that mesopelagic fish also undertake daily vertical migration during the winter in west Norwegian fiords suggest that there are other explanations than bioenergetics. It is more likely that vertical migration extends the time available for visual feeding while minimizing the visibility towards predators. This is also consistent with the camouflage coloration in mesopelagic fish and that juveniles can stay in shallower water than adults because they are smaller, often transparent and thus less visible than adults. There is a difference of a factor of 15 in metabolic rates between species that live at the surface and those that come no shallower than 800 m. This difference is found to be too great to be explained by decreases in temperature, oxygen content, decrease in food availability or increase in pressure. Comparative analyses of fish from different regions show similar depth trends even in isothermal regions (e.g., the Antarctic) for species which live at similar depths but at different oxygen concentrations. Several lines of research indicate that the metabolic decline is related to a reduction in locomotory abilities with increasing depth. It is suggested that the higher metabolic rates at shallower depths in groups with image-forming eyes is the result of selection action to favor the use of information on predators or prey at long distances when ambient light is sufficient. Hence, good locomotory abilities will be beneficial in order to escape predators. This idea is supported by the fact that major gelatinous groups that lack image-forming eyes do not show a decline in metabolic rate with depth. Thus, the lower metabolic rates found in fish living deeper where visibility is lower, result from the relaxation of selection for locomotory abilities, and is not a specific adaptation to environmental factors at great depths. If so, high

(c) 2011 Elsevier Inc. All Rights Reserved.

754

MESOPELAGIC FISHES

metabolic rate in the surface waters indicates a metabolic cost of predation risk because good locomotory abilities require high metabolism. At greater depths the predation risk is much lower and the need for locomotory abilities decreases.

See also Fiordic Ecosystems. Fish Feeding and Foraging. Fish Locomotion. Fish Migration, Vertical. Fish Predation and Mortality. Fish Reproduction. Fish Schooling. Fish Vision. Large Marine Ecosystems.

Further Reading Andersen NR and Zahuranec BJ (eds.) (1977) Oceanic Sound Scattering Prediction. New York: Plenum Press. Balin˜o B and Aksnes DL (1993) Winter distribution and migration of the sound scattering layers, zooplankton and micronecton in Masfjorden, western Norway. Marine Ecology Progress Series 102: 35--50.

Childress JJ (1995) Are there physiological and biochemical adaptations of metabolism in deep-sea animals. Trends in Ecology and Evolution 10: 30--36. Farquhar GB (1970) Proceedings of an International Symposium on Biological Sound Scattering in the Ocean. MC Report 005. Maury Center for Ocean Science. Washington, DC: Dept. of the Navy. Giske J, Aksnes DL, Balin˜o B, et al. (1990) Vertical distribution and trophic interactions of zooplankton and fish in Masfjorden, Norway. Sarsia 75: 65--81. Gjøsæter J and Kawaguchi K (1980) A Review of the World Resources of Mesopelagic Fish. FAO Fish. Tech. Paper No. 193. Rome: FAO. Kaartvedt S, Knutsen T, and Holst JC (1998) Schooling of the vertically migrating mesopelagic fish Maurolicus muelleri in light summer nights. Marine Ecology Progress Series 170: 287--290. Kristoffersen JB and Salvanes AGV (1998) Life history of Maurolicus muelleri in fjordic and oceanic environments. Journal of Fish Biology 53: 1324--1341. Marshall NB (1971) Exploration in the Life of Fishes. Cambridge, MA: Harvard University Press. Rosland R (1997) Optimal responses to environmental and physiological constraints: evaluation of a model for a planktivore. Sarsia 82: 113--128.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES P. B. Rhines, University of Washington, Seattle, WA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1717–1730, & 2001, Elsevier Ltd.

Introduction Mesoscale eddies are energetic, swirling, timedependent circulations about 100 km in width, found almost everywhere in the ocean. Several modern observational techniques will be used to profile these ‘cells’ of current, and to describe briefly their impact on the physical, chemical, biological, and geophysical aspects of the ocean. The ocean is turbulent. Viewed either with a microscope or from an orbiting satellite, the movements of sea water shift and meander, and eddying motions are almost everywhere. These unsteady currents give the ocean a rich ‘texture’ (Figure 1). If you stir a bathtub filled with ordinary water, it will

quickly be populated with eddies: whirling, unstable circulations that are chaotically unpredictable. There is also a circulation of the water with larger scale, that is, broader and deeper movements. The ‘mission’ of the eddies is to fragment and mix the flow, and to transport quantities like heat and trace chemicals across it. In a remarkably short time (considering the smallness of viscous friction in water) the energy in the swirling basin will have greatly diminished. The bath will also cool much more quickly than one would estimate, based on simple conduction of heat across the fluid into the air above. One may think about the fineness of the pattern of fluid motion in analogy to the resolution of an image on a computer screen. In a bathtub, the fluid eddies have scales from about 1 mm to 1 m, hence spanning a thousandfold range of sizes. In the oceans, the smallest circulations are also a few millimeters in size, but the largest are of the order of 10 000 km in diameter: this represents a range in scale of about 1010 between the smallest and the largest. There is

90 N

Latitude

45

0

_ 45

S _ 90 0

180

90

360

270

Longitude, ˚E

_ 12.00

_ 8.00

_ 4.00

0.00

4.00

8.00

12.00 (CMS)

Figure 1 Sea surface elevation (cm) from the Topex-Poseidon and ERS satellites, 25 March 1998. The mean sea surface elevation for this time of year has been subtracted, so that only ‘anomalies’ from normal conditions are shown. The speckled pattern shows mesoscale eddies almost everywhere. In addition there are larger-scale patterns associated with El Nin˜o (where the Equatorial Pacific has more level sea surface than normal) and large bands of high and low sea surface at middle latitudes. These may be associated with climate variability. The home web site for Topex-Poseidon satellites is http://topex-www.jpl.nasa.gov

(c) 2011 Elsevier Inc. All Rights Reserved.

755

756

MESOSCALE EDDIES

thus room for many sizes of motion, each with a distinct dynamical nature: from tiny eddies that strongly feel viscosity, to ‘mesoscale eddies’ that strongly feel the Earth’s rotation, to great ‘gyres’ of circulation filling entire oceans that feel also the curvature of the Earth. At scales in between are also numerous types of wave motion. Mesoscale eddies are whirling and localized yet they densely populate the ocean. Typically 100 km across, their size varies with latitude and other factors of their environment: energy level, nearly bottom topography, and the nature of their generation. Eddies need not necessarily be round, with circular streamlines. They are often generated by unstable meandering of an intense current like the Gulf Stream. In this case the waving deflection of the Stream is itself a form of latent eddy, which may eventually grow and ‘break’ to form a circular eddy (as an ocean surface wave grows and ‘breaks’ at a beach). Eddies are important because they have so much kinetic energy, and because they can transport momentum and trace water properties. They have deep ‘roots’ that often reach 5 km or more downward, carrying energy and momentum to the seafloor. They are responsible for the irreversible mixing of waters with different properties. Mesoscale eddies are typically as energetic as the concentrated currents that give birth to them. They may owe their existence to several sources other than meandering of strong currents: for example, direct generation by winds or cooling at the sea surface; flow over a rough seafloor or past islands and coastal promentories; or generation by mixing or waves of smaller scale.

‘Geography’ of Mesoscale Eddies Before describing the ‘physics’ of mesoscale eddies, we should discuss their ‘geography.’ A satellite image of the surface of the global ocean can be assembled from many orbits, as the Earth turns below. A particularly basic measurement is that of the height of the sea surface. If ordinary waves are averaged out, we are left with a surface smooth to the eye yet varying by a meter or so relative to the ‘geoid,’ which determines the gravitational horizon (the geoid itself is permanently distorted by seafloor topography and, by itself, yields useful approximate maps of the seafloor elevation). Small variations in height of the sea surface correspond to small variations in pressure in the ocean below. Lines of constant pressure (isobars) are approximate lines of flow, or streamlines for horizontal circulation. If one subtracts from this field the time-averaged sea surface height the result (Figure 1) is a dramatic

display of time-varying mesoscale eddies: they are nearly everywhere. Additionally one sees in this image from the Topex-Poseidon and European Remote Sensing satellites the large-scale variation of the sea surface along the Equator in the Pacific Ocean. This anomalous state is characteristic of El Nin˜o, when the Trade Winds fail to blow westward with normal intensity. Usually the winds pile up water at the west end of the Equator, but if they are absent the sea ‘sloshes’ back, one-quarter of the way round the Earth, toward South America. Animations of this field can be seen on the World Wide Web (for example, at http://topex-www.jpl.nasa.gov), and many of the features are seen to move westward. Mesoscale eddies (as currents at the ocean surface) are particularly apparent in Figure 1 along the paths of intense, major ocean currents. These delineate the Antarctic Circumpolar Current round Antarctica, which has a ‘saw-tooth’ form, flowing south-eastward across the South Indian and Pacific Oceans, and jogging northward where it encounters major seafloor ridges or gaps (at the Campbell Plateau south of New Zealand and the Drake Passage between South America and Antarctica, for example). In each subtropical ocean there are western boundary currents like the Gulf Stream and Kuroshio, which are marked by time-dependent energy after they leave the coasts and flow eastward and poleward. The jetlike equatorial currents show fine-scale energy that is more related to meandering than to separated, circular eddies. The westward flow in the low subtropical latitudes develops eddies in mid-ocean. Altimetry measurement has a large ‘footprint’ that misses eddies smaller than about 50 km in diameter. From direct measurements in the sea we know that the texture of the circulation includes mesoscale eddies smaller than this, particularly at high latitudes. Orbiting satellites do more for us than produce images. Freely drifting instruments on the sea surface, and at great depth below the surface, tell us ‘where the water goes.’ These can be tracked by satellites and acoustic networks. Rather than delineating a smooth pattern of general circulation, drifters in the North Atlantic (Figure 2) show a tangle of tracks, with intense mesoscale eddies causing the gyrelike circulation to be nearly obscured. This region of the Atlantic involves the subpolar gyre, circulating counterclockwise north of 481 N latitude, and the subtropical gyre, circulation clockwise to the south. The Gulf Stream leaves the US coast at Cape Hatteras, in the southwest corner of the figure. It flows east-northeast and rounds the Grand Banks of Newfoundland, flowing north to about the latitude of Newfoundland, where it separates from the coast again, joining the subpolar

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

757

70

65

60

Latitude

55

50

45

40

35

30 70

60

50

40

30

20

10

0

Longitude Figure 2 Tracks of drifting buoys (‘drifters’) on the sea surface, launched during 1996–1999 by Dr. P.P. Niiler, analyzed by J. Cuny. Tracks are colored corresponding to the ‘box’ (dashed lines) they were launched in. The tracks show intense eddy activity, superimposed on the general circulation. Typical duration of a track is 200 days. There is a mean movement of surface waters counterclockwise around this pattern; the Gulf Stream dominates the yellow tracks moving from west to east, and progressing into the other boxes. The purple tracks from the north-east move quickly westward, round the Labrador Sea (the north-west box) in strong boundary currents. This figure symbolizes the challenge of describing the ocean circulation in the presence of mesoscale eddies. There are several kinds of drifting floats, many of which also move vertically to record profiles of temperature, salinity and other properties; these involve some remarkable new technologies. Examples of web sites showing surface and deep-ocean currents using Lagrangian drifters include www.http://www.whoi.edu/science/PO/dept and www.http://flux.ocean.washington.edu/

gyre. The intense boundary current running westward around Greenland is clearly visible as the drifters invade from the east. The kinetic energy associated with eddies exceeds that in the time-averaged currents by factors ranging from 1.5 or so (in the jetlike current cores) to 50 or more (in the ‘quiet’ regions far from intense mean currents). Radiometers on satellites record images at many different wavelengths; in the infrared (typically between 3.7 and 13 mm wavelength), and at even longer wavelengths of ‘microwaves,’ the radiation is strongly related to the temperature of the water at the sea surface (sea surface temperature, SST). Images in visible light show the texture of ocean color, which is strongly correlated with biological activity. These same images record ‘sun glitter’ patterns that are textured by ocean currents. Radiometers typically cannot resolve features less than a kilometer wide, though visible-light imaging can distinguish features down to tens of meters. Satellites actively transmitting beams of radiation can sense the sea

surface elevation, slope, and roughness. Fine ripples and sharp surface wave crests give other sensors (as with synthetic aperature radar (SAR) satellites) resolution down to 20 m or so. These measurements tell us much about the surface currents and winds just above the sea surface. Using SST sensors we now zoom in on a smaller region of ocean. SST patterns are shaped also by the movement of heat in ocean currents. The Gulf Stream (Figure 3) is visible as a warm, red band with sharp edges, carrying tropical heat northward on the west side of the North Atlantic. It shows a warm mesoscale eddy breaking off its northern edge. There are also many features evident of finer scale than was visible using the altimeter data (Figure 1). As with the global pictures of sea surface elevation, SST satellite images can be viewed as animations (e.g., www.nesdis.noaa.gov/). This involves removing the obstacle of clouds (though some sensors, like the radiometers in the TRMM (Tropical Rainfall Measuring Mission) satellite can see SST right through

(c) 2011 Elsevier Inc. All Rights Reserved.

758

MESOSCALE EDDIES

40˚N

36˚N

74˚W

70˚W

Figure 3 Sea surface temperature (SST) patterns in the Gulf Stream and adjacent warm waters of the Sargasso Sea (to the south) and cold, shallow water on the continental shelf. Color indicates temperature, ranging from purple, blue, green, yellow, orange to red as one moves from cold to warm. The Gulf Stream is the narrow, deep red feature flowing rapidly from south-west (left) to north-east. Its instability spawns mesoscale eddies. Each gridded box is one degree wide, and hence the north–south size of each box is 111 km.

clouds). Viewing these animations, the trained eye will see a wealth of phenomena, from the swing of the seasons, to boundary currents, tropical instability waves, upwelling of cold waters at the coasts and Equator, and ubiquitous mesoscale eddies. Fritz Fuglister of Woods Hole Oceanographic Institution, once an artist during the Great Depression, pioneered the mapping of Gulf Stream eddies with painstaking ship surveys. It was a task befitting his training, and the ‘false color’ renditions used here to show temperature, are surely a high form of natural art. As well as being a warm current, the Gulf Stream is also a front separating the warm (red, orange) saline tropical waters of the Sargasso Sea to the south, from the fresher, colder (green, blue, purple) subpolar waters to the north. Despite the time of year (August), waters flowing south from the Labrador Current chill the coastal region as far south as Cape Hatteras. The Gulf Stream front was first mapped in 1768 by Benjamin Franklin, whose cousin Timothy Folger was familiar with it, as a site where whales could be found.

The roundish feature breaking off the north wall of the Gulf Stream in Figure 3 is an example of an eddy formed by instability of a current and its associated temperature front. This instability can draw its energy from two sources: the kinetic energy of the current or the gravitational potential energy of the tilted stratification. As the instability grows, the Stream meanders wildly. Like an oxbow in a sinuous river, it can break off and become an isolated eddy. Here the eddies are sometimes called ‘rings’ because they are like rings of Gulf Stream water enclosing a trapped, foreign water mass. Meanders toward the north thus break off on the north side of the Stream and form warm eddies (relative to the cold waters around them). Conversely, southward meanders break off, encapsulating cold water to form ‘cold’ rings that wander south-westward and are often absorbed back into the Gulf Stream. The net effect is an exchange of water across the front: the Gulf Stream is a ‘mixer.’ Biological communities are strongly affected by ocean circulation and eddies. In the East Australia

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

Current (Figures 4 and 5) the color of the sea surface can be used to estimate chlorophyll concentration (Figure 4) in green plants (phytoplankton). This current is, like the Gulf Stream, a western boundary current. The sea surface temperature for the same region at approximately the same time is shown in Figure 5. The two figures show the differing texture of the two properties, temperature and phytoplankton (plant growth). Temperature is strongly affected by the atmosphere, which erases the memory of SST patterns. Biological activity can persist for longer times, and hence the patterns show streakiness – longer persistence of fine details.

Baroclinic and Barotropic Eddies Eddies produced by the shearing motion of a current or by its store of gravitational potential energy are part of a life cycle of energy transformation. There is a natural evolution of the eddies toward greater width, and toward greater vertical penetration. With the right circumstances the cycle can continue until the eddies reach to the seafloor with nearly identical horizontal currents at every depth. This is known as a ‘barotropic’ state, whereas currents that decrease or increase with depth are termed ‘baroclinic.’ Baroclinic currents obey a balance of Coriolis forces and pressure forces in the horizontal, and gravity and pressure forces in the vertical: this is known as the ‘thermal’ wind balance.’ It establishes a close relationship between horizontal variations in fluid density (as in an ocean front separating warm water from cold) and vertical variations in current 150

759

velocity (as in a current whose velocity decreases as one moves downward from the sea surface). It is a key connection which, for more than a century, has allowed oceanographers to infer currents from observations of the temperature and salinity in the ocean (for temperature and salinity and pressure together determine the fluid density). Thus, for example, the Gulf Stream front, which in cross-section has tilted lines of constant density, is the site of strong vertical variations in current. These relationships are visible in Figure 6, showing a cross-section of potential temperature and salinity in the northern Atlantic. These high-resolution data show the upper layer of warm water that dominates the southern and eastern parts of the section, floating on a bed of much colder, denser water. The sloping surfaces of constant temperature and salinity are evidence of thermal wind velocities associated with the general circulation, and the smaller-scale wiggles show mesoscale eddies. The seafloor topography is dominated by the Mid-Atlantic Ridge. Near Cape Farewell, Greenland (the left end of the section), the subpolar waters reach right to the surface. The salinity section shows the warm water to be saline (of subtropical origin), while the deeper and more northern waters are of much lower salinity, owing to the sources of fresh water at high latitude. (Plots like this can be seen, or made to order, using software available at http://odf.ucsd.edu/OceanAtlas (the Ocean Atlas system) or http://www.awibremerhaven.de/GEO/ eWOCE (the Ocean Data View system).) The lower plots show the vertical profiles of potential temperature versus depth, and potential temperature versus salinity, 155

0.5

40˚S

0.3

_

0.7

Chlorophyll-a (mg m 3)

1.0

0.2

150˚E

155˚E

Figure 4 Chlorophyll-a concentration inferred from ocean color, SeaWIFS satellite. This is the East Australia Current along the coast of New South Wales. Note the richer content of finely textured eddies. Characteristically, ocean color and other ‘tracers’ can develop a more finely filamented structure than can temperature, whose patterns are erased by heat exchange with the atmosphere. Latitude and longitude (the parallels 401S latitude, 1501E and 1551E longitude) are shown (http://www.marine.csiro.au/~lband/SEAWIFS/).

(c) 2011 Elsevier Inc. All Rights Reserved.

760

MESOSCALE EDDIES

22˚C

20˚C 200 m 18˚C 40˚S

16˚C

14˚C

12˚C

150˚E

Copyright 1997 CSIRO Marine Research Figure 5 Sea surface temperature (1C) in the East Australia Current, showing a field of anticyclonic eddies. This is approximately the same region and time as in Figure 4, with warm waters in the north flowing from the tropics, meeting cold waters of the Southern Ocean. White regions are clouds. Temperature scale shown at right. (This image is from http://www.marine.csiro.au/~lband/ SEAWIFS/; there are many web sites providing SST imagery, for example http://www.rsmas.miami.edu/groups/rrsl/, http:// www.eln˜ino.noaa.gov/, and http://fermi.jhuapl.edu/avhrr/sst.html).

for the entire dataset (with colors indicating the very low values of dissolved silicate in this highly ventilated part of the world ocean). When we see eddies in the surface temperature we can thus infer that there will be variations in the currents from one vertical level to the next. Typically, warm eddies appear in cross-section as depressions in surfaces of constant temperature, while cold eddies are ‘domes’ of deep water elevated toward the surface. The sea surface has upward deflection opposite to that of the underlying density layers (provided the currents diminish as one moves downward from the surface). Usually this is the case, and this fits the picture of warm anticyclonically rotating eddies and cold cyclonically rotating eddies. In the Northern Hemisphere, cyclonic means counterclockwise, and the reverse in the Southern Hemi sphere. There are exceptions to this rule, typically occurring when the eddies are generated deep beneath the surface. Capping off this description of the vertical variation in ocean currents, we note that the sea surface elevation reveals not only the existence and shape of surface ocean current patterns but also their sense of rotation. The ‘lows’ in sea surface elevation are lowpressure cells beneath, and hence are cyclonic, while ‘highs’ in sea surface elevation are anticyclonic.

These difficult dynamical connections take on practical significance when one considers the biology of the ocean. Nutrients are richly abundant deep in the ocean, yet they need to be drawn up to the sunlit surface waters to produce chlorophyll-rich phytoplankton. Stable density layering of the oceans, however, provides a strong barrier to vertical movement of water. Anything that can lift deep water nearer the surface is likely to promote life, and this is just what cold, cyclonic eddies do.

Formation of Eddies Eddies and thermal wind balance are also strongly in evidence in the coastal zones of the ocean. The long stretch of the eastern Pacific, shown in Figure 7, extending from California to Washington, shows cold waters upwelling where the north winds of summer blow surface waters offshore. The southward-flowing California Current, and narrower upwelling region are strongly unstable, and mesoscale eddies grow rapidly. Nutrient-rich cold waters promote growth right through the entire food chain, from plankton to whales and sea birds. Eddies act to exchange water between the shallow continental shelf and deeper ocean to the west.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

761

˚C

0

17.5 Depth (m)

1000

15 12.5

2000

10 3000 7.5 5

60˚W

40˚W

2.5

20˚W

˚W

30˚W

20˚W

0˚

(A)

80˚W

40˚W

0˚

(PSU) 36.25

36

1000

Depth (m)

35.75 35.5 35.25

3000

35 34.75 34.5

5000 40˚W

(B)

20˚W

30˚W

Silicate (mol/kg)

Silicate (mol/kg)

40

Depth (m)

30

20 3000

15

30

Tpot-0 (˚C)

1000

40

20

10

10

10 5 0

5000 5 (C)

10 Tpot-0 (˚C)

0

15

34.5

35

(D)

35.5

36

Salinity (PSU)

Figure 6 Section view of (A) temperature and (B) salinity along a ship-track in the northern Atlantic (the vee-shaped path shown in the map inset), from the WOCE hydrographic program. Surfaces of constant fluid density have a form broadly similar to the temperature and salinity surfaces. Also shown are salinity along the same section (lower left), vertical profiles of potential temperature (lower center) and plots of potential temperature against salinity (lower right). (Plot courtesy of Dr. Rainer Schlitzer.)

(c) 2011 Elsevier Inc. All Rights Reserved.

762

MESOSCALE EDDIES

Columbia River

46˚N

WASHINGTON OREGON

44˚N

Cape Blanco

OREGON CALIFORNIA

42˚N

T(˚C) 18 17

Cape Mendocino

16 15 14

40˚N

13 12 11

Figure 8 Eddies formed by cooling a rotating fluid in the laboratory. The dark central disk sits at the water surface and cools it, mimicking a region of cooling to the atmosphere. Coriolis forces give the thermal convection form, initially as small plumes (a few hundred meters across in the ocean), subsequently as mesoscale eddies with scale of 10–100 km, which dominate the scale model experiment here (oceanic flows and waves can be studied using scale models in the laboratory (e.g., Geophysical Fluid Dynamics Laboratory, University of Washington, http:// www.ocean.washington.edu/research/gfd/gfd.html).

Pt. Arena

38˚N

36˚N

128˚E

126˚E

124˚E

122˚E

Figure 7 Coastal upwelling and eddies in the California Current. (Oregon State University.). Blue (cold) coastal waters are drawn up from below the surface, and are rich with nutrients. New instruments enable us to do ‘cat-scans’ of the upper ocean using instruments ‘flying’ behind a rapidly moving vessel, tethered with a cable (e.g., http://www.oce.orst.edu/research).

Eddies formed by convection can be seen over most of the Earth, but they are particularly energetic in the cold, high latitudes. A laboratory experiment (Figure 8) shows mesoscale eddies generated by cooling of the water surface. Rotation of the fluid organizes the eddies, which are much bigger than the convective plumes directly generated by cooling. In the Labrador Sea, cold winds from the Canadian Arctic sweep over the water and cool it intensively (at a rate exceeding 800 W per m2 of sea surface, in a cold-air outbreak, and averaging 300 W m2 for an entire winter month).

Eddies formed directly by winds blowing on the sea surface are thought to occur widely, and yet the large size of wind patterns is not well-matched to the small, roughly 50 km diameter of mesoscale eddies. However, near ocean boundaries, wind forcing can have demonstrable effect on eddies (for example, the westward Trade Winds spilling across the lowlands of Central America create a strong eddy-rich circulation in the eastern Pacific). Larger-scale eddies, more in tune with wind forcing take on the characteristics of Rossby waves (see below). Eddies formed by flow over an irregular seafloor are common, and can be identified in tracks of floats and drifters. These range across the spectrum of turbulent sizes, all the way to the grand scale setting the path of the Antarctic Circumpolar Current. Eddies formed by flow past an irregular coastline are seen widely. When fluid flows past a cylindrical island, it sheds a regular pattern of eddies with alternating rotation direction. This is known as a Karman vortex street. The interesting thing is that, when the same experiment is done in a laboratory, the regularity of the vortex street disappears as the flow is made stronger or the cylinder is made larger. At the much greater scales of oceanic flow it is at first surprising that the turbulence regime is not encountered. The likely reason is that fluid motions restricted to two dimensions cannot fragment their energy into a full state of turbulence as readily as can a fluid with full freedom to move in all three

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

dimensions. Thus, your kitchen sink looks more turbulent than does a much larger ocean basin.

The Physical Properties of Eddies Some basic physical effects If it were not for the Earth’s rotation, its associated Coriolis forces, and its spherical shape and (or) complex bottom topography, the eddies shown in these figures would be much larger in scale. To discuss these effects we need to review some of the basic physics of the ocean. On the great scale of the circulation of the oceans, there are several physical forces at work, particularly buoyancy forces and Coriolis forces. Buoyancy arises because both the water temperature and the concentration of dissolved salts (called the ‘salinity’ – kg of dissolved salt per kg of sea water) affect the density (expressed as the mass of 1 m3 of sea water). Coriolis forces arise ultimately from the rotation of the Earth. Buoyancy produces a layered ocean, with dense fluids beneath less dense fluid. A measure of its importance is the buoyancy frequency or Brunt–Vaisala frequency, N, measured in radians per second. If a region of sea water were lifted upward and then released, it would settle back to its original depth, bobbing about it with a frequency N (see Inverse Modeling of Tracers and Nutrients). The bobbing period (2p/N) varies from a few minutes in the upper ocean to several hours at great depth. Stable stratification greatly limits vertical motion of the fluid, for tremendous energy is required to lift fluid against gravity. Yet the deep ocean is ‘ventilated’ at high latitude. Cold air from the continents and the Arctic is particularly effective at cooling the ocean, making the waters dense enough to sink. Coriolis forces greatly restrict the motion of the fluid oceans and atmosphere. Their importance is measured simply by the rotational frequency, O, of the planet (2p per day). At locations other than the poles, this effect is diminished by the sine of the latitude (y); hence the important frequency, say f, is 2O sin(y) which is just equal to the frequency of a Foucault pendulum. Horizontal structure and size A number of factors are at work determining the diameters of mesoscale eddies. One central idea is that if the buoyancy forces and Coriolis forces are of similar strength, the width, call it l, will be approximately given by l ¼ NH=f , where N and f are as defined above, and H is the vertical scale of the eddy. This is known as the Rossby deformation radius, after Carl Gustav Rossby, a pioneer in both oceanography and atmospheric sciences. For eddies with vertical scale

763

comparable with the ocean depth, the size l ranges from a few hundred kilometers in the tropics to 10 km or so at high latitudes. This great range of variation comes from the tendency for the highlatitude ocean to have weaker density stratification (small N) and larger Coriolis frequency f. l also represents the horizontal distance traveled by a simple internal gravity wave in a half-pendulum day. The same dynamical eddies exist in the atmosphere, yet are much larger in horizontal scale. They are the basic high- and low-pressure cells seen on weather maps. Their 1000 km diameter (roughly) is also estimated by l, which is much larger because the buoyancy frequency of the atmosphere is so much greater on average, than that of the ocean. Vertical structure The oceans are full of threedimensional structures. The general circulation involves ‘arteries’ of flow, often narrow horizontally (say, 50 km wide) and vertically (say, 1 km or less, thick). Cross-sections of velocity or trace properties marking the circulation illustrate this. Because eddies are often spawned as instabilities of major currents, they too may be three-dimensional, and of limited extent in the vertical. Such structures, which are termed ‘baroclinic,’ may have a range of vertical scales, H. In addition, both the general circulation and eddies can exist in a form with no variation of horizontal current from the top of the ocean to the seafloor. These ‘tall’ currents and eddies, termed barotropic, are distinct and important. They disturb the density field only slightly, and hence are invisible in classic hydrographic sections. For this reason, they were not well understood or observed by early oceanographers. Barotropic flows have a signature at the sea surface, but otherwise have no gravitational potential energy. Tall, barotropic eddies evolve rapidly and are strongly associated with Rossby waves. Rossby waves; potential vorticity Finally, the shape of the planet is important. Its nearly spherical form causes Coriolis effects to change with latitude, and this leads to a new, rather exotic phenomenon known as Rossby waves. Water on a spinning planet is endowed with a ‘stiffness’ along lines parallel with the planet’s axis. This stiffness does many things to the circulation, tending to restrict motion to lie east and west. More generally, in the presence of valleys and ridges on the seafloor, currents can circulate freely along curves of constant sin(y) divided by depth. These are simply curves of constant ocean depth, if we measure the depth parallel to the Earth’s axis rather than vertically. Such pathways of freely flowing water are known as ‘geostrophic contours.’

(c) 2011 Elsevier Inc. All Rights Reserved.

764

MESOSCALE EDDIES

The physics of mesoscale eddies is described well by an exotic property of the fluid: the potential vorticity. We are interested in many different things in fluids: their velocity, temperature, density, salinity, etc., but certain of these properties are particularly illuminating. The fluid density, for example, is active in determining buoyant forces in the fluid. After correcting for pressure effects, the density is also a marker of fluid motion, so long as mixing and diffusive effects can be ignored: it stays the same (after that correction), if one follows a moving parcel of fluid. Maps of this corrected ‘potential’ density both give us dynamical information about currents and show us something about the mass distribution of the oceans. Potential vorticity also has the property that it remains constant, as we follow a parcel of fluid, until mixing or external forces or heating is felt. The quantity describes the ‘spin’ of a fluid parcel, including the rotation of the Earth, and also including a measure of the thickness of the fluid layer. From knowledge of the field of potential vorticity, one can calculate much about the currents and displacement of the mass field of the ocean. As a more general definition, geostrophic contours become lines of constant potential vorticity, which cover surfaces of constant (potential) density. This same stiffness imparted by the Earth’s rotation produces wave motions if fluid is pushed across, rather than along, geostrophic contours. These Rossby waves are ‘information carriers’ that help to form the general circulation. They are themselves unsteady currents, whose patterns radiate

Figure 9 Rossby waves in a laboratory simulation of the circulation of an ocean centered on the North Pole. The waves are visible as undulations of the dye line, and are propagating eastward away from the wave source. Nevertheless, the wavecrests seem to move westward (clockwise). The source of the wave motion is a small oscillating cylinder at the lower left (black band). Geophysical Fluid Dynamics Laboratory, University of Washington.

principally horizontally from where they are generated. A laboratory experiment, Figure 9, shows Rossby waves in a basin centered on a virtual North Pole. All of the motion in this experiment is generated by a small, oscillating body in the lower left of the figure (beneath the black rectangle). The Rossby waves are seen as wavy deflections of the central band of dye. Constant–latitude circles become marked with colored dyes as east–west currents develop in response to the Rossby waves’ shaping of the general circulation. These Rossby waves are ‘weak’ eddies. If they ‘break,’ that is, deform the basic fluid greatly and irreversibly, they fulfill our picture of turbulence: chaotic, with active stirring and mixing of trace properties, like the colored dyes here. The ‘rotational stiffness’ that makes the waves possible also greatly limits the north–south movement of fluid. Thus, the polar cap in this experiment is virtually unmixed, and is chemically isolated from the lower latitudes. This is just the physics at work in the atmosphere, in defining the polar ozone depletion zones (‘ozone holes’). Rossby waves, and their more violent cousins the mesoscale eddies, help to set the fundamental force balances of the general circulation. They redistribute the momentum of ocean currents horizontally (as in Figure 9), establishing and reshaping currents in horizontal planes. But the ocean is three-dimensional, and these waves and eddies are also active in transferring momentum downward from the sea surface. In Figure 1 the Antarctic Circumpolar Current is driven eastward by the strongest sustained winds on Earth. It thus becomes the greatest of ocean currents (in terms of transport and potential and kinetic energy). The eastward force of the winds is balanced by pressure forces exerted by the ridges and gaps of the seafloor topography. To connect these opposing forces, Rossby waves and eddies are active in transporting momentum downward. Elsewhere in the world ocean, eddies also provide essential communication of momentum downward from the surface. They drive deep gyres known as inertial recirculations, and establish the form of the deep roots of currents like the Gulf Stream. Mesoscale eddies also stir and mix the potential vorticity field. With weak ocean currents, the geostrophic contours, or free-flow pathways, tend to lie east and west. In order to develop the great gyres of circulation, with substantial north–south flow, the ocean has to reorganize its potential vorticity field accordingly. This is accomplished both by eddy activity and by the dynamical reshaping of the large-scale oceanic density field. We have argued that long waves, involving much subsurface activity, are an important cousin of

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

mesoscale eddies. Such waves are particularly visible in the tropical oceans. El Nin˜o is an interaction between atmosphere and oceans. While it recurs somewhat unpredictably, in ways not yet fully understood, oceanic wave propagation along the Equator adds a ‘delayed memory’ to the process (such propagation is clearly visible in satellite altimeter and SST animations; see www.pmel.noaa.gov). Precursors to El Nin˜o are recognizable, and give roughly six months of predictability at present. Sea surface temperature anomalies on 18 January 1999, during La Nin˜a (the opposite phase to El Nin˜o) involved an unusually cold eastern tropical Pacific and warm core in the subtropical North Pacific (Figure 10). The pattern is decorated with mesoscale eddies of much smaller scale and unstable waves on the equatorial westward jet. It is interesting that as one moves from high latitude toward the Equator, the Rossby scale, l, increases markedly and mesoscale eddies become larger and more wavelike. Energy sources in the strong Equatorial current

765

system give rise to tropical instability waves, which appear to play an important role in both dynamics and biology.

Modeling Techniques Figures 8 and 9 showed laboratory simulations of mesoscale eddies and Rossby waves. Some, but not all, physical effects active in ocean circulation can be modeled in the fluids laboratory. A persistent problem is the exaggerated effect of friction and molecular diffusion of heat and salt in the small scale of a model. Beginning in the 1970s, computers were developed with enough speed and memory to solve adequately the physical equations of motion, using methods of numerical approximation. Fully turbulent flows that have eluded theoreticians for hundreds of years have suddenly become accessible to ‘numerical experiments.’ These experiments have problems analogous of those in the fluids laboratory: limited resolution of fine details. Yet analysis of

Equator

Figure 10 Sea surface temperature map (showing the temperature anomaly, or difference between the temperature and its seasonal mean at each point) in the eastern Pacific Ocean for 18 January 1999 (http://podaac.jpl.nasa.gov/sst/intro.html). The low (purple, blue) temperatures along the Equator represent the unusually strong easterly (that is, westward) winds associated with La Nin˜a. Mesoscale eddies appear in middle latitudes, and also as tropical instability waves along the Equatorial circulation. Both satellite observations and in situ instruments, moored or drifting or shipborne, contribute to our understanding of the tropical oceans (e.g., http://www.pmel.noaa.gov).

(c) 2011 Elsevier Inc. All Rights Reserved.

766

MESOSCALE EDDIES

the flow is far easier in a computer model than a laboratory model: everything about the computermodeled flow is measurable. We have pointed out the range of length scales needed to describe oceanic motions, from 1 mm to 10 billion times that large. Computer models currently describe a range of scales, typically, of only a thousandfold; with these we can simulate a flow that is a few hundred eddies ‘wide.’ What does this mean in terms of representing the global ocean? Computer experiments were originally developed for the atmospheric fluid, with weather prediction a principal goal. There are many similarities between atmospheric and oceanic circulations and eddies, but there are also striking differences. Particularly, Rossby’s deformation scale, l, is much smaller in the ocean than the atmosphere. l is an estimate of the diameter of mesoscale eddies, and this means that the great high- and low-pressure patterns on a weather map, with cyclonic and anticyclonic systems typically having 1000 km diameter, are dynamically similar to 100 km wide ocean eddies. The texture in Figure 1 is much more fine-grained than that of a weather map showing atmospheric pressure patterns. These eddies are the most energetic features of the circulation and must be resolved by the models for

many purposes. Computer models of the ocean are thus global in extent yet need gridpoint spacing significantly less than l. Ten-kilometer spacing of gridpoints is typical in a modern ‘eddy resolving’ ocean model (with vertical resolution of a few hundred meters), and even this is not enough to resolve fully eddies and boundary currents. Atmospheric models with grids having ten times greater spacing are considered to have ‘high resolution.’ There is another striking, nearly devastating, obstacle to ocean modeling: evolution of the circulation is much slower than with comparable features of the atmospheric circulation. As we know from experience, the atmosphere adjusts to changes in solar heating after a period of a few months. Its weather features are rapidly destroyed by friction at the ground after just a few days. Thus a 10-year simulation of the atmosphere is very long indeed. However, the oceanic circulation responds much more slowly to changes in forcing, and this requires much longer simulation experiments. If the winds or solar heating at the sea surface are changed, the ocean will begin to respond quickly. Rossby waves will transmit the changes across major oceans in a few weeks. The tall, barotropic part of the flow will begin to adjust. Currents along the western boundaries and the

Figure 11 Temperature at 160 m depth, from the Parallel Ocean Processing model (http://vislab-www.nps.navy.mil/~braccio/) of the Naval Postgraduate School and Los Alamos National Laboratory. This is a snapshot during a single day in January 1992. The numerical model is driven by observed meterological winds and is also ‘restored’ back toward surface ocean observations. Notice the fine pattern of mesoscale eddies superimposed on the warm, (red) subtropical gyres and the cool (blue) high-latitude circulations. Grid points of this computer model are spaced 1/61 of latitude apart, giving reasonable coverage of mesoscale eddies at low and middle latitudes. This map of temperature differs from Figure 10, which shows only the anomaly field.

(c) 2011 Elsevier Inc. All Rights Reserved.

MESOSCALE EDDIES

767

Equator will quickly change. But the deeper, more baroclinic flow will take more than 10 years to adjust, and the great, global meridional overturning circulation will not fully redistribute heat, salt, and trace chemicals and biological fields for several thousand years. These key features – the smallness of strong ocean currents and their slow evolution in time – can be seen in animations of computer model runs (for example at http://vislab-www.nps.navy.mil/ ~braccio/ for the Naval Postgraduate School and Los Alamos National Laboratory POP model; or www. http://panoramix.rsmas.miami.edu/micom/ for the University of Miami isopycnal ocean model). These simulations (Figure 11) use the full power of our largest computers, and are wonderful renditions of an entire world of ocean physics. Computer simulations of weather and climate have to be run for many thousands of years if they are to encompass the full range of oceanic adjustment. In practice this cannot yet be done with 10 km grid spacing, and climate modelers instead use coarse resolution (typically 100–400 km grid spacing) and simulate the action of mesoscale eddies. They do this with exaggerated friction, and diffusion of heat and salinity that is much larger than in reality. These ‘sticky, conductive’ oceans may provide models of some of the important oceanic transport of heat and fresh water, and have much interesting structure, but they lack the full detail of both boundary currents and mesoscale eddies. The art of ‘parametrizing’ the effects of eddies so as to allow their neglect in detail is an active area of current research.

As a member of the huge family of turbulent motions, eddies contribute to the stirring and mixing of the oceans, to the creation of its basic, layered density field, and to its general circulation. The fundamental physics of eddies is expressed in terms of its potential vorticity, which is a tracerlike property that ‘moves with the fluid.’ The distribution of potential vorticity can be turned into knowledge of the currents and fluid density variations. The smallness and great energy of mesoscale eddies, the great thermal and chemical capacity of the oceans, and the slowness of the circulation conspire to challenge computer models, but rapidly increasing computer power is producing ever better representations of the ocean’s fabric. At present, rather short-lived experiments (a few decades duration) can be carried out that resolve the global field of eddies, intense currents, and wind-driven gyres, whereas the slower features important to long-term climate change cannot be examined while also resolving mesoscale eddies. Nevertheless, several important physical processes like turbulent mixing, convection, upper mixed layer dynamics, and interaction with complex bottom topography are not yet well simulated by computer models. Many of the important applications of physical circulation in the oceans involve vertical motion: for biological communities, for transport of trace gases and their exchange with the atmosphere, for ocean/atmospheric climate interaction. This vertical motion of the fluid is particularly difficult to predict without fully resolving the detail of mesoscale – and smaller – features.

Conclusion

See also

We have argued that mesoscale eddies contain large kinetic energy, comparable with that of the timeaveraged ocean circulation. Eddies are crucial to the transport of heat, momentum, trace chemicals, biological communities, and the oxygen and nutrients relating to life in the sea. They are also active in air– sea interaction, both through response to weather and in shaping the patterns of warmth that drive the entire atmospheric circulation.

General Circulation Models. Ocean Circulation. Rossby Waves. Satellite Remote Sensing of Sea Surface Temperatures. Satellite Remote Sensing SAR.

Further Reading Summerhayes CP and Thorpe SA (1996) Oceanography, An Illustrated Guide. New York: Wiley.

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION the open ocean, with the main impacted areas being estuaries, fjords, rias, and their adjoining shelf seas (Figure 1). In the marine environment, metals such as iron, vanadium, copper, and zinc are essential for certain biochemical reactions in organisms, but even in moderately contaminated estuaries these metals contribute to stress in marine biota. By virtue of their toxic and bioaccumulative properties both cadmium and mercury are regarded as ‘Black List’ substances, while lead is on the ‘Grey List’. These elements have little or no biochemical function and, while tolerable in minute quantities, exhibit toxic effects above critical concentrations. Mercury has a complex marine chemistry and exists in various forms, such as inorganic mercury, organically complexed mercury (with natural dissolved organic carbon), as a dissolved gas, Hg0, and as the methylated species monomethyl mercury (MMHg) and dimethyl mercury (DMHg). Both MMHg and DMHg are present in the water column in sediments and in the tissues of

G. E. Millward and A. Turner, University of Plymouth, Plymouth, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1730–1737, & 2001, Elsevier Ltd.

Introduction Marine pollution has been defined as ‘the introduction by man, directly or indirectly, of substances or energy to the marine environment resulting in deleterious effects such as hazards to human health; hindrance of marine activities, including fishing; impairment of the quality for the use of sea water; and reduction in amenities’ (GESAMP, 1990). Approximately 45% of people on Earth live within 150 km of the coast and marine pollution occurs as a consequence of increases in population density and industrialization. The problems of marine pollution are generally limited to nearshore waters rather than

Weathering

RIVER

Air_ sea exchange

Industry Md

Mp FBI

Discharge

Dumping /spillage Advection

ESTUARY Ms

Md Adsorption, desorption (K D) M

Coastal front

Advection

Uptake, dissolution Mb

p

Uptake decay

Deposition

Trophic transfer (bioaccumulation) Cross-frontal mixing

Upwelling

SHELF SEA OCEAN

Sediment

Interstitial water Resuspension infusion Ms Accretion Diagenesis, bioturbaton Ms Mi

Continental slope

Figure 1 Processes affecting the transport and biogeochemistry of metal pollutants in estuaries and shelf seas. FBI ¼ fresh water– brackish water interface. Metal compartments are designated. Md, dissolved; Mp, suspended particulate; Ms, sediment; Mi, interstitial water; Mb, biogenic particulate.

768

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION

marine organisms. Thus, depending on their physicochemical state or bioavailability, metals will impact upon different parts of the marine food web and in some cases bioaccumulation and/or biomagnification occurs, which may, ultimately, expose humans to a potential health hazard. When attempting to assess the biogeochemical pathways and health impact of metals it is crucial to determine the total concentration accurately and where possible to identify and quantify the physical and chemical forms, or species. The analytical determination of metals in sea water has had a difficult history and many measurements reported in the literature prior to about 1985 should be treated with caution. Major strides have been made in the minimization of contamination during sample collection, storage, and preparation and in the application of sensitive analytical techniques, sometimes coupled with methods for the separation of metal species. The concentrations of dissolved metals have been revised downwards in recent years as a consequence of the introduction of these advances, together with improvements in analytical quality assurance, including appropriate use of certified reference materials. The sources and pathways of metals through the coastal environment are complex (see Figure 1). Interfacial processes play a key role in their passage from the land to the sea. In estuaries the composition of river water may be modified by physicochemical processes at the fresh water–brackish water interface (FBI), where strong gradients of salinity, temperature, concentration and type of suspended particulate matter (SPM), pH and dissolved oxygen exist. Metal

769

exchanges, between the dissolved and particulate phases, take place under the influence of these gradients and this process is quantified by the partition coefficient (KD, 1 kg1) (eqn [1]). 

 Mp KD ¼ ½Md 

½1

Here [Mp] is the metal concentration of SPM in nmol kg1 (or mg kg1), and [Md] is the dissolved metal concentration in nmol l1 (or mg l1). Coastal sediments can contain elevated concentrations of dissolved metals in their interstitial waters which may be exchanged across the sediment–water interface via molecular diffusion or by resuspension and in soft sediments by enhanced diffusion due to bioturbation from burrowing organisms (Figure 1). The mercury cycle is complicated by the fact that microbial activity, in sediments and the water column, can produce DMHg and Hg0, both of which are volatile and can exchange across the air–sea interface.

Anthropogenic and Natural Inputs Dissolved and particulate metals in rivers and estuaries are derived from natural weathering process in the catchment area, and reflect the geological composition of the watershed (see Table 1 for the crustal abundance of selected metals) and the local climatic conditions. Natural concentrations of metals can be augmented in catchment areas that are mineralized, and there may be a significant anthropogenic perturbation downstream because of mineral extraction

Table 1 Fluxes of metals to the atmosphere from natural and anthropogenic sources. The interference factor is the ratio of the anthropogenic flux to the natural flux. The generic term ‘combustion’ refers to various combinations of coal, oil, and wood combustion and refuse incineration Metal

Crustal abundance (nmol g  1)

Atmospheric emission rate (t y1)

Natural Cadmium

Copper Mercury Lead Zinc

2

510 0.4

1.4

28 2.5

Anthropogenic 7.6

35 3.6

80

12

332

2000

45

131

Major uses of metals and their compounds Interference factor 5.3

1.2 1.5 27 3.0

(c) 2011 Elsevier Inc. All Rights Reserved.

Nonferrous metal production; cement/fertilizer manufacture; combustion Nonferrous metal production; biocides; combustion Chlorine cells; gold mining operations; combustion Petroleum additive; nonferrous metal production; combustion Nonferrous metal production; steel/ iron manufacturing; cement production

770

METAL POLLUTION

and processing. In densely populated regions, metals originate from a wide range of industrial, domestic and agricultural uses, and their inputs into river systems have increased significantly over the past two centuries. Regulated dredging and dumping of metal pollutants at sea, inadvertent spills, and illegal discharge all add to the complexity of anthropogenic inputs to the aquatic environment. Thus, our ability to unravel natural versus anthropogenic inputs is often complicated by the significant and uncontrolled human perturbation of catchments and their river systems. Only where metal compounds are entirely of anthropogenic origin, such as tributyl tin, can the human impact be evaluated. In the case of lead, however, it has been possible to identify man-made inputs via the application of inductively coupled plasma mass spectrometry (ICP-MS) to the determination of lead isotopic ratios (e.g., 206Pb to 207 Pb) which have distinct signature in leaded gasoline. Because a significant proportion of the marine environment has been altered by anthropogenic activities, natural concentrations values for dissolved and particulate metals are difficult to obtain unambiguously. Baseline values are often assumed from analyses of samples from remote systems that are considered to be ‘pristine’ or from metal analyses of sediment horizons dated as being prior to the industrial revolution. Another approach to assessing man’s impact on the global ocean is to compare the rates of metal emission to the atmosphere from natural and anthropogenic sources. In Table 1 the ‘interference factor’ is 41 for all metals, with relatively high values for lead and cadmium, suggesting that there is a significant anthropogenic alteration of their natural cycles. Macrotidal. estuaries have strong internal cycles and particles may be retained within the system for years, and in large estuaries for decades. Thus, estuaries are a significant repository for metals, although no systematic inventories of the sediment metal burden have been made. Suspended particles advecting from estuaries into shelf seas are trapped in the coastal margin and estimates show that B90% of the fluvial suspended load (and associated metals) of the Mississippi, St. Lawrence, Rhoˆne, and rivers in the south east of the United States is deposited in the coastal margin. Early diagenesis of deposited material may result in release of metals into sediment pore waters and, since the dissolved phase is generally considered to be more bioavailable, the composition of interstitial waters could be more important in the overall toxicity of the sediments than is their total metal content. Accurate quantification of interstitial water composition in

estuaries and shelf seas is hindered because of the heterogeneous distribution of sediment texture and because a satisfactory method has not yet been developed for application in shallow waters that are highly dynamic. Particulate metals deposited in the coastal margin are slowly advected onto the continental slope by seabed currents and wave action. Sediment diagenesis and diffusion releases dissolved metals into the oceanic water column, where they may be involved in upwelling processes at the shelf break (Figure 1). Comparisons of the relative magnitudes of the combined river and atmospheric fluxes with the upwelling flux suggests that the latter is greater by a factor 2 for copper, of 2–7 for zinc, and of about 10 for cadmium. In contrast for inorganic mercury the upwelling flux to shelf seas is half the magnitude of the combined river and atmospheric input, while for methylated mercury the main source to shelf seas is the upwelling flux.

Distributions The temporal and spatial distributions of dissolved metal pollutants are highly dependent on two important processes in the coastal boundary zone. Local hydrodynamics. Water is dispersed in estuaries and coastal waters according to the local hydrodynamics, which can be characterized by the flushing time. The dispersion and dilution of metal pollutants from point and diffuse sources, and therefore their range of concentrations, will be affected by the flushing time. A flushing time of 0.5 y for coastal waters is typical for the North Sea and Irish Sea, a value of 2 y is representative of the waters around Bermuda, while a value of 5 y is representative of a semi-enclosed sea such as the Baltic. Waters with longer flushing times will register a slower response to changes in metal inputs, whereas changes in metal concentrations will be detected earlier in waters with shorter flushing times. Particle–water interactions. Metal partitioning between the dissolved and particulate phases is a crucial factor because solutes are transported in a different way to particles. The latter experiences gravitational settling and aggregation, as well as advection and mixing. The concentrations of SPM and the types of SPM play a significant role on the fraction of metal carried in the particulate phase. Lead has a relatively high KD, largely owing to its tendency to complex with carboxyl and phonolic groups that dominate the surfaces of natural particles. In contrast, the relatively low KD for cadmium is the result of its ability to complex with chloride

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION

(TMZ)

Estuary Shelf sea Ocean 1.0

Cd 0.8

3

Zn

10 4

10

0.6

fd

Cu 0.4

Hg, Pb

0.2

10

0 0.1

5

10 6

10

7

1

10 _ SPM (mg l 1 )

100

1000

Figure 2 The fraction of metal in the dissolved phase (fd) as a function of the concentration of suspended particulate matter (SPM). The bands at the top of the diagram represent the general ranges of SPM concentrations from the open ocean through to the estuarine turbidity maximum zone (TMZ). The numbers next to the lines are values of the partition coefficients, KD, used to estimate the value of fd. Metals are associated with their typical KD values and are shown next to the appropriate line.

Table 2 Distributions of concentrations of dissolved cadmium, mercury, and lead (pmol l1) in rivers, estuaries, the English Channel and the North Atlantic Location

Cadmium

Mercury

Lead

River background Scheldt, Belgium Seine, France English Channel N. Atlantic Surface N. Atlantic Deep

90 180–800 900–1 800 100–130 1–10 350

2–20 3.4–14 2.5–40 1.5–2.5 1–7 1

1 000 240–810 o2 300–16 000 115–150 100–150 20

ions in sea water. Figure 2 shows the fraction of metals in the dissolved phase (fd) calculated as a function of SPM concentration, using representative KD values for sea water. In estuaries and coastal waters, significant fractions of the metals are associated with particulate matter, but, as the SPM concentration declines through the coastal margin and into the ocean, the dissolved phase assumes more importance Because of their reactivity with particles many metals have short residence times in the coastal ocean, in the range 50–1000 y. Distributions in Estuaries and Coasts

Table 2 illustrates the trends in the concentrations of dissolved metals from river to ocean. The Scheldt

771

and Seine estuaries have important anthropogenic sources of the metals compared to the riverine inputs. Estuarine chemistry in the Scheldt is complicated by the discharge of a high organic load that renders the waters of the upper estuary anoxic during most of the year. Thus, the concentrations of dissolved cadmium, mercury and lead in the water column can be relatively low as a result of the formation of sparingly soluble metal sulfides. However, the sediment interstitial waters of the Scheldt can contain up to 25 000 pmol 11 of dissolved cadmium. In the Scheldt and Seine estuaries, analyses of mercury speciation have shown the presence of the Hg0 form and microbial mediation appears to have transformed inorganic mercury into MMHg and DMHg. In assessing the distributions of dissolved metals in the coastal margin, their concentrations should be normalized with respect to salinity because a higher concentration at one location may be due to lower salinity (or greater fluvial influence). Distributions of dissolved metals in estuarine waters can vary linearly with salinity, the slope of which is dependent on the relative dissolved metal concentrations in the river and the sea, i.e., the ‘end-member’ concentrations. An example of conservative behavior, in which the concentration of dissolved cadmium varies in proportion to the amount of mixing between river water and sea water, is shown in Figure 3A for the Humber estuary plume. The observed behavior for cadmium is due to its affinity for the dissolved phase, even in turbid waters. The temporal change in the slope also shows that the distribution of the metal is highly responsive to changing inputs between different flow regimes. Total dissolved mercury (Figure 3B) displays non-conservation behaviour with a maximum concentration as a result of inputs from point sources along the banks of the Humber Estuary. Dissolved lead also behaves nonconservatively (Figure 3C) and the scatter of data arises because of diffuse atmospheric inputs. The dissolved lead is maintained at relatively low concentrations by its propensity to react with particles (Figure 2). Distribution in the North Atlantic Ocean

Dissolved cadmium has a higher concentration in the deep waters of the North Atlantic Ocean owing to its uptake by phytoplankton in surface waters and recycling at depth; it exhibits nutrient-like behavior. Dissolved mercury has almost no gradient through the water column, because of a significant loss of Hg0 to the atmosphere. However, in the deep waters of the North Atlantic, higher concentrations of MMHg and DMHg have been detected, possibly as a

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION

1.2

250 200

0.8

_1

[Hg] (pmol l )

0.6 0.4

150

100

50

0.2

0

0 28 (A)

30

32 Salinity

36

34

32 Salinity

30

28 (B)

36

1.0

3

0.8

_1

[Pb] (winter) (nmol l )

34

2

0.6 0.4

1 0.2

_1

_1

[Cd] (nmol l )

1.0

[Pb] (spring) (nmol l )

772

0

0 28

30

(C)

32 Salinity

34

36

Figure 3 Concentrations of dissolved metals in the Humber Estuary as a function of salinity in winter with high fluvial input (solid symbols) and spring with reduced fluvial input (open symbols): (A) cadmium; (B) total mercury; (C) lead.

consequence of remineralization of phytoplankton. In contrast, lead has higher concentrations in the surface waters of the Atlantic, reflecting an atmospheric input. The strength of the atmospheric lead source appears to be declining as a result of a decrease in the use of leaded petrol. The monitoring of dissolved lead in the waters around Bermuda over a 15-year period shows that lead concentrations have decreased significantly. Since the SPM concentrations in these waters are low, the changes in dissolved lead concentrations are controlled almost exclusively by the flushing time of the water and the changing lead inputs. Following the decline in the atmospheric input, these waters have relaxed to near-background concentrations of dissolved lead (Figure 4).

Environmental Impact The sediments of coastal regions reflect the long-term accumulation of metal contamination. Assessments of the anthropogenic component of metals in sediments require that the grain size must be accounted for and the corrected metal concentration compared

with an uncontaminated reference material. For sediments with similar grain sizes, normalization is achieved with respect to a major element that is unaffected by anthropogenic inputs, such as aluminum, lithium, or rubidium. The enrichment factor (EF) is then defined as in eqn [2]. 

   Mp = A1p EF ¼ ½Mr =½A1r 

½2

Here [Mp] and [Mr] are the metal concentrations in particulate matter and in crustal rock, respectively, and [Alp] and [Alr] are the concentrations of aluminum (or any suitable reference element) in particulate matter and crustal rock, respectively. Table 3 lists EFs for SPM or fine sediment in contrasting estuaries. Enrichment factors are close to unity for the baseline sediment and in the ‘pristine’ Lena Estuary, while the greatest EF values are encountered for cadmium in the Rhine (impacted by the production of phosphate fertilizers) and the Scheldt, and for copper in Restronguet Creek (impacted by historical mining activity). The general sequence of EFs

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION

_

0

50

[Pb] (pmol kg 1) 100

150

200

0 100

773

Table 4 Mean concentration (nmol g1) of metals in Fucus spp. in estuaries and coastal waters Location

Cadmium

Copper

Lead

Zinc

Hardangefjord, Norway Humber Estuary, UK Restronguet Creek, UK Baseline (Western Irish Sea)

63 46 7.3 4.4

330 680 14 000 60

220 39 140 6

21 000 8 500 26 000 1 100

1993

200 1989 300

1984

Depth (m)

400 500 600 1979

700 800 900 1000

Figure 4 Vertical profiles of dissolved lead in the north Atlantic near Bermuda. The low point at 200 m in 1984 could not be accounted for by the authors, even though the analysis had been replicated and checked, and was thought to be a residual from the deep mixed layer from the previous winter. (Reprinted from Wu and Boyle (1997), copyright 1997, with permission from Elsevier Science.)

Table 3 Enrichment factors calculated according to eqn [2] for metals in suspended particulate matter or fine sediment from estuarine and coastal environments Location

Cadmium

Copper

Lead

Zinc

Rhine Estuary, The Netherlands Scheldt Estuary, Belgium Seine Estuary, France Humber Estuary, UK Restronguet Creek, UK Lena Estuary, Russia Baseline (Norwegian Coastal Sediments)

310

21

84

23

38 18 N/A N/A N/A 0.6

3.1 3.1 2.1 88 0.8 0.2

8.9 9.6 7.4 28 1.4 2.2

4.9 3.3 2.4 28 1.1 1.9

N/A, not available.

is Cd4Pb4Cu, Zn; reflecting the relative significance of anthropogenic inputs to the estuarine environment and modification by different particle– water reactivities. The Scheldt Estuary has a history of metal pollution; and in comparison with other estuaries, Baeyens (see Further Reading) classifies the Scheldt as ‘moderately polluted for all metals in the dissolved phase and fairly highly polluted in the

particulate phase, especially for cadmium.’ Efforts are being made to reduce the concentrations of cadmium in SPM in the Scheldt, and in the low-salinity region concentrations have declined from 400 nmol g1 in 1978 to 79 nmol g1 in 1995. Metal contamination can also be registered in indicator organisms (that is, organisms that are able to accumulate metals rather than regulate them), affording a measure of the contamination of the marine food chain. Across a broad range of phyla (from macroalgae to dophins), copper, zinc, and possibly cadmium exhibit a relatively low spread of concentrations, indicating efficient regulation of these metals. For lead, the concentrations are lower in vertebrates, which may be due to effective regulation or reduced bioavailability of the metal. Mercury is exceptional because of its biomagnification along the food chain as a result of its being present mainly as methyl mercury which is eliminated slowly from the organism. Mercury is retained by long-lived species, such as seals and dolphins, owing to their biochemical ability to isolate mercury as mercuric selenide granules. Fucus is a representative indicator species and baseline concentrations of selected metals occur in samples from an uncontaminated area of the Western Irish Sea (Table 4). Concentrations of copper, lead, and zinc are generally highest in industrialized estuaries (e.g., Humber) and fiords (e.g., Hardangefjord) and those that drain mineralized catchments and old mine workings (e.g., Restronguet Creek). Elevated concentrations of metals impact the growth, respiration, reproduction, recruitment and species diversity of marine organism, for example in Restronguet Creek the absence of bivalves has been ascribed to high levels of copper and zinc which prevent the settlement of juvenile bivalves.

Human Health The toxicity of metals depends on their rate of excretion from an organism and their chemical form. The adverse effects of metals on human health were

(c) 2011 Elsevier Inc. All Rights Reserved.

774

METAL POLLUTION

recognized in the 1950s and 1960s following catastrophic events involving mercury. Inorganic mercury is normally excreted by humans and poses little hazard to the general public. However, organic mercury compounds, such as methylated forms, are not readily excreted. Methylated mercury compounds can pass to all tissues in the body after absorption from the gastrointestinal tract. They can cross diffusion barriers and penetrate membranes, such as the blood–brain barrier (causing irreversible brain damage) and the placenta (rendering methylated mercury concentrations in fetal blood higher than those in the mother). The most significant outbreak of neurological, and often fatal illnesses occurred among the residents of Minamata in Japan. Chemical companies had discharged or dumped tonnes of mercury compounds into Minamata Bay for decades and these accumulated in the tissues of shellfish and fish. Consequently, large doses were passed onto the local, fish-eating population. Eventually the Bay was sealed off with nets to prevent organisms contaminated with mercury from escaping and affecting other areas. Over several decades biogeochemical processes and hydrodynamic flushing of mercury from the Bay have resulted in concentrations of mercury falling below government standards. Presently, the World Health Organization (WHO) regards a tolerable daily intake of total mercury (inorganic þ organic) to be 50 mg d1 for an adult of 70 kg. Human exposure to high concentrations of cadmium are rare and current concern centers around the chronic toxicity caused by long-term exposure to low levels of the metal. Bone disorders are one manifestation of chronic cadmium exposure. Cadmium is present in all tissues of adults, with the most significant amounts found in the liver and kidney, and the concentrations tend to increase with age. The WHO regards a tolerable daily intake of cadmium to be 70 mg d1 for an adult of 70 kg.

Conclusions Despite coastal waters in the vicinity of urban and industrial regions being contaminated with metals, there exists no evidence of significant pollution that poses a threat to human health, except on a local scale (and usually in shellfish) or where control has been poor. Since the dominant temporary or ultimate sink for metal contaminants is the sediment, an important goal of current research is to understand the mechanisms and extent to which contaminants are extracted by organisms (i.e., contaminant bioavailability) and transferred within the marine food chain.

Glossary Advection Horizontal water motion. Bioavailable metals Dissolved and particulate metals that are accessible to organisms during normal metabolic activity. Bioaccumulative metals Metals that can be regulated and reside in the organism and are added to over its life. Biomagnified metals Metals that are not regulated by organisms that can acquire an even larger body burden of metals. Bioturbation Reworking of bottom sediment by burrowing marine organisms. Diagenesis Release of particulate metals into the dissolved phase under suboxic conditions. Flushing time The time required for an existing body of water to be exchanged with surrounding water. Upwelling Vertical, upward movement of water at the shelf break, often tidally induced.

See also Aeolian Inputs. Anthropogenic Trace Elements in the Ocean. Antifouling Materials. Atmospheric Input of Pollutants. Estuarine Circulation. Land– Sea Global Transfers. Metalloids and Oxyanions. Ocean Margin Sediments. Pollution: Effects on Marine Communities. Pore Water Chemistry. Refractory Metals. Regional and Shelf Sea Models. River Inputs. Temporal Variability of Particle Flux. Transition Metals and Heavy Metal Speciation. Shelf Sea and Shelf Slope Fronts.

Further Reading Baeyens W (1998) Evolution of trace metal concentrations in the Scheldt estuary (1978–1995). A comparison with estuarine and ocean levels. Hydrobiologia 366: 157--167. Clark RB (1998) Marine Pollution 4th edn. Oxford: Clarendon Press. Ebinghaus R (ed.) (1997) Regional and Global Cycles of Mercury: Source, Fluxes and Mass Balances. NATO Series, Amsterdam: Kluwer Press. GESAMP: Group of Experts on Scientific Aspects of Marine Pollution (1990) The State of the Marine Environment. Nairobi, Kenya: United Nations Environment Programme. Langston WJ and Bebianno MJ (eds.) (1998) Metal Metabolism in Aquatic Environments. London: Chapman and Hall. Lowry R, Cramer RN, and Rickards LJ (1992) North Sea Project CD ROM and Users Guide. Swindon: British

(c) 2011 Elsevier Inc. All Rights Reserved.

METAL POLLUTION

Oceanographic Data Centre, Natural Environment Research Council of the United Kingdom. Mantoura RFC, Martin J-M, and Wollast R (1991) Ocean Margin Processes in Global Change. Dahlem Workshop Reports No. 9. Chichester: Wiley. Oslo and Paris Commissions (1993) North Sea Quality Status Report – 1993. Fredensborg, Denmark: Olsen & Olsen.

775

Salbu B and Steinnes E (eds.) (1995) Trace Elements in Natural Waters. Boca Raton, FL: CRC Press. Sindermann CJ Ocean Pollution: Effects on Living Resources and Humans. Boca Raton, FL: CRC Press. Wu J and Boyle EA (1997) Lead in the western North Atlantic Ocean: completed response to leaded gasoline phaseout. Geochimica et Cosmochimica Acta 61: 3279--3283.

(c) 2011 Elsevier Inc. All Rights Reserved.

METALLOIDS AND OXYANIONS G. A. Cutter, Old Dominion University, Norfolk, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1737–1745, & 2001, Elsevier Ltd.

Introduction The concentrations and distributions of dissolved trace elements (typically called trace ‘metals,’ though not all trace elements are metals) in the world’s oceans are due to a complex interaction between their purely chemical behavior (e.g., acid/base properties, oxidation state, solubility), the way in which they are delivered to the ocean (atmosphere, rivers, submarine hydrothermal vents), biological reactions, and water circulation (e.g., currents). To organize this somewhat chaotic and confusing situation, the kinds of trace element behavior are classified into four types: conservative, nutrient-like or recycled, scavenged, and hybrid or mixed. A conservative trace element behaves like the major dissolved elements that make up the bulk of the ocean’s salinity (e.g., Naþ). These elements are only effected by the physical processes of mixing, or the addition (dilution) or removal (evaporation) of water. Since there are no chemical or biological reactions that affect these elements, they have rather uniform concentrations with ocean depth. In contrast, the nutrientlike trace element is taken up by phytoplankton in surface waters during photosynthesis (like the nutrient nitrate), and this organic matter-bound element begins to gravitationally settle into deep waters. However, organic matter is a precious commodity in the open and deep sea, so many levels of the food web (bacteria to zooplankton) consume this organic detritus, releasing some fraction of the bound trace element back into the water column. This recycling makes the nutrient-like trace element concentration lower at the surface and higher at depth, with the exact shape of the profile depending on the rate at which it is recycled. Many dissolved trace elements have high charge to atomic radius ratios, and electrostatically adsorb to particle surfaces; this process is loosely termed ‘scavenging.’ Thus, scavenged elements have distributions that depend on the number and type of particles (e.g., clay, phytoplankton) and the mode of introduction (e.g., atmosphere, hydrothermal vents). An excellent example of this interaction is given by lead, which is

776

very particle-reactive and is introduced from atmosphere, with the resulting distribution showing a surface maximum and rapid decrease with depth. Finally, many trace elements display features of both scavenged and nutrient-like elements, with the distribution of the micronutrient iron being a good example (particle reactive, but also recycled). Most of these classifications were developed for the metals that are cations (positively charged) in solution, but there are elements in periodic groups IVA, VA and B, VIA and B, VIIA, and VIII that actually form oxygen-containing anions. In general, these elements display their maximum potential oxidation state in sea water, and in aqueous solution undergo hydrolysis (e.g., Moþ6 þ 4H2O-MoO2 4 þ 8Hþ). The metalloid elements (antimony, Sb; arsenic, As; germanium, Ge; selenium, Se; tellurium, Te) all exist as oxyanions, as do the transition metals chromium (Cr), molybdenum (Mo), osmium (Os), rhenium (Re), tungsten (W), and vanadium (V). In addition to existing as anions, most of these trace elements can be found in multiple oxidation states (e.g., As(III) and As(V)), ensuring that the oxyanions probably have the most diverse behaviors of any trace elements in the ocean. Interestingly, the form in which an oxyanion exists in sea water, the ‘chemical speciation,’ strongly affects its biological and chemical reactivity. In this review, the oceanic distributions of each element will be discussed in terms of its purely chemical properties, known biological behavior, and general geochemical considerations (inputs and outputs). The focus will be primarily on the dissolved ions rather than those associated with particles, since these are free to move with the water molecules and are available to the first trophic level in the ocean, phytoplankton. Because deeper waters in the Pacific Ocean are much older and have undergone more mixing than those in the Atlantic, most data will come from the Pacific to focus on the biological and chemical processes affecting the element, and not the mixing of different water masses. In addition, all concentrations will be expressed in fractions of a mole per liter rather than as mass per liter. This allows direct comparisons between elements (i.e., atom to atom) and is consistent with the principles of chemical and biological reactions (e.g., to make CO2, it takes one atom of C and 2 atoms of O). As trace elements, the concentrations units will be nanomoles per liter (nmol l1; 109 mol l1), picomoles per liter (pmol l1; 1012 mol l1), and femptomoles per liter

(c) 2011 Elsevier Inc. All Rights Reserved.

METALLOIDS AND OXYANIONS

(fmol l1; 1015 mol l1). Nevertheless, low concentrations do not mean that the oxyanions are unimportant as either essential or toxic compounds, or as useful ocean tracers; these points will be highlighted below. To provide a logical order to this presentation, the periodic table will be followed from left to right.

The Elements Vanadium

In oxygenated sea water with an average pH of 8, vanadium should be found in the þ 5 oxidation state, which undergoes hydrolysis to form vanadate, HVO4 2 . In sea water with no oxygen (‘anoxic’) such as found in the Black Sea, V(V) can be reduced to V(IV) which is more reactive than V(V), and as a consequence, anoxic sediments have elevated concentrations of vanadium relative to sediments underlying oxic waters. Thus, sedimentary vanadium may act as a historical tracer of anoxic conditions. Vanadium has also been studied in sea water because it is enriched in fossil fuels and may be a potential pollutant. In this respect, vanadium does not have any established biological function, although the chemistries of phosphate and vanadate are similar, and _1

777

therefore vanadate might be taken up into soft tissues (e.g., lipids) along with phosphate. The depth profile of dissolved vanadium in the North Pacific Ocean (Figure 1A) shows a surface concentration of B32 nmol l 1 and an increase into deep waters to B36 nmol l 1. This slight surface depletion has also been observed at other locations in the Pacific and Atlantic Oceans and, based on their similarity to the depth profiles of phosphate, it appears that vanadium is taken up by phytoplankton in surface waters. This type of behavior is also found in estuaries where river and sea waters mix, and both phosphate and vanadium show removal. This means that processes at the ocean margins (e.g., in estuaries) reduce the amount of vanadium entering the oceans from rivers. In contrast, hydrothermal vents do not appear to be substantial sources or sinks of vanadium to the deep ocean. Chromium

Owing to its use in many industrial processes and its high toxicity, considerable attention has been paid to chromium in group VIA. The two primary oxidation states of chromium are þ 6 and þ 3, which hydrolyze in water to form chromate, CrO4 2 , and Cr(OH)3, respectively. _1

_

1 Vanadium (nmol l ) Molybdenum (nmol l ) Tungsten (pmol l ) 50 40 80 120 0 75 25 0 10 20 30 40 0

_1

Rhenium (pmol l ) 0 10 20 30 40 50

0

Depth (m)

1000

2000

3000

4000

5000 (A)

(B)

(C)

(D)

Figure 1 (A) Dissolved vanadium in the North Pacific Ocean, 111N, 1401W. (Data from Collier RW (1979) Particulate and dissolved vanadium in the North Pacific Ocean. Nature 309: 441–444.) (B) Dissolved molybdenum in the North Pacific Ocean, 301N, 1591500 W. (Data from Sohrin Y, Isshiki K and Kuwamoto T (1987) Tungsten in North Pacific waters. Marine Chemistry 22: 95–103.) (C) Dissolved tungsten in the North Pacific Ocean, 301N, 1591500 W. (Data from Sohrin Y, Isshiki K and Kuwamoto T (1987) Tungsten in North Pacific waters. Marine Chemistry 22: 95–103.) (D) Dissolved rhenium in the North Pacific Ocean, 241160 N, 1691320 W. (Data from Colodner D, Sachs J, Ravizza G, Turekian K, Edmond J and Boyle E (1993) The geochemical cycle of rhenium: a reconnaissance. Earth and Planetary Science Letters 117: 205–221.)

(c) 2011 Elsevier Inc. All Rights Reserved.

778

METALLOIDS AND OXYANIONS

Thermodynamic calculations show that chromate is the expected form in oxygenated sea water, while the insoluble Cr(III) species would predominate in very low-oxygen (so called ‘suboxic’) or anoxic waters. However, it is important to note that thermodynamic calculations only predict elemental speciation at equilibrium (when the rates of formation and destruction are balanced), but they do not consider the rates of conversion themselves. For example, Cr(III) should not exist in oxygenated sea water, but its rate of oxidation to Cr(VI) is slow (days to months), meaning that Cr(III) can persist in oxic water (‘kinetic stabilization’). In the eastern North Pacific Ocean, Cr(VI) displays a surface concentration of B3 nmol l 1 (Figure 2A), but then decreases rapidly to a minimum of 1.7 nmol l1 at 300 m depth and increases below this to levels of 4–5 nmol l1 in the deeper waters. While chromate appears to display a mixture of scavenged and nutrient-like behavior, the Cr(VI) minimum occurs at the same depth as the widespread suboxic zone in the eastern Pacific.

1

2

3

4

5

Molybdenum

Because many trace elements such as iron function as essential nutrients in the ocean, considerable attention was paid to molybdenum, since it is a cofactor in the nitrogen-fixing enzyme nitrogenase,

_1

_1

_1

Chromium(III) (nmol l )

Chromium(VI) (nmol l ) 0

Indeed, at other sites in the North Pacific without a suboxic layer, Cr(VI) has only nutrient-like profiles. Thus, the data in Figure 2A suggest Cr(VI) to Cr(III) reduction, and correspondingly, Cr(III) shows a maximum at the same depth (Figure 2B), although the increase in Cr(III) (B0.6 nmol l 1) is not as great as the Cr(VI) depletion (B1.3 nmol l 1). This is likely due to the higher reactivity of Cr(III), which would be scavenged by particles, decreasing its concentration. While this all might seem in agreement with thermodynamic predictions, the existence of Cr(III) in fully oxygenated surface and deep waters (Figure 2B), argues that Cr(III) is kinetically stabilized (slow to oxidize).

0

0.5

1.0

Osmium (fmol l ) 20

0

40

60

0

1000

Depth (m)

2000

3000

4000

5000 (A)

(B)

(C)

Figure 2 Dissolved Cr(VI) (A) and Cr(III) (B) in the eastern North Pacific Ocean, 231N, 1151W. (Data from Murray JW, Spell B and Paul B (1983) The contrasting geochemistry of manganese and chromium in the eastern tropical Pacific Ocean. In: Wong CS et al. (eds) Trace Metals in Seawater, NATO Conference Services 4: Marine Science vol. 9, pp. 643–668. New York: Plenium Press.) (C) Dissolved rhenium in the eastern North Pacific Ocean, 91460 N, 1041110 W. (Data from Woodhouse OB, Ravizza G, Falkner KK, Statham PJ and Peucker-Ehrenbrink B (1999) Osmium in seawater: vertical profiles of concentration and isotopic composition in the eastern Pacific Ocean. Earth and Planetary Science Letters 173: 223–233.)

(c) 2011 Elsevier Inc. All Rights Reserved.

METALLOIDS AND OXYANIONS

and therefore essential for many algae and ecosystems. As a group VIA element like chromium, in oxic sea water dissolved molybdenum should be found in the þ 6 oxidation state as the hydrolysis product molybdate, MoO4 2 , but in anoxic waters it is reduced to Mo(IV), which in the presence of hydrogen sulfide forms insoluble MoS2. In fact, molybdenum is enriched in anoxic sediments by this mechanism and, like vanadium, can be used as a sediment tracer of past anoxia. In spite of its crucial biological role, molybdenum in the oxic ocean shows remarkably conservative behavior (Figure 1B), with no surface depletion and the highest concentration (B105 nmol l 1) of any of the trace elements examined here. This does not mean that phytoplankton are not taking it up, but rather this uptake is trivial compared to its inputs. Hydrothermal vents also do not remove molybdenum, and the only waters where molybdenum removal is observed are in anoxic basins such as the Black Sea.

779

for dating very ancient (450 billion years) sediments (i.e., the 187Re/187Os ratio). In oxic sea water, Re(VII) is the stable oxidation state and after hydrolysis exists as the relatively unreactive perrhenate ion, ReO 4 . In the North Pacific Ocean (Figure 1D), as well as in the Atlantic, the profile of dissolved rhenium is quite conservative, with an average of 44.370.3 pmol l1. Using the same type of arguments used for tungsten, the very low concentration of crustal rhenium but relatively high sea water concentration (sea water rhenium and tungsten are nearly identical, but crustal tungsten is B3000 times more abundant than rhenium) suggests that rhenium is very unreactive in the entire ocean system. Nevertheless, in anoxic systems such as the Black Sea, rhenium does show substantial decreases in concentration (nonconservative behavior) that have been attributed to removal at the surface of anoxic sediments. However, in the modern ocean these anoxic systems are too rare to substantially alter the distributions of rhenium in the water column.

Tungsten

Completing the group VIA transition metal series, tungsten is chemically similar to molybdenum and is found in oxic sea water as W(VI) in the form of tungstate, WO4 2 . Owing to difficulties in determining tungsten in sea water, there are few depth profiles for this dissolved element. In the North Pacific Ocean (Figure 1C), tungsten displays slightly higher concentrations in surface waters (average of 58 pmol l1) compared to deep waters (average of 49 pmol l1). Dissolved tungsten appears to have a slightly scavenged type of profile, with the surface enrichment likely due to the deposition of terrestrial dust (aerosols). In crustal rocks that are the source of most elements to the ocean by weathering, the abundance of tungsten is about one-third that of molybdenum, but its sea water concentration is over 1000 times less (compare Figures 1B and C). Since tungsten shows no strong removal in the open ocean, this observation suggests that tungsten must be removed in the coastal ocean. Indeed, profiles of tungsten in estuaries shows strong removal like that of iron; molybdenum shows no such strong removal in estuaries. Thus, while tungsten shows nearly conservative behavior in the open ocean, when the whole land–ocean system is considered, tungsten actually has a very active removal, which lowers its sea water concentration. Rhenium

Interest in rhenium is for rather esoteric reasons, primarily because one of its radioactive isotopes, and that of its periodic table neighbor osmium, are useful

Osmium

From the prior discussion of rhenium, it would seem logical to consider the oceanic behavior of the group VIII element osmium in terms of its use as a dating tool. In addition, the ratio of two of its isotopes (187Os/186Os) can trace inputs from extraterrestrial sources (e.g., meteors) and terrestrial sources (i.e., crustal weathering) to the oceans, which are recorded in marine sediments. There is some debate about the exact form of osmium in sea water, but thermodynamic calculations suggest that Os(VIII) as would be stable in oxic sea water. DeH3OsO2 6 terminations of dissolved osmium in sea water are very difficult, especially since osmium concentrations are over 1000 times lower than those of rhenium. Indeed, the profile of osmium in the North Pacific (Figure 2C) indicates not only that osmium is found at very low concentrations (38 fmol l1) in surface waters, but also that its distribution is quite dynamic with depth (minimum at 460 m and rising to 51 fmol l1 in deep waters). This profile is from the eastern North Pacific, which has the distinct suboxic layer, and bears a striking resemblance to that of Cr(VI) in the same region (Figure 2A). Thus, osmium may have nutrient-like behavior that is also affected by oxidation–reduction reactions (i.e., reduced to a more particle-reactive (but unidentified) form in the suboxic zone). In this respect, profiles in more oxygenated waters of the Atlantic and Indian Oceans show no such depletion in the upper water column. The nutrient-like distribution does not mean that it is used as a nutrient or nutrient substitute, but rather

(c) 2011 Elsevier Inc. All Rights Reserved.

780

METALLOIDS AND OXYANIONS

that it is only carried along in the organic matter cycle. Germanium

Germanium is a group IVB metalloid, and therefore chemically quite different from the transition metals we have been considering so far. Being directly below silicon in the periodic table, germanium has quite similar chemistry and early studies of phytoplankton that make up their structural skeletons (‘tests’) of biogenic silica (e.g., diatoms), showed that they take up germanium along with silicon in a relatively constant atomic Ge : Si ratio of 106 : 1, a value that is nearly identical to the ratio in crustal rocks. Dissolved germanium exists as germanic (H3GeO4) in sea water just as silicon is found as silicic acid (H3SiO4). The depth profile of dissolved inorganic germanium in the North Pacific Ocean (Figure 3A) is undoubtedly nutrient-like, and not surprisingly exactly the same as that of silicon (i.e., uptake in the surface by siliceous phytoplankton; recycling at depth by the slow dissolution of biogenic silica). This _1

Inorganic Ge (pmol l ) 0

25 50 75 100 125

_1

MethylGe (nmol l ) 0 100 200 300 400

0

covariation is the primary reason why there is interest in marine germanium. The Ge : Si ratio derived from crustal weathering inputs to the ocean has a different value from that from hydrothermal vent fluids, and since it appears that the Ge : Si ratio in siliceous organisms records the water column value, the ancient record of crustal weathering versus hydrothermal inputs (i.e., plate spreading) to the oceans can be obtained. Unfortunately, it was discovered that there are methylated forms of germanium in sea water (monomethylgermanic acid, MMGe; dimethylgermanic acid, DMGe) which actually have higher concentrations than inorganic germanium, and for which there are no known silicon analogues. Thus, the two cycles seem to diverge, threatening the usefulness of the Ge : Si tracer. However, the distributions of these two methylated forms are very conservative in the open ocean (Figure 3B) and in estuaries. Indeed, methylgermanium is essentially inert and hence can build up to the observed ‘high’ concentrations. The source of these compounds still has not been found, although some production in anoxic, organic-rich waters has been documented. Nevertheless, the methylgermanium compounds do not really participate in the germanium cycle, and thus it seems that the Ge : Si ratio can still be used as a weathering versus hydrothermal input tracer. Arsenic

1000

Depth (m)

2000

3000

MMGe DMGe

4000

5000 (A)

(B)

Figure 3 (A) Dissolved inorganic germanium in the North Pacific Ocean, 251N, 1751E. (Data from Froehlich PN Jr and Andreae MO (1981) The marine geochemistry of germanium: ekasilicon. Science 213: 205–207.) (B) Dissolved methylgermanium compounds in the North Pacific Ocean, 251N, 1751E. MMGe is monomethylgermanic acid and DMGe is dimethylgermanic acid. (Data from Lewis BL, Froelich PN and Andreae MO (1985) Methylgermanium in natural waters. Nature 313: 303– 305.)

The group VB element arsenic is usually linked with toxicity, and indeed most studies of this element are driven by such concerns. However, arsenic’s toxicity is strongly affected by its chemical form and by the actual organisms being exposed. In oxygenated sea water, As(V) in the form of arsenate (HAsO4 2 ) is the stable form and, because of its nearly identical chemical properties to that of the nutrient phosphate, is highly toxic to phytoplankton. Interestingly, As(III), which can be found in anoxic waters as arsenite (As(OH)3), is not toxic to phytoplankton but is highly toxic to higher organisms such as zooplankton and fish. While this might seem irrelevant (fish do not live in anoxic waters), many phytoplankton have a mechanism to detoxify arsenate by reducing it to arsenite and releasing it to the oxic water column. Other phytoplankton can methylate arsenate to form monomethyl- and dimethylarsenates (MMAs and DMAs, respectively), which are nontoxic. All of these processes make the marine arsenic cycle quite complicated, a feature that is common for all of the metalloid elements. In the North Pacific Ocean (Figure 4A) arsenate has nutrient-like behavior, with depletion in the surface and recycling at depth as for

(c) 2011 Elsevier Inc. All Rights Reserved.

METALLOIDS AND OXYANIONS

0

Arsenic _1 (nmol l ) 5 10 15 20 25

Methylarsenic _1 (nmol l ) 0

Antimony _1 (nmol l )

0.5 1.0 1.5 2.0

0 0.1 0.2 0.3 0.4 0.5

781

Antimony(V) _1 (pmol l ) 0 0.5 1.0 1.5 2.0

0

1000

Depth (m)

2000

3000 As(III) As(V)

MMAs DMAs

Sb(III) MMSb

4000

5000 (A)

(B)

(C)

(D)

Figure 4 (A) Arsenate (As(V)) and arsenite (As(III)), and (B) monomethylarsenate (MMAs) and dimethylarsenate (DMAs) in the North Pacific Ocean, 301460 N, 1631300 W. (Data from Andrese MO (1979) Arsenic speciation in seawater and interstitial waters: the influence of biological–chemical interactions on the chemistry of a trace element. Limnology and Oceanography 24: 440–452.) (C) Dissolved antimony(III) (Sb(III)) and monomethyl antimonate (MMSb), and (D) antimony(V) in the South Atlantic Ocean, 171S, 251W. (Data from Cutter GA, Cutter LS, Featherstone AM and Lohrenz SE (2001). Antimony and arsenic biogeochemistry in the western Atlantic Ocean. Deep-Sea Research, in press.)

phosphate; this is consistent with their chemistries and biochemistries. In surface waters, biologically produced arsenite and methylated arsenic compounds have their maximum concentrations (Figures 4A and B), and quickly decrease with depth. Again, these distributions are consistent with established biological processes, and therefore they are not due to scavenging (i.e., input from the atmosphere and then adsorption to particles). Moreover, data for methylated arsenic and As(III) at many sites in the world’s oceans show that the amounts of these arsenic forms are roughly the inverse of the phosphate concentration – lower phosphate, higher methylarsenic and As(III). This is consistent with laboratory studies of phytoplankton where, under low-phosphate conditions, more arsenate is taken up and more methylated or reduced arsenic is produced in response to this stress. In addition, the fact that As(III) is found in oxygenated sea water at all (i.e., only stable in anoxic waters) demonstrates that its rate of oxidation is slow enough (half-life of months) that it can build up to almost 20% of the total dissolved arsenic in surface waters. Antimony

Antimony is a group VB metalloid like arsenic but it has more metallic character and the chemistry of

antimony is quite different than that of phosphorus or arsenic. Sb(V) is not as strong a Lewis acid as As(V) and in oxic sea water the stable form would be antimonate (Sb(OH)6  ), while Sb(III), like As(III), would be Sb(OH)3 in anoxic waters. Antimony also has methylated forms analogous to those of arsenic (i.e., MMSb, DMSb), although only MMSb has been found in the open ocean. Antimony is not as toxic as arsenic, and since it is used as a plasticizer and is enriched in fossil fuels, most of the interest in antimony has concerned its use as a pollution tracer (e.g. from the burning of plastics). Most of the data for dissolved antimony in the open ocean are from the Atlantic; the profiles in Figure 4C and D are typical for this element. The major form of dissolved antimony is antimonate, and it displays a profile consistent with mild scavenging (i.e., maximum of 1.5 nmol l1 at the surface due to atmospheric or riverine input, lower concentrations below the surface layer via adsorption onto particles, some recycling near the sediment–water interface). Measurements of antimony in atmospheric particles (aerosols) and rain show that atmospheric input can explain the surface antimony maximum. Thus, the concentration and behavior of antimony are quite different from those of arsenic. However, MMSb and Sb(III) are found in the surface waters (although the concentration of SbIII is only 0.02 nmol l1), like the

(c) 2011 Elsevier Inc. All Rights Reserved.

782

METALLOIDS AND OXYANIONS

equivalent arsenic forms, but their concentrations do not negatively correlate with the concentration of phosphate (i.e., do not appear to be a result of detoxification reactions). Indeed, it is not clear what mechanisms are producing these forms of dissolved antimony, and bacterial production of MMSb (bacteria are good methylators) and the photochemical reduction of Sb(V) to Sb(III) cannot be ruled out. Selenium

Selenium is a group VIB metalloid just below sulfur in the periodic table, and its chemistry and biochemistry is very similar to those of sulfur. The interest in this element is based on the fact that this trace element is both essential (e.g., a cofactor in antioxidant enzymes) and toxic, with the chemical form of the element strongly influencing its beneficial or toxic properties. As for sulfur, the most stable oxidation state in oxygenated sea water is Se(VI) as selenate (SeO4 2 ), while under suboxic conditions selenite (HSeO3  ) would predominate. Selenium forms insoluble elemental Se(0) in anoxic waters, whereas sulfur exists as sulfide (S(II)). Nevertheless, there are numerous organic forms of selenide (Se(-II)) such as selenomethionine, that could be bound in soluble peptides (to be referred to as ‘organic selenide’). There have also been recent measurements of the dissolved gas dimethylselenide ((CH3)2Se) in surface ocean, which results in a natural selenium input to the atmosphere. Laboratory and field studies have shown that selenite appears to be the most biologically preferred form of dissolved selenium by Selenium(IV) _1 (nmol l ) 0

0.5 1.0 1.5 2.0

Selenium(VI) _1 (nmol l ) 0

0.5 1.0 1.5 2.0

phytoplankton, while selenate is only taken up in the absence of selenite; data on the bioavailability of organic selenides suggest that these forms are the least available to marine phytoplankton. Depth profiles for dissolved selenium in the eastern North Pacific (Figure 5A–C) are from the same region where the chromium and osmium profiles (Figure 2A–C) were obtained, and the water column from B200 to 800 m is suboxic. In surface waters, both selenite (Figure 5A) and selenate (Figure 5B) are very depleted, and then show nutrient-like profiles with increasing depth; this is consistent with biotic uptake of both forms, incorporation into organic matter (as organic selenides), and subsequent recycling. In contrast, organic selenide (Figure 5C) has a maximum at the surface and in the suboxic zone. Laboratory and field studies using organic matter show that selenium is primarily bound in proteins as organic selenide. When this organic matter degrades, dissolved organic selenide is released, which then sequentially oxidizes to selenite and then very slowly to selenate. This process explains how these unstable forms are introduced to the water column, with kinetic stabilization allowing them to persist. In the suboxic zone, organic selenide is stabilized and a maximum can develop. Tellurium

Tellurium is a group VIB metalloid like selenium, but its lower position in the periodic table suggests that it has considerably more metallic character than selenium. Thus, Te(VI) exists as Te(OH)6 but, unlike

Organic Se(II) _1 (nmol l ) 0

0.5 1.0 1.5 2.0

Tellurium(IV) _1 (pmol l ) 0 0.1 0.2 0.3 0.4 0.5 0

Tellurium(VI) _1 (pmol l ) 0.5 1.0 1.5 2.0

0

Depth (m)

1000

2000

3000

4000

5000

(A)

(B)

(C)

(D)

(E)

Figure 5 (A) Selenium(IV) (selenite), (B) selenium(VI) (selenate), and (C) organic selenide (organic Se(II)) in the eastern North Pacific Ocean, 181N, 1081W. (Data from Cutter GA and Bruland KW (1984) The marine biogeochemistry of selenium: a re-evaluation. Limnology and Oceanography 29: 1179–1192.) (D) Tellurium(IV) and (E) tellurium(VI) in the eastern North Pacific Ocean, 71N, 781400 W. (Data from Lee DS and Edmond JM (1985) Tellurium species in seawater. Nature 313: 782–785.)

(c) 2011 Elsevier Inc. All Rights Reserved.

METALLOIDS AND OXYANIONS

Se(VI), Te(IV) is the stable form as Te(OH)4. There are few data for this element in the ocean, and profiles of tellurium in the eastern North Pacific (Figures 5D and E) show that its concentrations are B1000 times less than those of selenium and both forms of tellurium show strongly scavenged behavior. It is interesting to note that the most abundant form of tellurium is Te(VI), but it is the least thermodynamically stable. Although there are numerous biological reasons to expect reduced species in oxic waters (i.e., as for arsenic and selenium), this observation is somewhat difficult to explain. The elevated concentrations of Te(VI) at the surface, relative to Te(IV), suggest that the atmospheric or riverine inputs of this element are enriched in this form, but virtually no atmospheric data are available to confirm this speculation.

783

categories. Finally, the types of behavior found for these elements are strongly affected by both chemical and biological processes, and thus it is appropriate to say that they are ‘biogeochemically’ cycled.

See also Aeolian Inputs. Anthropogenic Trace Elements in the Ocean. Atmospheric Input of Pollutants. Carbon Cycle. Conservative Elements. Hydrothermal Vent Fluids, Chemistry of. Marine Silica Cycle. Metal Pollution. Nitrogen Cycle. Platinum Group Elements and their Isotopes in the Ocean. Pollution: Effects on Marine Communities. Refractory Metals. River Inputs. Transition Metals and Heavy Metal Speciation.

Conclusions Overall, the classic types of behavior described for trace elements are displayed by the oxyanions. Rhenium, molybdenum, and methylated germanium have conservative, salinity-like distributions with depth. This does not mean that they are truly unreactive, but rather that the reactions affecting them are minor. Vanadium, chromium, osmium, arsenic, and selenium show nutrient-like behavior, although this does not necessarily mean they are biologically required. Tungsten, antimony, and tellurium show scavenged behavior to some degree. Hybrid/mixed behavior was originally defined for trace elements with a single chemical form, but most of the oxyanions can exist in multiple chemical forms, creating a new type of hybrid distributions – oxidation–reduction behavior superimposed upon the other three

Further Reading Bruland KW (1992) Trace elements in seawater. In: Riley JP Chester R (eds.) Chemical Oceanography, vol. 8, pp. 157–220. London: Academic Press. Donat JR and Bruland KW (1995) Trace elements in the oceans. In: Salbu B and Steinnes E (eds.) Trace Elements in Natural Waters, pp. 247--281. Boca Raton, FL: CRC Press. Libes SM (1992) An Introduction to Marine Biogeochemistry. New York: Wiley. Millero FJ (1996) Chemical Oceanography, 2nd edn. Boca Raton, FL: CRC Press. Pilson MEQ (1998) An Introduction to the Chemistry of the Sea. Reading, NJ: Prentice Hall. Stumm W and Morgan JJ (1996) Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters. New York: Wiley.

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATES AND CLIMATIC EFFECTS B. U. Haq, Vendome Court, Bethesda, MD, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1757–1763, & 2001, Elsevier Ltd.

Introduction

Hydrate Stability and Detection

Natural gas hydrates are crystalline solids that occur widely in marine sediment of the world’s continental margins. They are composed largely of methane and water, frozen in place in the sediment under the dual conditions of high pressure and frigid temperatures at the sediment–water interface (Figure 1). When the breakdown of the gas hydrate (also known as clathrate) occurs in response to reduced hydrostatic pressure (e.g., sea level fall during glacial periods), or an increase in bottom-water temperature, it causes dissociation of the solid hydrate at its base, creating a zone of reduced sediment strength that is prone to structural faulting and sediment slumping. Such sedimentary failure at hydrate depths could inject large quantities of methane (a potent greenhouse gas) in the water column, and eventually into the atmosphere, leading to enhanced greenhouse warming. Ice core records of the recent geological past show that climatic warming occurs in tandem with rapid increase in atmospheric methane. This suggests that catastrophic release of methane into the atmosphere during periods of lowered sea level may have been a causal factor for abrupt climate change. Massive injection of methane in sea water following hydrate

Figure 1 A piece of natural gas hydrate from the Gulf of Mexico. (Photograph courtesy of I. MacDonald, Texas A & M University.)

784

dissociation during periods of warm bottom temperatures are also suspected to be responsible for major shifts in carbon-isotopic ratios of sea water and associated changes in benthic assemblages and hydrographic conditions.

Gas hydrate stability requires high hydrostatic pressure (45 bars) and low bottom-water temperature (o71C) on the seafloor. These requirements dictate that hydrates occur mostly on the continental slope and rise, below 530 m water depth in the low latitudes, and below 250 m depth in the high latitudes. Hydrated sediments may extend from these depths to c. 1100 m sub-seafloor. In higher latitudes, hydrates also occur on land, in association with the permafrost. Rapidly deposited sediments with high biogenic content are amenable to the genesis of large quantities of methane by bacterial alteration of the organic matter. Direct drilling of hydrates on the Blake Ridge, a structural high feature off the US east coast, indicated that the clathrate is only rarely locally concentrated in the otherwise widespread field of thinly dispersed hydrated sediments. The volume of the solid hydrate based on direct measurements on Blake Ridge suggested that it occupies between 0 and 9% of the sediment pore space within the hydrate stability zone (190–450 m sub-seafloor). It has been estimated that a relatively large amount, c. 35 Gt (Gigaton ¼ 1015 g), of methane carbon was tied up on Blake Ridge, which is equal to carbon from about 7% of the total terrestrial biota. Hydrates can be detected remotely through the presence of acoustic reflectors, known as bottom simulating reflectors (BSR), that mimic the seafloor and are caused by acoustic velocity contrast between the solid hydrate above and the free gas below. However, significant quantities of free gas need to be present below the hydrate to provide the velocity contrast for the presence of a BSR. Thus, hydrate may be present at the theoretical hydrate stability depths even when no BSR is observed. Presence of gas hydrate can also be inferred through the sudden reduction in pore-water chlorinity (salinity) of the hydrated sediments during drilling, as well as through gas escape features on land and on the seafloor. Global estimates of methane trapped in gas hydrate reservoirs (both in the hydrate stability zone

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATES AND CLIMATIC EFFECTS

and as free gas beneath it) vary widely. For example, the Arctic permafrost is estimated to hold anywhere between 7.6 and 18 000 Gt of methane carbon, while marine sediments are extrapolated to hold between 1700 and 4 100 000 Gt of methane carbon globally. Obtaining more accurate global estimates of methane sequestered in the clathrate reservoirs remains one of the more significant challenges in gas hydrate research. Another major unknown, especially for climatic implications, is the mode of expulsion of methane from the hydrate. How and how much of the gas escapes from the hydrate zone and how much of it is dissolved in the water column versus escaping into the atmosphere? In a steady state much of the methane diffusing from marine sediments is believed to be oxidized in the surficial sediment and the water column above. However, it is not clear what happens to significant volumes of gas that might be catastrophically released from the hydrates when they disintegrate. How much of the gas makes it to the atmosphere (in the rapid climate change scenarios it is assumed that much of it does), or is dissolved in the water column? Hydrate dissociation and methane release into the atmosphere from continental margin and permafrost sources and the ensuing accelerated greenhouse heating also have important implications for the models of global warming over the next century. The results of at least one modeling study play down the role of methane release from hydrate sources. When heat transfer and methane destabilization process in oceanic sediments was modeled in a coupled atmosphere–ocean model with various input assumptions and anthropogenic emission scenarios, it was found that the hydrate dissociation effects were smaller than the effects of increased carbon dioxide emissions by human activity. In a worst case scenario global warming increased by 10–25% more with clathrate destabilization than without. However, these models did not take into account the associated free gas beneath the hydrate zone that may play an additional and significant role as well. It is obvious from drilling results on Blake Ridge that large volumes of free methane are readily available for transfer without requiring dissociation.

Hydrate Breakdown and Rapid Climate Change The temperature and pressure dependency for the stability of hydrates implies that any major change in either of these controlling factors will also modify the zone of hydrate stability. A notable drop in sea

785

level, for example, will reduce the hydrostatic pressure on the slope and rise, altering the temperature– pressure regime and leading to destabilization of the gas hydrates. It has been suggested that a sea level drop of nearly 120 m during the last glacial maximum (c. 30 000 to 18 000 years BP) reduced the hydrostatic pressure sufficiently to raise the lower limit of gas hydrate stability by about 20 m in the low latitudes. When a hydrate dissociates, its consistency changes from a solid to a mixture of sediment, water, and free gas. Experiments on the mechanical strength behavior of hydrates has shown that the hydrated sediment is markedly stronger than water ice (10 times stronger than ice at 260 K). Thus, such conversion would create a zone of weakness where sedimentary failure could take place, encouraging low-angle faulting and slumping on the continental margins. The common occurrence of Pleistocene slumps on the seafloor have been ascribed to this catastrophic mechanism and major slumps have been identified in sediments of this age in widely separated margins of the world. When slumping occurs it would be accompanied by the liberation of a significant amount of methane trapped below the level of the slump, in addition to the gas emitted from the dissociated hydrate itself. These emissions are envisaged to increase in the low latitudes, along with the frequency of slumps, as glaciation progresses, eventually triggering a negative feedback to advancing glaciation, encouraging the termination of the glacial cycle. If such a scenario is true then there may be a built-in terminator to glaciation, via the gas hydrate connection. In this scenario, the negative feedback to glaciation, can initially function effectively only in the lower latitudes. At higher latitudes glacially induced freezing would tend to delay the reversal, but once deglaciation begins, even a relatively small increase in atmospheric temperature of the higher latitudes could cause additional release of methane from nearsurface sources, leading to further warming. One scenario suggests that a small triggering event and liberation of one or more Arctic gas pools could initiate massive release of methane frozen in the permafrost, leading to accelerated warming. The abrupt nature of the termination of the Younger Dryas glaciation (some 10 000 years ago) has been ascribed to such an event. Modeling results of the effect of a pulse of ‘realistic’ amount of methane release at the glacial termination as constrained by ice core records indicate that the direct radiative effects of such an emission event may be too small to account for deglaciation alone. However, with certain combinations of methane, CO2, and heat transport changes, it may be possible to simulate

(c) 2011 Elsevier Inc. All Rights Reserved.

786

METHANE HYDRATES AND CLIMATIC EFFECTS

changes of the same magnitude as those indicated by empirical data. The Climate Feedback Loop

The paleoclimatic records of the recent past, e.g., Vostock ice core records of the past 420 000 years from Antarctica, show the relatively gradual decrease in atmospheric carbon dioxide and methane at the onset of glaciations. Deglaciations, on the other hand, tend to be relatively abrupt and are associated with equally rapid increases in carbon dioxide and methane. Glaciations are thought to be initiated by Milankovitch forcing (a combination of variations in the Earth’s orbital eccentricity, obliquity and precession), a mechanism that also can explain the broad variations in glacial cycles, but not the relatively abrupt terminations. Degassing of carbon dioxide from the ocean surface alone cannot explain the relatively rapid switch from glacials to interglacials. The delayed response to glacially induced sea level fall in the high latitudes (as compared with low latitudes) is a part of a feedback loop that could be an effective mechanism for explaining the rapid warmings at the end of glacial cycles (also known as the Dansgaard-Oeschger events) in the late Quaternary. These transitions often occur only on decadal to centennial time scales. In this scenario it is envisioned that the low-stand-induced slumping and methane emissions in lower latitudes lead to greenhouse warming and trigger a negative feedback to glaciation. This also leads to an increase in carbon dioxide degassing for the ocean. Once the higher latitudes are warmed by these effects, further release of methane from near-surface sources could provide a positive feedback to warming. The former (methane emissions in the low latitudes) would help force a reversal of the glacial cooling, and the latter (additional release of methane from higher latitudes) could reinforce the trend, resulting in apparent rapid warming observed at the end of the glacial cycles (see Figure 2). The record of stable isotopes of carbon from Santa Barbara Basin, off California, has revealed rapid warmings in the late Quaternary that are synchronous with warmings associated with DansgaardOeschger (D-O) events in the ice record from Greenland. The energy needed for these rapid warmings could have come from methane hydrate dissociation. Relatively large excursions of d13 C (up to 5 ppm) in benthic foraminifera are associated with the D-O events. However, during several brief intervals the planktonics also show large negative shifts in d13 C (up to 2.5 ppm), implying that the

entire water column may have experienced rapid 12C enrichment. One plausible mechanism for these changes may be the release of methane from the clathrates during the interstadials. Thus, abrupt warmings at the onset of D-O events may have been forced by dissociation of gas hydrates modulated by temperature changes in overlying intermediate waters. For the optimal functioning of the negative–positive feedback model discussed above, methane would have to be constantly replenished from new and larger sources during the switchover. Although as a greenhouse gas methane is nearly 10 times as effective as carbon dioxide, its residence time in the atmosphere is relatively short (on the order of a decade and a half), after which it reacts with the hydroxyl radical and oxidizes to carbon dioxide and water. The atmospheric retention of carbon dioxide is somewhat more complex than methane because it is readily transferred to other reservoirs, such as the oceans and the biota, from which it can re-enter the atmosphere. Carbon dioxide accounts for up to 80% of the contribution to greenhouse warming in the atmosphere. An effective residence time of about 230 years has been estimated for carbon dioxide. These retention times are short enough that for cumulative impact of methane and carbon dioxide through the negative–positive feedback loop to be effective methane levels would have to be continuously sustained from gas hydrate and permafrost sources. The feedback loop would close when a threshold is reached where sea level is once again high enough that it can stabilize the residual clathrates and encourage the genesis of new ones. Several unresolved problems remain with the gas hydrate climate feedback model. The negative–positive feedback loop assumes a certain amount of time lag between events as they shift from lower to higher latitudes, but the duration of the lag remains unresolved, although a short duration (on decadal to centennial time scales) is implied by the ice core records. Also, it is not clear whether hydrate dissociation leads to initial warming, or warming caused by other factors leads to increased methane emissions from hydrates. Data gathered imply a time lag of c. 50 (710) years between abrupt warming and the peak in methane values at the Blling transition (around 14 500 years Bp), although an increase in methane emissions seems to have begun almost simultaneously with the warming trend (75 years). However, this does not detract from the notion that there may be a built-in feedback between increased methane emissions from gas hydrate sources and accelerated warming. If smaller quantities of methane released from hydrated sediments are

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATES AND CLIMATIC EFFECTS

787

Glaciation, and / or sea-level fall

Low_mid latitudes Reduced hydrostatic pressure on shelf /slope

Hydrate melting near the base

Low-angle faulting and largescale slumping

Release of methane into atmosphere

Reversal of glaciation and warming of high latitudes

Negative feedback to glacial cooling

High latitudes Dissociation of permafrost and shallow margin hydrates

Positive feedback to warming

Relatively abrupt termination of glacial cycle

Figure 2 The negative–positive feedback loop model of sea level fall, hydrate decomposition, and climate change (reversal of glaciation and rapid warming) through methane release in the low and high latitudes. (Adapted with permission from Haq, 1993.)

oxidized in the water column, initial releases of methane from dissociated hydrates may not produce a significant positive shift in the atmospheric content of methane. However, as the frequency of catastrophic releases from this source increases, more methane is expected to make it to the atmosphere. And, although the atmospheric residence time of methane itself is relatively short, when oxidized, it adds to the greenhouse forcing of carbon dioxide. This may explain the more gradual increase in methane, and is not inconsistent with the short temporal difference between the initiation of the warming trend and methane increase, as well as the time lag between the height of warming trend and the peak in methane values. Although there is still no evidence to suggest that the main forcing for the initiation of deglaciation is to be found in hydrate dissociation, once begun, a positive feedback of methane emissions from hydrate sources (and its by-product, carbon dioxide) can only help accelerate the warming trend.

Gas Hydrates and the Long-term Record of Climate Change

Are there any clues in the longer term geological record where cause and/or effect can be ascribed to gas hydrates? One potential clue for the release of significant volumes of methane into the ocean waters is the changes in d13 C composition of the carbon reservoir. The d13 C of methane in hydrates averages c.  60 ppm; perhaps the lightest (most enriched in 12 C) carbon anywhere in the Earth system. It has been argued that massive methane release from gas hydrate sources is the most likely mechanism for the pronounced input of carbon greatly enriched in 12C during a period of rapid bottom-water warming. The dissolution of methane (and its oxidative by-product, CO2) in the sea water should also coincide with increased dissolution of carbonate on the seafloor. Thus, a major negative shift in d13 C that occurs together with an increase in benthic temperature (bottom-water warming) or a sea level fall event

(c) 2011 Elsevier Inc. All Rights Reserved.

788

METHANE HYDRATES AND CLIMATIC EFFECTS

(reducing hydrostatic pressure) may provide clues to past behavior of gas hydrates. A prominent excursion in global carbonate and organic matter d13 C during the latest Paleocene peak warming has been explained as a consequence of such hydrate breakdown due to rapid warming of the bottom waters. The late Paleocene–early Eocene was a period of peak warming, and overall the warmest interval in the Cenozoic when latitudinal thermal gradients were greatly reduced. In the latest Paleocene bottom-water temperature also increased rapidly by as much as 41C, with a coincident excursion of about 2 to 3 ppm in d13 C of all carbon reservoirs in the global carbon cycle. A high resolution study of a sediment core straddling the Paleocene–Eocene boundary concluded that much of this carbon-isotopic shift occurred within no more than a few thousand years and was synchronous in oceans and on land, indicating a catastrophic release of carbon, probably from a methane source. The late Paleocene thermal maximum was also coincident with a major benthic foraminiferal mass extinction and widespread carbonate dissolution and low oxygen conditions on the seafloor. This rapid excursion cannot be explained by conventional mechanisms (increased volcanic emissions of carbon dioxide, changes in oceanic circulation and/or terrestrial and marine productivity, etc.). A rapid warming of bottom waters from 11 to 151C could abruptly alter the sediment thermal gradients leading to methane release from gas hydrates. Increased flux of methane into the ocean–atmosphere system and its subsequent oxidation is considered sufficient to explain the  2.5 ppm excursion in d13 C in the inorganic carbon reservoir. Explosive volcanism and rapid release of carbon dioxide and changes in the sources of bottom water during this time are considered to be plausible triggering mechanisms for the peak warming leading to hydrate dissociation. Another recent high-resolution study supports the methane hydrate connection to latest Paleocene abrupt climate change. Stable isotopic evidence from two widely separated sites from low- and southern high-latitude Atlantic Ocean indicates multiple injections of methane with global consequences during the relatively short interval at the end of Paleocene. Modeling results, as well as wide empirical data, suggest warm and wet climatic conditions with less vigorous atmospheric circulation during the late Paleocene thermal optimum. The eustatic record of the late Paleocene–early Eocene could offer further clues for the behavior of the gas hydrates and their contribution to the overall peak warm period of this interval. The longer term trend shows a rising sea level through the latest

Paleocene and early Eocene, but there are several shorter term sea level drops throughout this period and one prominent drop straddling the Paleocene– Eocene boundary (which could be an additional forcing component to hydrate dissociation for the terminal Paleocene event). Early Eocene is particularly rich in high-frequency sea level drops of several tens of meters. Could these events have contributed to the instability of gas hydrates, adding significant quantities of methane to the atmosphere and maintaining the general warming of the period? These ideas seem testable if detailed faunal and isotopic data for the interval in question were available with at least the same kind of resolution as that obtained for the latest Paleocene interval.

Timing of the Gas Hydrate Development

When did the gas hydrates first develop in the geological past? The specific low temperature–high pressure requirement for the stability of gas hydrates suggests that they may have existed at least since the latest Eocene, the timing of the first development of the oceanic psychrosphere and cold bottom waters. Theoretically clathrates could exist on the slope and rise when bottom-water temperatures approach those estimated for late Cretaceous and Paleogene (c. 7–151C), although they would occur deeper within the sedimentary column and the stability zone would be relatively slimmer. A depth of c. 900 m below sea level has been estimated for the hydrate stability zone in the late Palecene. If the bottom waters were to warm up to 221C only then would most margins of the world be free of gas hydrate accumulation. The implied thinner stability zone during warm bottomwater regimes, however, does not necessarily mean an overall reduced methane reservoir, since it also follows that the sub-hydrate free gas zone could be larger, making up to the hydrate deficiency. Prior to late Eocene there is little evidence of large polar ice caps, and the mechanism for short-term sea level changes remains uncertain. And yet, the Mesozoic–Early Cenozoic eustatic history is replete with major sea level falls of 100 m or more that are comparable in magnitude, if not in frequency, to glacially induced eustatic changes of the late Neogene. If gas hydrates existed in the pre-glacial times, major sea level falls would imply that hydrate dissociation may have contributed significantly to climate change and shallow-seated tectonics along continental margins. However, such massive methane emissions should also be accompanied by prominent d13 C excursions, as exemplified by the terminal Paleocene climatic optimum.

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATES AND CLIMATIC EFFECTS

The role of gas hydrate as a significant source of greenhouse emissions in global change scenarios and as a major contributor of carbon in global carbon cycle remains controversial. It can only be resolved with more detailed studies of hydrated intervals, in conjunction with high-resolution studies of the ice cores, preferably with decadal time resolution. A better understanding of gas hydrates may well show their considerable role in controlling continental margin stratigraphy and shallow structure, as well as in global climatic change, and through it, as agents of biotic evolution.

See also Paleoceanography, Climate Models in.

Further Reading Dickens GR, O’Neil JR, Rea DK, and Owen RM (1995) Dissociation of oceanic methane hydrate as a cause of the carbon isotope excursion at the end of the Paleocene. Paleoceanography 10: 965--971. Dillon WP Paul CK (1983) Marine gas hydrates II. Geophysical evidence. In: Cox JS (ed.) Natural Gas Hyrate. London: Butterswarth, pp. 73–90. Haq BU (1998) Gas hydrates: Greenhouse nightmare? Energy panacea or pipe dream? GSA Today, Geological Society of America 8(11): 1--6.

789

Henriet J-P and Mienert J (eds.) (1998) Gas Hydrates: Relevance to World Margin Stability and Climate Change, vol. 137. London: Geological Society Special Publications. Kennett JP, Cannariato KG, Hendy IL, and Behl RJ (2000) Carbon isotopic evidence for methane hydrate instability during Quaternary interstadials. Science 288: 128--133. Kvenvolden KA (1998) A primer on the geological occurrence of gas hydrates. In: Henriet J-P and Mienert J (eds.) Gas Hydrates: Relevance to World Margin Stability and Climate Change, vol. 137, pp. 9--30. London: Geological Society, Special Publications. Max MD (ed.) (2000) Natural Gas Hydrates: In: Oceanic and Permafrost Environments. Dordrecht: Kluwer Academic Press. Nisbet EG (1990) The end of ice age. Canadian Journal of Earth Sciences 27: 148--157. Paull CK, Ussler W, and Dillon WP (1991) Is the extent of glaciation limited by marine gas hydrates. Geophysical Research Letters 18(3): 432--434. Sloan ED Jr (1998) Clathrate Hydrates of Natural Gases. New York: Marcel Dekker. Thorpe RB, Pyle JA, and Nisbet EG (1998) What does the ice-core record imply concerning the maximum climatic impact of possible gas hydrate release at Termination 1A? In: Henriet J-P and Mienert J (eds.) Gas Hydrates: Relevance to World Margin Stability and Climate Change, vol. 137, pp. 319--326. London: Geological Society, Special Publications.

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATE AND SUBMARINE SLIDES J. Mienert, University of Tromsø, Tromsø, Norway & 2009 Elsevier Ltd. All rights reserved.

Introduction Methane hydrate is an ice-like substance composed of water molecules forming a rigid lattice of cages that enclose single molecules of gas, mainly methane. They occur in vast quantities beneath the ocean floor, and they are stable under high-pressure and lowtemperature conditions existing along many continental margins (Figure 1). In siliciclastic margins, sediment sliding and slumping can generate simple or complex slide scars and seabed morphologies (Figure 2). The gravity-driven movements of sediments masses vary in volume from a few cubic kilometers to several thousands of cubic kilometers. Giant slides reach run-out distances close to 1000 km, such as the Storegga slide on the midNorwegian margin. Average slope angles are low and vary typically between 21 and 41. A slide is distinguished from a slump based on the Skampton ratio h/l, where h is the thickness of the sliding block or depth and l is the length of the slide. Most observed submarine mass movements appear to be

translational with Skampton ratios o0.15. Slumps are rotational with Skampton ratios 40.33 and may coexist with translational slides creating mingled slope-failure generations. Seabed bathymetry and reflection seismic data from continental slopes and rises of the Atlantic, Pacific, and Indian Oceans and the Black Sea reveal major slide scars and slumps overlying gas hydrate deposits. Slides and slumps shape the morphology of world ocean margins and represent a major mechanism for transferring enormous amounts of sediment material over a relatively short time, between hours and years, into the deep sea. In most of the past slope-failure events, the trigger mechanisms of the slides and slumps are unknown and lead to speculations about various mechanisms. Many of the large slides are associated with surprisingly low slope angles (o41), much less than are required to destabilize the seafloor under static conditions. It appears that failure mechanisms have the highest possibility of triggering slumps during sea-level lowstands. However, this is not true for some of the largest slides known on continental margins; for example, the Storegga and Traenadjupet slides on the mid-Norwegian margin occurred during the last 8000 years within a period of rapid sea-level rise. The role of methane hydrate in the development of slides and slumps is an intriguing question in ocean

Seismic detection

Methane hydrate stability zone

0

0

Seafloor

Water temperature

8

12

16

BSR

0.4 Methane hydrate phase boundary

Methane hydrate stability

Geothermal gradient Base of gas hydrate

Methane + water

Water t or en im d e e S +s s te a ga r yd ee Fr sh

0.8

o afl

Water Sediment

Sediment

Gas-enchanced reflections

Ga

HSZ −20 0 20 Temperature (°C)

1.2

Depth (km)

Pressure (MPa)

4

1.6 Continental slope

Figure 1 Methane hydrate stability zone as a function of water temperature, geothermal and pressure gradient. A methane hydrate I structure is shown on the upper left and a methane hydrate sample on the upper right. The hydrate stability zone (HSZ) increases with water depth (pressure) while the base of the hydrate stability zone (BHSZ) is evident from the seismic detection by a bottom simulating reflector (BSR). The hydrate occurrence zone (HOZ) where hydrates occupy the pore space of sediments lies between the seafloor and the BSR. Adapted from Kayen RE and Lee HJ (1991) Pleistocene slope instability of gas-hydrate laden sediment on the Beaufort Sea margin. Marine Technology 10: 32–40.

790

(c) 2011 Elsevier Inc. All Rights Reserved.

0°

10° N

70°

N

65°

N

N

Sc

an

di

na

via

65°

791

20° E

70°

10° E

W

METHANE HYDRATE AND SUBMARINE SLIDES

60°

N Ormen Lange

Havsule

0°

Solsikke

Figure 2 Extent of one of the largest known submarine slides on continental margins and today’s 4000 km2 gas hydrate province inferred from BSR distributions (yellow). The slide affected an area of approximately 95 000 km2 and removed 2500–3500 km3 of sediments. The main slide scar is shown on the inlet figure with the main headwall at the continental shelf and several headwalls and individual glide planes downslope.

and climate research. The interest stems mainly from the hypothesis that large releases of methane from marine gas hydrate reservoirs generate continental margin instability. If methane hydrate dissociates, gas and fresh water are released. Based on the amount of dissociated hydrate, this volume increase

can generate pore-fluid overpressure and zones of reduced shear strength; thus glide planes develop in continental margin sediments. These zones develop critical preconditions for a slope to fail. Continental margin instability leading to slides or slumps and abrupt release of methane, a greenhouse gas, may be

(c) 2011 Elsevier Inc. All Rights Reserved.

792

METHANE HYDRATE AND SUBMARINE SLIDES

sufficiently large enough to trigger major climatic changes. Substantial uncertainties exit in the amount of gas hydrate in space and time, and in the dynamics of dissociation processes controlled mainly by ocean pressure (sea level) and temperature (hydro- and geothermal), and material properties of the subseafloor. They lead to many developments of regional and global scenarios and a basis for a continuing debate. One of the main debates centers on the explanation of observations of massive injections of isotopically light carbon into the oceans and atmosphere in Cenozoic and Mesozoic times, and whether the observations in the geological record are sufficient to propose a major release of methane from gas hydrate. Accumulating seabed observations from various areas along continental margins indicate massive slumping during late Cenozoic and Mesozoic times. Many have wondered which processes could account for giant slumps (41000 km2), aside from the classical preconditioning factors and trigger mechanisms such as oversteepening of slopes, rapid sediment loading, and, most importantly, earthquakes. The idea of linking methane hydrate and slumps became widespread after 1981 following a discovery of a slump/gas hydrate area on the eastern US continental margin. Further observations where slump locations coincide with present gas hydrate reservoirs suggested linkages, which in turn are considered as possible agents for rapid releases of methane. But the picture is more complex because climatically induced increases in sedimentation rate and changes in sediment type can create zones of weakness. The geological evidence to discriminate the mechanisms is generally missing after an offshore slide or slump has occurred. From the existing knowledge, it is difficult to extract any overarching role of methane hydrate in slope stability and slumping. One solution may arise from investigations of slump timing and frequency. Since the temperature– pressure regime on continental margins controls the stability of methane hydrates on a large scale, either of these can alter the stability of hydrates. This knowledge implies that methane hydrate reservoirs in continental margins should react vigorously and almost simultaneously with global change in the ocean environments, that is, sea level and temperature. Yet, current knowledge of all geophysically inferred gas hydrate occurrences and all drilled gas hydrate reservoirs underlines that we have no reason to consider hydrates to be either continuously or homogenously distributed in continental margins. Geophysical and geochemical data demonstrate that methane hydrate accumulations are a result of fluid migration and their pathways. Attempts to verify

coupled processes between hydrate dissociation and slumps from geological records are ongoing and are notably important for understanding climate-induced geohazards. It requires a knowledge of where, how much, and how far accumulations of gas hydrate extend in continental margins. Equally important is an adequate knowledge of changes in slope stability during hydrate formation and dissociation processes in a variety of geological environments from high to low latitudes.

Methane Hydrate Stability and Seismic Detection Methane hydrates belong to compounds named ‘clathrates’ after the Latin clathratus, meaning encaged. Gases that form hydrates occur mainly in two distinct structures: structure I hydrates form with natural gases containing molecules smaller than propane. They are found in situ in continental margin sediments with biogenic gases that consist mostly of methane but rarely contain carbon dioxide and hydrogen sulfide. Structure II hydrates form from thermogenic gases that, in addition to methane, incorporate molecules larger than ethane but smaller than pentane. The Gulf of Mexico is an area of massive but local development of structure II hydrates, where salt tectonics and related structures control focused fluid flow from hydrocarbon reservoirs. It is therefore of less relevance for extended slides and slumps. For oceanic methane hydrates and slumps, structure I hydrates are considered to be important. The methane hydrate stability depends mainly on pressure (P), temperature (T), and the presence of gas in excess of solubility as a function of P and T. The stability of hydrates is more susceptible to changes in temperature than in pressure. Methane hydrate equilibrium conditions are complex, depending also on the texture of the host sediments, pore water salinity, and type of gas entrapped. Laboratory results show that sand is more conducive to hydrate formation than silt and clay. Hydrate formation in sands starts at a lower pressure (shallower water depth) compared to silt and clay. Beneath the hydrate stability zone (HSZ), the solubility of gas controls the amount of free gas. Lateral and vertical variations in texture and mineralogical properties of sediment contribute to a heterogeneous distribution of hydrate in the HSZ and free gas beneath it. It is generally accepted that the depth of the HSZ can often be found by seismic detection of the base of the hydrate stability zone (BHSZ). The base is generally inferred from the presence of a

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATE AND SUBMARINE SLIDES

bottom-simulating seismic reflector (BSR), though the absence of a BSR does not rule out the presence of hydrates. A phased-reversed seismic reflector originates at this hydrate–gas phase boundary due to a distinct acoustic impedance contrast. The impedance contrast is caused by the low p-wave velocity of gas-bearing sediments beneath the higher p-wave velocities of hydrate-bearing sediments. The bottomsimulating behavior of the hydrate–gas phase boundary suggests a direct dependence on geothermal gradients, that is, temperature. Geological controls on hydrate formation may explain the oftenobserved discrepancies between the calculated BHSZ and the calculated and observed BSR depth. Observations of BSR’s beneath slump scars became an argument for inferring a link between methane hydrates and slumps, but this is clearly not sufficient as a sole indicator.

Natural Trigger Mechanisms of Slides from Geological Records Most submarine landslides occur unobserved on the seabed along continental margins. We, therefore, deduct the processes involved on the basis of seabed bathymetry, sub-seabed paleomorphology, and sediment cores a long time after an event takes place. Only a very few submarine landslide events are directly documented. They caused submarine cable breaks or started as an underwater slope failure and retrogressively approached the coastline. Much less information exists about their trigger mechanisms, and the most widely used example for a continental slope failure is the 1929 Grand Banks earthquake. It appeared to trigger a slump that subsequently developed into a turbidity current. Another condition that is most commonly associated with submarine slides is the rapid accumulation of thick sedimentary deposits over under-consolidated sediments, which by generating excess pore pressure reduces the effective stress that holds the sediment grains together. The magnitude of a combined instantaneous loading and a progressive excess pore pressure buildup appear to generate weakened sediment layers. The resulting reduction in shear strength allows sediments to move down very gentle slopes (o11). Good examples are the well-studied areas of the Mississippi River delta. On formerly glaciated continental margins such as the passive Norwegian margin, thick layers of siliceous ooze of Miocene age exist in the sub-seabed. Their low density provides conditions for high compressibility and thus pore pressure builds up when subjected to rapid sediment loading during ice sheet advances of

793

Pleistocene times. Cyclic sedimentation rates with high inputs of dense material during glacial times and low sediment input of more permeable material during interglacial times may explain the long-term instability of this margin. Megaslides seem to occur at a frequency of approximately 100 ky after each of the major ice ages since the onset of continental shelf glaciations at 500 ka. It is speculated that earthquake loading is the final trigger for the initiation of the slides at the end of each glaciation. Steeper slopes or internally derived seepage forces might aid to fracturing the seabed leading to slope failures. Examples come from salt tectonics and flow of salt and mud in the sub-seabed of the Gulf of Mexico and the western Mediterranean of the Nile River delta. Up to now, the analysis of the forces of gravity and earthquake loading have shown that they are often not great enough to be the sole cause of failure along the slopes of continental margins. The controls on most identified slides are still obscure though earthquakes have been documented for a few events. Gas hydrate dissociation is another factor that is becoming more frequently used to decipher what controls slope failure.

Methane Hydrate Decomposition as a Natural Trigger Mechanism In fully saturated structure I hydrates there is one molecule of methane present for every six molecules of water. In theory, when methane hydrate dissociates this results in an increase of volume and pressure buildup, because 1 m3 of methane hydrate dissociation develops into 164 m3 of methane gas at standard temperature and pressure conditions. The effect of dissociation pore pressure depends on water depth; if we decompose 1 m3 of gas hydrate at 900 m water depth (90 atm) we will get 1.82 m3 of free gas while at 1500-m water depth (150 atm) only 1.09 m3 of free gas develops. Shallow-water methane hydrate reservoirs are expected to develop a larger volume of free methane and are therefore able to generate higher excess pore pressure per unit volume than deep-water methane hydrate reservoirs depending on the rate of dissociation and the permeability of the sediment. As a result, upper continental margin slopes are more vulnerable to gas hydrate dissociation than lower slopes. Dissociaition of methane hydrates into liquid and free gas can build up excess pore pressure that reduces the shear strength of the sediments, thereby causing a natural trigger mechanism for a slope to fail. A change in gas solubility generated by variations in temperature and/or pressure is generally ignored

(c) 2011 Elsevier Inc. All Rights Reserved.

794

METHANE HYDRATE AND SUBMARINE SLIDES

as a factor in methane hydrate and slumps. A geological model for a retrogressive (bottom-up) failure surface generated by hydrate dissolution due to gas solubility, and a sketch for a progressive (top-down) failure surface by the hydrate dissociation due to a temperature increase illustrates the resulting differences in slope instabilities (Figure 3). There are numerous fundamental problems remaining with our observations and theories. We know that the degree of hydrate saturation in the pore space of sediments will determine the total volume and pore-pressure increase during decomposition. Observations indicate a heterogeneous distribution of hydrates within the hydrate stability zone (HSZ) in which, except for massive hydrates, the space occupied by hydrates is usually low (o5%) with local peaks (up to o15%). Thus, there appears to be no continuous or homogenous methane hydrate occurrence zone (HOZ) leading to a glide plane. The thickness of the HSZ increases with water depth, and theoretical gas– hydrate-bearing sediments could occur in a zone of B400 m below the seafloor (mbsf) in deep water of B2000 m.

(a)

Quantifying in situ distributions of methane hydrates in three dimensions is difficult because of the effect of the heterogeneous distribution sediment type and texture. Laboratory experiments and the physics involved in modeling hydrate formation and dissociation are complex, and simplified assumptions have been made which take into consideration the time dependence of the processes. The hydrate formation rates beneath the seabed depend on the fluid flow from beneath the HSZ. Here, it is the availability of water and gas which in turn depends on diffusion rates and seepage conditions. On the other hand, hydrate dissociation requires only temperature increase and/or pressure reduction, but these occur at different rates in nature. To add to this complexity, the dissipation rates of excess gas and water and thus excess pore pressure depend on the diffusion coefficient and permeability of the sub-seabed. Let us consider a simple hypothetical scenario. A drop in sea level during glacial times reduces the hydrostatic pressure, leading to a thinning of the HSZ and a potential thickening of the free gas zone beneath. As a consequence, the potential for

Methane hydrates and slides

0

Warming

1

Progressive 2 1000

Dissociation of gas hydrates due to ocean warming

3

r oo afl HSZ

Pressure

Se

2000 Retrogressive 3

SZ

BH

2500

1

as

(b)

Cyclic sedimentation and slides

g ree

2

F

4 3

Dissolution of gas hydrate due to sea-level increase

2

or

Seaflo

1

High-density deposits Low-density deposits

Figure 3 (a) Methane hydrates affected by ocean warming on the upper continental margin leading to progressive slide mechanisms while methane hydrates affected by sea level changes have the potential for retrogressive slide mechanisms. (b) Cyclic sedimentation with high density and less permeable sediments over less density and more permeable sediments causes overpressurized layers and characteristic glide planes indicating retrogressive sliding (see also Figure 2).

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATE AND SUBMARINE SLIDES

pore-pressure buildup and slumps increases. The emission of methane as part of the slump process would become greater with the frequency of slumps. Methane releases would enhance the greenhouse gases and provide a negative feedback to growing glaciations, thus becoming a potential terminator to glaciations during Quaternary times. The opposite can be drawn from methane releases during interglacial times. Ocean warming is superior to sea level rise affecting gas hydrate stability, methane hydrate melting, and the potential for slump processes and methane release increases. Such positive feedback may create an amplifier for global warming. Yet, like many other negative–positive feedback loop models, such classical scenarios remain to be proven and are mostly conjectural.

Methane Hydrates in Regions of Slope Instability: Searching for a Link The most needed information is how excess pore pressure induced by the dissociation of methane hydrate is linked to the spatial and temporal changes in methane hydrate concentrations in continental margins. Such key information is far from being obtained. There is hardly any direct evidence that the observed slides and slumps are triggered by methane hydrate dissociation. Up to now, observations and possible connections between methane hydrate and slumps exist in selected areas of nearly all ocean margins. Evidence is provided from slumps of the eastern US Atlantic continental slope; slides and slumps of the continental rise and slope of SW Africa; giant slides and slumps of the continental slope of Norway; slides and slumps of the continental slope

795

on the Alaska Beaufort Sea; slides on the northern California continental margin; slides on the southern Caspian Sea margins; slumps of the western continental margin of India; slope failures along the Chile– Peru margins; slope failures in the northwestern Sea of Okhotsk; slope failures on the Amazon fan of the Brazil margin; slope failures on the northern Black Sea margin; slope failures on the Gulf of Cadiz margin (Figure 4). Long-term instability is evident from the New Jersey margin of the East Coast of the United States. Within the Paleogene, four periods – near the Cretaceous– Tertiary boundary, Palaeocene–Eocene boundary, top of Lower Eocene, and Middle Eocene – show slope failures. They are identified in reflection seismic records. Since all of the events occur close to major sea level lowstands, methane hydrate dissociation accompanied by excess pore pressure is regarded as a potential factor for the initiation of the slides and slumps. Linking methane hydrates with slope failures also remains speculative during Neogene times. Dating the event of many of the slope failures is still a problem and therefore large uncertainties exist in relating slides and slumps to a specific ocean’s sea level and/or temperature condition. It has been suggested that for the past 45 ky, glacial-period slope failures occur mainly in low latitudes associated with sea level lowstand. It is during periods of change that hydrate is most likely to dissociate or to experience dissolution, depending on which direction of sea level and/or water temperature leads. This case is used to propose that reduced hydrostatic pressure triggered dissociation of methane hydrate and slope failure. In contrast, ocean warming during rising sea level conditions in the northern high latitudes is used

Methane hydrate and slumps/slides

Figure 4 Global distribution of major submarine landslides overlying areas of today’s known gas hydrate provinces.

(c) 2011 Elsevier Inc. All Rights Reserved.

796

METHANE HYDRATE AND SUBMARINE SLIDES

to argue for dissociation of methane hydrate contributing to slope failure. Another approach for dating slides is based on turbidite sequences deposited in ocean basins that are a consequence of slides and slumps. Dating of turbidites resulted in an equally complex picture. At the Norwegian margin, the inferred slides occurred during (1) periods of sea level lowstand or an advance or presence of the Fennoscandian Ice Sheet at the shelf break, or (2) at periods of rapid sea level rise accompanied by ocean warming. Seabed side-scan sonar and bathymetry data from the northern rim of the giant Storegga slide on the mid-Norwegian margin show cracks in the seabed upslope. The BSR, which crosses bedding planes and is projected toward the shelf edge pinches out at the upper slope where the HSZ intersects the seabed in the area of the cracks. Other empirical support of BSR occurrences in areas of slumps exists particularly from the west coast off India, Nankai trough, and the US Atlantic continental slope.

A Link between Slide Headwall and Excess Pore Pressure Ocean conditions and hence the stability of methane hydrate along most continental margins have changed significantly since the Last Glacial Maximum (LGM). In terms of methane hydrate stability, both the pressure related to eustatic sea level changes and bottom water temperature (BWT) changes in intermediate water masses have to be addressed. The hypothetical scenario mentioned before is replaced by an actual scenario for the Norwegian margin, where detailed information about ocean and seabed conditions exists. The combined effect of eustatic sea level rise of B120 m and BWT increases of B5 1C, documented since the last LGM, change hydrate stability conditions along the entire continental slope. In water depth of 4800 m, BHSZ deepens slightly due to the sea level rise. No temperature increase is indicated for deeper seabed. In contrast, the effect of sea level rise on the HSZ is, for localities above 800-m water depth, more than compensated by a rise in BWT. The combination of two such opposing or counteracting effects in terms of gas hydrate stability conditions (sea level rise is favorable for HSZ while BWT rise is not) predicts a distinct reduction in the extent and thickness of the HSZ in shallower water between 300 and 700 m since the LGM. Such results illustrate that shallower water depth areas are more sensitive to changes in sea level and water temperature than deeper localities. From this we may conclude that deeper water areas are less sensitive and methane hydrate

dissociation may trigger slumps favorably along the upper continental margins. Combining the dissociation model and the observation of a BSR along reflection seismic profiles crossing the Storegga headwall supports the hypothesis of excess pore pressure buildup at slide headwalls and thus slide progradation (from upper to lower slope). However, it contradicts many geomorphologic observations along continental margins including the well-studied Norwegian margin, which indicates retrogressive sliding (from the lower to the upper slope). Discussions about the seemingly opposing results are ongoing and lead to suggestions that excess pore pressure can also result from methane hydrate dissolution. This argument is based on thermodynamic calculations. Models show that due to a temperature and pressure increase, hydrate dissolution may start at the top of the hydrate occurrence zone. Simulation also documents that this process, due to a change in gas solubility, can occur at the origin of a retrogressive failure initiated at the lower slope. No conclusions exist as yet but the effect of dissolution, that is, the gas entering the pore water, and dissociation on deep and shallow water hydrates appear to be of importance for understanding all aspects of triggering retrogressive versus progressive slides.

Methane Blasts in the Past? Current debates concern the explanation of massive injections of isotopically light carbon to the oceans and atmosphere in Cenozoic and Mesozoic times but also beyond. Distinct negative d13C excursions measured in samples from various locations reflect a rapid injection of massive quantities of 12C-rich carbon to the ocean and atmosphere. Support for a hypothesis of methane blasts from gas hydrate dissociation has grown since agreements among scientists that other reservoirs cannot supply a sufficient mass of carbon over the relatively short geological timescales represented by the d13C spikes. Abrupt negative d13C excursions are reported from the Neoproterozoic; the early Paleozoic; the Permian/ Triassic; the Triassic/Jurassic; the late and early Jurassic; the early Cretaceous (Aptian); the Paleocene– Eocene boundary; and the Quaternary. The Quaternary has the advantage to be documented by both high-resolution climate ice core and ocean sediment core records. The Quaternary methane within the trapped air in the Greenland and Antarctica ice cores, and the d13C excursions from both planktonic and benthic foraminifera created a heated and ongoing debate about the sources of methane. What

(c) 2011 Elsevier Inc. All Rights Reserved.

METHANE HYDRATE AND SUBMARINE SLIDES

caused the rapid and large atmospheric methane increase in the atmosphere during interstadials, and the negative d13C excursions in the Santa Barbara Basin offshore California? What caused the rapid shifts to positive excursions during stadials? Which sources can react fast enough and at the same time are so large that shifts reach several hundred ppbv (parts per billion by volume) in atmospheric methane and several per mil ( 1% to 6%) in ocean carbon isotopes? Though indirect methane measurements of atmospheric methane concentration from ice cores go back more than a hundred thousand years, direct atmospheric measurements of methane only go back to 1970. Coming from the last glaciation into the deglaciation, the atmospheric methane as recorded in ice cores increased from c. 400 to 700 ppbv, but from 1800 to 1990 it reached 1700 ppbv. It is still increasing and eventually most methane (CH4) will end up as CO2 and water. Evidently, CH4 and CO2 changes do track most of the rapid climate shifts. Up to now, terrestrial measurements of the global methane budget provide clear hints that the main drivers of methane appear to be the terrestrial sources (natural, anthropogenic) and sinks (soil microbes). A general consensus exists that continental wetlands are a major methane source. At the same time there is a major unknown, which is the role of the deep and shallow ocean environment where microbes are thriving on methane, and where methane hydrate dissociation and slumps may release significant amounts of methane. These processes are now becoming recognized as part of a complex feedback system. The suggested methane blasts of the past were certainly helpful in increasing our awareness for a holistic approach.

Discussion Massive and rapid releases of methane from hydrates via slumping may have the potential for entering the atmosphere. The natural formation of methane hydrate beneath the oceans, however, may buffer the volume of methane entering the ocean and atmosphere. Our present observations from continental margin methane hydrate, BSRs, and slumps are still insufficient to quantify any of these major processes on a global or even regional scale. A breakthrough may come from a less obvious field, the methane records from ice cores. They may open new pathways for one of the two competing explanations for abrupt CH4 increases. One working hypothesis uses the sudden release of methane from marine hydrates while the other explanation makes the terrestrial

797

biosphere responsible for rapid methane releases. Developments of today’s arguments and theories for linkages of methane hydrate, slumps, and climate greatly depend on an understanding of the Quaternary atmospheric CH4 record coming from ice cores. Equally important are ocean records such as negative d13C spikes in fossil records of planktonic and benthic foraminifera, and the age and frequency of slumps associated with gas hydrate reservoirs. Quaternary ice core records may soon allow confirmation or rejection of the proposed scenarios for methane release from ocean margins due to the distinct deuterium/hydrogen record of marine hydrates. There is hope that dDCH4 provides a means of discriminating between methane sources from land or marine methane hydrate reservoirs. In Cenozoic and earlier times, the geological records of slumps and negative d13C excursions are less clear. An underlying understanding about the role of both methane hydrate and slumps for methane releases from past ocean hydrate reservoirs will remain a challenge for the years to come. An answer as to whether or not the dissociation/dissolution of massive methane hydrates contributed to slumping and rapid increases of atmospheric greenhouse gases would give us an important perspective on how likely it is that we shall encounter such geohazards in the future. It is not unreasonable to expect that the amount of methane hydrates and related pore pressure buildup during dissociation and dissolution processes in slide-dominated ocean margins is soon to be quantified. Today’s evaluations show that an ocean warming leading to a temperature increase at the BHSZ can lead to dissociation of gas hydrates and generation of high excess pore pressure. Sea level fall can generate similar effects. More research will establish whether this is a viable phenomenon leading to assessments of today’s methane hydrate and slope stability. Improved techniques for deciphering the origin of methane from ice cores, quantifications of methane records in oceans from isotopically light carbon spikes, and determinations of the timing of submarine slides and slumps in Cenozoic times will provide new information for the development of theories (models, hypotheses, etc.) for this potentially dynamic part of ocean margins.

See also Carbon Dioxide (CO2) Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Cenozoic Oceans – Carbon Cycle Models. Methane Hydrates and Climatic Effects. Slides, Slumps, Debris Flows and Turbidity Currents.

(c) 2011 Elsevier Inc. All Rights Reserved.

798

METHANE HYDRATE AND SUBMARINE SLIDES

Further Reading Carpenter G (1981) Coincident sediment slump/clathrate complexes on the US Atlantic continental slope. GeoMarine Letters 1: 29--32. Dickens GR (2001) The potential volume of oceanic methane hydrates with variable external conditions. Organic Geochemistry 32: 1179--1193. Hampton MA, Lee HJ, and Locat J (1996) Submarine landslides. Reviews of Geophysics 34(1): 33--60. Henriet JP and Mienert J (eds.) (1998) Geological Society, London, Special Publications No. 137: Gas Hydrates – Relevance to World Margin Stability and Climate Change. London: Geological Society. Higgins JA and Schrag DP (2006) Beyond methane: Towards a theory for the Paleocene–Eocene thermal maximum. Earth and Planetary Science Letters 245: 523--537. Katz ME, Cramer BS, Mountain GS, Katz S, and Miller KG (2001) Uncorking the bottle: What triggered the Paleocene/Eocene thermal maximum methane release? Paleoceanography 16: 549--562. Kayen RE and Lee HJ (1991) Pleistocene slope instability of gas-hydrate laden sediment on the Beaufort Sea margin. Marine Technology 10: 32--40. Kennett JP, Cannariato KG, Hendy IL, and Behl RJ (2003) American Geophysical Union, Special Publication, Vol. 54: Methane Hydrates in Quaternary Climate Change – The Clathrate Gun Hypothesis, 216pp. Washington, DC: American Geophysical Union.

Maslin M, Owen M, Day S, and Long D (2004) Linking continental-slope failures and climate change: Testing the clathrate gun hypothesis. Geology 32: 53--56. McIver RD (1982) Role of naturally occurring gas hydrates in sediment transport. AAPG Bulletin 66: 789--792. Mienert J, Vanneste M, Bu¨nz S, Andreassen K, Haflidason H, and Sejrup HP (2005) Ocean warming and gas hydrate stability on the mid-Norwegian margin at the Storegga slide. Marine and Petroleum Geology 22: 233--244. Mulder TH and Cochonat P (1996) Classification of offshore mass movements. Journal of Sedimentary Geology 66: 43--57. Paull CK, Buelow WJ, Ussler W, III, and Borowski WS (1996) Increased continental margin slumping frequency during sea-level lowstands above gas hydratebearing sediments. Geology 24: 143--146. Sloan ED and Koh C (2008) Clathrate Hydrates of Natural Gases, 3rd edn., Chemical Industries Series/119. Boca Raton, FL: CRC Press, Taylor and Francis Group. New York: Dekker. Sower T (2006) Late quaternary atmospheric CH4 isotope record suggests marine clathrates are stable. Science 311(5762): 838--840. Sultan N, Cochonat P, Foucher JP, and Mienert J (2004) Effect of gas hydrates melting on seafloor slope instability. Marine Geology 213: 379--401. Xu W and Ruppel C (1999) Predicting the occurrence, distribution, and evolution of methane gas hydrate in porous marine sediments. Journal of Geophysical Research 104(B3): 5081--5096.

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROBIAL LOOPS M. Landry, University of Hawaii at Manoa, Department of Oceanography, Honolulu, HI, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1763–1770, & 2001, Elsevier Ltd.

Introduction The oceans contain a vast reservoir of dissolved, organically complex carbon and nutrients. At any given point in time, most of this dissolved organic matter (DOM) is refractory to biological utilization and decomposition. However, a significant flow of material cycles rapidly through a smaller labile pool and supports an important component of the food web based on bacterial production. The recapturing of this otherwise lost dissolved fraction of production by bacteria and its subsequent transfer to higher trophic levels by a chain of small protistan grazers was initially called the microbial loop by Farooq Azam and co-workers in the early 1980s. The following decades have been charecterized by remarkable discoveries of previously unknown lifeforms and by major advances in our understanding of trophic pathways and concepts involving the seas’ smallest organisms. In modern usage, the term microbial food web incorporates the original notion of the microbial loop within this broader base of microbially mediated processes and interactions.

Perspectives on an Evolving Paradigm Although early studies date back more than a century, our understanding of the microbial ecology of the seas initially advanced slowly relative to other aspects of biological oceanography largely because of inadequate methods. Prior to the mid-1970s, for example, the simple task of assessing bacterial abundance in sea water was done indirectly, by counting the number of colonies formed when sea water was spread thinly over a nutritionally supplemented agar plate. We know now that about one marine bacterium in a thousand is ‘culturable’ by such methods. At the time, however, the low counts on media plates, typically tens to hundreds of cells per ml, were consistent with the then held view that sea water would not support a large and active assemblage of free-living bacteria. The role of bacteria was therefore assumed to be that of decomposers of

organically rich microhabitats such as fecal pellets or detrital aggregates. Coincidentally, early deficiencies in phytoplankton production estimates, due to trace metal contaminates and toxic rubber springs in water collection devices, were giving systematic underestimates of primary production, particularly in the low-nutrient central regions of the oceans that we now know to be dominated by microbial communities. For such regions, the low estimates of bacterial standing stocks and primary production were mutually consistent, reinforcing the notion of the central oceans as severely nutrient-stressed ‘biological deserts’ with sluggish rates of community growth and activity. Even so, there were early signs from both coastal and open-ocean studies of much greater microbial potential – one from newly developed sea water analyses of ATP (adenosine triphosphate), the shortlived energy currency of all living organisms; another from respiration (oxygen utilization) measurements of whole and size-fractioned sea water samples. Such measurements implied much larger concentrations of life-forms and metabolic rates than could be explained by the planktonic plants and animals that were being studied intensively in the 1970s. Recently developed methods for measuring bacterial production in the oceans, based on the uptake of radioisotope-labeled thymidine precursor for nucleic acid synthesis, were also giving results inconsistent with conventional wisdom. Thus, by the late 1970s, the stage was set for a revolutionary new paradigm, the microbial loop. Epifluorescence microscopy was perhaps the tool that most facilitated scientific discovery and general acceptance of the new paradigm. Here, for the first time, one could directly visualize, with the aid of fluorescent stains, the multitude of coccoid, rod, and squiggle-shaped forms that comprised the typical bacterial assemblage of about a million cells per ml. Autoradiography, a technique by which the cellular uptake of radioisotope-labeled substrates is developed on sensitive films, confirmed that most of these cells were living and active. Epifluorescence microscopy (EPI) also opened the door to discovering and studying other components of the microbial community. Synechococcus, a genus of small photosynthetic coccoid cyanobacteria, was soon recognized as a ubiquitous and important component of the marine phytoplankton from its characteristic shape and phycobilin accessory pigments, which glow orange under the EPI standard

(c) 2011 Elsevier Inc. All Rights Reserved.

799

800

MICROBIAL LOOPS

blue-light excitation. The same excitation wavelength causes the chlorophyll in photosynthetic cells to fluoresce bright red, allowing purely heterotrophic cells to be easily distinguished from chlorophyllcontaining cells of similar size and shape. This distinction was critical in demonstrating a clear food web coupling for bacterial production via small phagotrophic (i.e., particle-consuming) colorless flagellates, and it provided an important technique for identifying and quantifying trophic connections using fluorescently labeled beads and cells as tracers to assess grazer uptake rates. Such studies also had the unintended effect of illustrating the blurry distinction between pure autotrophy and heterotrophy within the microbial assemblage. For example, many of the small photosynthetic flagellates (containing chlorophyll) in the oceans have been observed to consume bacterialsized particles. Similarly, many, if not most, of the larger pigmented dinoflagellates follow a ‘mixed’ mode of nutrition involving photosynthesis and phagotrophy. In a phenomena known as kleptoplastidy, common forms of ciliated protozoa have also been shown to retain (literally ‘steal’) the chlorophyll-containing plastids of their prey and use them as functional photosynthetic units for a day or longer. The widespread occurrence of mixotrophy, in all of its various forms, has consequently emerged as one of the important findings related to the microbial food web. One notable discovery of the mid-1980s was that of Prochlorococcus, a tiny photosynthetic bacterium now known to be one of the most important primary producers in the tropical oceans, and probably on the planet. Although it seems remarkable that such an important organism could have escaped detection for more than a century of oceanographic investigation, this advance was again only made possible by new methods. In this case, the facilitating technology was a laser-based optical instrument, the flow cytometer, developed by medical research for the rapid analysis of individual cells in a narrowly focused fluid stream. The application of this new approach in the ocean sciences was quick to reveal high concentrations of the dimly red-fluorescing (chlorophyllcontaining) Prochlorococcus, which could not be distinguished from nonpigmented bacteria by standard EPI techniques. Herein lies one of the problems of epifluorescence microscopy, the confounding of significant populations of the autotrophic Prochlorococcus cells with heterotrophic bacteria. Some early reports of heterotrophic biomass greatly exceeding autotrophs in surface waters of the tropical oceans were a consequence of this methodological artifact.

Even among the heterotrophic bacteria, there have been discoveries of fundamental importance. For example, kingdom-specific molecular probes have recently shown that a significant fraction of the ‘bacteria’ from EPI counts are not true Bacteria at all, but lesser-known prokaryotes of the Kingdom Archaea. These organisms have been known to inhabit extreme environments such as hot springs and the interstices of salt crystals, but their presence in more typical oceanic habitats raises many interesting questions about the roles of their unique metabolic systems in ocean biogeochemistry. The application of powerful molecular methods to the ocean sciences in the 1990s has signaled the beginning a new era in marine microbial ecology with the potential to reveal the full spectrum of microbial population and physiological and metabolic diversity. As a natural extension of new methods to visualize and characterize community components of decreasing size, the 1990s have also seen a growing recognition of the importance of viruses in marine microbial communities. Based on a combination of fluorescent staining methods and electron microscopy, viruses are now clearly the most numerous component of the microbial community, exceeding bacteria in most cases by an order of magnitude. While not technically ‘real’ organisms, in the sense of having independent metabolic or reproductive capabilities, viruses can be significant vectors of bacterial and algal mortality and therefore have important implications for the functioning and energy flows in marine microbial communities. Recent studies have shown rates of microbe infection and viral turnover that would account for the loss of about half of bacterial production. The typical hostspecificity of viruses and their spread by densitydependent encounter frequency suggest that they also have a major role in maintaining bacterial diversity, by selectively punishing the most successful competitors.

Organization of the Microbial Food Web For the sake of representing diagrammatically the trophic connections among marine microbes, and indeed as a practical strategy for most ecological studies, it is necessary to compress the known complexities of microbial communities into a few functional categories (Figure 1). There is no clear cut-off between particulate and dissolved organic matter in the oceans, for example, but rather a spectrum of material ranging from low molecular weight (lowMW) amino acids and sugars to large complex

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROBIAL LOOPS

801

Mesozooplankton Size (μm) Total primary producers >20

10

5

2

0.2 Bacteria Viruses Labile

Refractory DOM

Nutrients N, P, Si, Fe Nutrients flows DOM flows Particulate flows Mixotrophy Figure 1 Conceptual representation of the microbial food web showing flows of nutrients and dissolved organics and interactions among various size classes of bacteria and protists. Primary producers (left side) are shown as chlorophyll-containing cells and heterotrophic (right side) cells are drawn without pigments. Mixotrophy is shown as feeding of some of the pigmented cells on smaller prey organisms.

molecules, to colloids, viruses and submicrometer remnants of previously consumed biota, and to wispy strands that link a fragile gelatinous matrix of living and dead material. Labeled simply as DOM in Figure 1, this material can cycle between refractory and labile pools by slow physical winnowing and leaching, extracellular enzymatic cleavage, or accelerated photochemical oxidation, the latter principally from enhanced ultraviolet radiation in near-surface waters. Although the relative magnitudes are debatable and likely to vary seasonally and regionally, inputs to the DOM box come from virtually all components of the marine plankton. Phytoplankton leak low-MW compounds across porous membranes, and they also produce the sugary products of

photosynthesis in excess when nutrients are insufficient for cell growth. Many larger consumers feed sloppily, producing DOM and particulate fragments as they grind and rip their food with silica-tipped teeth. Both large and small consumers excrete lowMW organics, often as a significant fraction of their total metabolism, and release organics in the form of incompletely digested material. Lastly, whole cells are fragmented into numerous components in the operationally defined DOM size range during the final stages of the viral lytic cycle. DOM from all of these various sources provides the substrate that fuels the growth of marine heterotrophic bacteria. The original microbial loop was patterned on the notion that consumers typically feed on prey

(c) 2011 Elsevier Inc. All Rights Reserved.

802

MICROBIAL LOOPS

organisms that are a factor of 10 less in cell size (length or equivalent spherical diameter), a view that fitted nicely with the traditional decadal size classifications of marine plankton. Accordingly, picoplankton (0.2–2 mm cells, including most prokaryotes) would be fed upon by nano-sized protists (2–20 mm cells, typically flagellates), and they, in turn, by microheterotrophs (20–200 mm cells, typically ciliates). From laboratory feeding experiments as well as from size-fraction manipulations of natural communities, however, it has become increasingly clear that many protists feed optimally on prey much closer to their own body size. The most common bacterivores in the oceans, small naked flagellates, are typically only 3–6 times the size of their average bacterial prey. Heterotrophic and mixotrophic dinoflagellates generally prey on cells of even larger relative size, some using extracellular capture and digestion to handle preferred prey larger than their own body size. Compared to the original microbial loop paradigm, the effect of compacting mean predator–prey size relationships for flagellates can add at least one more level of intermediate consumer between bacteria and the largest protozoa in the grazing chain. As will be presented in more detail below, complex relationships and feedbacks among bacteria and protists and their different degrees of availability to higher-order consumers argue for a broad view of the microbial food web, rather than a narrow focus on a single loop element. This brings us to the matter of definition. Are all single-celled organisms to be included in this web, or are there size or functional reasons to exclude some? Figure 1 is organized from the perspective of organisms described in the original microbial loop paradigm, the bacteria and the grazer chain. It therefore includes the autotrophic organisms that would constitute the main food items of some protistan grazers. Notably absent are the very large phytoplankton, principally large solitary cells or long diatom chains that figure so prominently in classical descriptions of the seasonal bloom cycles of temperate and boreal oceans. Such cells would not be readily available to protistan grazers because of their size, spines or other defensive strategies. They also function differently from smaller primary producers, being more intimately related to export flux from the euphotic zone by aggregate formation and direct cell sinking or by incorporation into the fast-sinking fecal pellets of large metazoan consumers. Thus, while all planktonic organisms are related in a sense by trophic linkages and feedbacks to dissolved organics and nutrients, we specifically exclude these larger primary producers from the microbial assemblage. The division is functional and roughly follows size,

distinguishing the direct flow of primary production to a network of metazoan consumers as opposed to that going primarily to protists. It is important to observe that the division is only loosely related to taxonomic groupings. For example, the dominant diatoms in many open ocean regions are tiny (o10 mm) pennate cells and well within the size range that can be grazed efficiently by ciliates and large flagellates. By the same token, large clumpforming filaments of the nitrogen-fixing cyanobacteria Trichodesmium spp. appear to be directly utilized only by certain harpacticoid copepods.

Food Web Transfers The original descriptions of the microbial loop by Williams and Azam and co-workers carefully observed that the transfer of bacterial production to higher trophic levels by a chain of protistan consumers was likely to be inefficient. Nonetheless, there was early speculation that this newly discovered pathway could represent a significant link or energy bonus to higher levels as opposed to a metabolic sink. This view was bolstered by evidence of high gross growth efficiencies (e.g., carbon growth ¼ 50– 70% of carbon substrate uptake) for bacteria under optimal laboratory conditions and observations that the growth efficiencies of small protists were also much higher and less sensitive to food concentration that those of metazoan consumers, like copepods. There is little support these days for significant transfer of DOM uptake by bacteria through the protistan grazing chain. For one, experimental studies, now available from many marine ecosystems, suggest an average bacterial growth efficiency of about 20% on naturally occurring organic substrates. In other words, 5 moles of dissolved organic carbon are needed to produce 1 mole C of bacteria, the remainder being metabolized to inorganic carbon. Virus-induced cell lysis, the so-called viral shunt to DOM, represents a further loss of potential production to trophic transfer. Lastly, each step in the grazing chain, operating at about 30% efficiency, takes its toll. Assuming half of bacterial mortality goes to viral lysis and half to small bacterivorous flagellates, less than 0.3% (0.2  0.5  0.3  0.3  0.3 ¼ 0.0027) of the DOM carbon uptake would be transferred past the largest protistan consumers in Figure 1. While such calculations can diminish one’s expectations for supporting significant fish production from bacteria per se, we must take a broader view of microbial contributions to plankton energy flows. In Figure 1, for example, bacteria do not constitute the

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROBIAL LOOPS

single source of materials to large heterotrophic protists and the animals that ultimately feed on them. At each step along the grazing chain, production of flagellates and ciliates is well supplemented by consumption of appropriate sizes of photosynthetic organisms. Two-way flows between autotrophic and heterotrophic compartments due to mixotrophs further complicate these interactions, making it difficult to view the contribution of individual components and loops in isolation of the others. To make matters even more complex, metazoan consumers do not as a rule wait patiently to siphon off only those resources that make it to the largest size categories of the microbial grazing chain. Mucus-net feeding pelagic tunicates, like appendicularians and salps, short-circuit the grazing chain by efficiently exploiting bacterial-sized particles, or at least the smallest size categories of nanoplankton. Somewhat less appreciated, the early developmental stages of planktonic crustaceans like copepods and euphausiids, as well as the larvae of benthic invertebrates in coastal areas, typically feed efficiently on the smallest size fractions of the microbial assemblage, graduating to larger prey as they grow. Regardless of the feeding habits of the adults, therefore, the developmental success of these organisms may depend on interactions with the smallest microbial size fractions.

803

significantly with phytoplankton for the uptake of dissolved nutrients. Their superior competitive ability comes, of course, from the high surface area to volume ratios of bacterial cells. Their demand derives from their relatively high nutrient requirements for cell growth compared to phytoplankton. For example, while phytoplankton grow optimally with a carbon to nitrogen ratio (C:N) of about 7, the typical ratio for bacteria is 5. Compared to phytoplankton, bacteria also seem to require iron at about twice the concentration relative to carbon. We can appreciate the interplay among growth efficiencies, nutrient richness of available substrates, and bacteria as a source or sink for recycled nutrients from Figure 2. High growth efficiency allows bacteria to use more nutrients for growth, with less recycled. Reducing the nutrient content (increasing C:N) of available DOM has a similar effect. Where the dissolved substrates are nutritionally rich, growth efficiency is likely to be high, so nutrient release will be positive but depressed. On the other hand, when the C:N ratio of DOM is too high to satisfy nutrient needs for growth, bacteria will seek limiting elements in inorganic pools. It is precisely in this latter case that we run across an interesting paradox of microbial interrelationships. Phytoplankton respond to nutrient-limited conditions by producing photosynthate sugars in excess of their growth needs and excreting them to the external environment. Since these sugars are

Nutrient Cycling

0.15 Release

Gross growth efficiency 20% 25% 30% 50%

0.1 0.05 0

Uptake

The lack of efficient transfer of bacterial production through the long protistan grazing chain suggests that the microbial loop must be extremely important in remineralization and nutrient cycling. This has clearly emerged as one of the primary functions of the microbial food web and has brought particularly a new understanding to the ecology of the open oceans where microbial interactions predominate. Such systems are often limited by primary nutrients such as nitrogen or phosphorus, and in some cases trace elements like iron. Their common characteristic is the relatively high turnover rate of small primary producers supported by the efficient recycling of nutrients. Without this positive feedback of nutrient recycling, the available resources would rapidly be locked into the standing biomass of plankton, and new growth and production would stop. Even though bacteria grow inefficiently on naturally available DOM and help to solubilize and degrade particulate organics with extracellular enzymes, they are typically not the major remineralizers in the seas. In fact, they often compete

_ 0.05 _ 0.1 5

10

15 20 25 C:N ratio of DOM substrates

30

Figure 2 Uptake or release of dissolved inorganic nitrogen (DIN) by bacteria as a function of gross growth efficiency (GGE) and the carbon:nitrogen ratio (C:N) of dissolved organic substrates. Figure shows that bacteria can act as decomposers (releasing nutrients) when substrates are nutrient rich (low C:N). With increasing C:N or increasing carbon growth efficiency, bacteria will show a deficit (negative) in nutrients from DOM and will compete with phytoplankton in nutrient uptake. In comparision, bacterivorous flagellates feeding on relatively nutrient-rich bacteria should always serve as significant decomposers, recycling about 75% of ingested N as DIN or small particulates at GGE ¼ 30%.

(c) 2011 Elsevier Inc. All Rights Reserved.

804

MICROBIAL LOOPS

easily assimilated by bacteria but are devoid of associated nutrients (i.e., high C:N), the effect is to enhance bacterial demand for inorganic nutrients and hence their competition for limiting substrate with phytoplankton. According to some analyses, if one considers the many indirect consequences of this enhancement effect on the microbial network, phytoplankton could ‘win’ in the end by stimulating grazing on bacteria and subsequent nutrient remineralization. However, bacterivorous mixotrophs have an inherent advantage under such conditions since they benefit both from direct consumption of nutrient-rich bacteria and by the stimulation of bacterial growth and nutrient cycling from the DOM released by true autotrophs. At the same time, photosynthetic carbon production allows mixotrophs to utilize ingested nutrients for growth at very high efficiency, releasing little back to the environment compared to pure heterotrophs. As one might imagine, bacterivorous mixotrophs become more important and can even dominate over pure autotrophic or heterotrophic flagellates in systems of increasing oligotrophy.

Regional Patterns and Variations Bacterial abundance and biomass vary in the oceans, within and between regions, but the range of variability is generally less than that observed for larger components of the food web. This is because bacteria and small algae in the microbial food web can be contained within certain limits by fast-growing protistan predators. In contrast, larger phytoplankton, such as diatoms, enjoy a substantial growth rate advantage over slow-responding mesozooplankton consumers, allowing them to increase explosively when light and nutrient conditions become optimal. Such cells give many regions of the oceans their characteristic seasonal blooms. Plankton blooms are generally short-lived phenomena because the larger components of the food web are not good at retaining nutrients in the surface waters once the water column is seasonally stratified. Large cells aggregate and sink when nutrients are exhausted, and mesozooplankton export nutrientrich material from surface waters as compact, fastsinking fecal pellets. Increasing nutrient stress naturally favors smaller competitors for the limiting resource and the smaller consumers that feed on them. Therefore, a declining bloom will evolve through various successional states toward a microbially dominated community in which production, grazing, and nutrient remineralization are more tightly coupled. One such transitional state is likely

to be a period in which the bacterial carbon demand overshoots the concurrent production of phytoplankton. This occurs during the declining stages of the bloom, when senescent phytoplankton produce carbohydrates in excess, when sick phytoplankton cells lyse, or when diminished antibacterial chemical defenses of phytoplankton allow bacteria to more effectively exploit the DOM accumulated during the bloom. Reduced carbon supply and enhanced mortality to bacterivorous protists and viruses will rapidly bring the bacteria back into balance. If we consider the full range of variability in the oceans, it is possible to find very rich coastal ecosystems or events with chlorophyll concentrations of 30 mg l1 or more and bacterial abundances exceeding 107 cells ml1. However, such extremes are relatively rare. Average concentrations are about two orders of magnitude lower for phytoplankton chlorophyll and 10-fold lower for bacteria. Particularly in the open oceans, many regions share similar characteristics with regard to general low levels of bacterial and phytoplankton standing stocks. If one looks, for example, at mean levels of bacterial abundance and biomass in tropical and subtropical seas, they vary quite little, regardless of whether the regions are relatively rich and productive (Arabian Sea), iron-limited (equatorial Pacific), or extremely oligotrophic (subtropical Pacific) (Table 1). As an indication of the selective pressures for small primary producers in such systems, photosynthetic bacteria (Prochlorococcus and Synechococcus) usually account for 40–50% of total bacterial biomass, with the relative abundance shifting toward Synechococcus in richer more-productive systems or seasons. As we move toward higher latitudes, photosynthetic bacteria decline in importance relative to eukaryotic phytoplankton, and bacteria as a group become increasingly more heterotrophic. Prochlorococcus are largely absent from the plankton at water temperatures below 121C, while Synechococcus are rare in polar waters. The importance of the microbial communities in the ecology of the oceans derives from the flow of nutrients and fixed carbon through them. This is in addition to the many roles that bacteria have with respect to chemical transformations and ocean geochemistry. When one takes into account the low growth efficiencies of bacteria and the deficiencies of the carbon-14 method, typical bacterial production estimates on the order of 5–20% of 14C-bicarbonate uptake (Table 1) are consistent with a carbon demand of about 50% of primary production. In addition, protistan grazers directly consume 50–90% of phytoplankton cellular growth in the open oceans, and often half or more in coastal waters. Through

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROBIAL LOOPS

805

Table 1 Representative estimates of bacterial cell abundances, carbon biomass, and percentage heterotrophic cell biomass for several regions of the open oceans according to recent studies Region

Central Equatorial Pacific, 11S–11N Subtropical Pacific, 231N Arabian Sea, NE Monsoon Tropical Atlantic, 5–201N Subtropical Atlantic, 301N Subarctic Atlantic, 50–601N Southern Ocean bloom, 471S, 61W Southern Ocean, 601S, 1701W

Abundance (10 3 cells ml 1)

Biomass (mg C l 1)

T ( 1C)

Hbact

PRO

SYN

28 25 27 27 22 12 4 0

670 550 850 750 500 1500 2000 300

140 210 70 210 70 20 ND 0

7 1 90 12 7 40 ND 0

14 14 22 18 9 23 24 4

% Hetero

Chla (mg l 1)

BP/PP

59 47 47 51 66 79 – 100

0.26 0.06 0.40 0.41 0.07 0.87 3.0 0.4

0.17 0.22 0.05 0.15 0.16

T (1C) and Chla are mean environmental temperature and total phytoplanton chlorophyll a. BP/PP is the ratio of bacterial (heterotrophic) production to total primary production. Total bacterial biomass is for combined populations of heterotrophic cells (Hbact), Prochlorococcus (PRO) and Synechococcus (SYN) determined from cell counts and mean carbon contents of 12, 35 and 100  10  15 g C per cell. ND, not determined.

these two routes, most of the organic production of the oceans is dissipated in the microbial food web.

Conclusion The last 20–30 years have been a period of unprecedented discovery relating to the microbial ecology of the oceans. Largely ignored only a short while ago, microbial food web interactions are now central to our understanding of energy and nutrient flows in the oceans. The ubiquitous and self-regulating microbial community provides the foundation upon which the rest of the food web operates. Though sometimes overridden by the dynamics of larger bloom-forming organisms, it emerges as the dominant trophic pathway in most open-ocean regions and the end point of community succession when nutrients become limiting. In contrast to the ecology of larger organisms in the seas, we know little about the dynamics and unique contributions of microbes at the species level. This remains an exciting area of research for the future.

Glossary Autotrophs Primary producers, organisms that utilize only inorganic carbon for metabolic synthesis. Bacterivory Consumption of bacteria. DOM Dissolved organic matter; operationally, all organic matter that can pass through a filter with 0.2 mm pores. Gross growth efficiency For a given organism, the efficiency of conversion of carbon intake into new carbon growth.

Heterotrophs Organisms that utilize organic sources of carbon (particulate or dissolved) for metabolic synthesis. Microplankton Planktonic organisms in the size range 20–200 mm; includes single-celled as well as multicellular organisms. Mixotrophic Organisms with a mixed mode of nutrition, typically combining the ability to derive significant nutrition from photosynthesis as well as feeding directly on other organisms (or dissolved substrates). Nanoplankton Planktonic singled-celled organisms in the size range 2–20 mm. Oligotrophic System characterized by low concentrations of nutrients and plankton biomass. Picoplankton Planktonic singled-celled organisms in the size range 0.2–2 mm.

See also Bacterioplankton. Photochemical Processes. Phytoplankton Blooms. Primary Production Distribution. Primary Production Methods. Primary Production Processes.

Further Reading Azam F, Fenchel T, Gray JG, Meyer-Reil LA, and Thingstad T (1983) The ecological role of water-column microbes in the sea. Marine Ecology Progress Series 10: 257--263. Fenchel T (1987) Ecology of Protozoa. The Biology of Free-living Phagotrophic Protists. Madison, WI: Brock/ Springer Science Tech.

(c) 2011 Elsevier Inc. All Rights Reserved.

806

MICROBIAL LOOPS

Fuhrman JA (1999) Marine viruses and their biogeochemical and ecological effects. Nature 399: 541--548. Hobbie JE and Williams PJ leB (eds.) (1984) Heterotrophic Activity in the Sea, NATO Conference Series. IV, vol. 15. New York: Plenum Press. Partensky F, Hess WR, and Vaulot D (1999) Prochlorococcus, a marine photosynthetic prokaryote of global significance. Microbiology and Molecular Biology Reviews 63: 106--127.

Raven JA (1997) Phagotrophy in phototrophs. Limnology and Oceanography 42: 198--205. Thingstad FT (1998) A theoretical approach to structuring mechanisms in the pelagic food web. Hydrobiologia 363: 59--72. Williams PJ leB (1981) Incorporation of microheterotrophic processes into the classical paradigm of the planktonic food web. Kieler Meeresforsch. Sondh 5: 1–28.

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROPHYTOBENTHOS G. J. C. Underwood, University of Essex, Colchester, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1770–1777, & 2001, Elsevier Ltd.

Introduction Microphytobenthos is a descriptive term for the diverse assemblages of photosynthetic diatoms, cyanobacteria, flagellates, and green algae that inhabit the surface layer of sediments in marine systems. Microphytobenthos occur wherever light penetrates to the sediment’s surface, and are abundant on intertidal mud and sandflats and in shallow subtidal regions. Microphytobenthic primary production may be high, matching that of phytoplankton in the overlying water column, yet this activity is compressed into a biofilm only a few millimeters thick. The relationship between irradiance and rates of microphytobenthic photosynthesis is fairly well understood, but new methods are revealing fine-scale effects of microspatial distribution within the vertical light profile and migration of cells throughout the diel illumination period. Patterns of biomass distribution and seasonal and spatial changes in species composition are well described, but studies differ on the relative importance of the factors influencing microphytobenthic biomass (irradiance, resuspension, nutrients, grazing, exposure, desiccation, etc.). Microphytobenthic biofilms play an important role in mediating the exchange of nutrients across the sediment–water interface, and microphytobenthos both stimulate and compete with various bacterial sediment processes. The presence of biofilms rich in extracellular polysaccharides alters the erosional properties of sediments, termed biostabilization. Types of Microphytobenthos

Sediment properties play a major role in determining the type of microphytobenthic assemblage present in a particular environment. Sediments consisting of fine silts and clays (less that 63 mm) are termed cohesive sediments. The fine nature of such material and the lack of suitable attachment points result in assemblages dominated by motile microphytobenthic species. These are termed ‘epipelic’ biofilms (epipelic: living on mud), and the microphytobenthos are sometimes termed ‘epipelon.’ Sediments consisting of larger particles, silty sands, and sands are

noncohesive, with greater pore space, and are generally more often disturbed. Growing attached to individual sand and silt particles are found ‘epipsammic’ taxa (epipsammic: living on sand). Epispammic assemblages usually contain a substantial proportion of epipelic taxa as well. Epipelic biofilms The commonest epipelic microphytobenthos are biraphid diatoms, with the genera Navicula, Gyrosigma, Nitzschia and Diploneis usually well represented (Table 1). In fine sediment habitats, light penetration is very limited and, in order to photosynthesize, cells need to be able to position themselves at the sediment surface. Biraphid diatoms move by excreting extracellular polymeric substances (EPS) from the raphe slit present in each of the silica cell walls (valves) that make up the cell. Cyanobacterial filaments move by gliding and nonflagellated euglenids move by amoeboid movement. In dense biofilms of epipelic diatoms, the concentrations of EPS can become high (200–300 mg g1 dry weight of sediment), providing a carbon source to the sediment system. High concentration of EPS can increase the force needed to erode sediments, termed ‘biostabilization.’ Epipelic biofilms can be very extensive on intertidal estuarine mudflats, where they can contribute up to 50% of estuarine carbon budgets. Epipsammic assemblages The ‘epipsammon’ are generally nonmotile, or only partially mobile. Diatoms are the major constituents, with araphid and monoraphid genera common (e.g., Opephora, Achnanthes, Amphora, and Cocconeis) (Table 1). Table 1 Genera of photoautotrophs commonly found in microphytobenthic communities Algal group

Epipelic

Epipsammic

Cyanobacteria

Oscillatoria

Bacillariophyta

Navicula Amphora Fallacia Staurophora Gyrosigma Pleurosigma Nitzschia Diploneis Cylindrotheca Euglena

Oscillatoria Microcoleus Spirulina Opephora Raphoneis Achnanthes Cocconeis Fragilaria Navicula Nitzschia Amphora

Euglenophyta Chlorophyta

(c) 2011 Elsevier Inc. All Rights Reserved.

Euglena Many

807

MICROPHYTOBENTHOS

Primary Production Photosynthesis

Microphytobenthos are photoautotrophic organisms. Hourly rates of primary production are high, Table 2 Daily and annual rates of primary production for epipelic and epipsammic microphytobenthos from a number of different habitats Site

Epipelon Ems-Dollard, Netherlandsa Tagus Estuary, Portugalb North Inlet, SC, USAc Langebaan Lagoon, South Africad Epipsammon Langebaan Lagoon, South Africad Laholm Bay, Swedene Ria de Arosa, Spainf Weeks Bay, AL, USAg a

Daily production (mg C m2 d1)

Annual production (g C m2 a1)

600–1370 5–32 (h1) – 17–69

62–276 47–178 56–234 253 (mud)

17–69

63 (sand)

10–200 – 10–750

0.3–20 54 90.1

with annual primary production ranging between 0.3 and 234 g C m2 a1 (Table 2). Different techniques are used to measure microphytobenthic primary production: (1) oxygen exchange across the sediment–water interface; (2) (14C)bicarbonate uptake in intact biofilms; (3) (14C)bicarbonate uptake in slurries; (4) oxygen production within biofilms using oxygen microelectrodes; and more recently (5) modulated chlorophyll a fluorescence techniques (Figure 1). These techniques all measure slightly different aspects of photosynthesis, making intercomparisons difficult. Oxygen exchange measurements on intact biofilms measure net community production, and if the oxygen uptake rate (negative) in the dark is subtracted from the net community production, then a measure of gross oxygen production is obtained (assuming that respiration in the light and dark do not differ). (14C)Bicarbonate uptake into intact biofilms measures net photosynthesis, and tends to underestimate carbon fixation rates as it is not possible to measure accurately the specific 14C activity within the thin photosynthetically active layer. In noncohesive sediments, percolation of sea water of known specific 14 C activity into the biofilm through the application of a slight vacuum to the bottom of a sediment core results in higher estimates of carbon fixation. Percolation techniques cannot be used with cohesive

Uptake Community respiration Net community production Gross oxygen production

Dark (heterotrophic) uptake Light uptake Net photosynthesis Potential net photosynthesis

O2 concentration profile (gross/net)

Colijn, F & De Jonge, V (1984) Marine Ecology Progress Series 14: 185–196. b Brotas, V & Catarino, F (1995) Netherlands Journal of Aquatic Ecology, 29: 333–339. c Pinckney, JL (1994) In: Biostabilization of Sediments (ed. WE Krumbein, DM Paterson and LJ Stal). Universita¨t Oldenburg, Oldenburg. pp. 55–84. d Fielding P et al. (1988) Estuarine Coastal shelf Science. 27: 413–426. e Sundba¨ck, K & Jo¨nsson, B (1988) Journal of Experimental Marine Biology and Ecology 122: 63–81. f Varela, M & Penas, E (1985) Marine Ecologoy Progress Series 25: 111–119. g Schreiber, RA & Pennock, JR (1995) Ophelia 42: 335–352.

Gross photosynthesis (bars)

Efficiency of PSII-estimated electron transport rate

Efflux 14 Oxygen microOxygen C uptake: electrodes intact sediments/slurries exchange

Epipsammic cells attach themselves to sand particles by a pad or short stalk of EPS, though many cells are also capable of movement. Filamentous and colonial cyanobacteria (Oscillatoria, Microcoleus), coccal green algae and motile flagellates and chlorophytes are common in epipsammic assemblages. Thus epipsammic assemblages often have a greater taxonomic diversity (at the level of algal groups). Light penetration is greater into sandy sediments, which are also disturbed by tidal and wind-induced currents. Cells are therefore frequently mixed within the sediment photic zone, and the requirement for motility is less. Indeed, in highly mixed systems, nonattached, motile taxa may be absent, and only attached species are found, often within depressions present on the surface of sand grains, where they receive protection from abrasion.

Fluorescence techniques

808

Figure 1 Techniques used to determine microphytobenthic primary production measure different aspects of photosynthesis in microphytobenthic biofilms, and have varying scales of vertical and horizontal resolution. Thus comparison of data needs to be made with care. PSII, photosystem II.

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROPHYTOBENTHOS

sediments. Oxygen exchange and 14C methods require the microphytobenthic community to be submerged and this may underestimate intertidal primary production, where the majority of the photosynthesis occurs during low tide exposure. 14C slurry techniques are a rapid method for measuring photosynthetic parameters, with photosynthesis versus light curves generated in a ‘photosynthetron.’ However, existing microgradients in the sediment are destroyed in slurries, and this technique therefore measures maximum potential primary production, in the absence of structure within the biofilm. Oxygen microelectrodes measure gross primary production rates at small-scale (100–200 mm) depth intervals down a profile into the sediment. Construction of complete photosynthesis profile curves is time-consuming; the time taken to generate sufficient replicate production profiles is greater than some of the temporal properties of the biofilm (e.g., endogenous vertical migration). To avoid this, production rates can be calculated from the profile of oxygen concentration with depth under a fixed irradiance, assuming diffusion and porosity coefficients. Net oxygen production can be calculated from the slope in oxygen concentrations out of the sediment, but with exposed sediments this can be problematic. Significant amounts of variation in oxygen production profiles can be due to patchiness in the distribution of microphytobenthic biomass. Oxygen microelectrodes are an

809

important tool for measuring the microspatial distribution of photosynthesis within sediments and response of photosynthesis to environmental variation, but scaling-up of these measurements to larger areal rates is contentious. Variable fluorescent techniques measure the activity of the photosystem II (PSII) reaction centre, thus providing an estimate of the rate of production of electrons by the water-splitting system of PSII (electron transport rate). Being noninvasive, fluorescence techniques can be used to rapidly and repeatedly measure in situ activity. As oxygen is a product of the water-splitting process, there is a relationship between oxygen production and PSII electron transport rate (ETR), and also reasonable linearity between 14C-fixation rates and ETR, especially in sediment slurries. Thus fluorescence techniques can provide an indirect (but nondestructive and rapid) measurement of microphytobenthic primary production. However, the relationship between ETR and oxygen evolution or 14C fixation can become nonlinear at high irradiances, and vertical migration of cells within the biofilm can complicate the interpretation of results. Variable fluorescence measurements can also be made on single cells, using a modified fluorescence microscope and image analysis techniques, allowing the photosynthetic response of single cells within a mixed population to be measured in undisturbed biofilms (Figure 2).

Figure 2 Fluorescence images of intact epipelic microphytobenthic biofilms, showing patchiness at a microscale in cell distribution, and differences in cell size (E ¼ Euglena sp.; P ¼ Pleurosigma angulatum; P1 ¼ Plagiotropis vitrea; D ¼ Diploneis didyma; Pg ¼ Petrodictyon gemma; S ¼ Staurophora sp.; N ¼ small Navicula species). Fluorescence imaging techniques can calculate the photosynthetic efficiency of individual cells within the biofilm, allowing taxonomic differences to be determined. (Images courtesy of ARM Hanlon, University of Essex.)

(c) 2011 Elsevier Inc. All Rights Reserved.

810

MICROPHYTOBENTHOS

Light Penetration and Photosynthesis

There are substantial spatial and temporal gradients in light availability in microphytobenthic habitats. Irradiance can exceed 2000 mmol photons m2 s1 on exposed intertidal sediments, while in clear shallow water sufficient light can penetrate to depths of 20–30 m, permitting microphytobenthic growth. Steep gradients of irradiance occur within sediments, where the attenuation of light is rapid (attenuation coefficients (k) between 1 and 3.5 mm1 for sandy and cohesive sediments, respectively). Thus the euphotic zone depth in sediments (1% of incident light) is usually much less than 1 cm (less than 2 mm in cohesive muddy sediments) (Figure 3). Light intensity just beneath the sediment surface, particularly at wavelengths 4700 nm can be greater than the incident light, owing to backscatter effects within the sediment. The spectral quality of light also changes

_

_1

Gross oxygen production (pmol O2 cm 3 h )

(A) 0

2

4

6

8

10

0

Depth (µm)

_ 500

_ 1000 O2 concentration vs depth O2 production Light % vs depth

_ 1500

_ 2000 0

100

300 500 200 400 _1 Oxygen concentration (µmol l ) _

(B) 0

600

_

Gross oxygen production (pmol O2 cm 3 h 1) 4 2 6 8

10

0

Depth (µm)

_ 500 _ 1000 O2 concentration vs depth O2 production Light % vs depth

_ 1500

_ 2000 0

300 500 400 200 100 _1 Oxygen concentration (µmol l )

within sediments and is further modified by increased light attenuation of specific wavelengths (particularly blue and red) due to absorption by microalgal photopigments. There is a fairly clear relationship between biomass-normalized primary production (mg C(mg Chla)1 h1, termed PB) and irradiance in microphytobenthic systems, up to saturating irradiances (PB max ). Irradiance accounts for between 30% and 60% of the variability in primary production, and biomass explains another 30–40%. Within cohesive sediments the majority of photosynthesis occurs within the top 200–400 mm of the sediment. In sandy sediments, where light penetration is greater, gross photosynthesis can occur deeper than this (up to 2 mm) (Figure 3) and may even show a biomodal distribution owing to distinct vertical separation of diatoms and cyanobacterial layers. Isolated microphytobenthos (i.e., in slurries, lens tissue preparations or cultures) reach PB max at light intensities between 100 and 800 mmol photons m2 s1 and show photoinhibition of PB at higher light intensities. Depth-integrated rates of sediment photosynthesis obtained from in situ oxygen microelectrode measurements saturate at higher irradiances than slurries and show little or no evidence of photoinhibition. In undisturbed sediments, the peak of gross oxygen production occurs deeper in the sediment at high light intensities (41200 mmol photons m2 s1) than at lower light intensities, mainly because of migration of the bulk of the microalgal population down into the sediment away from high surface irradiance. Microphytobenthos are sensitive to light intensity and UVB radiation, with surface biomass varying with irradiance. Some subtidal assemblages are shade-adapted and migrate down into the sediment at midday to avoid high light levels. Taxonomic differences occur with regard to positioning with the light field. The euglenophyte Euglena deses commonly occurs on intertidal flats and at high irradiance occurs on the surface of sediments, with epipelic diatoms underneath. Mixed assemblages of filamentous cyanobacteria and epipelic diatoms also show vertical positioning, with cyanobacteria positioned beneath the diatom layer. The ability of cells to migrate away from high irradiance allows microphytobenthos to respond to the light climate and position themselves at optimal irradiances.

600

Figure 3 Typical oxygen concentration profiles and rates of gross oxygen production in an epipelic (A) and epipsammic (B) microphytobenthic biofilm. Light attenuation is less steep in sandier sediments, and thus oxygen production occurs to a greater depth.

Vertical migration Following disturbance and/ or deposition of fresh sediment, microphytobenthos need to reposition themselves back within the euphotic zone to photosynthesize. In many intertidal habitats the microphytobenthos exhibit endogenous rhythms of vertical migration, with

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROPHYTOBENTHOS

migration remaining in synchrony with the daily shift of the tidal cycle within the diel light frame. These endogenous rhythms can be maintained for between 3 and 4 days in the absence of any light or tidal stimuli. Intertidal sites are also subject to varying patterns of diel illumination periods. The shifting pattern of tidal exposure (E55 min per day) within diel light curves, and the fortnightly cycle of spring and neap tides can result in periods when microphytobenthos are exposed to very high irradiance during low tide at solar noon (exceeding 2000 mmol photons m2 s1) and at other times (several days for some regions of the intertidal) when little or no light reaches the sediment surface. Thus cells need to be able to cope with periods of darkness, when they rely on intracellular carbon storage compounds (glucans) as an energy source. Temperature effects on microphytobenthic photosynthesis The temperature of a mudflat can change rapidly during a tidal emersion period, at up to 2–31C h1, with daily ranges of 201C and seasonal ranges between 0 and 351C. There is a clear relationship between PB max and temperature, with an optimal temperature for intertidal diatoms of 251C, while at temperatures above 251C there can be significant inhibition (30%) of microphytobenthic photosynthesis, particularly on upper shore intertidal regions. This can lead to reductions of biomass on upper shores. Extracellular Polysaccharide (EPS) Production and Sediment Biostabilization

Epipelic and epipsammic diatoms produce EPS either during motility or as an attachment structure. Microphytobenthos also excrete surplus photoassimilated carbon as carbohydrates when they are nutrient limited and subject to high irradiance. In diatom-rich biofilms, between 20% and 40% of the extracellular carbohydrate material present is polymeric, i.e., EPS. The remainder consists of nonpolymeric material, mainly simple sugars, leachates, and other photoassimilates. These low molecular weight exudates are rapidly utilized by bacteria, and may play a significant role in the ecology of cohesive sediments by providing bacteria with a readily available carbon source. Carbohydrate concentrations in sediments are much more a function of the epipelic rather than epipsammic (attached) diatom biomass, and within more mixed assemblages of photosynthetic microorganisms (cyanobacterial mats, high-saltmarsh algal assemblages), the close relationship between colloidal carbohydrate and chlorophyll a concentrations present in diatomdominated sediments is not present.

811

In epipelic diatoms, production of EPS requires between 0.1% and 16% of photosynthetically fixed carbon. EPS is produced both during illuminated and darkened periods. In conditions of darkness, the relative amounts of EPS produced increase, possibly linked to increased cellular motility. Extrusion of EPS by pennate diatoms is an active metabolic process, as vesicles filled with polymeric material are transported from the Golgi body to the raphe. Motility is generated by the extrusion of polymers through the cell membrane within the raphe, and the polymer strand is moved along the raphe by actin fibres. In dark conditions, internal storage carbohydrates (glucans) are metabolized to provide the carbon and energy sources needed to produce EPS. These mechanisms provide a route for the production of extracellular carbohydrate material into the surrounding sediments. The EPS produced by microphytobenthos diatoms binds together sediment particles and can form smooth surface layers. The binding strength of exopolymers varies with chemical composition and the degree of cross-linkage; and as polymers dehydrate during tidal exposure, their binding strength increases. Thus during tidal exposure there is an increase in concentrations (due to diatom photosynthesis and motility) and a reduction in sediment water content. This can significantly increase the critical shear required for sediment erosion (by up to 300%), when the tide covers the site. In epipsammic biofilms, sand particles can be stuck together by pads and fibers of EPS, as well as by filaments of cyanobacteria. These processes all result in biostabilization, and the presence of microphytobenthos significantly changes the sedimentological properties of their habitat.

Distribution and Biomass Small-scale Heterogeneity in Microphytobenthos

Microphytobenthos show a high degree of spatial heterogeneity in biomass and species composition. This patchiness occurs on a scale of micrometers to many tens of meters. There are also patterns of vertical distribution within sediments, with the bulk of the active biomass (determined as chlorophyll a) found within the top few millimeters of cohesive sediments, and the top centimeter of sandy sediments. However, viable cells and chlorophyll a can be isolated from deeper layers, up to 10–15 cm. Given the shallow photic depth in most sediments, only the algae in the uppermost depths of the sediments will be photosynthetically active. Yet many microphytobenthos can survive prolonged periods (2–3 weeks) of darkness, and there is some limited

(c) 2011 Elsevier Inc. All Rights Reserved.

812

MICROPHYTOBENTHOS

evidence of heterotrophy. Thus buried cells may, if mixed back to the surface, resume photosynthesis. Large-scale Heterogeneity

Sediment type is a major determining factor in the abundance and biomass of microphytobenthos. Sandy silts and sands support significantly lower concentrations of microalgal biomass than sites with fine cohesive sediments (chlorophyll a concentrations ranging from 1 to 560 mg m2 or 0.1–460 mg g1 sediment). As sediment grain size increases, the proportion of epipelic, motile taxa decreases and microphytobenthic assemblages in intertidal sands consist predominantly of smaller epipsammic taxa. Sands tend to have lower nutrient concentrations and are more frequently resuspended than are cohesive sediments, and all of these characteristics contribute toward lower microphytobenthic biomass. On intertidal mudflats, microphytobenthic biomass tends to be greater toward the upper shore. Lower shore sediments have a higher water content and are less stable than sediments at the middle and upper shore, partly owing to the energy of tidal flow and regular resuspension of sediments. Periods of illuminated exposure are shorter on the low intertidal where light penetration is restricted by highly turbid estuarine waters. Thus low shore microphytobenthos are probably light limited (in terms of available photoperiod per 24 h), while biomass accumulation is prevented by frequent disturbance. At higher tidal heights on a shore, the pattern of illuminated emersion periods and reduced resuspension contribute to create conditions favorable for epipelic microphytobenthos. However, upper shore stations are also subject to greater desiccation and temperature effects, the effect of which can be increased by long periods of exposure during neap tide periods. These factors usually result in a unimodal distribution of biomass across an intertidal flat, with the peak somewhere between mid-tide level and mean high water neap tide level, and not necessarily at the highest bathymetric level. In subtidal habitats, microphytobenthic biomass tends to decrease with increasing water depth owing to increasing light limitation. However, very shallow (o1 m) sediments in exposed situations are more prone to mixing and disturbance due to wave action or tidal flows, and thus biomass decreases in such sites. Temporal Variation

In temperate latitudes, increases in epipelic microphytobenthic biomass tend to occur during the summer months. However, peaks of biomass also occur frequently at other times of the year, and in

many estuarine systems epipelic diatom assemblages are less seasonally influenced than are phytoplankton communities. High temporal variability in biomass is a common feature of epipelic microphytobenthos, with biomass dependent on local environmental changes such as erosion and deposition events, desiccation linked to tidal exposure and weather conditions, and periods of rapid growth. Rapid doubling times (1–2 days) permit microphytobenthos to increase rapidly in density during favorable conditions. Subtidal microphytobenthos are not subjected to the extremes of exposure present on the intertidal, with irradiances ranging from very high during exposure to virtually nil during immersion in turbid overlying water. Subtidal microphytobenthos tends to show greater degrees of seasonality, with peaks of biomass and activity following the annual pattern of irradiance.

Response of Microphytobenthos to Nutrients Nutrient Limitation

The potential for nutrient limitation of microphytobenthos depends in part on the sediment type concerned. Fine cohesive sediments usually have high organic matter contents, with high rates of bacterial mineralization and high porewater concentrations of dissolved nutrients, while sand flats are more oligotrophic. There is therefore an increased possibility that microphytobenthos inhabiting sediments of a larger grain size will be nutrient limited. The spatial distribution of sediments within estuaries is also pertinent to whether nutrient limitation will occur, in that many estuaries exhibit significant nutrient gradients along their length and areas of extensive mudflats supporting microphytobenthos may coincide with regions of high nutrient concentration. There are few experimental data showing nutrient effects on intertidal microphytobenthos independently of other covarying factors that also affect primary production and biomass (shelter, salinity, etc.). Nutrient enrichment experiments on mudflats have found no consistent short-term pattern of increased photosynthesis or biomass, though long-term reductions (over 16 years) in nutrient inputs in estuaries have been shown to result in declines in biomass. In contrast, enrichment experiments in subtidal epipsammic microphytobenthos and cyanobacterial mats in nutrient-poor habitats have shown varying degrees of stimulation of microphytobenthic photosynthesis and biomass. It is generally considered that epipelic microalgae are not nutrient limited and that they obtained nutrients

(c) 2011 Elsevier Inc. All Rights Reserved.

MICROPHYTOBENTHOS

both from within the sediment, particularly during migration during tidal immersion, and from the overlying water. However, the term ‘nutrient limitation’ includes both Liebig-type limitation (on final biomass) and short term (Monod type) effects on rates of photosynthesis/growth. To what extent short-term nutrient dynamics within biofilms influence the rate of photosynthesis and growth of microphytobenthos is not known. As porewater concentrations of many nutrients (e.g., ammonium, phosphate) increase with depth within the sediment, cells exhibiting vertical migration may obtain nutrients when they have migrated away from the surface. Although this seems logically sound, there is as yet no experimental evidence to support this hypothesis. The nutrient environment is important in determining species composition. Ammonium concentrations in sediments influence the distribution of diatom species in both saltmarshes and mudflats. Concentrations of ammonium between 500 and 1000 mmol l1 are selective for some taxa of microalgae, with the toxic effects of ammonia being enhanced in high pH conditions. Sediment organic content and tolerance to sulfide also influence the species composition of microalgal biofilms. Given the steep gradients in pH, oxygen and sulfide within fine cohesive sediments, these are likely to be important selective factors determining the species diversity of microphytobenthic assemblages.

813

the high oxygen concentrations present in the surface sediments during photosynthesis. Photosynthetic production of oxygen increases the depth of the surface oxic layer, which increases the oxidation of vertically diffusing reduced molecules such as sulfide, ammonium, and phosphate. Thus the export fluxes of these compounds across the sediment–water interface can be significantly reduced compared to the fluxes under dark conditions (Figure 4). Assimilation by the algae of nutrients (ammonium, nitrite, nitrate, phosphate, CO2, dissolved organic carbon) diffusing into the biofilm both from overlying water and from deeper layers within the sediments also alters the pattern of exchange fluxes. On intertidal mudflats, microphytobenthic biofilms develop an ammonium demand during periods of tidal exposure and photosynthesis that persists for up to 4 h after tidal cover. Bacterial denitrification (the reduction of nitrate to nitrogen gas) is an important process in coastal sediments as it is the only mechanism by which nitrogen can be permanently removed from the marine environment. Microphytobenthos influence denitrification in a number of ways (Figure 4). The position of the microphytobenthos on the surface of the sediment allows them to assimilate nitrate from the overlying water column and reduce the amount diffusing into the sediment. Denitrification is an anaerobic process, and photosynthetic oxygen production increases the depth in the sediments at which it can occur, thereby increasing the diffusional path length for nitrate. By these processes, microphytobenthos reduce denitrification of nitrate from the water column. However, oxygen production can stimulate nitrification (production of nitrate from ammonium), and this nitrate can then be denitrified. Stimulation of this ‘coupled nitrification–denitrification’ pathway can be particularly significant,

Interaction between Microphytobenthos and Nutrient Cycling

The activity of microphytobenthos biofilms can significantly affect the fluxes of nutrients across the sediment–water interface (Figure 4). This is due to assimilation of nutrients by the algae from the overlying water and underlying porewaters, and to

N2 and N2O

+

Overlying water

NH4

2_

PO4 /SiO3

_

NO3

_

MPB biofilm

Aerobic zone

_

+ NH4

NO3

Anaerobic zone +

NH4

2_

_

PO4 /SiO3

NO3

_

Figure 4 Interactions between microphytobenthos and nutrient fluxes across the sediment–water interface. Open arrows indicate fluxes/processes stimulated by the action of microphytobenthic primary production. Solid arrows represent processes reduced by microphytobenthic activity.

(c) 2011 Elsevier Inc. All Rights Reserved.

814

MICROPHYTOBENTHOS

especially in low-nutrient environments. By these processes, microphytobenthos influence the nutrient dynamics of shallow water sediments (Figure 4). These processes will be affected by spatial and temporal (both diel and seasonal) differences in biofilm biomass, activity, and species composition. There is some evidence to suggest that differences in species composition effect the ability of biofilms to sequester C and N compounds from the overlying water.

See also Benthic Boundary Layer Effects. Geomorphology. Marine Mats. Microbial Loops. Nitrogen Cycle. Phytobenthos. Primary Production Methods. Salt Marshes and Mud Flats. Sandy Beaches, Biology of.

Further Reading Admiraal W (1984) The ecology of estuarine sedimentinhibiting diatoms. Progress in Phycology Research 3: 269--322. Decho AW (1990) Microbial exopolymer secretions in ocean environments: their role(s) in food webs and marine processes. Oceanography and Marine Biology Annual Reviews 28: 73--153.

MacIntyre HL, Geider RJ, and Miller DC (1996) Microphytobenthos: the ecological role of the ‘secret garden’ of unvegetated, shallow-water marine habitats. I. Distribution, abundance and primary production. Estuaries 19: 186--201. Miller DC, Geider RJ, and MacIntyre HL (1996) Microphytobenthos: the ecological role of the ‘secret garden’ of unvegetated, shallow-water marine habitats. II. Role in sediment stability and shallow water food webs. Estuaries 19: 202--212. Paterson DM (1994) Microbial mediation of sediment structure and behaviour. In: Stal LJ Caumette P (eds.) NATO ASI Series vol. G35. Microbial Mats pp. 97–109. Berlin: Springer Verlag. Round FE, Crawford RM, and Mann DG (1990) The Diatoms, Biology and Morphology of the Genera. Cambridge: Cambridge University Press. Sullivan MJ (1999) Applied diatom studies in estuarine and shallow coastal environments. In: Stoermer EF and Smol JP (eds.) The Diatoms: Applications for the Environmental and Earth Sciences, pp. 334--351. Cambridge: Cambridge University Press. Sundba¨ck K, Nilsson C, Nilsson P, and Jo¨nsson B (1996) Balance between autotrophic and heterotrophic components and processes in microbenthic communities of sandy sediments: a field study. Estuarine and Coastal Shelf Science 43: 689--706. Underwood GJC and Kromkamp J (1999) Primary production by phytoplankton and microphytobenthos in estuaries. Advances in Ecological Research 29: 93--153.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY M. R. Perfit, Department of Geological Sciences, University of Florida, Gainsville, FL, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1778–1788, & 2001, Elsevier Ltd.

Introduction The most volcanically active regions of our planet are concentrated along the axes of the globe, encircling mid-ocean ridges. These undersea mountain ranges, and most of the oceanic crust, result from the complex interplay between magmatic (i.e., eruptions of lavas on the surface and intrusion of magma at depth) and tectonic (i.e., faulting, thrusting, and rifting of the solid portions of the outer layer of the earth) processes. Magmatic and tectonic processes are directly related to the driving forces that cause plate tectonics and seafloor spreading. Exploration of mid-ocean ridges by submersible, remotely operated vehicles (ROV), deep-sea cameras, and other remote sensing devices has provided clear evidence of the effects of recent magmatic activity (e.g., young lavas, hot springs, hydrothermal vents and plumes) along these divergent plate boundaries. Eruptions are rarely observed because of their great depths and remote locations. However, over 60% of Earth’s magma flux (approximately 21 km3 year1) currently occurs along divergent plate margins. Geophysical imaging, detailed mapping, and sampling of midocean ridges and fracture zones between ridge segments followed by laboratory petrologic and geochemical analyses of recovered rocks provide us with a great deal of information about the composition and evolution of the oceanic crust and the processes that generate mid-ocean ridge basalts (MORB). Mid-ocean ridges are not continuous but rather broken up into various scale segments reflecting breaks in the volcanic plumbing systems that feed the axial zone of magmatism. Recent hypotheses suggest that the shallowest and widest portions of ridge segments correspond to robust areas of magmatism, whereas deep, narrow zones are relatively magmastarved. The unusually elevated segments of some ridges (e.g., south of Iceland, central portion of the Galapagos Rift, Mid-Atlantic Ridge near the Azores) are directly related to the influence of nearby mantle

plumes or hot spots that result in voluminous magmatism. Major differences in the morphology, structure, and scales of magmatism along mid-ocean ridges vary with the rate of spreading. Slowly diverging plate boundaries, which have low volcanic output, are dominated by faulting and tectonism whereas fast-spreading boundaries are controlled more by volcanism. The region along the plate boundary within which volcanic eruptions and high-temperature hydrothermal activity are concentrated is called the neovolcanic zone. The width of the neovolcanic zone, its structure, and the style of volcanism within it, vary considerably with spreading rate. In all cases, the neovolcanic zone on mid-ocean ridges is marked by a roughly linear depression or trough (axial summit collapse trough, ASCT), similar to rift zones in some subaerial volcanoes, but quite different from the circular craters and calderas associated with typical central-vent volcanoes. Not all mid-ocean ridge volcanism occurs along the neovolcanic zone. Relatively small (o1 km high), near-axis seamounts are common within a few tens of kilometers of fast and intermediate spreading ridges. Recent evidence also suggests that significant amounts of volcanism may occur up to 4 km from the axis as off-axis mounds and ridges, or associated with faulting and the formation of abyssal hills. Lava morphology on slow spreading ridges is dominantly bulbous, pillow lava (Figure 1A), which tends to construct hummocks (o50 m high, o500 m diameter), hummocky ridges (1–2 km long), or small circular seamounts (10s–100s of meters high and 100s–1000s of meters in diameter) that commonly coalesce to form axial volcanic ridges (AVR) along the valley floor of the axial rift zone. On fast spreading ridges, lavas are dominantly oblong, lobate flows and fluid sheet flows that vary from remarkably flat and thin (o4 cm) to ropy and jumbled varieties (Figure 1). Although the data are somewhat limited, calculated volumes of individual flow units that have been documented on mid-ocean ridges show an inverse exponential relationship to spreading rate, contrary to what might be expected. The largest eruptive units are mounds and cones in the axis of the northern Mid-Atlantic Ridge whereas the smallest units are thin sheet/lobate flows on the East Pacific Rise. Morphologic, petrologic, and structural studies of many ridge segments suggest they evolve

(c) 2011 Elsevier Inc. All Rights Reserved.

815

816

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

Figure 1 Examples of different morphologies, surface textures and sediment cover on lava flows on the northern East Pacific Rise. Digital images were taken from heights of 5–10 m above the seafloor using the Woods Hole Oceanographic Instution’s camera system. The dimensions of the photographs are approximately 4.5 m  3.0 m. (A) Pillow lava. (B) Hackly or scrambled flow. (C) Lobate lava. (D) Lineated sheet flow. (E) Ropy sheet flow. (F) Collapse structure in lobate flows. (G) A young flow contact on top of older flows. (H) Heavily sediment covered lobate flows with small fissure. Images from Kuras et al. 2000.

through cycles of accretion related to magmatic output followed by amagmatic periods dominated by faulting and extension.

Magma Generation Primary MORB magmas are generated by partial melting of the upper mantle; believed to be composed of a rock type termed peridotite which is primarily composed of the minerals olivine, pyroxenes (enstatite and diopside), and minor spinel or garnet.

Beneath ridges, mantle moves upward, in part, due to convection in the mantle but possibly more in response to the removal of the lithospheric lid above it, which is spreading laterally. Melting is affected by the decompression of hot, buoyant peridotite that crosses the melting point (solidus curve) for mantle material as it rises to shallow depths (o100 km), beneath the ridges. Melting continues as the mantle rises as long as the temperature of the peridotite remains above the solidus temperature at a given depth. As the seafloor spreads, basaltic melts formed in a broad region (10s to 100s of kilometers) beneath

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

the ridge accumulate and focus so that they feed a relatively narrow region (a few kilometers) along the axis of the ridge (Figure 1). During ascent from the mantle and cooling in the crust, primary mantle melts are subjected to a variety of physical and chemical processes such as fractional crystallization, magma mixing, crustal assimilation, and thermogravitational diffusion that modify and differentiate the original melt composition. Consequently, primary melts are unlikely to erupt on the seafloor without undergoing some modification. Picritic lavas and magnesian glasses thought to represent likely primary basalts have been recovered from a few ocean floor localities; commonly in transform faults (Table 1). MgO contents in these basalts range from B10 wt% to over 15 wt% and the lavas typically contain significant amounts of olivine crystals. Based on comparisons with high-pressure melting experiments of likely mantle peridotites, the observed range of compositions may reflect variations in source composition and mineralogy (in part controlled by pressure), depth and percentage melting (largely due to temperature differences), and/or types of melting (e.g., batch vs. fractional).

Ocean Floor Volcanism and Construction of the Crust Oceanic crust formed at spreading ridges is relatively homogeneous in thickness and composition compared to continental crust. On average, oceanic crust is 6–7 km thick and basaltic in composition as compared to the continental crust which averages 35–40 km thick and has a roughly andesitic composition. The entire thickness of the oceanic crust has not been sampled in situ and therefore the bulk composition has been estimated based on investigations of ophiolites (fragments of oceanic and back-arc crust that have been thrust up on to the continents), comparisons of the seismic structure of the oceanic crust with laboratory determinations of seismic velocities in known rock types, and samples recovered from the ocean floor by dredging, drilling, submersibles, and remotely operated vehicles. Rapid cooling of MORB magmas when they come into contact with cold sea water results in the formation of glassy to finely crystalline pillows, lobate flows, or sheet flows (Figure 1). These lava flows typically have an B0.5–1 cm-thick outer rind of glass and a fine-grained, crystalline interior containing only a few percent of millimeter-sized crystals of olivine, plagioclase, and more rarely clinopyroxene in a microscopic matrix of the same minerals. MORB lavas erupt, flow, and accumulate to form the

817

uppermost volcanic layer (Seismic Layer 2A) of ocean crust (Figure 2). Magmas that do not reach the seafloor cool more slowly with increasing depth forming intrusive dikes at shallow levels (0.5–3 km) in the crust (layer 2B) and thick bodies of coarsely crystalline gabbros and cumulate ultramafic rocks at the lowest levels (3–7 km) of the crust (layer 3) (Figure 2). Although most magma delivered to a MOR is focused within the neovolcanic zone, defined by the axial summit collapse trough or axial valley, off-axis volcanism and near-axis seamount formation appear to add significant volumes of material to the uppermost crust formed along ridge crests. In some portions of the fast spreading East Pacific Rise, off-axis eruptions appear to be related to syntectonic volcanism and the formation of abyssal hills. Near-axis seamount formation is common along both the East Pacific Rise and medium spreading rate Juan de Fuca Ridge. Even in areas where there are abundant offaxis seamounts they may add only a few percent to the volume of the extrusive crust. More detailed studies of off-axis sections of ridges are needed before accurate estimates of their contribution to the total volume of the oceanic crust can be made. Oceanic transform faults are supposed to be plate boundaries where crust is neither created nor destroyed, but recent mapping and sampling indicate that magmatism occurs in some transform domains. Volcanism occurs in these locales either at short, intratransform spreading centers or at localized eruptive centers within shear zones or relay zones between the small spreading centers.

Mid-ocean Ridge Basalt Composition Ocean floor lavas erupted along mid-ocean ridges are low-potassium tholeiites that can range in composition from picrites with high MgO contents to ferrobasalts and FeTi basalts containing lower MgO and high concentrations of FeO and TiO2, and even to rare, silica-enriched lavas known as icelandites, ferroandesites and rhyodacites (Table 1). In most areas, the range of lava compositions, from MgO-rich basalt to FeTi basalt and ultimately to rhyodacite, is generally ascribed to the effects of shallow-level (low-pressure) fractional crystallization in a subaxial magma chamber or lens (Figure 2). A pronounced iron-enrichment trend with decreasing magnesium contents (related to decreasing temperature) in suites of genetically related lavas is, in part, what classifies MORB as tholeiitic or part of the tholeiitic magmatic suite (Figure 3).

(c) 2011 Elsevier Inc. All Rights Reserved.

(c) 2011 Elsevier Inc. All Rights Reserved.

50.64 1.43 15.17 10.45 0.19 7.53 11.62 2.51 0.11 0.14 99.61 7.70 2148

50.49 1.78 14.55 10.87 0.20 7.22 11.58 2.74 0.13 0.17 99.62 7.49 2303

SiO2 TiO2 Al2O3 FeO * MnO MgO CaO Na2O K2O P2O5 Sum K/Ti N¼

50.41 1.54 14.75 11.19 nd 7.49 11.69 2.28 0.10 0.14 99.60 6.24 867

Galapagos 50.03 1.28 15.97 9.26 0.15 8.06 12.21 2.68 0.08 0.13 99.73 6.10 623

Seamounts 48.80 0.97 17.12 8.00 0.14 10.28 11.93 2.32 0.03 0.07 100 3.0 10

Pacific Picritic 50.61 2.36 13.30 13.61 0.23 5.92 10.43 2.74 0.16 0.22 99.39 7.04 706

Pacific Ferrobasalt 55.37 2.10 12.92 13.11 0.21 3.64 8.05 3.33 0.44 0.40 99.40 13.80 97

Pacific High-silica

50.10 1.86 15.69 9.78 0.19 7.00 11.17 3.04 0.43 0.24 99.37 22.26 304

Pacific

Enriched

51.02 1.46 15.36 9.56 0.18 7.31 11.54 2.52 0.36 0.19 99.31 23.67 972

Atlantic

6.93 10.90 3.16 0.66 0.32 99.14 32.77 65

49.17 1.94 16.86 9.21

Galapagos

50.19 1.74 16.71 8.77 0.15 6.80 10.67 3.38 0.75 0.33 99.34 39.05 197

Seamounts

Analyses done by electron microprobe on natural glasses at the Smithsonian Institution in Washington, D.C. (by W. Melson and T.O’Hearn) except the picritic samples that were analyzed at the USGS in Denver, Co. Enriched MORB in this compilation are any that have K/Ti values greater than 13. High-silica lavas have SiO2 values between 52 and 64. K/Ti ¼ (K2O/TiO2)  100. N ¼ number of samples used in average. FeO * ¼ total Fe as FeO.

Atlantic

Pacific

Normal

Average compositions of normal and enriched types of basalts from mid-ocean ridges and seamounts

Oxide wt%

Table 1

818 MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

819

Figure 2 Diagrammatic three-dimensional representation of oceanic crust formed along a fast-spreading ridge showing the seismically determined layers and their known or inferred petrologic composition. Note that although most of the volcanism at midocean ridges appears to be focused within the axial summit trough, a significant amount of off-axis volcanism (often forming pillow mounds or ridges) is believed to occur. Much of the geochemical variability that is observed in MORB probably occurs within the crystal–liquid mush zone and thin magma lens that underlie the ridge crest. The Moho marks the seismic boundary between plutonic rocks that are gabbroic in composition and those that are mostly ultramafic but may have formed by crystal accumulation in the crust.

14

4.0 TiO2 3.0

12

CaO

10 8

2.0

6

1.0

4 2 18

20

Al2O3

FeO total 15

16

10

14

5

12 10

0 0 1 2 3 4 5 6 7 8 9 MgO (wt %)

0 1 2 3 4 5 6 7 8 9 MgO (wt %)

Figure 3 Major element variations in MORB from the Eastern Galapagos Spreading Center showing the chemical trends generated by shallow-level fractional crystallization in the oceanic crust. The rocks range in composition from basalt to ferrobasalt and FeTi basalt to andesite.

Although MORB are petrologically similar to tholeiitic basalts erupted on oceanic islands (OIB), MORB are readily distinguished from OIB based on their comparatively low concentrations of large ion

lithophile elements (including K, Rb, Ba, Cs), light rare earth elements (LREE), volatile elements and other trace elements such as Th, U, Nb, Ta, and Pb that are considered highly incompatible during melting of mantle mineral assemblages. In other words, the most incompatible elements will be the most highly concentrated in partial melts from primitive mantle peridotite. On normalized elemental abundance diagrams and rare earth element plots (Figure 4), normal MORB (N-type or NMORB) exhibit characteristic smooth concave-down patterns reflecting the fact that they were derived from incompatible element-depleted mantle. Isotopic investigations have conclusively shown that values of the radiogenic isotopes of Sr, Nd, Hf and Pb in NMORB are consistent with their depleted characteristics and indicate incompatible element depletion via one or more episodes of partial melting of upper mantle sources beginning more than 1 billion years ago. Compared to ocean island basalts and lavas erupted in arc or continental settings, MORB comprise a relatively homogeneous and easily distiguishable rock association. Even so, MORB vary from very depleted varieties (D-MORB) to those containing moderately elevated incompatible element abundances and more radiogenic isotopes. These less-depleted MORB are called E-types (E-MORB)

(c) 2011 Elsevier Inc. All Rights Reserved.

820

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

100

Andesite

Chondrite normalized

FeTi

MORB 10

1 La Ce Pr Nd

Sm Eu Gd Tb Dy Ho

Y

Er Tm Yb

Element

Figure 4 Chondrite-normalized rare earth element (REE) abundances in a suite of cogenetic lavas from the Eastern Galapagos Spreading Center (also shown in Figures 3 and 6). Increasing abundances of REE and the size of the negative europium anomaly from MORB to andesite are consistent with evolution of the suite primarily by fractional crystallization. Concave-down patterns are an indication of their ‘normal’ depleted chemical character (N-MORB).

or P-types, indicative of an ‘enriched’ or ‘plume’ component (Table 1) typically associated with intraplate ‘hot spots’. Transitional varieties are classified as T-MORB. Enriched MORB are volumetrically minor on most normal ridge segments, but can comprise a significant proportion of the crust around regions influenced by plume magmatism such as the Galapagos Islands, the Azores, Tristan, Bouvet, and Iceland.

Mineralogy of Mid-ocean Ridge Basalts The minerals that crystallize from MORB magmas are not only dependent on the composition of the melt, but also the temperature and pressure during crystallization. Because the majority of MORB magmas have relatively similar major element compositions and probably begin to crystallize within the uppermost mantle and oceanic crust (pressures less than 0.3 GPa), they have similar mineralogy. Textures (including grain size) vary depending on nucleation and crystallization rates. Hence lavas, that are quenched when erupted into sea water, have few phenocrysts in a glassy to cryptocrystalline matrix. Conversely, magmas that cool slowly in subaxial reservoirs or magma chambers form gabbros that are totally crystalline (holocrystalline) and composed of well-formed minerals that can be up to a few centimeters long. Many of the gabbros recovered from the ocean floor do not represent melt compositions but rather reflect the accumulation of crystals and percolation of melt that occurs during convection,

deformation and fractional crystallization in the mush zone hypothesized to exist beneath some midocean ridges (Figure 2). These cumulate gabbros are composed of minerals that have settled (or floated) out of cooling MORB magmas and their textures often reflect compaction, magmatic sedimentation, and deformation. MOR lavas may contain millimeter-sized phenocrysts of the silicate minerals plagioclase (solid solution that ranges from CaAl2Si2O8 to NaAlSi3O8) and olivine (Mg2SiO4 to Fe2SiO4) and less commonly, clinopyroxene (Ca[Mg,Fe]Si2O6). Spinel, a Cr-Al rich oxide, is a common accessory phase in more magnesian lavas where it is often enclosed in larger olivine crystals. Olivine is abundant in the most MgO-rich lavas, becomes less abundant in more evolved lavas and is ultimately replaced by pigeonite (a low-Ca pyroxene) in FeO-rich basalts and andesite. Clinopyroxene is only common as a phenocryst phase in relatively evolved lavas. Titanomagnetite, ilmenite and rare apatite are present as microphenocrysts, although not abundantly, in basaltic andesites and andesites. Intrusive rocks, which cool slowly within the oceanic crust, have similar mineralogy but are holocrystalline and typically much coarser grained. Dikes form fine- to medium-grained diabase containing olivine, plagioclase and clinopyroxene as the major phases, with minor amounts of ilmenite and magnetite. Gabbros vary from medium-grained to very coarse-grained with crystals up to a few centimeters in length. Because of their cumulate nature and extended cooling histories, gabbros often exhibit layering of crystals and have the widest mineralogic variation. Similar to MORB, the least-evolved varieties (troctolites) consist almost entirely of plagioclase and olivine. Some gabbros can be nearly monomineralic such as anorthosites (plagioclaserich) or contain monomineralic layers (such as olivine that forms layers or lenses of a rock called dunite). The most commonly recovered varieties of gabbro are composed of plagioclase, augite (a clinopyroxene) and hypersthene (orthopyroxene) with minor amounts of olivine, ilmenite and magnetite and, in some cases, hornblende (a hydrous Fe-Mg silicate that forms during the latest stages of crystallization). Highly evolved liquids cool to form ferrogabbros and even rarer silica-rich plutonics known as trondhjemites or plagiogranites. The descriptions above pertain only to those portions of the oceanic crust that have not been tectonized or chemically altered. Because of the dynamic nature of oceanic ridges and the pervasive hydrothermal circulation related to magmatism, it is common for the basaltic rocks comprising the crust

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

Chemical Variability Although MORB form a relatively homogeneous population of rock types when compared to lavas erupted at other tectonic localities, there are subtle, yet significant, chemical differences in their chemistry due to variability in source composition, depth and extent of melting, magma mixing, and processes that modify primary magmas in the shallow lithosphere. Chemical differences between MORB exist on all scales, from individual flows erupted along the same ridge segment (e.g., CoAxial Segment of the Juan de Fuca Ridge) to the average composition of basalts from the global ridge system (e.g. Mid-Atlantic Ridge vs. East Pacific Rise). High-density sampling along several MOR segments has shown that quite a diversity of lava compositions can be erupted over short time (10s–100 years) and length scales (100 m to a few kilometers). Slow spreading ridges, which do not have steady-state magma bodies, generally erupt more mafic lavas compared to fast spreading ridges where magmas are more heavily influenced by fractional crystallization in shallow magma bodies. Intermediate rate-spreading centers, where magma lenses may be small and intermittent, show characteristics of both slow- and-fast spreading centers. In environments where magma supply is low or mixing is inhibited, such as proximal to transform faults, propagating rift tips and overlapping spreading centers, compositionally diverse and highly differentiated lavas are commonly found (such as the Eastern Galapagos Spreading Center, Figures 3, 4 and 6). In these environments, extensive fractional crystallization is a consequence of relatively cooler thermal regimes and the magmatic processes associated with rift propagation. Local variability in MORB can be divided into two categories: (1) those due to processes that affect

an individual parental magma (e.g., fractional crystallization, assimilation) and (2) those created via partial melting and transport in a single melting regime (e.g., melting in a rising diapir). In contrast, global variations reflect regional variations in mantle source chemistry and temperature, as well as the averaging of melts derived from diverse melting regimes (e.g. accumulative polybaric fractional melting). At any given segment of MOR, variations may be due to various combinations of these processes. Local Variability

Chemical trends defined by suites of related MOR lavas are primarily due to progressive fractional crystallization of variable combinations and proportions of olivine, plagioclase and clinopyroxene as a magma cools. The compositional ‘path’ that a magma takes is known as its liquid line of descent (LLD). Slightly different trajectories of LLDs (Figure 5) are a consequence of the order of crystallization and the different proportions of crystallizing phases that are controlled by initial (and subsequent changing) liquid composition, temperature, and pressure. In some MORB suites, linear elemental trends may be due to mixing of primitive magmas with more evolved magmas that have evolved along an LLD. Suites of MORB glasses often define distinctive LLDs that match those determined by experimental crystallization of MORB at low to moderate

ol-pl-

cpx

4.0 Na2O (wt %)

to be chemically altered and metamorphosed. When this occurs, the primary minerals are recrystallized or replaced by a variety of secondary minerals such as smectite, albite, chlorite, epidote, and amphibole that are more stable under lower temperature and more hydrous conditions. MOR basalts, diabases and gabbros are commonly metamorphosed to greenschists and amphibolites. Plutonic rocks and portions of the upper mantle rich in olivine and pyroxene are transformed into serpentinites. Oceanic metamorphic rocks are commonly recovered from transform faults, fracture zones and slowly spreading segments of the MOR where tectonism and faulting facilitate deep penetration of sea water into the crust and upper mantle.

821

3.0

ol-pl

Cayman Kane Clipperton AMAR Kolbeinsey

ol-pl-

cpx

ol-pl

2.0

ol-pl-cp

x

6.0

7.0 8.0 MgO (wt %)

9.0

Figure 5 MgO vs. Na2O in MORB from five different Ridge segments (Mid Cayman Rise in the Caribbean; near Kane Fracture Zone on the Mid-Atlantic Ridge, 231N; AMAR on the Mid-Atlantic Ridge around 371N; East Pacific Rise near the Clipperton Fracture Zone around 101N; Kolbeinsey Ridge north of Iceland. Lines are calculated Liquid Lines of Descent (LLDs) from high MgO parents. Bar shows where clinopyroxene joins plagioclase and olivine as a fractionating phase. Na8 is determined by the values of Na2O when the LLD is at MgO of 8 wt%. (Adapted with permission from Langmuir et al., 1992.)

(c) 2011 Elsevier Inc. All Rights Reserved.

822

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

pressures that correspond to depths of B1 to 10 km within the oceanic crust and upper mantle. Much of the major element data from fast-spreading ridges like the East Pacific Rise are best explained by low-pressure (B0.1 GPa) fractional crystallization whereas at slow-spreading ridges like the Mid-Atlantic Ridge data require higher pressure crystallization (B0.5–1.0 GPa). This is consistent with other evidence suggesting that magmas at fastspreading ridges evolve in a shallow magma lens or chambers and that magmas at slow-spreading ridges evolve at significantly greater depths; possibly in the mantle lithosphere or at the crust–mantle boundary. Estimated depths of crystallization correlate with increased depths of magma lens or fault rupture depth related to decreasing spreading rate. Cogenetic lavas (those from the same or similar primary melts) generated by fractional crystallization exhibit up to 10-fold enrichments of incompatible trace elements (e.g., Zr, Nb, Y, Ba, Rb, REE) that covary with indices of fractionation such as decreasing MgO (Figure 6) and increasing K2O concentrations and relatively constant incompatible

Figure 6 Trace element (Zr and Ce) versus MgO variation diagram showing the systematic enrichments of these highly incompatible elements with increasing fractionation in a suite of cogenetic lavas from the Eastern Galapagos Spreading Center.

trace element ratios irrespective of rock type. In general, the rare earth elements show systematic increases in abundance through the fractionation sequence from MORB to andesite (Figure 4) with a slight increase in light rare earth elements relative to the heavy-rare earth elements. The overall enrichments in the trivalent rare earth elements is a consequence of their incompatibility in the crystals separating from the cooling magma. Increasing negative Eu anomalies develop in more fractionated lavas due to the continued removal of plagioclase during crystallization because Eu partially substitutes for Ca in plagioclase which is removed during fractional crystallization. Global Variability

MORB chemistry of individual ridge segments (local scale) is, in general, controlled by the relative balance between tectonic and magmatic activity, which in turn may determine whether a steady-state magma chamber exists, and for how long. Ultimately, the tectonomagmatic evolution is controlled by temporal variations in input of melt from the mantle. Global correlation of abyssal peridotite and MORB geochemical data suggest that the extent of mantle melting beneath normal ridge segments increases with increasing spreading rate and that both ridge morphology and lava composition are related to spreading rate. The depths at which primary MORB melts form and equilibrate with surrounding mantle remain controversial (possibly 30 to 100 km), as does the mechanism(s) of flow of magma and solid mantle beneath divergent plate boundaries. The debate is critical for understanding the dynamics of plate spreading and is focused on whether flow is ‘passive’ plate driven flow or ‘active’ buoyantly driven solid convection. At present, geological and geophysical observations support passive flow which causes melts from a broad region of upwelling and melting to converge in a narrow zone at ridge crests. It has also been hypothesized that melting beneath ridges is a dynamic, near-fractional process during which the pressure, temperature, and composition of the upper mantle change. Variations in these parameters as well as in the geometry of the melting region result in the generation of MORB with different chemical characteristics. Differences in the major element compositions of MORB from different parts of the world’s oceans (global scale) have been recognized for some time. In general, it has been shown that N-type MORB from slow-spreading ridges such as the Mid-Atlantic Ridge are more primitive (higher MgO) and have

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

greater Na2O, Al2O3 and lower FeO and CaO/Al2O3 contents at given MgO values than lavas from medium- and fast-spreading ridges (Figure 7). A comparison of ocean floor glass compositions (over 9000) analyzed by electron microprobe at the Smithsonian Institution from major spreading centers and seamounts is presented in Table 1. The analyses have been filtered into normal (N-MORB) and enriched (E-MORB) varieties based on their K/Ti ratios (E-MORB [K2O/ TiO2]  100413) which reflect enrichment in the highly incompatible elements. These data indicate that on average, MORB are relatively differentiated compared to magmas that might be generated directly from the mantle (compare averages with picritic basalts from the Pacific in Table 1). Furthermore, given the variability of glass compositions in each region, N-MORB have quite similar average major element compositions (most elemental concentrations overlap at the 1-sigma level). E-MORB, are more evolved than N-MORB 4.5 Mid-Atlantic Ridge (1.0 GPa) East Pacific Rise (0.1 GPa)

4.0

Juan de Fuca Ridge (1.0 GPa)

Na 2O

3.5

3.0

2.5 2.0 18.0

Al 2 O3

16.0

14.0

12.0

10.0 3.0

4.0

5.0

6.0 7.0 MgO

8.0

9.0

10.0

Figure 7 Major element variation diagrams showing compositional ranges from different spreading rate ridges. Generally higher Na2O and Al2O3 concentrations in Mid-Atlantic Ridge (hatchured field) lavas in comparison to MORB from the Juan de Fuca (grey field) and East Pacific Rise (dark field) are shown. Lines show calculated liquid lines of descent at 0.1 and 1.0 GPa for parental magmas from each ridge.

823

from comparable regions of the ocean and there are a higher proportion of E-MORB in the Atlantic (31%) compared to the Pacific (12%) and Galapagos Spreading Center region (7%). Unlike the Atlantic where E-MORB are typically associated with inflated portions of the ridge due to the effects of plume– ridge interaction, East Pacific Rise E-MORB are randomly dispersed along-axis and more commonly recovered off-axis. As well as having higher K2O contents than N-MORB, E-MORB have higher concentrations of P2O5, TiO2, Al2O3 and Na2O and lower concentrations of SiO2, FeO and CaO. Positive correlations exist between these characteristics, incompatible element enrichments and more radiogenic Sr and Nd isotopes in progressively more enriched MORB. Direct comparison of elemental abundances between individual MORB (or even groups) is difficult because of the effects of fractional crystallization. Consequently, fundamental differences in chemical characteristics are generally expressed as differences in parameters such as Na8, Fe8, Al8, Si8 etc. which are the values of these oxides calculated at an MgO content of 8.0 wt% (Figure 5 and 8). When using these normalized values, regionally averaged major element data show a strong correlation with ridge depth and possibly, crustal thickness. MORB with high FeO and low Na2O are sampled from shallow ridge crests with thick crust whereas low FeO– high Na2O MORB are typically recovered from deep ridges with thin crust (Figure 8). This chemical/tectonic correlation gives rise to the so-called ‘global array’. Major element melting models indicate there is a strong correlation between the initial depth of melting and the total amount of melt formed. As a consequence, when temperatures are high enough to initiate melting at great depths, the primary MORB melts contain high FeO, low Na2O and low SiO2. Conversely, if the geothermal gradient is low, melting is restricted to the uppermost part of the upper mantle, and little melt is generated (hence thinner crust) and the basaltic melts contain low FeO, high Na2O and relatively high SiO2. Although the global systematics appear robust, detailed sampling of individual ridge segments have shown MORB from limited areas commonly exhibit chemical correlations that form a ‘local trend’ opposite to the chemical correlations observed globally (e.g., FeO and Na2O show a positive correlation). A local trend may reflect the spectrum of melts formed at different depths beneath one ridge crest rather than the aggregate of all the melt increments. Although the original hypothesis that global variations in MORB major element chemistry are a consequence of total extents of mantle melting and

(c) 2011 Elsevier Inc. All Rights Reserved.

824

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

3.5 Global MORB Regional Averages Na 8

3.0 2.5 2.0 (A)

Fe 8

11 10 9 8 (B) 0

1000 2000 3000 4000 5000 6000 Depth (m)

3.5

See also

Local trend

(C)

3.0 Na 8

Decreasing melting Shallower mean depth

7

2.5

depleted to moderately enriched varieties. The compositional variability in primary MORB result from combinations of differing source compositions, extents and styles of partial melting, and depths of melt formation. The moderately evolved composition of most MORB primarily reflects the effects of crystal fractionation that occurs as the primary melts ascend from the mantle into the cooler crust. Although MORB are relatively homogeneous compared to basalts from other tectonic environments, they exhibit a range of compositions that provide us with information about the composition of the mantle, the influence of plumes, and dynamic magmatic processes that occur to form the most voluminous part of the Earth’s crust.

Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Propagating Rifts and Microplates. Seamounts and Off-Ridge Volcanism.

2.0 Fe 8

Further Reading

1.5 6

7

8

9

10

11

12

Increasing melting Greater mean depth Figure 8 (A) and (B) Global correlations between regional averages of ridge axial depth and the Na8 and Fe8 of MORBs. Different groups of MORB are distinguished. &, Normal ridge segments; B, ridges behind island arcs; ’, ridges influenced by hot spots. (C) Global trend of Na8 vs. Fe8 due to differences in extents and depths of melting. Representative ‘Local trend’ is common along individual portions of some ridges. (Adapted with permission from Langmuir et al., 1992.)

mean pressure of extraction due to variations in mantle temperature, more recent evidence suggests that heterogeneity in the mantle also plays an important role in defining both global and local chemical trends. In particular, U-series data suggest some MDRB melts equilibrate with highly depleted mantle at shallow depths whereas others equilibrate with less depleted garnet pesidatite at depths greater than B80 km.

Conclusions Passive rise of the mantle beneath oceanic spreading centers results in the decompression melting of upwelling peridotite which gives rise to a spectrum of MORB compositions varying from extremely

Batiza R and Niu Y (1992) Petrology and magma chamber processes at the East Pacific Rise – 91300 N. Journal of Geophysical Research 97: 6779--6797. Grove TL, Kinzler RJ, and Bryan WB (1992) Fractionation of Mid-Ocean Ridge Basalt (MORB). In: PhippsMorgan J, Blackman DK, and Sinton J (eds.) Mantle Flow and Melt Generation at Mid-ocean Ridges, Geophys. Monograph 71, pp. 281--310. Washington, DC: American Geophysical Union. Klein EM and Langmuir CH (1987) Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92: 8089--8115. Kurras GJ, Fomari DJ, Edwards MH, Perfit MR, and Smith MC (2000) Volcanic morphology of the East Pacific Rise Crest 91 490 –520 N:1. Implications for volcanic emplacement processes at fast-spreading mid-ocean ridges. Marine Geophysical Research 21: 23--41. Langmuir CH, Klein EM, and Plank T (1992) Petrological systematics of mid-ocean ridge basalts: constraints on melt generation beneath ocean ridges. In: PhippsMorgan J, Blackman DK and Sinton J (eds) Mantle Flow and Melt Generation at Mid-ocean Ridges, Geophysics Monograph 71, Washington, DC: American Geophysical Union, p. 183–280. Lundstrom CC, Sampson DE, Perfit MR, Gill J, and Williams Q (1999) Insight, into mid-ocean ridge basalt petrogenesis: U-series disequilibrium from the Siqueiro, Transform, lamont seamounts, and East Pacific Rise Journal of Geophysical Research 104: 13035–13048.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE GEOCHEMISTRY AND PETROLOGY

Macdonald KC (1998) Linkages between faulting, volcanism, hydrothermal activity and segmentation on fast spreading centers. In: Buck WR, Delaney PT, Karson JA and Lagabrielle Y (eds) Faulting and Magmatism at Mid-ocean ridges, American Geophysics Monograph 106, Washington, DC: American Geophysical Union, p. 27–59. Nicolas A (1990) The Mid-Ocean Ridges. Springer Verlag, Berlin. Niu YL and Batiza R (1997) Trace element evidence from seamounts for recycled oceanic crust in the Eastern Pacific mantle. Earth Planet. Sci. Lett 148: 471--483. Perfit MR and ChadwickWW (1998) Magmatism at midocean ridges: constraints from volcanological and geochemical investigations. In: Buck WR, Delaney PT, Karson JA and Lagabrielle Y (eds), Faulting and Magmatism at Mid-ocean Ridges. American Geophysics Monograph 106, Washington, DC: American Geophysics Union, p. 59–115. Perfit MR and Davidson JP (2000) Plate tectonics and volcanism. In: Sigurdsoon H (ed.) Encyclopedia of Volcanoes. San Diego: Academic Press.

825

Perfit MR, Ridley WI, and Jonasson I (1998) Geologic, petrologic and geochemical relationships between magmatism and massive sulfide mineralization along the eastern Galapagos Spreading Center. Review in Economic Geology 8(4): 75--99. Shen Y and Forsyth DW (1995) Geochemical constraints on initial and final depths of melting beneath mid-ocean ridges. Journal of Geophysical Research 100: 2211--2237. Sigurdsson H (ed.) (2000) Encyclopedia of Volcanoes. Academic Press Sinton JM and Detrick RS (1992) Mid-ocean ridge magma chambers. Journal of Geophysical Research 97: 197--216. Smithsonian Catalog of Basalt Glasses [http://www.nmnh. si.edu/minsci/research/glass/index.htm] Thompson RN (1987) Phase-equilibria constraints on the genesis and magmatic evolution of oceanic basalts. Earth Science Review 24: 161--210.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE S. M. Carbotte, Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1788–1798, & 2001, Elsevier Ltd.

Introduction New crust is created at mid-ocean ridges as the oceanic plates separate and mantle material upwells and melts in response through pressure-release melting. Mantle melts rise to the surface and freeze through a variety of processes to form an internally stratified basaltic crust. Seismic methods permit direct imaging of structures within the crust that result from these magmatic processes and are powerful tools for understanding crustal accretion at ridges. Studies carried out since the mid 1980s have focused on three crustal structures; the uppermost crust formed by eruption of lavas, the magma chamber from which the crust is formed, and the Moho, which marks the crust-to-mantle boundary. Each of these three structures and their main characteristics at different mid-ocean ridges will be described here and implications of these observations for how oceanic crust is created will be summarized. The final section will focus on how crustal structure changes at ridges spreading at different rates, and the prevailing models to account for these variations. Seismic techniques employ sound to create crosssectional views beneath the seafloor, analogous to how X-rays and sonograms are used to image inside human bodies. These methods fall into two categories; reflection studies, which are based on the reflection of near-vertical seismic waves from interfaces where large contrasts in acoustic properties are present; and refraction studies, which exploit the characteristics of seismic energy that travels horizontally as head waves through rock layers. Reflection methods provide continuous images of crustal boundaries and permit efficient mapping of small-scale variations over large regions. Locating these boundaries at their correct depth within the crust requires knowledge of the seismic velocity of crustal rocks, which is poorly constrained from reflection data. Refraction techniques provide detailed information on crustal velocity structure but typically result in relatively sparse measurements that represent large spatial averages. Hence the types of information obtained from reflection and refraction

826

methods are highly complementary and these data are often collected and interpreted together. Much of what we know about the seismic structure of ridges has come from studies of the East Pacific Rise. This is a fast-spreading ridge within the eastern Pacific that extends from the Gulf of California to south of Easter Island. Along this ridge, seafloor topography is relatively smooth and seismic studies have been very successful at imaging the internal structure of the crust. Comparatively little is known from other ridges, in part because fewer experiments have been carried out and in part because, with the rougher topography, imaging is more difficult.

Seismic Layer 2A Early Studies

Seismic layer 2A was first identified in the early 1970s from analysis of refraction data at the Reykjanes Ridge south of Iceland. This layer of low compressional- or P-wave velocities (o3.5 km s1), which comprises the shallowest portion of the oceanic crust (Figure 1) was attributed to extrusive rocks with high porosities due to volcanically generated voids and extensive crustal fracturing. In the late 1980s a bright event corresponding with the base of seismic layer 2A was imaged for the first time using multichannel seismic reflection data. This event is not a true reflection but rather is a refracted arrival resulting from turning waves within a steep velocity gradient zone that marks the base of seismic layer 2A. Within this gradient zone P-wave velocity rapidly increases to velocities typical of seismic layer 2B (45.0 km s1) over a depth interval of B100–300 m (Figure 1A). The 2A event is seen in the far offset traces of reflection data collected with long receiver arrays (42 km) and has been successfully stacked, providing essentially continuous images of the base of layer 2A at mid-ocean ridges. The Geological Significance of the Layer 2A/2B Transition: Is it a Lithological Transition from Extrusives to Dikes or a Porosity Boundary within the Extrusives?

In most recent studies, layer 2A near the ridge axis is assumed to correspond with extrusive rocks and the base of layer 2A with a lithological transition to the sheeted dike section of oceanic crust. The primary evidence cited for this lithological interpretation comes from studies at Hess Deep in the equatorial

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

East Pacific Rise

Deep Sea Drilling Program

Hess deep

9° N

Hole 504B _

_1

2

3

Vp (km s ) 5 4

Vp (km s 1) 6

4 0.0

0.0 Layer 2A

Depth below seafloor (km)

827

Extrusives Mixed zone

0

6 Layer 2A Extrusives

Layer 2B

Mixed zone 0.8

0.8

5

Sheeted dikes

0.8 Sheeted dikes

Isotropic gabbros

1.6 Axial magma chamber

2.4

(A)

(B)

(C)

Figure 1 (A) Seismic velocity with depth for newly formed crust at the East Pacific Rise. Layer 2A, 2B, and the low velocities associated with the axial magma chamber (AMC) are identified. (Data from Vera EE, Buhl P, Mutter JC et al. (1990), Journal of Geophysical Research 95: 15529–15556). (B) Lithological cross section for the upper crust at Hess Deep derived from submersible observations. (Synthesis is from Francheteau J, Armijo R, Cheminee JL et al. (1992) Earth and Planetary Sciences Letters 111: 109– 121. (C) Comparison of P-wave velocities from in situ sonic logging within Deep Sea Drilling Hole 504B and the lithological units observed within the hole. (From Becker K et al. (1988) Proceedings of ODP, Initial reports, Part A, v.111, Ocean Drilling Program. College Station, TX).

eastern Pacific. In this area, observations of fault exposures made from manned submersibles show that the extrusive rocks are B300–400 m thick, similar to the thickness of layer 2A measured near the crest of the East Pacific Rise (compare Figures 1A and 1B). Other researchers have suggested that the base of layer 2A may correspond with a porosity boundary within the extrusive section associated with perhaps a fracture front or hydrothermal alteration. This interpretation is based primarily on observations from a deep crustal hole located off the coast of Costa Rica, which was drilled as part of the Deep Sea Drilling Program (DSDP). Within this hole (504B) a velocity transition zone is found that is located entirely within the extrusive section (Figure 1C). Here, a thin high-porosity section of rubbly basalts and breccia with P-wave velocities of B4.2 km s1 overlies a thick lower-porosity section of extrusives with higher P-wave velocities (5.2 km s1) (Figure 1). However, the relevance of these observations for the geological significance of ridge crest velocity structure is questionable. Crust at DSDP hole 504B is 5.9 My old and it is well established that the seismic velocity of the shallow crust increases with age owing to crustal alteration (see below). Indeed, the velocities within the shallowest extrusives at DSDP 504B (B4 km s1) are much higher than observed at

the ridge crest (2.5–3 km s1), indicating that significant crustal alteration has occurred (compare Figures 1A and 1C). Conclusive evidence regarding the geological nature of seismic layer 2A will likely require drilling or observations of faulted exposures of crust at or near the ridge crest made where seismic observations are also available. At present, the bulk of the existing sparse information favors the lithological interpretation and layer 2A is commonly used as a proxy for the extrusive crust. If this interpretation is correct, mapping the layer 2A/2B boundary provides direct constraints on the eruption and dike injection processes that form the uppermost part of oceanic crust. Characteristics of Layer 2A at Mid-ocean Ridges

Along the crest of the East Pacific Rise, layer 2A is typically 150–250 m thick (Figures 2 and 3). Only minor variations in the thickness of this layer are observed along the ridge crest except near transform faults and other ridge offsets where this layer thickens. Across the ridge axis, layer 2A approximately doubles or triples in thickness over a zone B2–6 km wide indicating extensive accumulation of extrusives within this wide region (Figures 4 and 5). This accumulation may occur through lava flows that travel

(c) 2011 Elsevier Inc. All Rights Reserved.

828

MID-OCEAN RIDGE SEISMIC STRUCTURE

up to several kilometers from their eruption sites at the axis, either over the seafloor or perhaps transported through subsurface lava tubes. Volcanic eruptions that originate off-axis may also contribute to building the extrusive pile. On the flanks of the East Pacific Rise the base of layer 2A roughly follows the undulating abyssal hill relief of the seafloor (Figure 6). Layer 2A is offset at the major faults that bound the abyssal hills. Superimposed on this undulating relief are smaller-scale variations in 2A

thickness (50–100 m) that may reflect local build-up of lavas through ponding at and draping of seafloor faults (Figures 5 and 6). Layer 2A is thicker and more variable in thickness (200–550 m) along the axis of the intermediate spreading Juan de Fuca Ridge, located in the northeast Pacific (see Figure 9). At this ridge, the sparse existing data suggest that layer 2A does not systematically thicken away from the ridge axis, and it appears that lavas accumulate within a narrower

Base of layer 2A

16° 05'

AMC

16° 00'

16° 00'

15° 55' AMC 15° 50' 15° 50'

15° 45'

15° 40'

line 1334

15° 40'

15° 35'

15° 30'

Seafloor

AMC

10 km

15° 30'

15° 25'

_ 105° 30'

_ 105° 25'

_ 105° 20'

_ 105° 15'

2.5

3.0

3.5

4.0

4.5

Time (s) 1.0

2.2 2.4

2.5

2.6 2.7 2.8 Depth (km)

2.9

3.0

3.3

3.9

Figure 2 Example of a multichannel seismic line collected along the axis of the East Pacific Rise showing the base of the extrusive crust (layer 2A) and the reflection from the top of the axial magma chamber (AMC). Right-hand panel shows the bathymetry of the ridge axis with the location of the seismic profile in black line. The dashed lines on the seismic section mark the locations of very small offsets that are observed in the narrow depression along the axis where most active volcanism is concentrated.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

Depth (km)

_ 2.0 _ 2.5

829

Seafloor

_ 3.0

Base of 2A

_ 3.5 _ 4.0 _ 4.5

AMC

_ 5.0 _ 19°

_ 20°

_ 21°

_ 18°

_ 17°

_ 15°

_ 16°

_ 14°

_ 13°

Latitude

Depth (km)

_ 2.0 _ 2.5

Orozco TF Clipperton TF

_ 3.0 _ 3.5 _ 4.0 _ 4.5 _ 5.0 9°

10°

11°

12°

13°

14°

15°

16°

17°

Latitude Figure 3 Cross section along the axis of the southern (top panel) and northern (bottom panel) East Pacific Rise showing depth to seafloor, the base of the extrusive crust, and the axial magma chamber reflection. This compilation includes results from all multichannel reflection data available along this ridge. Labeled arrows show the locations of transform faults. Other arrows mark the locations of smaller discontinuities of the ridge axis known as overlapping spreading centers. (Top panel, data from Hooft EE, Detrick RS and Kent GM (1997) Journal of Geophysical Research 102: 27319–27340. Bottom panel, data from Kent GM, Harding AJ, and Orcutt JA (1993) Journal of Geophysical Research 98: 13945–13696; Detrick RS, Buhl P, Vera E et al. (1987) Nature 326: 35–41; Babcock JM, Harding AJ, Kent GM and Orcutt JA (1998) Journal of Geophysical Research 103: 30451–30467; Carbotte SM, PonceCorrea G and Solomon A (2000) Journal of Geophysical Research 105: 2737–2759).

Southern East Pacific Rise 17° 28'S Base of layer 2A

Time (s)

Line 1106

3.5

Seafloor

4.0 4.5

AMC

0

5 km

Figure 4 Example of a multichannel seismic profile shot across the ridge axis. Figure shows the magma chamber reflection and the event from the base of layer 2A. (Data from Carbotte SM, Mutter JC and Wu L (1997) Journal of Geophysical Research 102: 10165–10184).

zone than at the East Pacific Rise. Along the slow spreading Mid-Atlantic Ridge, the extrusive section is built largely within the floor of the median valley. At all the ridges that have been surveyed to date, extrusive layer thicknesses measured beyond the axial region are very similar (350–600 m).

velocities gradually increase to levels typical of layer 2B by 20–40 Ma. Recent compilations of modern seismic data suggest that layer 2A velocities increase abruptly and at quite young crustal ages (o5 Ma), rather than the gradual change evident in the early data. This increase in the velocity of layer 2A is the primary known change in the seismic structure of oceanic crust with age. This change is commonly attributed to precipitation of low-temperature alteration minerals within cracks and voids in the extrusive section during hydrothermal circulation of sea water through the crust. The detailed geochemical and physical processes associated with how hydrothermal precipitation of minerals affects seismic velocities is poorly understood. However, infilling of voids with alteration minerals is believed to increase the mechanical competency of the crust sufficiently to account for the evolution of layer 2A with age.

Axial Magma Chamber

Evolution of Layer 2A with Age

Global compilations carried out in the mid-1970s showed that the velocity of layer 2A increases as the crust ages away from the ridge axis and that layer 2A

Early Studies

Drawing on observations of the crustal structure of ophiolites (sections of oceanic or oceanic-like crust

(c) 2011 Elsevier Inc. All Rights Reserved.

830

MID-OCEAN RIDGE SEISMIC STRUCTURE

exposed on land) and the geochemistry of seafloor basalts, geologists long believed that mid-ocean ridges were underlain by large, essentially molten, magma reservoirs. However, until the last 15 years, few actual constraints on the dimensions of magma chambers at ridges were available. Early seismic studies on the East Pacific Rise detected a zone of lower seismic velocity beneath the ridge axis as expected for a region containing melt. A bright reflector was also found indicating the presence of a sharp interface with high acoustic impedance contrast within the upper crust. In the mid 1980s an extensive seismic reflection and refraction experiment was carried out on the northern East Pacific Rise by researchers at the University of Rhode Island, Lamont-Doherty Earth Observatory, and Scripps Institution of Oceanography. This study imaged a bright subhorizontal reflector located 1–2 km below seafloor along much of the ridge. In several locations this reflector was found to be phase-reversed relative to the seafloor reflection, indicating it resulted from an interface with an abrupt drop in seismic velocity. Based on its reversed phase and high amplitude, this event is now recognized as a reflection from a largely molten region located at the top of what is commonly referred to as an axial magma chamber. Seismic refraction and tomography experiments show that this reflector overlies a broader zone that extends to the base of the crust within which seismic velocities are reduced relative to normal crust. At shallow depths this low-velocity zone is B2 km wide. It broadens and deepens beneath the ridge flanks and is B10 km wide at the base of the crust. Because of the relatively small velocity anomaly associated with much of this low-velocity zone (o1 km s1), this region is interpreted to be hot, largely solidified rock, and crystal mush containing only a few per cent partial melt. The Characteristics of the Axial Magma Chamber at Mid-ocean Ridges

Several seismic reflection studies have now been carried out along the fast-spreading East Pacific Rise imaging over 1400 km of ridge crest (Figure 3). A reflection from the magma chamber roof is detected beneath B60% of the surveyed region and can be traced continuously in places for tens of kilometers. This reflector is found at a depth of 1–2 km below seafloor and deepens and disappears toward major offsets of the ridge axis, including transform faults and overlapping spreading centers. Most volcanic activity along the East Pacific Rise is concentrated within a narrow depression, o1 km wide, which is interrupted by small steps or offsets that may be the

boundaries between individual dike swarms. In many places, the magma chamber reflector does not disappear beneath these offsets (Figure 2). However, changes in the depth and width of the reflector are often seen. Seismic tomography studies centered at 91300 N on the East Pacific Rise shows that a broader region of low velocities within the crust pinches and narrows beneath two small offsets. These results suggest that segmentation of the axial magma chamber may be associated with the full range of ridge crest offsets observed on the seafloor. Migration of seismic profiles shot perpendicular to the ridge axis reveals that the magma chamber reflection arises from a narrow feature that is typically less than 1 km in width (e.g., Figure 4). Refraction data and waveform studies of the magma chamber reflection suggest that it arises from a thin body of magma a few hundred to perhaps a few tens of meters thick, leading to the notion of a magma lens or sill. Initial studies assumed that this lens contained pure melt. However, recent research suggests that much of the magma lens may have a significant crystal content (425%) with regions of pure melt limited to pockets only a few kilometers or less in length along the axis. Possible magma lens reflections, similar to those imaged beneath the East Pacific Rise, have also been imaged along the intermediate spreading Juan de Fuca Ridge and Costa Rica Rift and at the back-arc spreading center in the Lau Basin. In these areas, reflectors 1–2.5 km wide and at 2.5–3 km depth are detected (see Figures 10 and 11). Diffractions from the edges of these reflectors are shallower than diffractions due to seafloor topography, indicating that these events clearly lie within the crust. However, there is some debate whether these reflections correspond with magma bodies. Refraction studies along the northern Juan de Fuca show no evidence for a low-velocity zone coincident with the intracrustal reflection. In addition, the Juan de Fuca and Costa Rica Rift data are too noisy to allow determination of the polarity of the event. Hence we cannot rule out the possibility that the shallow crustal reflections observed at these ridges are due to an abrupt velocity increase within the crust, perhaps associated with a frozen magma lens or a cracking front, rather than a velocity decrease associated with the presence of melt. Along the slow spreading Mid-Atlantic Ridge, evidence for magma lenses has been found in one location along the Reykjanes Ridge. Here an intracrustal reflection at a depth of B2–5 km is observed, similar to the depths of magma lens events observed beneath portions of the intermediate spreading ridges. The absence of magma lens reflections in seismic

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

17°10'

831

500 m

_ 17° 15'

_ 17° 20'

_ 17°25'

1106

1106

_ 17° 30'

_ 17° 35' _ 113°15'

_ 113° 05'

_ 113° 10'

2.4

2.6

2.8

_ 113°10'

_ 113°15'

3.0

3.2

_ 113° 05'

3.4

Depth (km) Figure 5 Illustration of the thickening of the seismically inferred extrusive crust (layer 2A) across the axis of the southern East Pacific Rise. Left-hand panel: Bathymetry map of the region with the location of cross-axis seismic lines shown in light line. The bold black line shows the location of the narrow depression along the ridge axis where most volcanic activity occurs. The black dots show the width of the region over which the seismically inferred extrusives accumulate as interpreted from the data shown in the right-hand panel. Righthand panel: Thickness of the extrusive crust inferred from the seismic data along each cross-axis line. Black dots mark the location where 2A reaches its maximum thickness away from the axis. Seismic line 1106 shown in Figure 4 is labeled. (Data from Carbotte SM, Mutter JC and Wu L (1997) Journal of Geophysical Research 102: 10165–10184).

data collected elsewhere along the Mid-Atlantic Ridge could be due to the imaging problems associated with the very rough topography of the seafloor typical at this ridge. However, there is also evidence from refraction data and seismicity studies that large, steady-state magma bodies are not present beneath

this ridge. Microearthquake data show that earthquakes can occur to depths of 8 km beneath parts of the Mid-Atlantic Ridge, indicating that the entire crustal section is sufficiently cool for brittle failure. In other areas, slightly reduced velocities within the crust have been identified, indicating warmer

(c) 2011 Elsevier Inc. All Rights Reserved.

832

MID-OCEAN RIDGE SEISMIC STRUCTURE

Normal faulting

Volcanism and faulting

Volcanism ~ 5 km

Volcanism and faulting 1_ 2 km

Normal faulting ≥ 5 km

Plate boundary zone _ 7× VE ~

Figure 6 Schematic representation of the accumulation of the extrusive crust on the fast-spreading East Pacific Rise. The extrusive layer gradual increases in thickness within a wide zone centered on the narrow depression that marks the innermost axis. Normal faults begin to develop at the edges of this volcanic zone but may be buried by the occasional lava flow that reaches this distance. Beyond this zone, large-scale normal faulting occurs that gives rise to the fault-bounded abyssal hills and troughs found on the ridge flanks. (Reproduced from Carbotte SM, Mutter JC, and Wu L (1997), Journal of Geophysical Research 102: 10165–10184).

temperatures and possibly the presence of small pockets of melt. The prevailing model for magma chambers beneath ridges (Figure 7), incorporates the geophysical constraints on chamber dimensions described above as well as geochemical constraints on magma chamber processes. At fast-spreading ridges (Figure 7A), the magma chamber is composed of the narrow and thin melt-rich magma lens that overlies a broader crystal mush zone and surrounding region of hot but solidified rock. The dike injection events and volcanic eruptions that build the upper crust are assumed to tap the magma lens. The lower crust is formed from the crystal residuum within the magma lens and from the broader crystal mush zone. At slow-spreading ridges (Figure 7B) a short-lived dikelike crystal mush zone without a steady-state magma lens is envisioned. At these ridges volcanic eruptions occur and the crystal mush zone is replenished during periodic magma injection events from the mantle.

Moho The base of the crust is marked by the Mohorovcic Discontinuity, or Moho, where P-wave velocities increase from values typical of lower crustal rocks (6.8–7.0 km s1) to mantle velocities (48.0 km s1). The change in P-wave velocity is often sufficiently abrupt that a subhorizontal Moho reflection is observed from which the base of the crust can be mapped. Depth to seismic Moho provides our best estimates of crustal thickness and is used to study how total crustal production varies in different ridge settings.

Characteristics of Moho at Mid-ocean Ridges

Reflection Moho is often imaged in data collected at the East Pacific Rise (Figure 8). In places it can be traced below the region of lower crustal velocities found at the ridge and occasionally beneath the magma lens reflection itself. Depths to seismic Moho indicate average crustal thicknesses of 6–7 km. There is no evidence for thickening away from the ridge crest, indicating that the crust acquires its full thickness within a narrow zone at the axis. The Moho reflection has three characteristic appearances on the East Pacific Rise: as a single, a diffuse, or a shingled event. These variations presumably reflect changes in the structure and composition of the crust-to-mantle transition such as are observed in ophiolites, where the base of the crust can vary from a wide band of alternating lenses of mafic and ultramafic rocks to an abrupt and simple transition zone. At the Mid-Atlantic Ridge, the base of the crust is not marked by a strong Moho reflection such as is imaged at the East Pacific Rise. Here an indistinct boundary is found that is absent in many places. Hence most of our information on crustal thickness at this ridge has come from seismic refraction studies. Detailed refraction surveys are available within a number of locations that show a clear pattern of thinner crust (by 1–4 km) toward transform faults and smaller ridge offsets. These results are interpreted to reflect focused mantle upwelling and greater crustal production within the central regions of ridge segments away from ridge offsets. Significant variations in crustal thickness are also observed along the East Pacific Rise. However, in the

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

833

Fast-spreading ridges Cross-axis volcanics

Depth (km)

1

dikes mush

3 gabbro

transition

zone

5 Moho

Along-axis

Deval

OSC

Depth (km)

volcanics dikes

1

melt mush crystal

mush

3

zone transition

5

Moho

0

2

6

4

8

10

Distance (km)

(A)

Slow-spreading ridges Cross-axis

Depth (km)

2

rift valley

4 6 Moho

gabbro

transition zone mush

8 10 (B)

5

0

5

10

Distance (km)

Figure 7 Schematic representation of the axial magma chamber beneath fast-spreading (A) and slow-spreading (B) ridges. At a fast-spreading ridge a thin zone of predominantly melt (black region) is located at 1–2 km below seafloor that grades downward into a partially solidified crystal mush zone. This region is in turn surrounded by a transition zone of solidified but hot rock. Along the ridge axis, the ‘melt’ sill and crystal mush zone narrows and may disappear at the locations of ridge discontinuities (labeled Deval and OSC in along-axis profile). At a slow-spreading ridge a steady-state melt region is not present. Here a dike-like mush zone forms small silllike intrusive bodies that crystallize to form the oceanic crust. (Reproduced from Sinton JA and Detrick RS (1991) Journal of Geophysical Research 97: 197–216.)

region with the best data constraints (91–101N) the spatial relationships are the opposite of those observed on the Mid-Atlantic Ridge. Within this region, crust is B2 km thinner, not thicker, within the central portion of the segment where a range of ridge crest observations indicate that active crustal accretion is focused. At this fast-spreading ridge the presence of a steady-state magma chamber and broad region of hot rock (Figure 7) may permit efficient redistribution of magma away from regions of focused delivery from the mantle. The absence of a steady-state magma chamber beneath the slow-

spreading Mid-Atlantic Ridge may prohibit significant along-axis transport of magma such that at this ridge thicker crust accumulates at the site of focused melt delivery.

Variations in Crustal Structure with Spreading Rate Spreading rate has long been recognized as a fundamental variable in the crustal accretion process. At slow-spreading ridges new crust is formed within a

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

pronounced topographic depression, whereas at fast spreading ridges a smooth and broad topographic high is found. Gravity anomalies indicate significant variations in crustal and mantle properties along the axis of slow-spreading ridges, whereas they are subdued and quite uniform at fast-spreading ridges. Magnetic anomalies are more complex and often difficult to identify at slow-spreading ridges and seafloor basalts are typically more primitive. These observations indicate that spreading rate plays an important role in how magma is segregated from the mantle and delivered to form new oceanic crust. Seismic techniques provide direct constraints on the significance of spreading rate for the distribution of magma within the crust, total crustal production, and the internal stratification of the crust resulting from the magmatic processes of crustal creation. Some aspects of the seismic structure of ridges are surprisingly similar at all spreading rates. The thickness of the extrusive layer away from the ridge axis is similar (B350–600 m), indicating that the total volume of extrusives produced by seafloor spreading is independent of spreading rate. Average crustal thickness is also comparable at all spreading ridges (6–7 km) and total crustal production does not appear to depend on spreading rate. However, the characteristics of crustal magma lenses and the pattern of accumulation of the extrusive layer are different at fast and slow ridges. Throughout the fast-spreading range (85–150 mm y1) the extrusive section is thin (B200 m) at the ridge axis (Figure 9)and accumulates away from the axis within a zone 2–6 km wide. At these rates, magma lenses are imaged beneath much of the axis and have similar widths (Figure 10) and

9°30'N

3

104°14' W SF

Time (s)

4 FT

AMC

5 I

M

6 7 5 km Figure 8 Multichannel seismic line crossing the East Pacific Rise at 91300 N showing the Moho reflection (labeled M). The seafloor (SF) magma chamber reflection (AMC) and other intracrustal reflections (FT, I) are labeled. (Reproduced from Barth GA and Mutter JC (1996) Journal of Geophysical Research 101: 17951–17975.)

are located at similar depths within the crust (Figure 11). At slower-spreading ridges (o70–80 mm y1) magma lenses appear to be present only intermittently. Where they are observed they lie at a deeper level within the crust (42.5 km) and form a second distinct depth population (Figure 11). The extrusive layer is thicker along the axis and does not systematically thicken away from the ridge. At these ridges the extrusive section appears to acquire its full thickness within the innermost axial zone. These differences in the accumulation of extrusives at fastspreading and slow-spreading ridges could reflect differences in lava and eruption parameters (e.g., eruptive volumes, lava flow viscosity, and morphology) that govern flow thicknesses and the distances lavas may travel from their eruption sites.

CRR

600

Average layer 2A thickness on-axis (m)

834

MAR 500 400 JdF 300

16N

13N 14S 9N 17S

200

100

0 0

50

100

150

200

_1

Full spreading rate (km My ) Figure 9 Thickness of the extrusive crust at the ridge axis versus spreading rate. For data obtained from detailed reflection surveys, average thickness is shown with black dots and standard deviations. East Pacific Rise data are labeled by survey location and are from (16N) Carbotte SM, Ponce-Correa G and Solomon A (2000) Journal of Geophysical Research 105: 2737–2759; (13N) Babcock JM, Harding AJ, Kent GM and Orcutt JA (1998) Journal of Geophysical Reseach 103: 30451–30467; (9N) Harding AJ, Kent GM and Orcutt JA (1993) Journal of Geophysical Research 98: 13925–13944; (14S) Kent GM, Harding AJ, Orcutt JA et al. (1994) Journal of Geophysical Research 99: 9097–9116; (17S) Carbotte SM, Mutter JC and Wu L (1997). Journal of Geophysical Research 102; 10165–10184. Costa Rica Rift (CRR) data are from Buck RW, Carbotte SM, Mutter CZ (1997) Geology 25: 935–938. Data from other ridges are derived from other seismic methods and are shown in stars. Data for the Mid-Atlantic Ridge (MAR) are from Hussenoeder SA, Detrick RS and Kent GM (1997), EOS Transactions AGU, F692. Data for Juan de Fuca Ridge (JdF) are from McDonald MA, Webb SC, Hildebrand JA, Cornuelle BD and Fox CG (1994) Journal of Geophysical Research 99: 4857–4873.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMIC STRUCTURE

RR ?

9NOSC

SEPR 14°_ 20°S NEPR 16°N

1 3000

S

Magma lens depth (km)

Magma lens width (m)

4000

835

East Pacific Rise JdF

2000

16N 9N 13N

1000

17S 14S

N RR

2

JdF CRR

NEPR 9°_13°

3

0 0

50

100

150

Lau

200

_1

Spreading rate (km My ) Figure 10 Width of magma lens reflections beneath ridges versus spreading rate. Average widths (black dots) and standard deviations are shown for regions where detailed reflection surveys have been carried out. East Pacific Rise data are labeled by survey location and are from (16N) Carbotte SM, Ponce-Correa G and Solomon A (2000) Journal of Geophysical Research 105: 2737–2759; (13N) Babcock JM, Harding AJ, Kent GM and Orcutt JA (1998) Journal of Geophysical Research 103: 30451–30467; (9N and 9NOSC) Kent GM, Harding AJ and Orcutt JA (1993) Journal of Geophysical Research 98: 13945–13970, 13971–13996 (the data labeled 9NOSC correspond with an unusually wide lens mapped near an overlapping spreading center); (14S) and (17S) from compilation of Hooft EE, Detrick RS and Kent GM (1997) Journal of Geophysical Research 102: 27319–27340. An estimate of lens width from wide-angle seismic data along the Reykjanes Ridge (RR) is shown in open star: Sinha MC, Navin DA, MacGregor LM et al. (1997) Philosophical Transactions of the Royal Society of London 355(1723): 233– 253. The Juan de Fuca (JdF) estimate is from Morton JL, Sleep NH, Normark WR and Tompkins DH (1987) Journal of Geophysical Research 92: 11315–11326.

What Controls the Depth at which Magma Chambers Reside at Ridges?

Two main hypotheses have been put forward to explain the depths at which magma chambers are found at ridges. One hypothesis is based on the concept of a level of neutral buoyancy for magma within oceanic crust. This model predicts that magma will rise until it reaches a level where the density of the surrounding country rock equals that of the magma. However, at ridges, magma lenses lie at considerably greater depths than the neutral bouyancy level predicted for magma if its density is equivalent to that of lavas erupted onto the seafloor (2700 kg m3). Either the average density of magma is greater, or mechanisms other than neutral buoyancy control magma lens depth. The prevailing hypothesis is that magma chamber depth is controlled by spreading rate-dependent variations in the thermal structure of the ridge. In

4 shallow lens

deep lens 5 0

20

40

60

80

100

120

140

160

_1

Full spreading rate (mm y ) Figure 11 Average depth of magma lens reflections beneath ridges versus spreading rate. The curved line shows the depth to the 12001C isotherm calculated from the ridge thermal model of Phipps Morgan J and Chen YJ (1993) Journal of Geophysical Research 98: 6283–6297. Data from different ridges are labeled Reykjanes Ridge (RR), Juan de Fuca Ridge (JdF), Costa Rica Rift (CRR), Lau Basin (Lau), northern and southern East Pacific Rise (NEPR and SEPR, respectively). (Reproduced from Carbotte SM, Mutter CZ, Mutter J and Ponce-Correa G (1998) Geology 26: 455–458.)

this model, a mechanical boundary such as a freezing horizon or the brittle–ductile transition acts to prevent magma from rising to its level of neutral buoyancy. Both of these horizons will be controlled by the thermal structure of the ridge axis. Compelling support for this hypothesis was provided by the inverse relation between spreading rate and depth to low-velocity zones at ridges apparent in early datasets. Numerical models of ridge thermal structure have been developed that predict systematic changes in the depth to the 12001C isotherm (proxy for basaltic melts) with spreading rates that match the first-order depth trends for magma lenses. This model predicts a minor increase in lens depth within the fast-spreading rate range and an abrupt transition to deeper lenses at intermediate spreading rates, consistent with the present dataset (Figure 11). However, there is no evidence for the systematic deepening of magma lenses within the intermediate to slow spreading range that is predicted by the numerical models (Figure 11). Instead, where magma lenses have been observed at these ridges, they cluster

(c) 2011 Elsevier Inc. All Rights Reserved.

836

MID-OCEAN RIDGE SEISMIC STRUCTURE

at 2.5–3 km depth. At these spreading rates there may be large local variations in the supply of magma from the mantle to the axis that control ridge thermal structure and give rise to shallower magma lenses than predicted from spreading rate alone. In light of recent observations, the role of neutral buoyancy may need to be reconsidered. If the magma lens is not a region of 100% melt, magma densities may be considerably higher than used in previous neutral buoyancy calculations. The magma lens may indeed lie at its correct neutrality depth, and the observed variation in magma lens depth may reflect changes in the density of the melt and crystal aggregate found within the lens.

See also Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Seamounts and Off-Ridge Volcanism. Seismic Structure.

Further Reading Buck WR, Delaney PT, Karson JA, and Lagabrielle Y (eds.) (1998) Faulting and Magmatism at Mid-Ocean Ridges, Geophysical Monograph 106. Washington, DC: American Geophysical Union. Detrick RS, Buhl P, Vera E, et al. (1987) Multichannel seismic imaging of a crustal magma chamber along the East Pacific Rise. Nature 326: 35–41. Jacobson RS (1992) Impact of crustal evolution on changes of the seismic properties of the uppermost oceanic crust. Reviews of Geophysics 30: 23--42. Phipps Morgan J and Chen YJ (1993) The genesis of oceanic crust: magma injection, hydrothermal circulation, and crustal flow. Journal of Geophysical Research 98: 6283--6297. Sinton JA and Detrick RS (1991) Mid-ocean ridge magma chambers. Journal of Geophysical Research 97: 197--216. Solomon SC and Toomey DR (1992) The structure of midocean ridges. Annual Review of Earth and Planetary Sciences 20: 329--364.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY D. R. Bohnenstiehl, North Carolina State University, Raleigh, NC, USA R. P. Dziak, Oregon State University/National Oceanic and Atmospheric Administration, Hatfield Marine Science Center, Newport, OR, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The mid-ocean ridge that divides this planet’s ocean basins represents a rift system where new seafloor is emplaced, cooled, and deformed. These processes facilitate hydrothermal circulation and the exchange

200°

80°

240°

280°

320°

of elements between the solid Earth and ocean, support complex ecosystems in the absence of sunlight, and form one of the most active and longest belts of seismicity on the planet (Figure 1). This article outlines the tools used to study earthquakes in the oceanic ridge and transform environments, discusses the underlying mechanisms that cause or influence seismicity in these settings, and explores the impacts of earthquakes on submarine hydrothermal systems. As shown below, earthquakes can be used to track a number of important physical processes, including tectonic faulting, subsurface diking, seafloor eruptions, and hydrothermal cracking. As such, studies of earthquake patterns in space and time have become

0°

40°

80°

120°

160°

KR

60° RR JdFR/ GR

40°

NMAR

20° CaR

NEPR 0° GSC

CIR

−20°

SEPR

ChR

SMAR

−40°

SWIR SEIR

−60° PAR −6000 −6750

−4500

−5250

−3000

−3750

−1500

−2250

−750

0

Earthquake depth (km) < 33 km

33 −150 km

> 150 km

Figure 1 Global map of seismicity from National Earthquake Information Center (NEIC) catalog, MZ5, 1980–2005. Depths o33 km (red), 33–150 km (blue), and 4150 km (green). Mid-ocean ridges and oceanic transforms are defined by narrow bands of shallow hypocenter earthquakes. Spreading centers and approximate full spreading rates: Southern East Pacific Rise (SEPR, B140 mm yr  1), Northern East Pacific Rise (NEPR, B110 mm yr  1) Pacific–Antarctic Ridge (PAR, B65 mm yr  1), Galapagos Spreading Center (GSC, B45–60 mm yr  1), Chile Rise (ChR, B50 mm yr  1), Northern Mid-Atlantic Ridge (NMAR, 25 mm yr  1), Southern Mid-Atlantic Ridge (SMAR, B30 mm yr  1), Carlsberg Ridge (CaR, B30 mm yr  1), Central Indian Ridge (CIR, B35 mm yr  1), Southwest Indian Ridge (SWIR, B15 mm yr  1), Southeast Indian Ridge (SEIR, B70 mm yr  1), Kolbeinsey/Mohns Ridges (KR, B15–20 mm yr  1), Reykjanes Ridge (RR, B20 mm yr  1), Juan de Fuca and Gorda Ridges (JdFR/GR, B60 mm yr  1). Earthquake data from http://earthquake.usgs.gov.

(c) 2011 Elsevier Inc. All Rights Reserved.

837

838

MID-OCEAN RIDGE SEISMICITY

fundamental in efforts to understand the seafloor spreading system within an integrated framework.

Methods of Monitoring Seismicity Seismicity in the oceanic ridge-transform environment is monitored using a combination of hydroacoustic technologies, which record water-borne acoustic phases associated with submarine earthquakes, and traditional seismic sensors, which record ground motion induced by compressional and shear waves propagating through the solid Earth (Figure 2). These technologies should be viewed as complementary, as each presents certain advantages and limitations. Seismometers

There are hundreds of seismometers deployed on the continents and islands across the globe. These instruments, operated primarily by governments and universities, form networks of sensors that can be used to monitor seismicity, as well as clandestine nuclear tests. However, for spreading centers that lie thousands of kilometers from the nearest seismic station, our ability to detect and locate earthquakes remains limited. Within the most remote ocean basins, global seismic networks consistently detect only earthquakes larger than roughly magnitude (M) 4.5–5, or ruptures having an approximate physical

scale of Z1 km. The signals generated by smaller events typically attenuate to below background noise levels as they traverse long solid-Earth paths between the source region and land-based sensors. Seismometers may also be deployed on or beneath the seafloor in containers that protect the instrumentation from the extreme pressure of the deep ocean (Figure 2). Seafloor instruments are commonly known as ocean bottom seismometers (OBSs). They typically are deployed from a surface ship, allowed to free-fall into position, and later located using acoustic ranging techniques. Upon retrieval, an acoustic switch is remotely triggered, causing the instrument package to release a set of anchor weights and rise buoyantly to the surface in the vicinity of a waiting ship. The detection capabilities of an OBS array depend on the number and distribution of stations. Deployment of a dozen or more instruments for year-long or multiyear observations are becoming increasingly common with recent improvements in technology. OBSs deployed at very local scales, arrays with apertures of only a few kilometers, may be used to monitor small cracking events (often with Mo0) in the vicinity of hydrothermal systems. Deployments of larger-aperture arrays, tens of kilometers across, are typically used to study volcanic and tectonic processes within a spreading segment and may provide a record of many earthquakes that would otherwise go undetected by land-based seismometers.

Hydrophones

Land-based seismometer

SOFAR Ocean bottom seismometers

1

10

100

1000

Horizontal scale (km) Figure 2 Technologies used to monitor seismicity within the mid-ocean ridge and oceanic transform environments. Horizontal scale indicates network aperture and approximate distances at which earthquakes are commonly monitored using the various methods.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

Hydrophones

Due to the combined dependence of ocean sound speed on pressure and temperature, much of the global ocean exhibits a low-velocity region known as the sound fixing and ranging (SOFAR) channel. Seismically generated acoustic energy may become trapped in the SOFAR, where it propagates laterally via a series of upward- and downward-turning refractions having little interaction with the seafloor or sea surface. The attenuation due to geometric spreading is cylindrical (R  1) for SOFAR-guided waves, making transmission significantly more efficient relative to solid-Earth phases that undergo spherical spreading loss (R  2). Water-borne, earthquake-generated acoustic signals were first observed on near-shore seismic stations. As these signals arrived after the faster propagating primary (P) and secondary (S) solidEarth phases, they were termed tertiary (T). Today T phases are monitored most commonly by hydrophones deployed directly within the SOFAR. The first systematic effort to use hydrophone data to produce a continuous catalog of mid-ocean ridge earthquakes began in the early 1990s when NOAA/ PMEL gained access to the US Navy’s Sound Surveillance System (SOSUS), a permanent network of bottom-mounted hydrophone arrays within the Northeast Pacific. This enabled the real-time detection of T waves generated by seafloor earthquakes and reduced the detection threshold for seismicity at the Juan de Fuca and Gorda Ridges by almost 2 orders of magnitude. The geometry of the SOSUS network also provided a better azimuthal distribution of stations than could be accomplished using land-based seismometers. The station geometry, combined with the existence of a well-defined velocity model of the oceans, also yielded significant improvements in location accuracy. Early successes using SOSUS facilitated the development of moored autonomous underwater hydrophones (AUHs) that could be used to monitor global ridge segments. In this design, the hydrophone sensor and instrument package are suspended within the SOFAR channel using a seafloor tether and foam flotation (Figure 2). These instruments have been deployed successfully along mid-ocean ridge spreading centers in the Atlantic and Pacific Oceans, as well as back-arc spreading systems of the Marianas (W. Pacific) and Bransfield Strait (Antarctica). A typical deployment consists of only six to seven instruments that monitor B201 along axis and provide a consistent record of earthquakes with M Z 2.53.0. Although regional hydrophone arrays can provide improvements in detection and location capabilities,

839

relative to land-based seismic stations, it is not presently possible to extract the focal mechanism directly from the T waveform. Similarly, only relative depth estimates are acquired through measurement of the T wave rise time, defined as the time between the onset of the signal and its amplitude peak.

Global to Regional Tectonic Patterns A view of global seismic patterns (Figure 1) shows ocean basin earthquakes to be narrowly focused along the spreading axis. Such events have shallow hypocenters and focal mechanisms that dominantly indicate normal faulting on moderately dipping (B451) ridge-parallel structures. Shallow-hypocenter earthquakes also cluster tightly along the oceanic transforms that accommodate the differential motion between offset ridge segments, with the sense of motion along these conservative plate boundaries being dominantly strike-slip on subvertical structures. These patterns reflect a stress regime arising from the motion of the plates and the geometry of their boundaries (Figure 3). Transform and Segment Boundary Seismicity

In the late 1960s, focal mechanism studies of oceanic transform earthquakes provided key evidence in support of plate tectonic theory. Early physiographic maps of the oceans had outlined the mid-ocean ridge system and identified places where this submarine mountain range appeared to be offset laterally. One might infer from these observations that the two ridge segments were once aligned and that their offset represents the cumulative displacement along the connecting fault system. In contrast, the then emerging theory of plate tectonics required the Fracture zone (inactive)

Plate A

Transform Abyssal hill faulting Spreading axis Plate B Figure 3 Schematic showing the style of faulting and seismicity associated with an idealized spreading center and transform plate boundary. Double black line indicates the axis of the oceanic spreading center. Thin gray lines show normal faults along the rift flanks. Arrows indicate the direction of relative motion between the plates. Focal mechanisms shown with shaded regions indicating compressive first motions.

(c) 2011 Elsevier Inc. All Rights Reserved.

840

MID-OCEAN RIDGE SEISMICITY

opposite sense of motion along these structures in order to accommodate the process of seafloor spreading along two ridge segments that maintain a near-constant offset through time. As Sykes showed in 1967, the first-motion direction of P waves radiated during transform earthquakes clearly supports the latter model. Today the orientation and slip direction of earthquakes are determined routinely for global events larger than M 5–5.5. These data, combined with detailed morphological observations, provide a more complete picture of the structure and seismicity of oceanic transforms and higher-order segment boundaries. Many transforms are segmented, stepping in an en echelon fashion, with pull-apart basins and more commonly intra-transform spreading centers defining the segmentation. As an earthquake propagates, offsets of sufficient size may arrest the rupture, limiting its spatial extent. The maximum size of a transform earthquake is observed to be a function of slip rate, with the largest transform earthquakes occurring at slow spreading rates (o40 mm yr  1) and having sizes of roughly M 7.5. At spreading rates 4100 mm yr 1, the maximum observed magnitude drops to below M 6.5. This reflects the thermal structure of oceanic lithosphere, with a shallower depth of seismogenic rupture at faster spreading rates, and potentially a greater tendency for transform segmentation. Earthquakes with an oblique sense of slip and orientation are sometimes observed at inside corners, the region where the active transform segment meets the ridge axis (Figure 3). Such events may represent slip on curvilinear abyssal hill faults that rotate up to 15–201 as they approach the transform. In other cases, inside corner earthquakes may be better depicted as a compound rupture, where strike-slip motion along the transform and dip-slip motion along an orthogonal abyssal hill fault occur simultaneously. In these cases, the radiation pattern cannot be well described by the double-couple force model used to represent the vast majority of global earthquakes. Transform seismicity is restricted principally to the region between the two active spreading centers, as no relative motion is required across the fault beyond this region. Fracture-zone scars, however, may be traced across ocean basins and continue to represent zones of weakness. In some areas, intra-plate stresses are sufficient to reactivate these structures, infrequently generating large earthquakes. Oceanic transforms exhibit similar kinematics to continental transforms, such as the San Andreas Fault in the United States or the Anatolian Fault in

Turkey. Oceanic transforms, however, show one major difference – an inventory of earthquakes along oceanic transforms documents a dramatic ‘slip deficit’, with the majority of all transform motion occurring aseismically. A prediction of the moment release rate along a transform can be obtained as P Mo/t ¼ nmlw, where n is the relative plate velocity, m is the shear modulus of the rock, l is the length of the transform, w is the down-dip width of the rupture, and t is the time period considered. Mo is the static seismic moment, which for an individual earthquake is the product of fault rupture area (lw), mean slip, and m. The width w is generally taken to correspond to the B600–6501 isotherm, as constrained by slip inversion results and laboratorybased deformation studies (Figure 4). When the sum of the observed seismic moment on a transform is compared with the expected moment, their ratio (or seismic coupling coefficient, a) is typically B1 for continental transforms, but much less than 1 along oceanic transforms. The average a for all oceanic transforms is roughly 0.25, indicating that three-fourths of all transform motion occur aseismically. Studies investigating the velocity dependence of a yield conflicting results. Such assessments are complicated by the presence of significant inter-transform variability in a even at a given spreading rate, uncertainty in the depth of seismic faulting, and the somewhat limited timescale of the observations. The most recent and detailed studies, however, suggest little-to-no systematic relationship with slip rate, or perhaps slightly lower a at fastspreading rates. Aseismic fault motions on the oceanic transforms are likely taken up by slow, creeping events. These slow ruptures are inefficient at producing seismic waves and so are termed quiet or silent earthquakes. It recently has been suggested that silent earthquakes may trigger large seismic quakes in neighboring portions of the transform that are more prone to stick-slip behavior. In this view, some seismic events on the transforms may be viewed as aftershocks of these silent earthquakes. At higher-order segment boundaries, propagating rifts and overlapping spreading centers accommodate ridge segmentation through a mechanism known as bookshelf faulting (Figure 5). Here the seafloor between the overlapping rift tips is deformed by simple shear, resulting in the rotation of the initially ridge-parallel seafloor fabric. The style of deformation is reminiscent of a stack of books on a shelf tipping over, with slippage between the ‘books’ occurring along the preexisting abyssal fault systems. Focal mechanisms indicate a strike-slip sense of motion, with one set of nodal planes trending

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

(a)

336°

338°

340°

342°

344°

346°

841

348°

2°

2°

0°

0°

he

Romanc

−2°

336°

338°

(b)

340°

342°

344°

Romanche transform

W 0

E

346°

Mantle

3/94 5/95

E 0

Ocean 8/92 400°C

10

400°C 700 °C

20

Depth (km)

Depth (km)

348°

W Chain transform

Crust 10

−2°

Chain

20

700°C Figure 4 (a) Map of the equatorial Mid-Atlantic Ridge showing the Romanche and Chain transform faults. The focal mechanisms obtained in the broad-band body-wave modeling are joined to their NEIC locations (red circles). (b) The centroid depths of the earthquakes are plotted as circles with symbol size proportional to magnitude. Solid symbols represent earthquakes with the bestresolved parameters, and open symbols those with fewer data or poorer fits. All the depths are accurate to within 3 km. Isotherms are calculated by using a half-space cooling model and averaging both sides of the transform. Slip inversions of the 1994 and 1995 Romanche earthquakes suggest that the former ruptured from near the surface to 20-km depth, and the latter from 10- to 25-km depth. This is consistent with the B6001 isotherm controlling the maximum depth of seismic slip on these transforms. The 1992 Chain earthquake is the only event with a centroid within the crust, but it was large enough to have ruptured into the upper mantle. Reprinted by permission from Macmillan Publishers Ltd: Nature (Abercrombie R and Ekstro¨m G (2001) Earthquake slip on oceanic transform faults. Nature 410: 74–77), copyright (2001).

parallel to an abyssal fabric that becomes progressively rotated within the deformation zone. This mode of deformation is most common at fast to intermediate spreading rates, with well-studied examples along the southern East Pacific Rise (4120 mm yr 1), Galapagos (B60 mm yr 1), and Lau Basin spreading centers (B100 mm yr 1). A spectacular slow-spreading example, however, can be observed within Iceland’s southern seismic zone. Spreading-center Earthquakes

When the oceanic ridge systems are monitored using global seismic stations (MZB4.5) or regional hydrophone arrays (MZB2.53) a sharp contrast in the frequency of mid-ocean ridge earthquakes is observed as function of spreading rate (Figures 1 and 6). Along fast-spreading (480 mm yr  1) axial highs, the ridge displays a dearth of small- to moderatemagnitude earthquakes. In contrast, such events are

abundant along the rift valleys of a slow-spreading (o40 mm yr 1) ridge system. These first-order patterns largely reflect the variable thickness of the brittle layer, which controls the predicted moment release, and seismic coupling coefficient of the rift-bounding normal fault systems. The predicted rate of seismic moment release can be estimated in a manner analogous to that described P for oceanic transforms: Mo/t ¼ nmlw/sin ycos y, where y is the fault dip, n is the full spreading rate, l is the length of the plate boundary, and w is the thickness of the seismogenic lithosphere. A comparison with the observed moment release at a slowspreading rift zone indicates that only 10–20% of the plate divergence is accommodated by seismic processes. This estimate agrees well with the amount of brittle strain accommodated by the normal fault populations of the abyssal plains. Hence, riftbounding normal fault systems at slow-spreading centers display a high-level of seismic coupling, with

(c) 2011 Elsevier Inc. All Rights Reserved.

842

MID-OCEAN RIDGE SEISMICITY

180° 30′

184° 00′ −18° 45′

−19° 00′

−19° 15′

−19° 30′

−19° 45′

−20° 00′

Figure 5 Central Lau Basin propagator with focal mechanics solutions. The Central Lau Basin Spreading Center is propagating to the south as spreading ceases along the northern tip of the Eastern Lau Spreading Center. Inset shows idealized case of bookshelf faulting, with progressive clockwise rotation of the abyssal fabric as the left-offset-propagating rift advances to the south. Focal mechanisms adapted from Wetzel et al. (1993) and the Global Centroid Moment Tensor Project (http://www.globalcmt.org). From Wetzel LR, Wiens DA, and Kleinrock MC (1993) Evidence from earthquakes for bookshelf faulting at large non-transform ridge offsets. Nature 362: 235–237.

the ratio of seismic fault slip to expected fault slip being close to 1. However, the emplacement of dikes along or near the rift axis takes up the majority (80–90%) of the plate separation. At fast-spreading ridges, commonly less than 1% of the predicted moment release is observed seismically along the rift. While most of the plate separation is accommodated by diking, as it was at slower rates, faulting studies indicated the fastspreading lithosphere undergoes an extension of B4–8%. The observed moment release is insufficient to account for this. Fast-spreading normal fault systems, therefore, appear to have low a and must accrue displacement during aseismic slip events or by

abundant microseismic activity too small to be detected by the hydrophones or global seismic stations. At intermediate spreading rates, the density of seismic events shows a first-order correspondence with the morphology of the ridge, with small- to moderate-magnitude earthquakes being abundant along rifted-spreading centers and comparatively rare at intermediate-rate axial highs (Figure 7). Intermediate-spreading-rate ridges can therefore assume both the morphologic and seismic characteristics of the fast- and slow-spreading end members. The maximum hypocentral depth of a rift zone earthquake decreases with increasing spreading rate, reflecting the temperature limits of brittle faulting

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

(a)

310°

315°

320°

325°

35°

35°

30°

30° MAR

25°

25°

20°

20°

843

and steeper thermal gradients at faster spreading rates (Figure 8(a)). OBS studies define additional intra- and inter-segment patterns. Within some slowspreading segments, the along-axis depth of microseismicity shallows near the center of a segment and deepens near its ends. This reflects the presence of a hotter lithosphere near segment center and is consistent with a three-dimensional (3-D) pattern of upwelling and melt focusing at slow-spreading rates. Studies on the Mid-Atlantic Ridge also show a positive correlation between the relief of the median valley and the maximum depth of microseismicity detected using OBSs (Figure 8(b)). This is consistent with stretching models that indicate greater fault offsets in regions of thicker brittle lithosphere.

Magmatism and Diking 1999−2003

15° 310° (b)

250°

315°

15° 325°

320°

255°

260°

265°

10°

EPR

10°

5°

5° GSC

0°

0°

−5°

−5°

−10°

1996−2002 250°

255°

260°

−10°

265°

Figure 6 Mid-ocean ridge seismicity viewed from regional hydrophone arrays. (a) Mid-Atlantic Ridge (B25 mm yr 1), (b) East Pacific Rise (B110 mm yr  1) and Galapagos Spreading Center (B45–60 mm yr 1). Events located using four or more hydrophones are shown. M Z 2.5  3 earthquakes are consistently detected by these arrays. Yellow stars indicate position of hydrophones in each array. Red arrows mark firstorder (transform) offsets of the ridge axis. Data from http:// www.pmel.noaa.gov.

Prior to an eruption, mid-ocean ridge melts accumulate in crustal-level chambers. At fast-spreading rates axial magma lenses are ubiquitous, being found beneath 460% of the axis that has been surveyed using multichannel reflection techniques. At slowspreading rates, crustal-level melt is absent beneath much of the ridge axis and melt bodies may be ephemeral or localized beneath axial volcanoes. At intermediate rates, the distribution of melt can range between the two extremes, with some axial highs showing nearly continuous along-axis magma chambers reminiscent of fast-spreading centers. The accumulation of melt within the crustal chamber elevates its pressure relative to the surrounding host rock. This inflation deforms the surrounding crust and triggers earthquakes within the vicinity of the chamber. For magma to leave the chamber, the pressure (P) within must exceed the sum of the minimum confining stresses (s3) at the chamber boundary and the tensile strength (T) of the rock. The dike’s orientation will be orthogonal to the least compression stress direction and therefore it should be subvertical and aligned parallel to the axis of spreading. Although cracking in the vicinity of the dike tip creates many small earthquakes, most events of sufficient size to be detected on global seismic or regional hydrophone arrays are thought to occur on preexisting fault surfaces. In the region above the dike, a narrow zone of ridge-normal extension exists where seismicity is localized (Figure 9). A broader zone of extension exists beyond the along-axis tips of the intrusion, where seismicity will be triggered in front of a laterally propagating dike. As the dike passes, the lithosphere will be compressed at depth within the region adjacent to the dike.

(c) 2011 Elsevier Inc. All Rights Reserved.

844

MID-OCEAN RIDGE SEISMICITY

(a)

(b) 230°

232°

229° 30′ 229° 40′ 233° 50′

234°

233° 05′ 233° 15′ 233° 25′ 42° 50′

45° 05′

50°

42° 45′ 45° 00′ 42° 40′

48°

44° 55′ 42° 35′

JdFR 44° 50′

42° 30′ 46°

Axial Vol.

44° 45′ 42° 25′ −1)

44°

Gorda

42°

Seafloor depth (km)

(c)

40°

JdF Ridge (60 mm yr −2.0

Gorda Ridge (60 mm yr−1) JdF

−2.5 −3.0 −3.5

Gorda 0

5

10 15 20 Distance along profile (km)

25

30

Figure 7 (a) SOSUS-detected earthquakes along the Gorda and Juan de Fuca Ridges in the Northeast Pacific, 1993–2005. Earthquakes located using four or more hydrophones are shown. M Z 2.5 earthquakes are consistently detected by SOSUS. (b) Enlarged view of portions of the Gorda and Juan de Fuca Ridge crests. (c) Cross-axis profiles from these regions.

(a) 0

Half-spreading rate (mm yr −1) 5 10 15 20

(b) 25

5 Maximum centroid depth

10

Inferred maximum depth of faulting

Max. depth of seismicity

Centroid depth (km)

2

4 5 6 7 8 9

15

35 N 23 N 29 N ctr 29 N,end 26 N

3

0

200 400 600 800 1000 1200 1400 1600 1800 2000 Cross-axis relief (m)

Figure 8 (a) Centroid depth of mid-ocean ridge earthquakes obtained from body waveform inversion of teleseismic arrivals vs. halfspreading rate. Maximum depth of seismic faulting is inferred to be less than twice the maximum centroid depth. (b) Relationship between cross-axis relief and maximum depth of seismicity. Maximum depth of seismicity was inferred from the focal depths of inner valley floor earthquakes, as recorded by OBS studies. (a) Reprinted by permission from Macmillan Publishers Ltd: Nature (Solomon SC, Huang PY, and Meinke L (1988) The seismic moment budget of slowly spreading ridges. Nature 334: 58–61), copyright (1988). (b) Reproduced from Barclay AH, Toomey DR, and Solomon SC (2001) Microearthquake characteristics and crustal VP/VS structure at the Mid-Atlantic Ridge, 351 N. Journal of Geophysical Research 106: 2017–2034 (doi:10.1029/2000JB900371), with permission from American Geophysical Union.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

Depth

Extension

Dike

Compression

by dike Figure 9 Cartoon showing regions of compression (white) and extension (shaded) associated with a vertically propagating dike. Thin black lines indicate normal faults, with ticks on the down thrown block. Seismicity (black dots) is promoted within the extensional region and inhibited elsewhere. The width of the failure zone scales with the depth to the top of the dike.

130° 00′ W

129° 40′ W

46.20

67

46.20 46.20

44

=0

46.20

md

Latitude (°N)

New lava flows

56

.27

46.20

46° 40′ N

46° 30′ N

(b)

33

1k

(a)

23.

Fault ‘locked’

Much of what we know about mid-ocean ridge dike injection comes from SOSUS observations of intermediate-rate-spreading centers in the NE Pacific. On 26 June 1993, soon after monitoring began, an intense episode of volcanic seismicity was recorded in real time along the CoAxial Segment of the Juan de Fuca Ridge (JdFR), marking the first observation of such an event (Figure 10). The seismic swarm started near 461 150 N and migrated northward during the next 40 h to B461 360 N, where majority of earthquake activity occurred during the following 3 weeks. Earthquakes propagated NNE at a velocity of 0.370.1 m s 1. The reservoir that acted as the dike’s source likely resided beneath, or to the south of, the initial swarm of earthquakes. The character and propagation velocity of this earthquake swarm were very similar to dike injections observed at Krafla and Kilauea Volcanoes. Seismicity went undetected by land-based seismic networks, suggesting earthquakes Mr4.0. Since the CoAxial activity, six additional magmatic episodes have been observed hydroacoustically on the Northeast Pacific spreading centers, and several distinctive characteristics have been identified: (1)

m s −1

Rif

t ax

is

Earthquakes triggered above dike

22

46.20 46° 20′ N

845

11

46.10 25 (c)

0 27

29

1

Jun.

3

5

7

Jul. 1993

46° 00′ N Axial volcano 45° 50′ N

12

8

Beams change

T wave rise time (s)

46° 10′ N

4

0 27

45° 40′ N 130° 10′ W

129° 50′ W

129° 30′ W

Jun.

29

1 Jul.

3

5

7

9

11

1993

Figure 10 (a) Bathymetric map of the CoAxial segment of the JdFR. Circles show the locations of swarm recorded during a 23-day period of intense activity. (b) Along-segment position of epicenters vs. time during the swarm. (c) T wave rise time vs. time during the swarm. Events show a decrease in rise time consistent with the shallowing of hypocenters as the dike approaches the seafloor, erupting fresh lava in the north. Reproduced from Dziak RP, Fox CG, and Schreiner AE (1995) The June–July 1993 seismo-acoustic event at CoAxial Segment, Juan de Fuca Ridge: Evidence for a lateral dike injection. Geophysical Research Letters 22: 135–138, with permission from American Geophysical Union.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

the sequences are swarms, lacking a dominant event, with a temporal history that cannot be described by a power-law decay in event rate (i.e., Omori’s law); (2) the total number of earthquakes during swarms exceeds 10% (4350 events/week) of the variance of long-term background JdFR–Gorda Ridge seismicity; (3) swarms have several episodes of intense activity that reach 50–100 earthquakes/hour; (4) swarms last from several (45) days to several weeks; (5) earthquakes may migrate up to tens of kilometers alongaxis following the lateral injection of magma through the crust; and (6) swarms may be accompanied by continuous, broad-band energy (3–30 Hz) interpreted as ‘intrusion tremor’, resulting from magma breaking through the crust. Autonomous hydrophone recordings from the north-central Mid-Atlantic Ridge (B25 mm yr 1) indicate that diking events are less frequent, consistent with the lower rates of plate separation. During more than 5 years of monitoring, only one probable volcanogenic swarm was detected along B2500 km of ridge axis. This activity was within the Lucky Strike segment (371 N), somewhat outside of the AUH array. Earthquake locations could be determined for 147 hydrophone-detected events, with 33 of sufficient size to be located by land-based seismometers (M 3.6–5.0). In terms of total moment release, this earthquake sequence was one of largest on the Mid-Atlantic Ridge in the last three decades. The activity displayed a swarm-like temporal

(b)

loor

lley f

18 200 Cum. number

16

EQs per day

14

Channelized flow

12 10 8

13 km

Flow toes

100 50 0

6 4

Volcanically overprinted faults

E

150

/da y

a Rift v

0.5 EQs/day

Qs

(a)

behavior and was accompanied by intrusion tremor; however, it was of short duration (o2 days) relative to previous episodes detected on the Northeast Pacific spreading centers. By far the largest episode of volcanogenic seismicity recorded on a mid-ocean ridge system occurred within the Arctic Ocean basin on the ultra-slowspreading (o5 mm yr 1) Gakkel Ridge in 1999. Seismic activity began in mid-January and continued vigorously for 3 months, with a reduced rate of activity continuing for 4 additional months (Figure 11). In total, 252 events were large enough (M44) to be recorded on global seismic networks (no hydroacoustic monitoring was in place). Subsequent geophysical studies of the Gakkel indicate the presence of a large, recently erupted flow and a volcanic peak directly in the area of seismic activity, with several large central volcanoes spectacled throughout an otherwise magma starved and heavily sedimented rift system. The duration of the 1999 activity suggests intrusion event(s) spanning several months and indicates a significant volume of magma beneath these central volcanoes. Because of their small magnitudes, earthquakes generated by diking events along fast-spreading ridge crests may be difficult to detect using hydroacoustic or global seismic techniques. Despite the greater rate of plate separation, during an 8-year period of hydroacoustic monitoring on the East Pacific Rise (101 S–101 N), only a handful of short-duration

2E

846

0

275 105 126 Days since 001, 1999

2 0

0

50

100 150 200 Days since 001, 1999

250

300

N

Figure 11 Seismovolcanic activity on the ultra-slow-spreading Gakkel Ridge near 851 N, 851 E. (a) Three-dimensional view from west (bottom) to east (top). The dark, reflective terrain is centered about a close-contoured high having a maximum vertical relief of 500 m. Red circles show the locations of epicenters, Jan.–Sep. 1999. (b) Histogram showing progression of the swarm through time, with each bar representing the number of events per day. Inset figure shows the cumulative number of events through time. There is a clear decrease in the rate of activity on 15 April (dot-dash line). Dashed line shows 6 May, when the USS Hawkbill passed over the area collecting the side scan imagery shown in (a). (a) Reprinted by permission from Macmillan Publishers Ltd: Nature (Edwards MH, Kurras GJ, Tolstoy M, Bohnenstieh DR, Coakley BJ, and Cochran JR (2001) Evidence of recent volcanic activity on the ultra-slow spreading Gakkel Ridge. Nature 409: 808–812), copyright (2001).

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

847

(hours to days) swarms containing a few tens to hundreds of locatable earthquakes were detected. They were primarily located along intra-transform spreading centers and near segment ends, where a somewhat colder, thicker lithosphere might be expected to support larger ruptures. In January of 2006, a submarine diking event and seafloor eruption were captured for the first time by a network of OBSs. The eruption was located within the well-studied 91500 N region of the East Pacific Rise. As part of a multidisciplinary study, a network of up to 12 OBSs was deployed at this site in October 2003. Analysis has shown a gradual ramp-up in activity during a more than 2-year interval prior to the eruption, with activity culminating in an intense period of seismicity that lasted B6 h on 22 January 2006. Although many of the OBS instruments were destroyed by the eruption, the event was also recorded by regional AUHs, and the joint analysis of these data is expected to further illuminate the process of dike injection at a fast-spreading ridge. It should be emphasized that these somewhat limited observations may not be characteristic of all eruptive activity at mid-ocean ridges. Although ridge scientists often refer to a volcano-tectonic cycle, given a sufficiently long period of observation it is likely that mid-ocean ridge eruptions, like their continental equivalents, will display a power-law scaling with respect to size and duration, with smaller episodes occurring more frequently than larger ones. The scaling exponent and maximum eruption size, however, should be expected to vary as a function of spreading rate, reflecting variability in the dynamics of each system and the size of its crustal reservoir. Likewise, eruption frequency is likely to show a long-term clustering, rather than the periodic behavior sometimes alluded to.

field, strain rates from cooling are estimated to be on the order of 10 6 yr 1, about 3 orders of magnitude higher than tectonic strain rates. In 1995, a vent-scale OBS study lasting 105 days was conducted at 91 500 N on the East Pacific Rise, where a series of temperature probes also were deployed within the high-temperature vent systems (Figure 12). This study recorded a swarm of 162 microearthquakes during a period of about 3 h. Hypocenters defined a subvertical column at a depth of 0.7–1.1 km (Figure 12). Four days later, temperature sensors within a high temperature (Bio9) vent began to record increasing fluid temperature of B1 1C/day, with this trend continuing for 7 days. Vent temperature then returned to background levels during the next B120 days. Comparisons with vent temperatures during more than 3 years of monitoring showed this postseismic fluctuation to be the largest recorded, making a strong argument that the earthquakes had perturbed the hydrothermal system. It was suggested that fluids rapidly penetrated the new fractures and extracted heat from the fresh rock. A decade later, multiyear OBS monitoring, in conjunction with in situ temperature and chemical and biological studies, has been reestablished in the 91 500 N region. The results of this ongoing work, combined with similar vent-scale OBS studies conducted at intermediate- and slow-spreading vent sites, show microseismicity to be ubiquitous in hydrothermal areas. Moreover, they suggest a more complex relationship between seismic activity and vent hydrology than could have been envisioned based on the 1995 study, with temperature and flow responses exhibiting variable amplitude, sign, and phase, and many swarms producing no detectable perturbation at the vent sites.

Hydrothermal Seismicity

Tidal Triggering

Within the mid-ocean ridge hydrothermal regime, earthquake activity may be triggered in response to the removal of heat, which leads to thermal contraction and subsequent cracking, or via hydrofracture when trapped pockets of fluid are heated. Such events are typically small, with rupture diameters of 1 m to tens of meters, and can only be recorded using local OBS arrays deployed within the high-temperature vent fields. Within the shallow crust the predicted mode of failure is for purely tensional (mode I) or mixed-mode fracture, rather than the shear failure observed for the largermagnitude spreading-center earthquakes. Given estimates of the hydrothermal heat flux within a vent

Solar and lunar tidal forces exert short period stress variations that act on the Earth. Beneath the oceans and in coastal areas, stress changes influencing earthquake occurrence reflect the combined effect of the direct Earth tide and indirect loading of the Earth by the ocean tides. Together these changes are on the order of 103  104 Pa, much less than the average stress drop of an earthquake or the strength of crustal rocks. Many studies have examined the role of tides in triggering seismicity around the globe. With the exception of some terrestrial volcanic areas, earthquakes and tides are generally not correlated or weakly correlated during periods when tidal fluctuations are at their largest amplitude.

(c) 2011 Elsevier Inc. All Rights Reserved.

848

MID-OCEAN RIDGE SEISMICITY

(a)

(b) Temperature (°C)

Seafloor

1

2

3

374 372 370 368 366 364 100

Melt lens # Events/day

Distance below seafloor (km)

0

0 1 −1 Distance across axis (km)

30 10 No data

No data

3 1

0

50

100 150 200 Julian days, 1995

250

300

15

2

10

0

5

−2

0 161

166

171

Ocean tide (m)

Number of earthquakes

Figure 12 (a) Ridge normal cross section of relocated microearthquakes associated with a 1995 swarm beneath the 91 500 N hydrothermal site on the East Pacific Rise. (b) Correlation between the microearthquake swarm and the fluid exit temperature at vent Bio9. Reproduced from Sohn RA, Hildebrand JA, and Webb SC (1999) A microearthquake survey of the high-temperature vent fields on the volcanically active East Pacific Rise (91 500 N). Journal of Geophysical Research 104: 25367–25378 (doi:10.1029/ 1999JB900263), with permission from American Geophysical Union.

176

Julian days, 1995 Figure 13 (a) Fifteen-day time histogram of the earthquake count in 2-h bins with the ocean tide time series superimposed. Earthquakes are triggered at periods of low ocean tide, when the seafloor is unloading and stress within the crust is most extensional. Data collected during an ocean bottom seismic experiment near 481 N, 2311 E on the Endeavour Segment of the intermediate-rate JdFR. Reproduced from Wilcock W (2001) Tidal triggering of microearthquakes on the Juan de Fuca Ridge. Geophysical Research Letters 28: 3999–4002 (doi:10.1029/2001GL013370), with permission from American Geophysical Union.

Similarly, analysis of hydrophone-derived and global seismic catalogs does not show a robust correlation between tidal phase and the occurrence times of small- to moderate-size mid-ocean ridge earthquakes. However, microseismicity (Mo2) within axial hydrothermal systems commonly shows a strong correlation, with earthquakes occurring

during periods of peak extensional stress. In some locations, such as the JdFR in the Northeast Pacific Ocean, the amplitude of the ocean tides is large and these peak extensional periods correspond to times of low tide, when the seafloor is unloaded as the height of the overlying water column reaches a minimum (Figure 13). In other areas, such as the

(c) 2011 Elsevier Inc. All Rights Reserved.

849

MID-OCEAN RIDGE SEISMICITY

91 500 N region of the East Pacific Rise, tidal changes in the height of the sea surface are minimal and the Earth tide stresses have been shown to dominate. One explanation for these observations comes from laboratory studies of rock failure, which indicate that earthquake triggering in response to a periodic loading is dependent on the period of the oscillation relative to the time it takes a slip event to nucleate. Earthquake nucleation time, which is inversely proportional to stressing rate, is estimated to be on the order of a year or more for typical tectonic settings. Consequently, these systems are largely insensitive to tidal stress changes of semi-diurnal frequency. In volcano-hydrothermal systems, however, cooling- and magma-induced stress changes elevate the stressing rate by several orders of magnitude. Therefore, earthquake nucleation times become

comparable to tidal periods, and microseismicity in this setting becomes susceptible to tidal influence.

Impacts on Hydrothermal Systems Earthquakes may disturb the hydrology of marine hydrothermal systems through several mechanisms. As described previously, microearthquakes occurring beneath a vent field can alter the local permeability, creating new fluid pathways that allow the migration of cold water into the reaction zone or redirect the escape of buoyant, heated waters. Since the precipitation of minerals within the up- and downflow zones is predicted to reduce permeability over time, earthquakes events may be critical in maintaining the longevity of hydrothermal systems.

(a) Cleft segment 44° 40′ N

Vent 1

Plume

NEIC locations

Temperature (°C)

(b)

280

2 Jun. 2000 11/26/99 mb= 4.2 Mw 6.2 mainshock 12/21/99 260 mb= 4.2 12/20/00 mb= 4.0

240 220

Mainshock NEIC location

Plume 200 09/18 02/13 1999 2000 44° 20′ N

07/12

12/08

05/04 2001

Date (month/day)

(c)

WA

TraBlan ns co for m

Gorda Ridge

OR

SO

SU

S lo

cati

ons

CA

44° 00′ N 130° 20′ W

130° 00′ W

129° 40′ W

Temperature (°C)

Jua n

de Fuc a Rid ge

280

270

260

250

2 Jun. 2000 Mw 6.2 mainshock

Vent 1 240 09/21 12/03 02/13 04/25 1999 2000 Date (month/day)

07/06

Figure 14 Map showing acoustically derived locations of foreshocks (red dots), mainshock (yellow star), and aftershocks (white dots) of 1–7 Jun. 2000 earthquake sequence on the Western Blanco Transform. Error bars on SOSUS earthquake locations represent one standard error. Note difference in location relative to NEIC seismic epicenters (black squares), which are scattered across the plate interior. Locations of Vent 1 and Plume hydrothermal vent sites along southern JdFR are also shown. (b) Twenty-three-month temperature record at Plume hydrothermal site. Fluid at Plume vent decreased 4.9 1C during the 18 h immediately following mainshock. Three additional temperature drops appear to be associated with much smaller (MB4.0  4.2) earthquakes along the western Blanco (blue dots in (a)). (b) Eleven-month temperature record at Vent 1 hydrothermal site shows a possible delayed response to the June 2000, M 6.2 earthquake. After 7 months of steady readings (27273 1C), probe temperature decreased slightly prior to the event, but then dropped markedly by 18 1C begining 7 days after the event. Reproduced from Dziak RP, Chadwick WW, Jr., Fox CG, and Embley RW (2003) Hydrothermal temperature changes at the southern Juan de Fuca Ridge associated with MW 6.2 Blanco Transform earthquake. Geology 31: 119–122, with permission from the Geological Society of America.

(c) 2011 Elsevier Inc. All Rights Reserved.

850

MID-OCEAN RIDGE SEISMICITY

Earthquakes also induce static stress changes, with amplitudes that decay with distance from the earthquake source and become insignificant at ranges beyond 1–2 rupture lengths. These changes trigger secondary earthquakes (aftershocks) and dilate (or compress) the surrounding lithosphere. At more remote distances, up to several tens of rupture length from the epicenter, the transient stresses associated with passing seismic waves also may impact the hydrothermal system. In these cases, earthquakeinduced ground shaking is thought to jar open some hydrothermal conduits and seal others. Observations in the marine environment indicate that seismicity may create fluctuations in vent flow, temperature, and chemistry with variable characteristics. For example, drops in temperature at vent sites on the southern JdFR have been observed to correlate with earthquakes on the Western Blanco Transform at distances 430 km (Figure 14). The sensitivity, phase, and amplitude of these changes, however, vary significantly even between closely spaced sites (Figures 14(a) and 14(b)). Given the distances involved, temperature changes correlated with moderate-size (MB4.0) earthquakes on the Blanco Transform likely reflect the systems response to dynamic stress transients, while much larger earthquakes (M46) also induced significant static changes. Importantly, earthquake-induced changes within mid-ocean ridge hydrothermal systems can dramatically affect the biological communities that derive their energy from chemosynthetic processes. For example, the sub-seafloor microbial community is extremely sensitive to temperature, oxygen content, pH, and other environmental variables. It thrives in subsurface zones where the hot hydrothermal fluid mixes with entrained seawater. Changes in crustal fluid temperatures of only a few degrees, or a minor alteration of crustal permeability, can cause the prevailing microbial species to weaken and encourage new species to thrive and become dominant. Similarly, changes in the effluent thermal flux of hydrothermal vents can enhance heat output changing the size and temperature of the thermal boundary layer, a zone just above the axial floor that supports abundant and diverse macrofaunal communities. Still more dramatic impacts may be associated with the emplacement of magma at depth or the eruption of lava onto the seafloor. The heat supplied by these systems feeds massive bacteria blooms (floc) and may generate event megaplumes within the water column, huge volumes of hydrothermal fluid enriched in reduced chemicals that rise up to 1 km above the seafloor.

See also Acoustics, Deep Ocean. Hydrothermal Vent Fluids, Chemistry of. Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Seamounts and Off-Ridge Volcanism.

Further Reading Abercrombie R and Ekstro¨m G (2001) Earthquake slip on oceanic transform faults. Nature 410: 74--77. Barclay AH, Toomey DR, and Solomon SC (2001) Microearthquake characteristics and crustal VP/VS structure at the Mid-Atlantic Ridge, 351 N. Journal of Geophysical Research 106: 2017--2034 (doi:10.1029/ 2000JB900371). Bird P, Kagan YY, and Jackson DD (2002) Plate tectonics and earthquake potential of spreading ridges and oceanic transform faults. In: Stein S and Freymueller JT (eds.) Geodynamics Series 30: Plate Boundary Zones, pp. 203--218. Washington, DC: American Geophysical Union. Boettcher MS and Jordan TH (2004) Earthquake scaling relations for mid-ocean ridge transform faults. Journal of Geophysical Research 109: B12302 (doi:10.1029/ 2004JB003110). Dziak RP, Chadwick WW, Jr., Fox CG, and Embley RW (2003) Hydrothermal temperature changes at the southern Juan de Fuca Ridge associated with MW 6.2 Blanco Transform earthquake. Geology 31: 119--122. Dziak RP, Fox CG, and Schreiner AE (1995) The June–July 1993 seismo-acoustic event at CoAxial Segment, Juan de Fuca Ridge: Evidence for a lateral dike injection. Geophysical Research Letters 22: 135--138. Edwards MH, Kurras GJ, Tolstoy M, Bohnenstieh DR, Coakley BJ, and Cochran JR (2001) Evidence of recent volcanic activity on the ultra-slow spreading Gakkel Ridge. Nature 409: 808--812. Fox CG, Matsumoto H, and Lau T-KA (2001) Monitoring Pacific Ocean seismicity from an autonomous hydrophone array. Journal of Geophysical Research 106: 4183--4206. Sohn RA, Hildebrand JA, and Webb SC (1999) A microearthquake survey of the high-temperature vent fields on the volcanically active East Pacific Rise (91 500 N). Journal of Geophysical Research 104: 25367--25378 (doi:10.1029/1999JB900263). Solomon SC, Huang PY, and Meinke L (1988) The seismic moment budget of slowly spreading ridges. Nature 334: 58--61. Sykes LR (1967) Mechanism of earthquakes and nature of faulting on the mid-oceanic ridges. Journal of Geophysical Research 72: 2131--2153. Tolstoy M, Cowen JP, Baker ET, et al. (2006) A seafloor spreading event captured by seismometers. Science 314: 1920--1922 (doi:10.1126/science.1137082).

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE SEISMICITY

Wetzel LR, Wiens DA, and Kleinrock MC (1993) Evidence from earthquakes for bookshelf faulting at large nontransform ridge offsets. Nature 362: 235--237. Wilcock W (2001) Tidal triggering of microearthquakes on the Juan de Fuca Ridge. Geophysical Research Letters 28: 3999--4002 (doi:10.1029/2001GL013370).

851

http://www.pmel.noaa.gov – National Oceanic and Atmospheric Administration’s Acoustic Monitoring Program. http://earthquake.usgs.gov – US Geological Survey Earthquake Hazards Program.

Relevant Websites http://www.globalcmt.org – Global CMT Web Page.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY K. C. Macdonald, Department of Geological Sciences and Marine Sciences Institute, University of California, Santa Barbara, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1798–1813, & 2001, Elsevier Ltd.

Introduction The mid-ocean ridge is the largest mountain chain and the most active system of volcanoes in the solar system. In plate tectonic theory, the ridge is located between plates of the earth’s rigid outer shell that are separating at speeds of B10 to 170 mm y1 (up to 220 mm y1 in the past). The ascent of molten rock from deep in the earth (B30–60 km) to fill the void between the plates creates new seafloor and a volcanically active ridge. This ridge system wraps around the globe like the seam of a baseball and is approximately 70 000 km long. Yet the ridge itself is onlyB5–30 km wide–very small compared to the plates, which can be thousands of kilometers across (Figure 1). Early exploration showed that the gross morphology of spreading centers varies with the rate of

plate separation. At slow spreading rates (10–40 mm y1) a 1–3 km deep rift valley marks the axis, while for fast spreading rates (490 mm y1) the axis is characterized by an elevation of the seafloor of several hundred meters, called an axial high (Figure 2). The rate of magma supply is a second factor that may influence the morphology of mid-ocean ridges. For example, a very high rate of magma supply can produce an axial high even where the spreading rate is slow; the Reykjanes Ridge south of Iceland is a good example. Also, for intermediate spreading rates (40–90 mm y1) the ridge crest may have either an axial high or rift valley depending on the rate of magma supply. The seafloor deepens from a global average of B2600 m at the spreading center to 45000 m beyond the ridge flanks. The rate of deepening is proportional to the square root of the age of the seafloor because it is caused by the thermal contraction of the lithosphere. Early mapping efforts also showed that the mid-ocean ridge is a discontinuous structure that is offset at right angles to its length at numerous transform faults tens to hundreds of kilometers in length. Maps are powerful; they inform, excite, and stimulate. Just as the earliest maps of the world in the sixteenth century ushered in a vigorous age of

Figure 1 Shaded relief map of the seafloor showing parts of the East Pacific Rise, a fast-spreading center, and the Mid-Atlantic Ridge, a slow-spreading center. Courtesy of National Geophysical Data Center.

852

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

853

}

Neovolcanic zone Fast (EPR 3° S)

Intermediate (EPR 21° N)

Slow (MAR 37° N)

VE ~ 4× AXIS (A)

20

10

0 km

(B)

10

20

30

(C)

Figure 2 Topography of spreading centers. (A) Cross-sections of typical fast-, intermediate-, and slow-spreading ridges based on high resolution deep-tow profiles. The neovolcanic zone is noted (the zone of active volcanism) and is several kilometers wide; the zone of active faulting extends to the edge of the profiles and is several tens of kilometers wide. After Macdonald et al. (1982). (B) Shaded relief map of a 1000 km stretch of the East Pacific Rise extending from 81 to 171N. Here, the East Pacific Rise is the boundary between the Pacific and Cocos plates, which separate at a ‘fast’ rate of 110 mm y1. The map reveals two kinds of discontinuities: large offsets, about 100 km long, known as transform faults and smaller offsets, about 10 km long, called overlapping spreading centers. Colors indicate depths of from 2400 m (pink) to 3500 m (dark blue). (C.) Shaded relief map of the Mid-Atlantic Ridge. Here, the ridge is the plate boundary between the South American and African plates, which are spreading apart at the slow rate of approximately 35 mm y1. The axis of the ridge is marked by a 1–2 km deep rift valley, which is typical of most slow-spreading ridges. The map reveals a 12 km jog of the rift valley, a second-order discontinuity, and also shows a first-order discontinuity called the Cox transform fault. Colors indicated depths of from 1900 m (pink) to 4200 m (dark blue).

exploration, the first high-resolution, continuous coverage maps of the mid-ocean ridge stimulated investigators from a wide range of fields including petrologists, geochemists, volcanologists, seismologists, tectonicists, and practitioners of marine magnetics and gravity; as well as researchers outside the earth sciences including marine ecologists, chemists, and biochemists. Marine geologists have found that many of the most revealing variations are to be observed by exploring along the axis of the active ridge. This along-strike perspective has

revealed the architecture of the global rift system. The ridge axis undulates up and down in a systematic way, defining a fundamental partitioning of the ridge into segments bounded by a variety of discontinuities. These segments behave like giant cracks in the seafloor that can lengthen or shorten, and have episodes of increased volcanic and tectonic activity. Another important change in perspective came from the discovery of hydrothermal vents by marine geologists and geophysicists. It became clear that in studies of mid-ocean ridge tectonics, volcanism, and

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

hydrothermal activity, the greatest excitement is in the linkages between these different fields. For example, geophysicists searched for hydrothermal activity on mid-ocean ridges for many years by towing arrays of thermisters near the seafloor. However, hydrothermal activity was eventually documented more effectively by photographing the distribution of exotic vent animals. Even now, the best indicators of the recency of volcanic eruptions and the duration of hydrothermal activity emerge from studying the characteristics of benthic faunal communities. For example, during the first deep-sea mid-ocean ridge eruption witnessed from a submersible, divers did not see a slow lumbering cascade of pillow lavas as observed by divers off the coast of Hawaii. What they saw was completely unexpected: white bacterial matting billowing out of the seafloor, creating a scene much like a midwinter blizzard in Iceland, covering all of the freshly erupted, glassy, black lava with a thick blanket of white bacterial ‘snow’.

Large-scale Variations in Axial Morphology; Correlations with Magma Supply and Segmentation The axial depth profile of mid-ocean ridges undulates up and down with a wavelength of tens of kilometers and amplitude of tens to hundreds of meters at fast and intermediate rate ridges. This same pattern is observed for slow-spreading ridges as well, but the wavelength of undulation is shorter and the amplitude is larger (Figure 3). In most cases, ridge axis discontinuities (RADs) occur at local maxima along the axial depth profile. These discontinuities include transform faults (first order); overlapping spreading centers (OSCs, second order) and higherorder (third-, fourth-order) discontinuities, which are increasingly short-lived, mobile, and associated with smaller offsets of the ridge (see Table 1 and Figure 4). A much-debated hypothesis is that the axial depth profile (Figures 3 and 5) reflects the magma supply along a ridge segment. According to this idea, the magma supply is enhanced along shallow portions of ridge segments and is relatively starved at segment ends (at discontinuities). In support of this hypothesis is the observation at ridges with an axial high (fast-spreading ridges) that the cross-sectional area or axial volume varies directly with depth (Figure 6). Maxima in cross-sectional area (42.5 km2) occur at minima along the axial depth profile (generally not near RADS) and are thought to correlate with regions where magma supply is robust. Conversely, small cross-sectional areas (o1.5 km2) occur at local depth maxima and are interpreted to reflect minima

Slow

25°N

3000 m

30°N 2 2 22

2 22

500 m

854

2 2

2 22

2

1

1

(A)

15°N

10°N

Fast 1

2

2

1

2 1

2

3000 m (B)

Superfast

20°S 2

15°S 2 2

2

2 2

1 2

3000 m (C) Figure 3 Axial depth profiles for (A) slow-spreading and (B) fast-spreading, and (C) ultrafast-spreading ridges. Discontinuities of orders 1 and 2 typically occur at local depth maxima (discontinuities of orders 3 and 4 are not labeled here). The segments at faster spreading rates are longer and have smoother, lower-amplitude axial depth profiles. These depth variations may reflect the pattern of magma delivery to the ridge.

in the magma supply rate along a given ridge segment. On slow-spreading ridges characterized by an axial rift valley, the cross-sectional area of the valley is at a minimum in the mid-segment regions where the depth is minimum. In addition, there are more volcanoes in the shallow midsegment area, and fewer volcanoes near the segment ends. Studies of crustal magnetization show that very highly magnetized zones occur near segment ends, which is most easily explained by a locally starved magma supply resulting in the eruption of highly fractionated lavas rich in iron. Multichannel seismic and gravity data support the axial volume/magma supply/segmentation hypothesis (Figure 6). A bright reflector, which is phasereversed in many places, occurs commonly (460% of ridge length) beneath the axial region of both the northern and southern portions of the fast- and ultrafast spreading East Pacific Rise (EPR). This reflector has been interpreted to be a thin lens of magma residing at the top of a broader axial magma reservoir. The amount of melt is highly variable alongstrike varying from a lens that is primarily crystal mush to one that is close to 100% melt. This ‘axial magma chamber’ (AMC) reflector is observed where the ridge is shallow and where the axial high has a

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

855

Table 1 Characteristics of segmentation. This four-tiered hierarcy of segmentation probably represents a continuum in segmentation

Segments Segment length (km) Segment longevity (years) Rate of segment lengthening (long term migration) mm y1 Rate of segment lengthening (short term propagation)mm y1

Discontinuities Type

Offset (km) Offset age (years)c Depth anomaly Off-axis trace

Order 1

Order 2

Order 3

Order 4

6007300a (4007200)b 45  106

140790 (50730) 0.5–5  106

20710 (15710?) B104–105

775 (775?) o103

0–50

(0.5–30  106) 0–1000

(?) Indeterminate:

(?) Indeterminate:

(0–30) 0–100

(0–30) 0–1000

no off-axis trace Indeterminate:

no off-axis trace Indeterminate:

(?)

(0–50)

no off-axis trace

no off-axis trace

Transform, large propagating rifts

Overlapping spreading centers (oblique shear zones, rift valley jogs) 2–30 0.5  106 (2  106) 100–300 (300–1000) V-shaped discordant zone Yes

Overlapping spreading centers (intervolcano gaps), devals 0.5–2.0

Devals, offsets of axial summit caldera (intravolcano gaps) o1

B0 30–100 (50–300) Faint or none

B0 0–50 (0–100?) None

Rarely

No?

(?) Yes, except during OSC linkage? (NA) Small reduction in volume (NA) Usually

(?) Rarely

Yes

Yes, except during OSC linkage? (NA) No, but reduction in volume (NA) Yes

Small reduction in volume? (NA) B50%

Yes

Yes

Yes (NA)

Often (NA)

430 40.5  106 (42  106) 300–600 (500–2000) Fracture zone

High amplitude magnetization?

Yes

Breaks in axial magma chamber? Breaks in axial lowvelocity zone? Geochemical anomaly? Break in hightemperature venting?

Always Yes (NA)

Values are 71 standard deviation. Where information differs for slow- versus fast-spreading ridges (o60 mm y1), it is placed in parentheses. c Offset age refers to the age of the seafloor that is juxtaposed to the spreading axis at a discontinuity. Updated from Macdonald et al. (1991). NA, not applicable; ?, not presently known as poorly constrained. a b

broad cross-sectional area. Conversely, it is rare where the ridge is deep and narrow, especially near RADs. A reflector may occur beneath RADs during events of propagation and ridge-axis realignment, as may be occurring now on the EPR near 91N. There is evidence that major-element geochemistry correlates with axial–cross-sectional area (Figure 7). On the EPR 131–211S, there is a good correlation between MgO wt% and cross-sectional area (high MgO indicates a higher eruption temperature and perhaps a greater local magmatic budget). The abundance of hydrothermal venting (as measured by

light transmission and backscatter in the water column and geochemical tracers) also varies directly with the cross-sectional area of the EPR. It is not often that one sees a correlation between two such different kinds of measurements. It is all the more remarkable considering that the measurements of hydrothermal activity are sensitive to changes on a timescale of days to months, while the cross-sectional area probably reflects a timescale of change measured in tens of thousands of years. On slow-spreading centers, such as the Mid-Atlantic Ridge (MAR), the picture is less clear. Seismic

(c) 2011 Elsevier Inc. All Rights Reserved.

856

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Fast S4 D4

Axial high

D2

S2 ~ 50 km ~ 20 km

S4

S2 ~ 1050 km ~5 km

S2

S3

D1 Transform fault

S1

D4 S4 ~ 5 −10 D3 km ~ 1 km S4

S3

D4

(A) Slow S2 ~ 20 km

D2

~ 20 km S2 S1

D1 Transform fault

S3 D3 ~ 10 S3 km ~5 km D2 Volcanoes in floor of rift valley

S3

S4

S3

S4

S4 ~2− 6 km S4 ~1 km

D4 D4 D3 D4

S4

(B)

Axial rift valley

Figure 4 A possible hierarchy of ridge segmentation for (A) fast-spreading and (B) slow-spreading ridges. S1–S4 are ridge segments or order 1–4, and D1–D4 are ridge axis discontinuities of order 1–4. At both fast- and slow-spreading centers, first-order discontinuities are transform faults. Examples of second-order discontinuities are overlapping spreading centers (OSCs) on fastspreading ridges and oblique shear zones on slow-spreading ridges. Third-order discontinuities are small OSCs on fast-spreading ridges. Fourth-order discontinuities are slight bends or lateral offsets of the axis of less than 1 km on fast-spreading ridges. This fourtiered hierarchy of segmentation is probably a continuum; it has been established, for example, that fourth-order segments and discontinuities can grow to become third-, second-, and even first-order features and vice versa at both slow- and fast-spreading centers. Updated from Macdonald et al. (1991).

and gravity data indicate that the oceanic crust thins significantly near many transform faults, even those with a small offset. This is thought to be the result of highly focused mantle upwelling near mid-segment regions, with very little along axis flow of magma away from the upwelling region. Focused upwelling is inferred from ‘bulls-eye’-shaped residual gravity anomalies and by crustal thickness variations documented by seismic refraction and microearthquake studies. At slow-spreading centers, melt probably resides in small, isolated, and very short-lived pockets beneath the median valley floor (Figure 5C) and beneath elongated axial volcanic ridges. An alternative view is that the observed along-strike variations in topography and crustal thickness can be accounted for by along-strike variations in mechanical thinning of the crust by faulting. There is no conflict between these models, so both focused upwelling and mechanical thinning may occur along each segment. One might expect the same to hold at fastspreading centers, i.e., crustal thinning adjacent to OSCs. This does not appear to be the case at 91N on the EPR, where seismic data suggest a thickening of the crust toward the OSC and a widening of the AMC reflector. There is no indication of crustal

thinning near the Clipperton transform fault either. And yet, as one approaches the 91N OSC from the north, the axial depth plunges, the axial cross-sectional area decreases, the AMC reflector deepens, average lava age increases, MgO in dredged basalts decreases; hydrothermal activity decreases dramatically, crustal magnetization increases significantly (suggesting eruption of more fractionated basalts in a region of decreased magma supply), crustal fracturing and inferred depth of fracturing increases (indicating a greater ratio of extensional strain to magma supply), and the throw of off-axis normal faults increases (suggesting thicker lithosphere and greater strain) (Figure 8). How can these parameters all correlate so well, indicating a decrease in the magmatic budget and an increase in amagmatic extension, while the seismic data suggest crustal thickening off-axis from the OSC and a wider magma lens near the OSC? One possibility is that mantle upwelling and the axial magmatic budget are enhanced away from RADs even at fast-spreading centers, but that subaxial flow of magma ‘downhill’ away from the injection region redistributes magma (Figure 5). This along-strike flow and redistribution of magma may be unique to spreading centers with an axial high

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

857

Axial depth profile

Axial depth (m)

2

3

2

2500

3 1

1

3000

3500

~50_100 km UPWELLING ASTHENOSPHERE

(A) Axis

Fast ~10_30 km

4

4

Seafloor along ridge axis

~2_5 km

Lens of melt

~ 2 km

Axial melt reservoir Hot rock (B)

Slow 4

4

4

4

4

4 Inner wall

(C) Figure 5 Schematic diagram of how ridge segmentation may be related to mantle upwelling (A), and the distribution of magma supply (B and C). In (A), the depth scale applies only to the axial depth profile; numbers denote discontinuities and segments of orders 1–3. Decompression partial melting in upwelling asthenosphere occurs at depths 30–60 km beneath the ridge. As the melt ascends through a more slowly rising solid residuum, it is partitioned at different levels to feed segments of orders 1–3. Mantle upwelling is hypothesized to be ‘sheetlike’ in the sense that melt is upwelling along the entire length of the ridge; but the supply of melt is thought to be enhanced beneath shallow parts of the ridge away from major discontinuities. The rectangle is an enlargement to show fine-scale segmentation for (B) a fast-spreading example, and (C) a slow-spreading example. In (B) and (C) along-strike cross-sections showing hypothesized partitioning of the magma supply relative to fourth-order discontinuities (4s) and segments are shown on the left. Acrossstrike cross-sections for fast- and slow-spreading ridges are shown on the right. Updated from Macdonald et al. (1991).

such as the EPR or Reykjanes where the axial region is sufficiently hot at shallow depths to facilitate subaxial flow. It is well documented in Iceland and other volcanic areas analogous to mid-ocean ridges that magma can flow in subsurface chambers and

dikes for distances of many tens of kilometers away from the source region before erupting. In this way, thicker crust may occur away from the midsegment injection points, proximal to discontinuities such as OSCs.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

2

Area (km )

858

6 5 4 3 2 1 0 _1

CLIPPERTON 90% 75% 10%

9

11

10

13

12

_ 2500

Axial depth

Depth (m)

_ 3000 _ 3500 _ 4000

AMC depth

_ 4500 _ 5000 9

11

10

13

12

Latitude (deg) Figure 6 Profiles of the along-axis cross-sectional area, depth, and axial magma chamber (AMC) seismic reflector for the EPR 91– 131N. The locations of first- and second-order discontinuities are denoted by vertical arrows (first-order discontinuities are named); each occurs at a local minimum of the ridge area profile, and a local maximum in ridge axis depth. Lesser discontinuities are denoted by vertical bars. There is an excellent correlation between ridge axis depth and cross-sectional area; there is a good correlation between cross-sectional area and the existence of an axial magma chamber, but detailed characteristics of the axial magma chamber (depth, width) do not correlate. Updated from Scheirer and Macdonald (1993) and references therein.

9

8

MgO content (wt %)

Based on studies of the fast-spreading EPR, a ‘magma supply’ model has been proposed that explains the intriguing correlation between over a dozen structural, geochemical and geophysical variables within a first-, second-, or third-order segment (Figure 9). It also addresses the initially puzzling observation that crust is sometimes thinner in the midsegment region where upwelling is supposedly enhanced. Intuitively, one might expect crust to be thickest over the region where upwelling is enhanced as observed on the MAR. However, along-axis redistribution of melt may be the controlling factor on fast-spreading ridges where the subaxial melt region may be well-connected for tens of kilometers. In this model, temporal variations in along-axis melt connectivity may result in thicker crust near mid-segment when connectivity is low (most often slowspreading ridges), and thicker crust closer to the segment ends when connectivity is high (most often, but not always the case at fast-spreading ridges). The basic concepts of this magma supply model also apply to slow-spreading ridges characterized by an axial rift valey. Mantle melting is enhanced beneath the midsegment regions. However, the axial region is colder (averaged over time) and along-strike redistribution of melt is impeded. Thus, the crust tends to be thickest near the midsegment regions and thinnest near RADs (Figures 8 and 9).

7

6

5

0

1

2

4

3

5

6

2

Area (km ) Figure 7 Cross-sectional area of the East Pacific Rise versus MgO content of basalt glass (crosses from EPR 5–141N, solid circles from 13–231S). There is a tendency for high MgO contents (interpreted as higher eruption temperatures and perhaps higher magmatic budget) to correlate with larger cross-sectional area. Smaller cross-sectional areas correlate with lower MgO and a greater scatter in MgO content, suggesting magma chambers which are transient and changing. Thus shallow, inflated areas of the ridge tend to erupt hotter lavas. Updated from Scheirer and Macdonald (1993) and references therein.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Segment end (discontinuity)

Segment 'center'

Segment end (discontinuity)

~2500

> 10

0

~3000 2. Cross-sectional 2 area (km )

5

10. Average lava age

2 3. Axial magma chamber occurrence (%)

90

11. Lava lake abundance (% area)

100

0

9

12. Lava domes abundance (% area)

4. MgO (wt %) 7

100

0

> 10

13. Sheet and lobate lava flows (% area)

0

100

< 50

30

14. Pillow lavas (% area)

5

7. Crustal thickness (km)

Youngest

Oldest

30

6. Crustal _1 magnetization (A m )

Segment 'center'

9. Earthquakes (>m = 2) per year

1. Depth (m)

5. No. vent 2 communities per km

859

100

0

7

15. Calculated fissure depth (m)

5

> 50

~ 400

100

16. Fissure density 2 ( no. per km )

8. Fault scarp height (m) 40

300

100

(A)

Figure 8 Schematic summary of along-axis variations in spreading center properties from segment end (discontinuity of order 1, 2, or 3) to segment mid-section areas for (A) fast-spreading ridges with axial highs and (B) slow-spreading ridges with axial rift valleys. A large number of parameters correlate well with location within a given segment, indicating that segments are distinct, independent units of crustal accretion and deformation. These variations may reflect a fundamental segmentation of the supply of melt beneath the ridge. (Less than 1% of the ridge has been studied in sufficient detail to create this summary.)

Fine-scale Variations in Ridge Morphology within the Axial Neovolcanic Zone The axial neovolcanic zone occurs on or near the axis of the axial high on fast-spreading centers, or within the floor of the rift valley on slow-spreading centers (Figure 5B and C, right). Studies of the widths of the polarity transitions of magnetic

anomalies, including in situ measurements from the research submersible Alvin, document that B90% of the volcanism that creates the extrusive layer of oceanic crust occurs in a region 1–10 km wide at most spreading centers. Direct qualitative estimates of lava age at spreading centers using submersibles and remotely operated vehicles (ROVs) tend to confirm this, as well as recent high-resolution seismic measurements that show that layer 2A (interpreted

(c) 2011 Elsevier Inc. All Rights Reserved.

860

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Segment end (discontinuity)

Segment 'center'

~2500

Segment end (discontinuity) > 50%

Axial volcanic ridges (% area)

Depth (m)

0%

~3500

Moderate

8 Crustal thickness (km)

Average lava age Old

3 High

1000 Scarp height (m)

Mantle density Low

100

High

Magnetization

MgO (wt %)

High

Low

Low

Hydrothermal vent abundance

Segment 'center'

Low

Very low

(B)

Figure 8 Continued

to be the volcanic layer) achieves its full thickness within 1–5 km of the rise axis. However, there are significant exceptions, including small volume offaxis volcanic constructions and voluminous off-axis floods of basaltic sheet flows. The axial high on fast- and intermediate-spreading centers is usually bisected by an axial summit trough B10–200 m deep that is found along approximately 60–70% of the axis. Along the axial high of fastspreading ridges, sidescan sonar records show that there is an excellent correlation between the presence of an axial summit trough and an AMC reflector as seen on multichannel seismic records (490% of ridge length). Neither axial summit troughs nor AMCs occur where the ridge has a very small crosssectional area. In rare cases, an axial summit trough is not observed where the cross-sectional area is large. In these locations, volcanic activity is occurring at present or has been within the last decade. For example, on the EPR near 91450 –520 N, a volcanic eruption documented from the submersible ALVIN was associated with a single major dike intrusion, similar to the 1993 eruption on the Juan de Fuca Ridge. Sidescan sonar records showed that an axial

trough was missing from 91520 N to 101020 N, and in subsequent dives it was found that dike intrusion had propagated into this area, producing very recent lava flows and hydrothermal activity complete with bacterial ‘snow-storms.’ A similar situation has been thoroughly documented at 171250 –300 S on the EPR where the axial cross-sectional area is large but the axial summit trough is partly filled. Perhaps the axial summit trough has been flooded with lava so recently that magma withdrawal and summit collapse is just occurring now. Thus, the presence of an axial summit trough along the axial high of a fast-spreading ridge is a good indicator of the presence of a subaxial lens of partial melt (AMC); where an axial summit trough is not present but the cross-sectional area is large, this is a good indicator of very recent or current volcanic eruptions; where an axial summit trough is not present and the cross-sectional area is small, this is a good indicator of the absence of a magma lens (AMC). In contrast to the along-axis continuity of the axial neovolcanic zone on fast-spreading ridges, the neovolcanic zone on slower-spreading ridges is considerably less continuous and there is a great deal of variation from segment to segment. Volcanic

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

861

MAGMA SUPPLY MODEL AST 2A

Depth Layer 2A Low-velocity zone

Melt lens

2A

Melt lens may be absent

Moho (A) High magma budget

(B) Moderate to low magma budget

(C) Very low, sporadic magma supply

Shallow water depth

Deeper water depth

Deep rift valley floor

Broad cross-sectional area

Small cross-sectional area

Wide valley floor

Wide low-velocity zone High MgO content High temperature Low density

Narrow low-velocity zone (except where 2 LVZs coalesce at an OSC) Lower MgO content Lower temperature Higher density

Narrow low-velocity zone

Melt lens ubiquitous, depth and width variable

Melt lens less common, highly variable width, depth where present, disrupted at RADs

Melt lens very rare

Thinner crust (melt flows 'downhill' away from mantle upwelling zone)

Thicker crust, but highly variable

Highly variable crustal thickness

Axial summit trough present unless in eruptive phase

Axial summit through rare

Hydrothermal venting abundant

Hydrothermal venting rare

Hydrothermal vents rare, fault controlled

Less crustal fissuring

More crustal fissuring

Extensive fissuring

Younger average lava age

Older average lava age

Highly variable but generally older lava age, mostly pillows.

Lower crustal magnetization

Higher crustal magnetization

Smaller throw on flanking fault scarps

Larger throw on flanking fault scarps

Large fault scarps, especially at inside corner highs

Layer 2A thinnest along axis

Layer 2A thinnest along axis

Layer 2A thick beneath axial volcanic ridges

Figure 9 Magma supply model for mid-ocean ridges (see references in Buck et al., 1998). (A) represents a segment with a robust magmatic budget, generally a fast-spreading ridge away from discontinuities or a hotspot dominated ridge with an axial high (AST is the axial summit trough). (B) represents a segment with a moderate magma budget, generally a fast-spreading ridge near a discontinuity or a nonrifted intermediate rate ridge. (C) represents a ridge with a sporadic and diminished magma supply, generally a rifted intermediate to slow rate spreading center (for along-strike variations at a slow ridge, see Fig. 8B).

contructions, called axial volcanic ridges, are most common along the shallow, mid-segment regions of the axial rift valley. Near the ends of segments where the rift valley deepens, widens, and is truncated by transform faults or oblique shear zones, the gaps between axial volcanic ridges become longer. The gaps between axial volcanic ridges are regions of older crust characterized by faulting and a lack of

recent volcanism. These gaps may correspond to finescale (third- and fourth-order) discontinuities of the ridge. Another important difference between volcanism on fast- and slow-spreading ridges is that axial volcanic ridges represent a thickening of the volcanic layer atop a lithosphere that may be 5–10 km thick, even on the axis. In contrast, the volcanic layer is

(c) 2011 Elsevier Inc. All Rights Reserved.

862

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Percentage of inward facing fault scarps

80

50

0

50

100 Spreading rate

150

Figure 10 Spreading rate versus percentage of fault scarps that are inward-facing (facing toward the spreading axis versus away from the spreading axis). A significant increase in the percentage of inward-facing scarps occurs at slower spreading rates.

usually thinnest along the axis of the EPR. Thus the axial high on fast-spreading ridges is not a thickened accumulation of lava, while the discontinuous axial volcanic ridges on slow-spreading ridges are. On both slow- and fast-spreading ridges, pillow and lobate lavas are the most common lava morphology. Based on laboratory studies and observations of terrestrial basaltic eruptions, this means that the lava effusion rates are slow to moderate on most mid-ocean ridges. High volcanic effusion rates, indicated by fossil lava lakes and extensive outcrops of sheet flow lava morphology, are very rare on slowspreading ridges. High effusion rate eruptions are more common on fast-spreading ridges and are more likely to occur along the shallow, inflated midsegment regions of the rise, in keeping with the magma supply model for ridges discussed earlier. Low effusion rate flows, such as pillow lavas, dominate at segment ends (Figure 8A). Very little is known about eruption frequency. It has been estimated based on some indirect observations that at any given place on a fast-spreading ridge eruptions occur approximately every 5–100 years, and that on slow-spreading ridges it is approximately every 5000–10 000 years. If this is true, then the eruption frequency varies inversely with the spreading rate squared. On intermediate- to fastspreading centers, if one assumes a typical dike width of B50 cm and a spreading rate of 5–10 cm y1, then an eruption could occurBevery 5–10 years. This estimate is in reasonable agreement with the occurrence of megaplumes and eruptions on the well-monitored Juan de Fuca Ridge. However, observations in sheeted dike sequences in Iceland and

ophiolites indicate that only a small percentage of the dikes reach the surface to produce eruptions. On fast-spreading centers, the axial summit trough is so narrow (30–1000 m) and well-defined in most places that tiny offsets and discontinuities of the rise axis can be detected (Table 1, Figure 2). This finest scale of segmentation (fourth-order segments and discontinuities) probably corresponds to individual fissure eruption events similar to the Krafla eruptions in Iceland or the Kilauea east rift zone eruptions in Hawaii. Given a magma chamber depth of 1–2 km, an average dike ascent rate of B0.1 km h1 and an average lengthening rate of B1 km h1, typical diking events would give rise to segments 10–20 km long. This agrees with observations of fourth-order segmentation and the scale of the recent diking event on the Juan de Fuca Ridge and in other volcanic rift zones. The duration of such segments is thought to be very short,B100–1000 years (too brief in any case to leave even the smallest detectable trace off-axis, Table 1). Yet even at this very fine scale, excellent correlations can be seen between average lava age, density of fissuring, the average widths of fissures, and abundance of hydrothermal vents within individual segments. In fact there is even an excellent correlation between ridge cross-sectional area and the abundance of benthic hydrothermal communities (Figure 8). A curious observation on the EPR is that the widest fissures occur in the youngest lava fields. If fissures grow in width with time and increasing extension, one would expect the opposite; the widest fissures should be in the oldest areas. The widest fissures are B5 m. Using simple fracture mechanics, these fissures probably extend all the way through layer 2A and into the sheeted dike sequence. These have been interpreted as eruptive fissures, and this is where high-temperature vents (43001C) are concentrated. In contrast to the magma rich, dike-controlled hydrothermal systems that are common on fast-spreading centers, magma-starved hydrothermal systems on slow-spreading ridges tend to be controlled more by the penetration of sea water along faults near the ridge axis. (See Hydrothermal Vent Deposits.)

Faulting Extension at mid-ocean ridges causes fissuring and normal faulting. The lithosphere is sufficiently thick and strong on slow-spreading centers to support shear failure on the axis, so normal faulting along dipping fault planes can occur on or very close to the axis. These faults produce grabens 1–3 km deep. In

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

863

Faulting along a slow-spreading rift valley axis

Segment center Small fault throws Close fault spacing

Segment center Thick crust Thin lithosphere

Isotherms

Segments ends Thin crust Thik lithosphere Large-throw faults Large fault spacing

Crust Mantle Lithosphere Asthenosphere

Strong upwelling, elevated isotherms at segment centres

Figure 11 A geological interpretation for along-axis variations in scarp height, and more closely spaced scarps near mid-segment on a slow-spreading center. Cross-section through segment center (top) shows more closely spaced, smaller-throw faults than at the segment ends (bottom). Focused mantle upwelling near the segment center causes this region to be hotter; the lithosphere will be thinner while increased melt supply creates a thicker crust. In contrast to fast-spreading centers, there may be very little melt redistribution along-strike. Near the segment ends, the lithosphere will be thicker and magma supply is less creating thinner crust. Along axis variations in scarp height and spacing reflect these along axis variations in lithospheric thickness. Amagmatic extension across the larger faults near segments ends may also thin the crust, especially at inside corner highs. Modified from Shaw (1992).

contrast, normal faulting along inclined fault planes is not common on fast-spreading centers within 72 km of the axis, probably because the lithosphere is too thin and weak to support normal faulting. Instead, the new thin crust fails by simple tensional cracking. Fault strikes tend to be perpendicular to the least compressive stress; thus they also tend to be perpendicular to the spreading direction. While there is some ‘noise’ in the fault trends, most of this noise can be accounted for by perturbations to the least compressive stress direction due to shearing in the vicinity of active or fossil ridge axis discontinuities. Once this is accounted for, fault trends faithfully

record changes in the direction of opening to within 731 and can be used to study plate motion changes on a finer scale than that provided by seafloor magnetic anomalies. Studies of the cumulative throw of normal faults, seismicity, and fault spacing suggest that most faulting occurs within 720–40 km of the axis independent of spreading rate. There is a spreading rate dependence for the occurrence of inward and outward dipping faults. Most faults dip toward the axis on slow-spreading centers (B80%), but there is a monotonic increase in the occurrence of outward dipping faults with spreading rate (Figure 10). Inward and outward facing faults are approximately equally abundant at very fast

(c) 2011 Elsevier Inc. All Rights Reserved.

864

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Rift valley

Transform fault

A Inside A′ Ridge corner transform intersection Rift valley axis

Inside corner high A′

A Detachment fault

Figure 12 Inside corner high at a slow-spreading ridge transform intersection. Extension is concentrated along a detachment fault for up to 1–2 million years, exposing deep sections of oceanic crust and mantle. The oceanic crust is thinned by this extreme extension; crustal accretion and magmatic activity may also be diminished.

spreading rates. This can be explained by the smaller mean normal stress across a fault plane that dips toward the axis, cutting through thin lithosphere, versus a fault plane that cuts through a much thicker section of lithosphere dipping away from the axis. Given reasonable thermal models, the difference in the thickness of the lithosphere cut by planes dipping toward versus away from the axis (and the mean normal stress across those planes) decreases significantly with spreading rate, making outward dipping faults more likely at fast-spreading rates (Figures 10 and 11). At all spreading rates, important along-strike variations in faulting occur within major (first- and second-order) spreading segments. Fault throws (inferred from scarp heights) decrease in the mid-segment regions away from discontinuities (Figures 8 and 11). This may be caused by a combination of thicker crust, thinner lithosphere, greater magma supply and less amagmatic extension away from RADs in the mid-segment region (Figure 11). Another possible explanation for along-strike variations in fault throw is along-strike variations in the degree of coupling between the mantle and crust. A ductile lower crust will tend to decouple the upper crust from extensional stresses in the mantle, and the existence of a ductile lower crust will depend on spreading rate, the supply of magma to the ridge, and proximity to major discontinuities. Estimates of crustal strain due to normal faulting vary from 10–20% on the slow-spreading MAR to B3–5% on the fast-spreading EPR. This difference may be explained as follows. The rate of magma supply to slow-spreading ridges is relatively low compared with the rate of crustal extension and

faulting, while extension and magma supply rates are in closer balance on fast-spreading ridges. The resulting seismicity is different too. In contrast to slowspreading ridges where teleseismically detected earthquakes are common, faulting at fast-spreading ridges rarely produces earthquakes of magnitude 44. Nearly all of these events are associated with RADs. The level of seismicity measured at fastspreading ridges accounts for only a very small percentage of the observed strain due to faulting, whereas fault strain at slow ridges is comparable to the observed seismic moment release. It has been suggested that faults in fast-spreading environments accumulate slip largely by stable sliding (aseismically) owing to the warm temperatures and associated thin brittle layer. At slower spreading rates, faults will extend beyond a frictional stability transition into a field where fault slip occurs unstably (seismically) because of a thicker brittle layer. Disruption of oceanic crust due to faulting may be particularly extreme on slow-spreading ridges near transform faults (Figure 12). Unusually shallow topography occurs on the active transform slip side of ridge transform intersections; this is called the high inside corner. These highs are not volcanoes. Instead they are caused by normal faults which cut deeply and perhaps all the way through oceanic crust. It is thought that crustal extension may occur for 1 to 2 million years on detachment faults with little magmatic activity. This results in extraordinary extension of the crust and exposure of large sections of the deep crust and upper mantle on the seafloor. Corrugated slip surfaces indicating the direction of fault slip are also evident and are called by some investigators, ‘megamullions.’ At distances of several tens of kilometers off-axis, topography generated near the spreading center is preserved on the seafloor with little subsequent change until it is subducted, except for the gradual accumulation of pelagic sediments at rates ofB0.5– 20 cm per thousand years. The preserved topographic highs and lows are called abyssal hills. At slow-spreading centers characterized by an axial rift valley, back-tilted fault blocks and half-grabens may be the dominant origin of abyssal hills (Figure 13), although there is continued controversy over the role of high-angle versus low-angle faults, listric faulting versus planar faulting, and the possible role of punctuated episodes of volcanism versus amagmatic extension. At intermediate-rate spreading centers, abyssal hill structure may vary with the local magmatic budget. Where the budget is starved and the axis is characterized by a rift valley, abyssal hills are generally back-tilted fault blocks. Where the magmatic budget is robust and an axial high is present,

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

865

Spreading direction

Planar or listric normal faults

~200 m ~2 km (A) Outward-dipping normal fault

Inward-dipping normal fault

(B) Flow fronts, elongate pillows

(C) Flow fronts Tensional fracture (D) Flow fronts or undraped outward-dipping normal faults Normal fault Volcanic growth faults

(E)

Figure 13 Five models for the development of abyssal hills on the flanks of midocean ridges. (A) Back-tilted fault blocks (episodic inward-dipping normal faulting off-axis). (B) Horst and graben (episodic inward- and outward-dipping faulting off-axis). (C) Whole volcanoes (episodic volcanism on-axis). (D) Split volcanoes (episodic volcanism and splitting on-axis). (E) Horsts bounded by inwarddipping normal faults and outward-dipping volcanic growth faults (episodic faulting off-axis and episodic volcanism on or near-axis).

the axial lithosphere is episodically thick enough to support a volcanic construction that may then be rafted away intact or split in two along the spreading axis, resulting in whole-volcano or split-volcano abyssal hills, respectively. Based on observations made from the submersible ALVIN on the flanks of the EPR, the outward facing slopes of the hills are neither simple outward dipping normal faults, as would be predicted by the horst/ graben model, nor are they entirely volcanic-constructional, as would be predicted by the split-volcano model. Instead, the outward facing slopes are ‘volcanic growth faults’ (Figure 14). Outward-facing scarps produced by episodes of normal faulting are buried near the axis by syntectonic lava flows originating along the axial high. Repeated episodes of dip–slip faulting and volcanic burial result in

Axial summit trough Hill Active faults Inactive volcanic growth faults ~ 6 km ~0.1 my

Syntectonic lava flows

'Volcanic growth fault' 0 km 0 my

Figure 14 Volcanic growth faults; cross-sectional depiction of the development of volcanic growth faults. Volcanic growth faults are common on fast-spreading centers and explain some of the differences between inward- and outward-facing scarps as well as the morphology and origin of most abyssal hills near fastspreading centers.

(c) 2011 Elsevier Inc. All Rights Reserved.

866

MID-OCEAN RIDGE TECTONICS, VOLCANISM, AND GEOMORPHOLOGY

Near-axis fault-bounded grabens

Grabens lengthen and link by crack propagation

Mean graben length increases with crustal age up to ~ 0.7_ 2 MY

Axis t1

Axis t2

Axis t3

Figure 15 Proposed time sequence of along-strike propagation and linkage of near-axis faults and grabens that define the edges of abyssal hills; time-averaged propagation rates are approximately 20–60 km per million years.

structures resembling growth faults, except that the faults are episodically buried by lava flows rather than being continuously buried by sediment deposition. In contrast, the inward dipping faults act as tectonic dams to lava flows. Thus, the abyssal hills are horsts and the intervening troughs are grabens with the important modification to the horst/graben model that the outward facing slopes are created by volcanic growth faulting rather than traditional normal faulting. Thus, on fast-spreading centers, abyssal hills are asymmetric, bounded by steeply dipping normal faults facing the spreading axis, and bounded by a volcanic growth faults on the opposing side (Figure 14). The abyssal hills lengthen at a rate of approximately 60 mm/y for the first B0.7 my by along strike propagation of individual faults as well as by linkage of neighboring faults (Figure 15).

See also Hydrothermal Vent Deposits. Hydrothermal Vent Fluids, Chemistry of. Mid-Ocean Ridge Seismic Structure. Propagating Rifts and Microplates.

Further Reading Buck WR, Delaney PT, Karson JA, and Lagabrielle Y (1998) Faulting and Magmatism at Mid-Ocean Rides. AGU Geophysical Monographs 106. Washington, DC: American Geophysical Union.

Humphris SE, Zierenberg RA, Mullineaux LS, and Thompson RE (1995) Seafloor Hydrothermal Systems: Physical, Chemical, Biological and Geochemical Interactions. AGU Geophysical Monographs 91. Washington, DC: American Geophysical Union. Langmuir CH, Bender JF, and Batiza R (1986) Petrological and tectonic segmentation of the East Pacific Rise, 51300 N–141300 N. Nature 322: 422--429. Macdonald KC (1982) Mid-ocean ridges: fine scale tectonic, volcanic and hydrothermal processes within the plate boundary zone. Annual Reviews of Earth and Planetary Science 10: 155--190. Macdonald KC and Fox PJ (1990) The mid-ocean ridge. Scientific American 262: 72--79. Macdonald KC, Scheirer DS, and Carbotte SM (1991) Mid-ocean ridges: discontinuities, segments and giant cracks. Science 253: 986--994. Menard H (1986) Ocean of Truth. Princeton, NJ: Princeton University Press. Phipps-Morgan J, Blackman DK, and Sinton J (1992) Mantle Flow and Melt Generation at Mid-ocean Ridges AGU Geophysical Monographs 71. Washington, DC: American Geophysical Union. Shaw PR (1992) Ridge segmentation, faulting and crustal thickness in the Atlantic Ocean. Nature 358: 490--493. Scheirer DS and Macdonald KC (1993) Variation in crosssectional area of the axial ridge along the East Pacific Rise. Evidence for the magmatic budget of a fastspreading center. Journal of Geophysical Research 98: 7871--7885. Sinton JM and Detrick RS (1992) Mid-ocean ridge magma chambers. Journal of Geophysical Research 97: 197--216.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE mantle matrix and buoyantly rises toward the surface, where it forms new, basaltic, oceanic crust. The crust and mantle cool at the surface by thermal conduction and hydrothermal circulation. This cooling generates a thermal boundary layer, which is rigid to convection and is the newly created edge of the tectonic plate. As the lithosphere moves away from the ridge, it thickens via additional cooling, becomes denser, and sinks deeper into the underlying ductile asthenosphere. This aging process of the plates causes the oceans to double in depth toward continental margins and subduction zones (Figure 2), where the oldest parts of plates are eventually thrust downward and returned to the hot underlying mantle from which they came. Mid-ocean ridges represent one of the most important geological processes shaping the Earth; they produce over two-thirds of the global crust, they are the primary means of geochemical differentiation in the Earth, and they feed vast hydrothermal systems

G. Ito and R. A. Dunn, University of Hawai’i at Manoa, Honolulu, HI, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Plate tectonics describes the motion of the outer lithospheric shell of the Earth. It is the surface expression of mantle convection, which is fueled by Earth’s radiogenic and primordial heat. Mid-ocean ridges mark the boundaries where oceanic plates separate from one another and thus lie above the upwelling limbs of mantle circulation (Figure 1). The upwelling mantle undergoes pressure-release partial melting because the temperature of the mantle solidus decreases with decreasing pressure. Newly formed melt, being less viscous and less dense than the surrounding solid, segregates from the residual

90

180 150 120 90 60 30

0

30

60

90

120 150

90

75

75 Iceland

60 Mid -At lan tic Rid ge

60

45 30

Pacif ic Ris e

15 0

East

15 30

45

Azores

30

Transform offset Ridge segment

15 0

Galapagos Spreading Center

15 30

Easter I. an Indi est

thw Sou e d i R g

45

45

60

60 75

180 150 120 90 60 30

0

30

8000 7000 6000 5000 4000 3000 2000 1000 0 1000 Topography (km)

60

2000

90

3000

120

4000

150

5000

75

6000

7000

8000

Figure 1 Map of seafloor and continental topography. Black lines mark the mid-ocean ridge systems, which are broken into individual spreading segments separated by large-offset transform faults and smaller nontransform offsets. Mid-ocean ridges encircle the planet with a total length exceeding 50 000 km. Large arrows schematically show the direction of spreading of three ridges discussed in the text.

(c) 2011 Elsevier Inc. All Rights Reserved.

867

868

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

(a)

Depth (km)

3 4 5 6

Plate HSC

7 50 100 Seafloor age (My)

0

Height

(b)

150

1

0.5

0

0

2

4

6 Width

8

10

12

Figure 2 (a) Global average (dots) seafloor depth (after correcting for sedimentation) and standard deviations (light curve) increase with seafloor age. Dashed curve is predicted by assuming that the lithosphere cools and thickens indefinitely, as if it overlies an infinite half space. Solid curve assumes that the lithosphere can cool only to a maximum amount, at which point the lithosphere temperature and thickness remain constant. (b) Temperature contours in a cross section through a 2-D model of mantle convection showing the predicted thickening of the cold thermal boundary layer (i.e., lithosphere, which beneath the oceans reaches a maximum thickness of B100 km) with distance from a mid-ocean ridge (left side). Midway across the model, small-scale convection occurs which limits the thickening of the plate, a possible cause for the steady depth of the seafloor beyond B80 Ma. (a) Modified from Stein CA and Stein S (1992) A model for the global variation in oceanic depth and heat flow with lithospheric age. Nature 359: 123–129, with permission from Nature Publishing Group. (b) Adapted from Huang J, Zhong S, and van Hunen J (2003) Controls on sublithospheric small-scale convection. Journal of Geophysical Research 108 (doi:10.1029/2003JB002456) with permission from American Geophysical Union.

that influence ocean water chemistry and support enormous ecosystems. Around the global ridge system, new lithosphere is formed at rates that differ by more than a factor of 10. Such variability causes large differences in the nature of magmatic, tectonic, and hydrothermal processes. For example, slowly spreading ridges, such as the Mid-Atlantic Ridge (MAR) and Southwest Indian Ridge (SWIR), exhibit heavily faulted axial valleys and large variations in volcanic output with time and space, whereas faster-spreading ridges, such as the East Pacific Rise (EPR), exhibit smooth topographic rises with more uniform magmatism. Such observations and many others can be largely understood in context of two basic processes: asthenospheric dynamics, which modulates deep temperatures and melt production rates; and the balance between heat delivered to the lithosphere, largely by magma migration, versus that lost to the surface by conduction and hydrothermal circulation. Unraveling the nature of these interrelated processes requires the integrated use of geologic mapping, geochemical and petrologic analyses, geophysical sensing, and geodynamic modeling.

Mantle Flow beneath Mid-ocean Ridges While the mantle beneath mid-ocean ridges is mostly solid rock, it does deform in a ductile sense, very slowly on human timescales but rapidly over geologic time. ‘Flow’ of the solid mantle is therefore often described using fluid mechanics. The equation of motion for mantle convection comes from momentum equilibrium of a fluid with shear viscosity Z and zero Reynolds number (i.e., zero acceleration): 
 rd Z rV þ rVT  rP þ r½ðz  2=3ZÞrdV þ ð1  fÞDrg ¼ 0

½1

The first term describes the net forces associated with matrix shear, where V is the velocity vector, rV is the velocity gradient tensor, and rVT is its transpose; the second term rP is the non-hydrostatic pressure gradient; the third term describes matrix divergence rdV (with effective bulk viscosity z) associated with melt transport; and the last term is the body force, with f being the volume fraction occupied by melt

(c) 2011 Elsevier Inc. All Rights Reserved.

869

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Imposed surface motion to simulate seafloor spreading 0

1500

Rigid, cool lithosphere

Depth (km)

20 40

1300 Hot ductile

18 1000

as thenosphere 14 6

60

10

2%/My

500

80 100 150

100

50 0 50 Across-axis distance from ridge (km)

100

0 150 Temp. (C)

Figure 3 Cross section of a 2-D numerical model that predicts mantle flow (shown by arrows whose lengths are proportional to flow rate), temperatures (shading), and melt production rate beneath a mid-ocean ridge (contoured at intervals of 2 mass %/My). In this particular calculation, asthenospheric flow is driven kinematically by the spreading plates and is not influenced by density variations (i.e., mantle buoyancy is unimportant). Model spreading rate is 6 mm yr 1.

and rr being the density contrast between the solid and melt. To first order, the divergence term is negligible, in which case eqn [1] describes a balance between viscous shear stresses, pressure gradients, and buoyancy. Locally beneath mid-ocean ridges, the spreading lithospheric plates act as a kinematic boundary condition to eqn [1] such that seafloor spreading itself drives ‘passive’ mantle upwelling, which causes decompression melting and ultimately the formation of crust (Figure 3). Independent of plate motion, lateral density variations can drive ‘active’ or ‘buoyant’ mantle upwelling and further contribute to decompression melting, as we discuss below. Several lines of evidence indicate that the upwelling is restricted to the upper mantle. The pressures at which key mineralogical transitions occur in the deep upper mantle (i.e., at depths 410 and 660 km) are sensitive to mantle temperature, but global and detailed local seismic studies do not reveal consistent variations in the associated seismic structure in the vicinity of mid-ocean ridges. This finding indicates that any thermal anomaly and buoyant flow beneath the ridge is confined to the upper mantle above the discontinuity at the depth of 410 km. Regional body wave and surface wave studies indicate that it could even be restricted to the upper B200 km of the mantle. Although some global tomographic images, based on body wave travel times, sometimes show structure beneath mid-ocean ridges extending down to depths of 300–400 km, these studies tend to artificially smear the effects of shallow anomalies below their actual depth extent. Further evidence comes from the directional dependence of seismic wave propagation speeds. This

seismic anisotropy is thought to be caused by latticepreferred orientation of olivine crystals due to mantle flow. Global studies show that at depths of 200–300 km beneath mid-ocean ridges, surface waves involving only horizontal motion (Love waves) tend to propagate slower than surface waves involving vertical motion (Rayleigh waves). This suggests a preferred orientation of olivine consistent with vertical mantle flow at these depths but not much deeper.

Mantle Melting beneath Mid-ocean Ridges Within the upwelling zone beneath a ridge, the mantle cools adiabatically due to the release of pressure. However, since the temperature at which the mantle begins to melt drops faster with decreasing pressure than the actual temperature, the mantle undergoes pressure-release partial melting (Figure 3). In a dry (no dissolved H2O) mantle, melting is expected to begin at approximately 60-km depth. On the other hand, a small amount of water in the mantle (B102 ppm) will strongly reduce the mantle solidus such that melting can occur at depths 4100 km. Detailed seismic studies observe low-velocity zones extending to depth of 100–200 km, which is consistent with the wet melting scenario. The thickness of oceanic crust times the rate of seafloor spreading is a good measure of the volume flux (per unit length of ridge axis) of melt delivered from the mantle. Marine seismic studies of the Mohorovicˇic´ seismic boundary (or Moho), which is often equated with the transition between the

(c) 2011 Elsevier Inc. All Rights Reserved.

870

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

gabbroic lower oceanic crust and the peridotitic upper mantle, find that the depth of the Moho is more or less uniform beneath seafloor formed at spreading rates of B20 mm yr  1 and faster (Figure 4). This observation indicates that the flux of melt generated in the mantle is, on average, proportional to spreading rate. What then causes such a behavior? Melt flux is proportional to the height of the melting zone as well as the average rate of (a)

upwelling within the melting zone. If the upwelling and melt production rate are proportional to spreading rate, then, all else being equal, this situation explains the invariance of crustal thickness with spreading rate. On the other hand, all else is not likely to be equal: slower spreading tends to lead to a thicker lithospheric boundary layer, a smaller melting zone, and a further reduction in melt production. The cause for the lack of decrease in

1

Rise

Axial relief (km)

0

Rift

1

2

Slow

Int.

Fast

3

(b)

10

8 Seismic crust (km)

Passive flow

Buoyant flow

6

4

Fast

Slow

2 Ultraslow 0 0

20

40

60

80

100

Full spreading rate (mm yr

120

140

160

1

)

Figure 4 (a) Axial morphology of ridges is predominantly rifted valleys (negative relief) along slow-spreading ridges and axial topographic highs (positive relief) along fast-spreading rates. The ultraslow spreading ridges (triangles) include the Gakkel Ridge in the Arctic and South West Indian Ridge, southeast of Africa. Dashed lines show conventional divisions between slow-, intermediate-, (labeled ‘Int.’), and fast-spreading ridges. (b) Seismically determined crustal thicknesses (symbols) are compared to theoretical predictions produced by two types of mantle flow and melting models: the passive flow curve is for a mantle flow model driven kinematically by plate spreading, the buoyant flow curve includes effects of melt buoyancy, which enhances upwelling and melting, particularly beneath slow-spreading ridges. Horizontal bars show a revised classification scheme for spreading rate characteristics. Reprinted by permission from Macmillan Publishers Ltd: Nature (Dick HJB, Lin J, and Schouten H (2003) An ultraslow-spreading class of ocean ridge. Nature 426: 405–412), copyright (2003).

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

(a)

(b)

Buoyancy-driven flow (R 169)

Depth (z /60 km)

Plate-driven flow (R  0.026)

871

0.0

0.2 0.4 Porosity (%)

0.6

0.0

0.5 1.0 1.5 Porosity (%)

Figure 5 Predictions from 2-D numerical models of mantle flow (white streamlines), melt retention (shading), and pressure-driven melt migration (black streamlines). (a) Fast spreading with high mantle viscosity (1020  1021 Pa s) impairs buoyant flow and leads to large pressure gradients, which draw melt from the broad melting zone toward the ridge axis. (b) Low mantle viscosity (1018  1019 Pa s) allows the low-density, partially molten rock to drive buoyant upwelling, which focuses beneath the ridge axis and allows melt to flow vertically. Black arrows show the predicted width of ridge-axis magmatism, which is still much broader than observed. Figure provided by M. Spiegelman (pers. comm., 2007). Also, see Spiegelman M (1996) Geochemical consequences of melt transport in 2-D: The sensitivity of trace elements to mantle dynamics. Earth and Planetary Science Letters 139: 115–132.

Depth (km)

0 20 40 60 150

100

50

4

4.1

0 Distance (km) 4.3 4.2 Vs (km s−1)

50

4.4

100

150

4.5

Figure 6 Cross-sectional tomographic image of the upper mantle shear wave velocity structure beneath the southern East Pacific Rise produced from Love wave data. Ridge axis is at x ¼ 0 km. At the top of the figure the high-velocity lithospheric lid is clearly evident, as well as its thickening with distance from the ridge. The low-velocity zone beneath the ridge is consistent with the presence of high temperatures and some retained melt. Body wave and Rayleigh wave studies indicate that the low-velocity zone extends even wider below than what is shown here. Adapted from Dunn RA and Forsyth DW (2003) Imaging the transition between the region of mantle melting and the crustal magma chamber beneath the southern East Pacific Rise with short-period Love waves. Journal of Geophysical Research 108(B7): 2352 (doi:10.1029/2002JB002217), with permission from American Geophysical Union.

crustal thickness with decreasing spreading rate for rates 4c. 20 mm yr  1 must involve other processes. A possible solution could have to do with the likelihood that the relative strengths of the platedriven (i.e., kinematic) versus buoyancy-driven mantle upwelling change with spreading rate. Let us examine two end-member scenarios for mantle flow and melting. Case 1 considers a situation in which mantle flow is driven entirely kinematically by the

separation of the lithospheric plates (Figures 3 and 5(a)). This passive flow scenario is predicted if the plate-driven component of flow overwhelms the buoyancy-driven part (i.e., the last term in eqn [1]). The other end-member possibility, case 2, considers a situation in which buoyant flow dominates over the plate-driven part. Lateral density variations, Dr, in the melting zone probably occur due to the presence of small amounts of melt, which has a lower density

(c) 2011 Elsevier Inc. All Rights Reserved.

872

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Ridge axis

Freezing zone

~10 km

Constant melt flux Net Buoyancy Dilation

Figure 7 (Left) Cross-sections of a 2-D model of melt migration, with shading showing porosity (black ¼ 0%, white ¼ 3.5%) and at four different times, increasing clockwise from the upper left. A constant melt percolation flux rises through the bottom of the box. A ‘freezing boundary’ represents the cooler lithosphere which slopes toward the ridge axis (upper left). (Right) Enlarged portion of red box. The freezing boundary halts the rise of melt and diverts melt parallel to it. The net pressure gradient driving the flow is caused by the two components shown by the arrows. The freezing boundary generates porosity waves that propagate away from the boundary. These waves are predicted mathematically to arise from eqns [1], [2], and two others describing conservation of melt and solid mass. Modified from Spiegelman M (1993) Physics of melt extraction: Theory, implications and applications. Philosophical Transactions of the Royal Society of London, Series A 342: 23–41, with permission from the Royal Society.

than the solid. A positive feedback can occur between melting and buoyant flow, such that melting increases buoyancy and upwelling, which leads to further melting (Figure 5(b)). Returning to the weak dependence on spreading rate for rates 420 mm yr 1; at least at fast-spreading rates, the plate-driven component of flow can be as strong or stronger than any buoyancy component. Thus, numerical models of fast-spreading ridges that do or do not include buoyancy predict similar crustal thicknesses and an insensitivity of crustal thickness to spreading rate (Figure 4). However, as spreading rate drops, buoyancy forces can become relatively important, so that – below slow–intermediate-spreading lithosphere – they generate ‘fast’ mantle upwelling. This fast upwelling is predicted to enhance melt production and compensate for the effects of surface cooling to shrink the melting zone; crustal thickness is therefore maintained as spreading rates decrease toward B20 mm yr 1. Observational evidence for the relative importance of plate-driven versus buoyant flow is provided by studies of body and surface wave data along the EPR. Tomographic images produced from these data reveal a broad region of low seismic wave speeds in the upper mantle (Figure 6), interpreted to be the region of melt production. To date, there is little indication of a very narrow zone of low wave speeds, such as that predicted for buoyant upwelling zone as depicted by case 2 (Figure 5(b)). These findings

support the predictions in Figure 4 that plate-driven flow is strong at fast-spreading ridges, such as the EPR. At spreading rates less than 20 mm yr 1, however, the melting process appears to change dramatically. Here, melt flux is not proportional to spreading rate (Figure 4). At these ultraslow rates, crustal thickness drops rapidly with decreasing spreading rate, suggesting a nonlinear decrease in magma flux. A leading hypothesis suggests that the melt reducing effects of the top-down cooling and corresponding shrinkage of the melting zone overwhelm the meltenhancing effects of any buoyancy-driven upwelling. Whatever the exact cause is, the large variability in crustal thickness seen at these spreading rates is one example of the large sensitivity of ridge-axis processes to surface cooling at slow or ultraslow spreading rates.

Melt Transport to Ridge Axes How melt is transported upward from the mantle source to the ridge is another long-standing problem. Seismic studies of oceanic crust indicate that the crust is fully formed within a few of kilometers of the axis of a ridge, requiring either a very narrow melting zone beneath the ridge or some mechanism that focuses melts from a broader melting zone to a narrow region at the ridge axis.

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

873

East Pacific Rise 2200

0 10 9 N

9 30'

12 30'

13 N

13 30'

3000

2800

20

3600 96 W

95 W

Mid-Atlantic Ridge 20

C 3000

0 3800

20 34 S

33 S

32 S

31 S 3000

20 3000 0 3800 20

25 S

26 S

0 3800

20 22 30' N

23 N

20

A x

0

Bathymetry (m) (-----------)

Mantle Bouguer anomaly (mgal) (-----------)

Cocos−Nazca Ridge 0

23 30' N 3000 3800

g

20

4600 28 N

29 N

30 N

Figure 8 Profiles of seafloor depth (dashed lines) and mantle Bouguer gravity anomaly (solid lines) taken along the axes of various mid-ocean ridges (as indicated in the figure). Arrows mark various ridge-axis discontinuities. Note anticorrelation of gravity and bathymetry along the MAR, indicating shallower bathymetry and thicker crust near the centers of ridge segments. The East Pacific Rise is a fast-spreading ridge, the Cocos–Nazca Ridge (also called the Galapagos Spreading Center) spreads at an intermediate rate, and the MAR is a slow-spreading ridge. Adapted from Lin J and Phipps Morgan J (1992) The spreading rate dependence of threedimensional mid-ocean ridge gravity structure. Geophysical Research Letters 19: 13–16, with permission from American Geophysical Union.

Laboratory experiments and theory show that melt can percolate through the pore space of the matrix (with volume fraction f) in response to matrix pressure gradients rP. The Darcy percolation flux is described by fðv  VÞ ¼ ðk=mÞrP

½2

where (v  V) is the differential velocity of melt v relative to the matrix V, m is melt viscosity, and k is the permeability of the porous matrix. Equation [1] shows that rP is influenced by both melt buoyancy and matrix shear: melt buoyancy drives vertical percolation while matrix shear can push melt

sideways. In mantle flow case 1, in which plate-driven flow dominates, the zone of melting is predicted to be very wide, requiring large lateral pressure gradients to divert melt 50–100 km sideways toward the ridge (Figure 5). For matrix shear to produce such large lateral gradients requires very high asthenospheric viscosities (1020  1021 Pa s). When buoyant flow dominates (case 2), the melting zone is much narrower and the low viscosities (1018 1019 Pa s) generate small lateral pressure gradients such that the melt rises mostly vertically to feed the ridge axis. Both of these cases, however, still predict zones of magmatism at the ridge axis that are wider than those typically observed.

(c) 2011 Elsevier Inc. All Rights Reserved.

874

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

4

ho ohoo MM

15

10

Depth (km)

0

8 12 5 0 Alo ng axis 5 (km )

6 6

10 15

12

0 (km) axis ross

Ac

2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

0.2

0.4

0.6 (km s−1)

Seismic P wave speed Figure 9 A perspective view of seafloor bathymetry and a seismic tomographic image of the East Pacific Rise (91 N latitude) magmatic system. The image is of the P wave velocity anomaly, relative to an average vertical velocity profile, contoured at 0.2 km s 1. The vertical planes show the continuity of the crustal magmatic system beneath the ridge (the significant low-velocity region centered beneath the ridge). The deep horizontal plane is located in the mantle just below the crust and shows that the crustal low-velocity region extends downward into the mantle. The mantle velocity anomaly is continuous beneath the ridge, but shows variations in magnitude and location that suggest variations in melt supply. Adapted from Dunn RA, Toomey DR, and Solomon SC (2000) Three-dimensional seismic structure and physical properties of the crust and shallow mantle beneath the East Pacific Rise at 91 300 N. Journal of Geophysical Research 105: 23537–23555, with permission from the American Geophysical Union.

An additional factor that can help further focus melts toward the ridge is the bottom of the cold lithosphere, which shoals toward the ridge axis. Theoretical studies predict that as melt rises to the base of the lithosphere, it freezes and cannot penetrate the lithosphere (Figure 7). But the steady percolation of melt from below causes melt to collect in a high-porosity channel just below the freezing boundary. The net pressure gradient that drives melt percolation is the vector sum of the component that is perpendicular to the freezing front, caused by matrix dilation as melt fills the channel, and the vertical component caused by melt buoyancy. Consequently, melt flows along the freezing front, upward toward the ridge axis.

Regional and Local Variability of the Global Mid-Ocean Ridge System Major differences in the regional and local structure of mid-ocean ridges are linked to the previously noted

processes that influence asthenospheric flow and the heat balance in the lithosphere. One example that highlights a mid-ocean ridge’s sensitivity to lithospheric heat balance is the overall shape or morphology of ridges at different spreading rates. At fast-spreading rates, the magmatic (heat) flux is high and this forms a hot crust with thin lithosphere. The mechanical consequences of both a thin lithosphere and relatively frequent magmatic intrusions to accommodate extension generate a relatively smooth, axial topographic ridge standing hundreds of meters above the adjacent seafloor (Figures 1 and 4). Slowspreading ridges, however, have proportionally smaller magma fluxes, cooler crust, and thicker lithosphere. The mechanical effects of a thick lithosphere combined with less-frequent eruptions to accommodate extension cause the axes of slowspreading ridges to be heavily faulted valleys, as deep as 2 km below the adjacent seafloor. Ridges spreading at intermediate rates show both morphologies, and appear to be sensitive to subtle fluctuations in magma supply. Seismic imaging of the crust reveals that melt

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Ocea

Depth (km)

nogra

0 2

Nontr ansfo

xis ge a

rm off

re Zo

ne

4 6 8 10 10 0 Acro ss a xis ( km)

10 20

20 10 1.1 0 ) 10 (km 0.9 20 axis g n 30 Alo 0.7 Vp (km s−1)

20

Depth (km)

Fractu

Moho

Rid

set

pher

875

0 2

30

0.5 0.3 0.1 0.1 0.3 0.5

Moho

0.7

4 6 8 10

30

20

20 10 Ac ros sa

0

0

xis

10

(km

)

20

30

20

s 10 axi ng o l A

10 ) (km

Figure 10 A perspective view of seafloor bathymetry and a tomographic image of the MAR (351 N latitude) magmatic system. The image is of the P wave velocity anomaly, relative to an average vertical velocity profile, contoured at 0.1 km s 1. The vertical planes reveal partially molten bodies (the low-velocity regions), which are discontinuous beneath the ridge. The deep horizontal plane is located in the mantle just below the crust and shows that the crustal low-velocity region extends downward into the mantle. Crustal thickness is also greatest at the center of the ridge (black line labeled ‘Moho’). The seismic image indicates that as magma rises in the mantle, it becomes focused to the center of the ridge segment where it then feeds into the crust. Adapted from Dunn RA, Lekic V, Detrick RS, and Toomey DR (2005) Three-dimensional seismic structure of the Mid-Atlantic Ridge at 351 N: Focused melt supply and non-uniform plate spreading. Journal of Geophysical Research 110: B09101 (doi:10.1029/2004JB003473), with permission from American Geophysical Union.

supply and crustal structure vary with spreading rate (Figure 4), geodynamic setting, as well as time. Another major characteristic is the variability in topography, gravity, and crustal thickness as a function of distance along different mid-ocean ridges. Along individual segments of fast-spreading ridges, topography, gravity, as well as crustal thickness are remarkably uniform, varying by less then B20% (Figures 8 and 9). Such relative uniformity probably indicates a steady magma supply from below as evident from seismic imaging of magma being stored over large distances along fast-spreading ridges (Figure 9). Quasi-steady-state crustal magmatic systems have been seismically shown to extend into the underlying mantle. That said, the variability that is present between and along fast-spreading ridge segments reveals some important processes. Like all mid-ocean ridges,

fast-spreading ridge segments are separated by largeqoffset fracture zones at the largest scale, and also by overlapping spreading centers (OSCs) at an intermediate scale. OSCs are characterized by the overlap of two en echelon ridge segments that offset the ridge by several kilometers. In one view, OSCs occur at the boundary between two widely separated, and misaligned, regions of buoyant mantle upwelling. An opposing model states that OSCs are mainly tectonic features created by plate boundary reorganization, below which mantle upwelling is primarily passively driven. A consensus on which hypothesis best explains the observations has not yet been reached. At still finer scale, segmentation is apparent as minor morphologic deviations from axial linearity (or ‘DEVALs’) at intervals of 5–25 km. Individual DEVAL-bounded segments of the EPR are associated with higher proportions of melt in the crustal and

(c) 2011 Elsevier Inc. All Rights Reserved.

876

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

(a)

10 E

11 E

12 E

13 E

14 E

15 E

1

m 9m

16 E

17 E

6.2 mm yr1

yr

3.

52 S re

ctu

e

ka

n zo

fra

a

Sh

Narrowgate magmatic segment

53 S

Joseph Mayes segment Magmatic segment 54 S (b)

16 E discontinuity

100 km

Scale ~ 1:500 000

52 S

53 S 15 0 5 10 15 20 25 30 35 40 45 Mantle Bouguer anomaly (5 mgal contours)

10 E

11 E

12 E

13 E

14 E

15 E

16 E

Figure 11 (a) Map of the Southwest Indian Ridge bathymetry (200-m contours). Ridge axis runs left to right across this figure. Large red arrows show relative direction of seafloor spreading (oblique to the ridge axis). Circles with black and white pattern indicate the locations and slip mechanisms of recorded earthquakes. Red and green dots indicate locations where crustal basaltic rocks and mantle peridotite rocks, respectively, were recovered. Significant amounts of mantle peridotite can be found at the seafloor along ultraslow-spreading ridges. (b) Mantle Bouguer gravity anomaly. The large gravity lows signify thick belts of crust and/or low-density mantle, and correspond to regions where basalts have been predominantly recovered. Reprinted by permission from Macmillan Publishers Ltd: Nature (Dick HJB, Lin J, and Schouten H (2003) An ultraslow-spreading class of ocean ridge. Nature 426: 405–412), copyright (2003).

upper mantle magmatic system. This suggests that the melt flux from the mantle is locally greater between DEVALs than at the boundaries. The cause is controversial and may even have more than one origin. One possible origin is small-scale mantle diapirism that locally enhances melt production. Such a hypothesis is deduced from deformation fabrics seen in ophiolites, which are sections of oceanic lithosphere that are tectonically thrust onto continents. Alternatively, melt production can be locally enhanced by mantle compositional heterogeneity. Still other possibilities involve shallower processes such as variability in melt transport. In stark contrast to fast-spreading ridges, slowspreading ridges show huge variability in topography,

gravity, and crustal thickness along individual spreading segments that are offset by both transform and nontransform boundaries (Figures 8, 10, and 11). The crust is usually thickest near the centers of ridge segments and can decrease by 50% or more toward segment boundaries (Figure 10). These and several other observations probably indicate strong alongaxis variability in mantle flow and melt production. For example, a recent seismic study reveals a large zone in the middle to lower crust at the center of a slow-spreading ridge segment with very low seismic wave speeds. This finding is consistent with locally elevated temperatures and melt content that extend downward into the uppermost mantle. Although the observations can be explained by several different

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Mantle viscosity (Pa s)

(a)

1020 2-D 1019

1018 3-D 1017

0

20 40 60 Half-spreading rate (mm yr −1)

(b)

~100 km

Hig

Depth

h

, de

Low

Ac

ros

hyd ra , h ted ydr ous

sa

xis

xis

ga

n Alo

Figure 12 Predictions of a 3-D numerical model of mantle flow and melting. (a) Predicted variability in crustal production along the model ridge is characterized as 3-D or 2-D if the variability is, respectively, larger than or less than an arbitrary threshold. Along-axis variation increases with decreasing mantle viscosity and with decreasing spreading rate. (b) Perspective view showing retained melt (shading, varying from 0% far from the axis to 1.8% at centers of columnar zones), mantle flow (small white arrows of length proportional to flow rate), temperature (white contours), and melt productivity (black contours). The large white arrow depicts a plate spreading slowly at a rate of 12 mm yr 1 away from the plane of symmetry at the ridge axis (right vertical plane). Melt retention buoyancy generates convective mantle upwellings in the lower part of the melting zone where viscosities are low (below the red line). In the upper portion of the melting zone (above the red line), viscosity is high, owing to the extraction of water from the solid residue. In this zone, plate-driven mantle flow dominates. Thus, essentially all of the along-axis variability is generated in the lower half of the melting zone. It is the thickness of this lower zone of melting that controls the wavelength of variability. Wavelengths of 50–100 km are typical along the MAR. Adapted from Choblet G and Parmentier EM (2004) Mantle upwelling and melting beneath slow spreading centers: Effects of variable rheology and melt productivity. Earth and Planetary Science Letters 184: 589–604, with permission from American Geophysical Union.

mantle flow and melt transport scenarios, a predominant view supports the hypothesis that subridge mantle flow beneath slow-spreading ridges is largely influenced by lateral density variations.

877

Causes for the major differences in along-axis variability between fast- and slow-spreading ridges have been explored with 3-D numerical models of mantle convection and melting. Models of only platedriven flow predict that the disruption of the ridge near a segment offset both locally reduces upwelling and enhances lithosphere cooling beneath it, both of which tend to somewhat reduce melt production near the offset. For a given length of segment offset, the size of the along-axis variability is smallest at the fastest spreading rates and increases with decreasing spreading rate. This prediction is broadly consistent with the observations; however, such models still underpredict the dramatic variability observed along many segments of slow-spreading ridges. Again, a consideration of both plate- and buoyancydriven flow provides a plausible solution. In the direction parallel to the ridge axis, variations in density can be caused by changes in temperature, retained melt, as well as solid composition due to melt extraction (melting dissolves high-density minerals and extracts high-density elements like iron from the residual solid). All three sources of buoyancy are coupled by the energetics and chemistry of melting and melt transport. As discussed above, models of fastspreading systems predict plate-driven flow to be most important such that buoyancy causes only subtle along-axis variations in melting (Figure 12). As spreading rate decreases, the relative strength of buoyant flow increases as does the predicted variability of melt production. Models that include buoyancy more successfully predict typical amplitudes of variations along slow-spreading ridges. Even more dramatic melt supply variations are observed at a few locations along ridges, which include the MAR at Iceland and near the Azores Islands, the Galapagos Spreading Center near the Galapagos Archipelago, and the EPR near Easter Island (Figure 1). These regions occur where ‘hot spots’ in the mantle produce so much magmatism that islands are formed. Iceland in fact is a location where a mid-ocean ridge is actually exposed above sea level. These ‘hot-spot-influenced’ sections of midocean ridges show elevated topography and enhanced crustal thickness over distances of many hundreds of kilometers. The most likely cause for these features are anomalously hot, convective upwellings that rise from depths at least as deep as the base of the upper mantle. Fluid dynamical studies show that plumes of rising mantle can arise from hot thermal boundary layers such as the core mantle boundary. When these hot upwellings eventually rise to the lithosphere, they expand beneath it and can enhance volcanism over large distances (Figure 13).

(c) 2011 Elsevier Inc. All Rights Reserved.

878

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Residual topography (km)

(a) 5 12 4 3 12 7 12 7 12 2 1 2 121 2 12 61212 7 2 7 22 2 22

3 2

North 1

1

South 9

10

11

0 1

(b) 5 4

12

30

12

1212

1 21 1 2

20

3

10

12 6 7

7 7 12 7

2 2 2 2 2 22

1

1000

12 12

9 10

500 0 500 Distance from Iceland (km)

(c)

11

1000

Plate spreading

0

is Ridge ax

100 200

400

Ax

is

200 rp en di cu

pe

0

la

100 0

200

300 is

g ax

Alon

400 (km)

Depth (km)

Crustal thickness (km)

40

300 500

r

Figure 13 (a) Observed residual topography (solid curve and circles) and (b) crustal thickness of Iceland and the MAR, compared to the predictions of a 3-D model of a hot mantle plume rising beneath the ridge (dashed). (c) Perspective view of potential temperatures (white 4c. 1500 1C) within the 3-D model. The vertical cross sections are along (right) and perpendicular (left) to the ridge. Viscosity decreases with temperature and increases at the dry solidus by 102 because water is extracted from the solid with partial melting. Thermal buoyancy causes the hot plume material to spread hundreds of kilometers along the MAR away from Iceland. Crustal thickness is predicted to be greatest above the hot plume and to decrease away from Iceland due to decreasing temperatures. Reproduced from Ito G and van Keken PE (2007) Hot spots and melting anomalies. In: Bercovici D (ed.) Treatise in Geophysics, Vol. 7: Mantle Dynamics. Amsterdam: Elsevier, with permission from Elsevier.

So far, we have discussed characteristics of the mid-ocean ridge system that are likely to be heavily influenced by differences in heat transport and mantle flow. At the frontier of our understanding of mantle processes is the importance of composition. Studies of seafloor spreading centers at back of the arcs of subduction zones reveal how important mantle composition can be to seafloor creation. For example, the Eastern Lau Spreading Center is characterized by rapid along-strike trends in many observations that are contrary to or unseen along normal mid-ocean ridges. Contrary to the behavior of mid-ocean ridges, as spreading rate increases

along the Eastern Lau back arc spreading system from a slow rate of B40 mm yr 1 in the south to an intermediate rate of B95 mm yr 1 in the north, the ridge axis changes from an inflated axial high to a faulted axial valley and the evidence for magma storage in the crust disappears. Coincident with this south-to-north variation, the crustal composition changes from andesitic to tholeiitic and isotopic characteristics change from that of the Pacific domain to more like that in the Indian Ocean. It is hypothesized that many of the along-strike changes along the Eastern Lau Spreading Center are produced by variable geochemical and petrological

(c) 2011 Elsevier Inc. All Rights Reserved.

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Thin back-arc crust

Thick back-arc crust

Remnant arc

879

Arc volcanic front

Ridge melting

'no

re sid ue

rm

Pa r

b

Am

 Normal

Reduced

Increased

er at W

n

letio

Dep

Mantle wedge characterisitcs relative to MORB source mantle



Melt productivity relative to MORs

t ien

Arc melting

t-m elt ed

tle

an

m al'

'B pa lob r t- s' re me of sid lte ue d

Solidus

Proximity to arc volcanic front Figure 14 Mantle composition in the wedge above a subducting slab can significantly affect melting beneath back-arc spreading centers. In this scenario, buoyant (partially melted) residual mantle from the arc region is rehydrated by water expelled from the subducting plate, becomes less viscous, and rises into the melting regime of the spreading center, where (because it has already been partially melted) it reduces the total melt production (middle panel). Arc volcanism occurs closer to the subduction zone and originates as fluids percolate from the subducting slab up into the hot mantle wedge and cause melting by reducing melting temperature. If seafloor spreading were closer to this arc melting zone, it would likely form ‘thick back-arc crust’ and an axial morphology that resembles fast spreading ridges even though spreading here could be slow. Bottom curves schematically show melt depletion and hydration trends in the mantle wedge and their hypothesized effects on the ridge melt productivity with distance from the arc volcanic front. MORS, mid-ocean ridges; MORB, mid-ocean ridge basalt. Adapted from Martinez F and Taylor B (2006) Modes of crustal accretion in back-arc basins: Inferences from the Lau Basin. In: Christie DM, Fisher CR, Lee S-M, and Givens S (eds.) Geophysical Monograph Series 166: Back-Arc Spreading Systems: Geological, Biological, Chemical, and Physical Interactions, pp. 5–30 (10.1029/ l66GM03). Washington, DC: American Geophysical Union, with permission from American Geophysical Union.

inputs influenced by subduction (Figure 14). From south to north, the distance of the ridge from the Tonga arc increases from 30 to 100 km and the depth to the underlying slab increases from 150 to 250 km. To the south, melt production is most likely enhanced by the proximity of the ridge to the arc which causes the ridge to tap arc volcanic melts (slab-hydrated); whereas to the north, melt flux is probably reduced by the absence of arc melts in

the ridge melting zone, but in addition, the mantle flow associated with subduction could actually deliver previously melt-depleted residue back to the ridge melting zone. Yet farther to the north, the ridge is sufficiently far away from the slab, such that it taps ‘normal’ mantle and shows typical characteristics of mid-ocean spreading centers. Similar hypotheses have been formed for other back-arc systems.

(c) 2011 Elsevier Inc. All Rights Reserved.

880

MID-OCEAN RIDGES: MANTLE CONVECTION AND FORMATION OF THE LITHOSPHERE

Summary The global mid-ocean ridge system is composed of the divergent plate boundaries of plate tectonics and it is where new ocean seafloor is continually created. Of major importance are the effects of plate motion versus buoyancy to drive asthenospheric upwelling, the balance between heat advected to the lithosphere versus that lost to the seafloor, as well as mantle compositional heterogeneity. Such interacting effects induce variations in the thickness of crust as well as local structural variability of mid-ocean ridge crests that are relatively small at fast-spreading ridges but become more dramatic as spreading decreases. Through examining this variability geoscientists are gaining an understanding of mantle convection and chemical evolution as well as key interactions with the Earth’s surface.

See also Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Seismicity. Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology. Seamounts and Off-Ridge Volcanism.

Further Reading Buck WR, Lavier LL, and Poliakov ANB (2005) Modes of faulting at mid-ocean ridges. Nature 434: 719--723. Choblet G and Parmentier EM (2004) Mantle upwelling and melting beneath slow spreading centers: Effects of variable rheology and melt productivity. Earth and Planetary Science Letters 184: 589--604. Dick HJB, Lin J, and Schouten H (2003) An ultraslowspreading class of ocean ridge. Nature 426: 405--412. Dunn RA and Forsyth DW (2003) Imaging the transition between the region of mantle melting and the crustal magma chamber beneath the southern East Pacific Rise with short-period Love waves. Journal of Geophysical Research 108(B7): 2352 (doi:10.1029/2002JB002217). Dunn RA, Lekic V, Detrick RS, and Toomey DR (2005) Three-dimensional seismic structure of the Mid-Atlantic Ridge at 351 N: Focused melt supply and non-uniform plate spreading. Journal of Geophysical Research 110: B09101 (doi:10.1029/2004JB003473). Dunn RA, Toomey DR, and Solomon SC (2000) Threedimensional seismic structure and physical properties of

the crust and shallow mantle beneath the East Pacific Rise at 91 300 N. Journal of Geophysical Research 105: 23537--23555. Forsyth DW, Webb SC, Dorman LM, and Shen Y (1998) Phase velocities of Rayleigh waves in the MELT experiment on the East Pacific Rise. Science 280: 1235--1238. Huang J, Zhong S, and van Hunen J (2003) Controls on sublithospheric small-scale convection. Journal of Geophysical Research 108: 2405 (doi:10.1029/2003 JB002456). Ito G and van Keken PE (2007) Hot spots and melting anomalies. In: Bercovici D (ed.) Treatise in Geophysics, Vol. 7: Mantle Dynamics. Amsterdam: Elsevier. Lin J and Phipps Morgan J (1992) The spreading rate dependence of three-dimensional mid-ocean ridge gravity structure. Geophysical Research Letters 19: 13--16. Martinez F and Taylor B (2006) Modes of crustal accretion in back-arc basins: Inferences from the Lau Basin. In: Christie DM, Fisher CR, Lee S-M, and Givens S (eds.) Geophysical Monograph Series 166: Back-Arc Spreading Systems: Geological, Biological, Chemical, and Physical Interactions, pp. 5--30. Washington, DC: American Geophysical Union (doi:10.1029/l66GM03). Phipps Morgan J, Blackman DK, and Sinton JM (eds.) (1992) Geophysical Monograph 71: Mantle Flow and Melt Generation at Mid-Ocean Ridges, p. 361. Washington, DC: American Geophysical Union. Phipps Morgan J and Chen YJ (1993) Dependence of ridge-axis morphology on magma supply and spreading rate. Nature 364: 706--708. Shen Y, Sheehan AF, Dueker GD, de Groot-Hedlin C, and Gilbert H (1998) Mantle discontinuity structure beneath the southern East Pacific Rise from P-to-S converted phases. Science 280: 1232--1234. Spiegelman M (1993) Physics of melt extraction: Theory, implications and applications. Philosophical Transactions of the Royal Society of London, Series A 342: 23--41. Spiegelman M (1996) Geochemical consequences of melt transport in 2-D: The sensitivity of trace elements to mantle dynamics. Earth and Planetary Science Letters 139: 115--132. Stein CA and Stein S (1992) A model for the global variation in oceanic depth and heat flow with lithospheric age. Nature 359: 123--129.

Relevant Websites http://www.ridge2000.org – Ridge 2000 Program.

(c) 2011 Elsevier Inc. All Rights Reserved.

MILLENNIAL-SCALE CLIMATE VARIABILITY J. T. Andrews, University of Colorado, Boulder, CO, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Analysis of Quaternary marine sediment cores has changed in emphasis several times over the last four decades. In particular, this has involved a change in focus from variations in proxy records on orbital or Milankovitch timescales (with recurring periodicities of c. 20 000, 40 000, and 100 000 years), to an interest in the sub-Milankovitch variability (Figure 1). In turn, this has frequently meant a change in the length of the record from several million years, to several tens of thousands of years (often the last glacial/deglacial cycle which extended from 120 000 years ago to the present). It has also meant an increased interest in sites with high rates of sediment accumulation (4 410 cm ky 1). (a)

(b) 0

History The ability to undertake millennial-scale climate reconstructions from marine sediments was conditioned by several requirements, which could not be met until the 1980s and early 1990s. An underlying rationale for this interest was that of the results from (c)

(d)

OS#

GPC-9

20

Although not precisely defined, the term ‘millennialscale climate variability’ is usually considered to cover events with periods of between 1000 and 10 000 years. Evidence for abrupt, millennial-scale changes in ocean sediments (Figures 1(a) and 1(b)) has resulted in a paradigm shift. The role of the oceans in abrupt climate forcing is now considered to be paramount, whereas under the Milankovitch scenario the role of the oceans was frequently considered subordinate to changes on land associated with the growth and decay of the large Quaternary ice sheets (Figure 2), which were principally driven by changes in high-latitude, Northern Hemisphere insolation (e.g., Figure 1(d)).

1 LGM

2

The Holocene

Cal yr ka

OS4

4

H-2

10 15

LGM

LGM

60

20 25

80

80

H-3 30

5 100

H-4 0

120

20

40

H-1

3

60

5 LGM

Younger Dryas 40

0

0

GPC-9

DSDP-609 The Holocene

0.2 0.4 0.6 0.8 Lithics/total

D 1

0

100 35

10 20 30 40 50 Carbonate weight %

120

Penultimate 6 glaciation 140

140 0

10 20 30 40 50 Carbonate weight %

390 420 450 480 W m−2 Jul. 65° N

Figure 1 Examples of millennial-scale oscillations in marine records (see Figure 2 for core locations). Data were taken from the National Oceanographic and Atmospheric Administration Paleoclimate Database (http://www.ncdc.noaa.gov/paleo). Age is given in thousands of years ago (ka) and the marine oxygen isotope stages (OS) are shown. The temporal location of Heinrich events (H-1–H-3), the Last Glacial Maximum (LGM), and the dramatic Younger Dryas cold event are shown. On the right-hand side is the summer insolation at 65 1 N for July. Notice the absence of millennial-scale variability in the Milankovitch forcing of global climate although the main peaks and troughs are picked out in the carbonate record (left-hand column).

(c) 2011 Elsevier Inc. All Rights Reserved.

881

882

MILLENNIAL-SCALE CLIMATE VARIABILITY

Arctic Ocean

Eurasian ice sheets

North American ice sheets

54

-0 09

-0

87

HU

75

HU

V29-204

DSDP-609

GPC-g

Ceara Rise Meltwater and IRD transport

Approx. southern limit of IRD

Northern Hemisphere ice sheets

Figure 2 The glacial world of the Last Glacial Maximum (LGM) showing routes for the export of fresh water (from meltwater, e.g., Gulf of Mexico), and the major iceberg-rafted detrital (IRD) sources for deposition in North Atlantic basins. The location of several cores mentioned in the article are shown.

the Greenland and Antarctic ice core records which showed remarkably rapid oscillations in a variety of proxies for the last 40 000–80 000 years. The advent of accelerator mass spectrometry (AMS) 14C dating

of small (2–10 mg) samples of foraminifera allowed many of the world’s deep-sea and shelf sediments to be directly dated up to a limit of between 30 000 and 40 000 years ago. Because of the small sample size

(c) 2011 Elsevier Inc. All Rights Reserved.

MILLENNIAL-SCALE CLIMATE VARIABILITY

required and the relatively fast turnaround, it became possible to obtain many dates on a core, and in some cases the density approached one date per thousand years. This technology also meant that sediment cores from environments with high rates of sediment accumulation could now be successfully dated; thus a variety of sediment environments from the ‘drifts’ around the North Atlantic, to glaciated shelves and fiords, could now be studied. In these environments, sediment accumulation rates (SARs) were often greater than 20 cm ky  1 and could reach rates as high as 2 m ky  1. In areas with these very high rates of sediment accumulation there was a clear need for improved coring technology such that cores of tens of meters in length could be obtained. Giant piston cores were thus developed with recoveries in the range of 20–60 m. An example is the Calypso system deployed from the French research vessel Marion Dufresne and employed as part of the IMAGES (International Marine Past Global Change) program. This allowed for very high-resolution studies (SAR of Z1 m ky  1, decadal resolution) if the cores recovered sediments with basal dates of 10 000 BP. If the cores were extracted from areas with more modest rates of sediment accumulation (410 and o100 cm ky 1), then millennial-scale studies would have been possible. However, one problem with longer temporal records was that they recovered sediments with ages greater than the radiocarbon limit of c. 45 ka. Records that extended back into marine or oxygen isotope stages 4 and 5 (Figure 1, OS column) could not be numerically dated per se, but their chronology had to be derived by correlation with other records. In some important but rare cases, the marine sediment is annually laminated and a chronology can be developed by counting the varves. The rationale for conducting high-resolution, millennial-scale studies of marine sediments has been largely driven by the need to ascertain if the abrupt climate changes recorded in the polar ice sheet records, particularly the millennial-scale Dansgaard– Oeschger (D–O) events, were evident in the ocean system (Figures 1(b) and 1(c)).

Examples of Millennial-scale Oceanographic Proxy Records In the last decade, the number of papers on millennialscale ocean variability has increased substantially. In all cases, some property of the sediment is measured and climatic variability deduced. The mostdocumented proxies for millennial changes in ocean climate and hydrography are: (1) changes in the

883

noncarbonated sand size (o2000 and 463 mm) fraction, the so-called iceberg-rafted detrital (IRD) fraction; (2) changes in the d18O of planktonic and benthic foraminifera which reflect both changes in the global ice volume, temperature, and meltwater volume; (3) changes in the d13C of marine carbonates which is a measure of productivity and water mass history and is used to trace variations in the production and circulation patterns of bottom water; (4) changes in the composition of faunas or floras which reflect the response of the biota to oceanographic changes; and (5) changes in the geochemical properties of the inorganic shell of organisms, or bulk sediment, which can be calibrated against climatic variables, such as sea surface temperature (SST). Usually, more than one of these parameters are measured, or a ratio, for example, lithics/ (lithics þ foraminifera), is used to develop a scenario of oceanic climate variability (Figure 1(b)).

Iceberg-rafted Detrital (Heinrich) Events Heinrich’s seminal paper on the occurrence, in cores off Portugal, of discrete IRD peaks during the last 60 000 years or so resulted in a wealth of data and hypotheses about ‘Heinrich events’. It is now believed that ‘armadas of icebergs’ were released on a quasi-periodic basis into the North Atlantic, with the major source area being the Hudson Strait. Hudson Strait is a large, deep trough, which drained a substantial fraction (2–4  106 km2) of the interior of the Laurentide Ice Sheet (Figure 2). In the Labrador Sea, and the areas south of Greenland and toward Europe, evidence for these armadas is dramatically visible in many parameters, but especially in the changes of the detrital carbonate content of the cores, derived from the erosion of the Paleozoic limestone that outcrops on the floors of Hudson Strait and Hudson Bay (Figure 3). These dramatic sedimentological events have been termed H-0, H-1, etc., and have the following radiocarbon ages (years ago): H-0 ¼ 10 000–11 000; H-1 ¼ 14 5007; H-2 ¼ 20 5007; H-3 ¼ 27 0007; H4 ¼ 34 0007 (Figures 1 and 3). Older H-events, H-5 and H-6, lie beyond the limits of 14C dating but have inferred ages of 48 000 and 60 000 years ago. Because of the 7error in the 14C dates, the duration of each of these IRD intervals is not well defined. Available dates indicate that they persisted for a few hundred to about a thousand years (Figure 1(b)) and have a quasi-periodic return interval averaging c. 6 ky. Studies in the North Atlantic indicate that during H-events there are coeval changes in other

(c) 2011 Elsevier Inc. All Rights Reserved.

884

MILLENNIAL-SCALE CLIMATE VARIABILITY

(a)

0

60 33 560 + 675

20 620 + 220

5

10

18 840 + 180

Carbonate

40

20 320 + 260

50

30

Mass suscept. (× 10−7)

14 530 + 90 and 14 400 + 205

HU87033-009

15

20

20

10

25

H-3? H-1

H-4

H-2

0

30 400

500

600

700

800

900

1000

1100

Depth (cm) 25

25

20

15

# > 2 mm

15

Total carb.

> 49 230

20

HU75-054

29 725 + 405

20 760 + 190 21 760 + 255

# >2 mm

10

10

5

5

0

H-2 H-3 0

H-6?

H-4? 200

Total carb. (%)

(b)

400 Depth (cm)

600

800

0

Figure 3 (a) Changes in the detrital carbonate content and magnetic susceptibility of core HU87033-009 from just north of the Hudson Strait outlet (see Figure 2). Note that the scale for mass magnetic susceptibility (10  7 m3 kg  1) is reversed because in this area the magnetite concentrations are diluted by the input of diamagnetic detrital carbonate. (b) Detrital carbonate and clasts 42 mm in HU75-054 from south of Davis Strait, northwestern North Atlantic (Figure 2). Note that the agreement between detrital carbonate events (primarily a measure of North Atlantic Heinrich events) and coarse ice-rafted detritus is far from perfect.

(c) 2011 Elsevier Inc. All Rights Reserved.

MILLENNIAL-SCALE CLIMATE VARIABILITY

parameters, with planktonic foraminiferal assemblages decreasing in numbers per gram but also becoming nearly entirely polar in composition. Benthic foraminifera show strongly decreased productivity. At the same time, the stable oxygen isotopic composition suggests an increase in surface meltwater. Controversy exists on the regional extent of H-events and the underlying mechanism(s) for discrete IRD events. Dating is clearly a critical issue as these millennial-scale events are of short duration and often date from times (o20 000 years ago) when the errors of the radiocarbon dates are measured in one to several hundred years. In addition, efforts to correlate millennial-scale oceanic H-events with abrupt events on land (or in ice cores) face the problem of correcting the marine dates for both changes in the ocean reservoir correction and 14C production rates. Thus ice core/ocean record correlations are often based on fitting the ‘wiggles’ of the proxies from the two systems. The extent of IRD events coeval with the main North Atlantic belt of iceberg-rafted materials (Figure 2) is a matter of debate. In the Nordic Seas, in the Labrador Sea, and in Baffin Bay, the IRD signal in cores is pervasive during the last glacial cycle and cannot be used per se to identify H-events (Figure 3(b)). In contrast, in the Labrador Sea, H-events are easily distinguished by the dramatic increase in detrital carbonate during these abrupt events (Figures 3(a) and 3(b)). These data are not surprising; in areas close to ice margins, the rafting of sediments in icebergs would be a persistent transport mechanism whereas the collapse or surge of a major outlet might be distinguished by an abrupt change in sediment provenance. It is also uncertain as to what extent small changes in IRD have any significance given the strong, stochastic nature of iceberg/ sediment relationships. The origins of these millennial-scale changes in ice sheet dynamics is considered to be attributable to either inherent glaciological mechanisms associated with changes at the bed of these former ice sheets, or alternatively researchers have argued that they represent climate forcing. The main issue of concern for glaciologists is how atmospheric processes could translate to the bed of large ice sheets at a rate compatible with millennial-scale variations. No plausible mechanism has been discovered. On the other hand, if coeval H-events are seen outside the North Atlantic, which has been suggested, then mechanisms are needed to transfer the process from a regional scale to a global scale. Two mechanisms have been invoked. They are not mutually exclusive. In the first scenario, it is posited that a collapse of the North American Ice Sheet

885

during an H-event causes a rapid rise in sea level of 1–5 m, which then triggers instabilities in other ice sheet margins, which have advanced toward the shelf break (in areas such as Norway, Greenland, and Iceland). As yet it is unclear whether these events around the North Atlantic affected the grounded margins of the Antarctic Ice Sheet and the West Antarctic Ice Sheet in particular. In the second scenario, the collapse of the Northern Hemisphere ice sheets, or the Laurentide (North American) Ice Sheet in particular, would result in the transport of large volumes of fresh water, in the form of icebergs and basal meltwater, to the North Atlantic. Isotopic changes in the d18O of planktonic foraminifera certainly occur during H-events (although foraminifera often disappear from the sediment during H-events, hence detailed records are sparse) and indicate that d18O values get lower, indicating the presence of a surface, low-salinity layer. If these waters are advected toward sites of vertical convection in the North Atlantic then both theory and observations indicate that this process will turn off or curtail the global thermohaline circulation. Thus the next question is whether sites beyond the normal limits of iceberg rafting and direct glacial impact show any evidence of millennial-scale climate oscillations in either the surface or deep waters.

Other Millennial-scale Proxies A variety of proxy data from deep-sea sediment drifts in the western North Atlantic, south of 351 N (Figure 2), indicate substantial, millennial-scale changes in the deep ocean (Figures 1(a) and 1(c)). Changes in the CaCO3 content of the sediment reflect the integration of carbonate production in surface waters, carbonate dissolution, resuspension and transport of continental margin sediments, and dilution with glacially and fluvially derived terrigenous sediments from the Canadian Maritimes and Eastern Canadian Arctic. Core GPC-9 was taken from a water depth of 4758 m on the Bahamas Outer Ridge. The CaCO3 record spectacularly captures oscillations in this proxy, which range over 48%, from lows during marine oxygen isotope stages (OS) 2 and 4 of B2% to peaks during OS 5 of c. 50% (Figure 1(a)). It was also shown that these oscillations were also evident in the stable isotopic composition of foraminifera, although the isotopic changes tended to lead the carbonate fluctuations by 1000 years or so during carbonate event D. Changes in the d13C may be linked with changes in the thermohaline circulation, such that a significant reduction in the formation of North Atlantic Deep

(c) 2011 Elsevier Inc. All Rights Reserved.

886

MILLENNIAL-SCALE CLIMATE VARIABILITY

Water (NADW) is indicated by low (d13C) in two benthic foraminifera genera. More recent work has concentrated on the events during OS 3 (Figure 1), as this was a period of extreme and abrupt oscillations in the Greenland ice core records. This interval includes Heinrich events 3, 4, and 5 (i.e., between B31 600 and 47 800) calendar years ago (Figure 1(b)). Variations in d13C of planktonic foraminifera along a transect from the south of Iceland (c. 601 N) to the Ceara Rise (c. 51 N) (Figure 2) have been examined. Cores from different water depths along the transect were used to reconstruct changes in water mass on millennial timescales. A critical question is the relationship between changes in the deep-sea circulation and ventilation and H-events. Is there a cause-and-effect relationship such that H-events result in a response in ocean circulation? Because these sites are outside the IRD belt of the North Atlantic, the correlation between actual IRD or carbonate-rich H-events (e.g., Figure 3) and ocean geochemical responses relies on the quality of the chronology. Within OS 3, the errors on AMS 14C dates are frequently between 7300 and 7500 years; hence the issue of a direct correlation to events lasting a mere 1000 years is of concern. However, the d13C records from the Ceara Rise indicate that cold, relatively fresh Antarctic Bottom Water (AABW, lower d13C), which underlies the warmer and saltier NADW (higher d13C), thickened by a factor of 2. The thickening of the AABW at the site began ‘several thousands of years’ prior to each H-event and extended for ‘several thousand years’ after each event. These intervals of expanded AABW were times of reduced NADW production. These intervals of reduced NADW production and associated reduction in the thermohaline circulation cannot be directly caused by ice sheet collapse and the presence of a freshwater cap over the northern North Atlantic. Further, in a core from the Bermuda Rise (near GPC-9) (Figure 2), reconstructed SST fluctuations of 2–5 1C have been shown. These SST estimates could be mapped directly onto the d18O oscillations from the ice cores at the Greenland Summit.

Millennial-scale Events in the Last 12 000 Years (The Holocene) A critical question for society is whether such rapid millennial-scale oscillations continued during the ‘Postglacial’ or Holocene period of the last 10 000 radiocarbon years (about 12 000 calendar years), and if so were they too associated with changes in the thermohaline circulation and episodic ice-rafting events? In general, data from the Greenland ice cores

indicate that climatic variability was substantially reduced during the Holocene. Temperature reconstructions from borehole and isotopic measurements indicate that temperatures at the summit of the ice sheet warmed dramatically by 16 1C at the onset of the Holocene. Over the last 10 000 years there have been temperature variations of c. 2–3 1C, and in the last 5000 years these are superimposed on a gradual, long-term temperature decrease. Based on the chronology of Holocene glacial readvances, a 25007 year cycle has been advocated. It is only in the last few years or so that researchers in the marine community have focused on producing high-resolution records from this most recent interval of Earth’s history. Cores have often been selected that have sufficiently high rates of sediment accumulation that sampling can resolve multicentury-, even multidecadal-, scale events. It has been proposed that variations in the numbers of hematitestained quartz grains at sites in the North Atlantic reflective a pervasive 1470-year ‘beat’ during the Holocene that are linked with variations in solar activity. Cycles with a similar periodicity have been reported from a variety of sedimentary archives including silt size (as a measure of current speed), sediment color, and the amount and composition of the sand fraction. Although the B1470-year cycles have been attributed to iceberg-rafting events, their magnitude in the records is not remotely at the scale of the Heinrich events. Indeed the ‘pervasive’ nature of this ice-rafting signal has been questioned on the basis of quantitative studies of quartz and plagioclase weight percentage data at sites from North Iceland and the Vorring Plateau.

Discussion: Importance and Mechanisms In the 1970s and 1980s, a common view of the global climate system on scales from 1 to 106 years was that there were systematic changes associated with the Milankovitch orbital variations, which effected insolation. Evaluation of changes in the global ice volume indicated dominant periodicities of 41 000 and c. 20 000 years, and in the last 0.7 million years a 100 000-year cycle became evident. At higher frequencies, the spectra of climatic variability was essentially blank between 20 000-year and the 22-year sunspot cycle. This absence of recurring periodicities suggested that global climate change within this range had no obvious or repetitive forcing function. The advent of the successful icecoring programs, especially the Greenland Ice Sheet boreholes, and the subsequent development of

(c) 2011 Elsevier Inc. All Rights Reserved.

MILLENNIAL-SCALE CLIMATE VARIABILITY

well-dated, multiproxy records of the atmosphere, led to a search for recurring frequencies between 1/20 000 and 1/22 cycles per year. This analysis suggested that there is a 6000–7000-year periodicity, which is approximately the same as the interval between the successive Heinrich events. Because of the largely unknown errors connected with the value of the ocean reservoir correction, and the conversion from radiocarbon years to sidereal years, the spacing between H-0 and H-4 was c. 5000, 8000, 7000, and 8000 years with uncertainties of several hundred years. However, each H-cycle was composed of several higher-frequency events, the D–O cycles, which had a recurrence interval of around 2000 years. In detailed records from cores in the North Atlantic, a series of D–O events are bundled with bounding H-events. These ‘packages’ show an overall saw-tooth decrease in warm surface water indicators over the course of a cycle, with a final abrupt and extreme minimum, which marks the onset of an H-event. This was rapidly followed by a dramatic rise in the warm-water proxies. Broecker referred to this pattern as a ‘Bond cycle’. The prevailing wisdom calls for these oscillations to be associated with changes in the thermohaline circulation, but there is the ‘chicken or egg’ syndrome. Changes in the thermohaline circulation are usually associated with changes in the saltiness of the surface waters. Thus the dramatic collapse of a large ice sheet, and the subsequent export of fresh water in the form of meltwater plumes and icebergs, is a legitimate mechanism for curtailing convective overturning at sites in the northern North Atlantic. An important question, presently unanswered, then becomes that how these changes are transmitted rapidly and at the millennial scale, synchronously through the atmosphere (to account for the observed rapid changes in ice sheet isotopes and precipitation chemistry), and within the oceans. There are several lines of evidence to suggest that one way in which the ocean circulation compensates for changes in the ‘deep’ thermohaline circulation is by the increased production of what has been termed ‘glacial intermediate water’. There is a dramatic decrease in the variability of most climate proxies in ocean sediments over the last 11 000 cal years. H- and D–O events characterize marine oxygen isotope stages 2, 3, and 4 (Figure 1) when the Earth was marked by extensive glaciation and sea levels were lowered between 40 and 110 m. High-resolution sampling of marine cores from deep ocean basins and continental margins which span one to several thousand of years indicate that changes in oceanography have taken place at millennial timescales over the present interglacial period.

887

A key question is whether all proxies will record the same oscillations? The notion that there are thresholds in the climate system suggests that not all events may be archived in marine sedimentary records. An example is the Holocene ice rafting of sediments. In some parts of the world oceans, measurable quantities of sediment could be rafting to ice-distal locations on and in sea ice. However, the sediment burden in sea ice is relatively light and, furthermore, the thickness of a typical multiyear pan of sea ice is measured in meters to a few tens of meters versus hundreds of meters for true icebergs. Hence, melting and erosion of sea ice results in a limited transport of sea-ice-rafted sediment when compared to iceberg rafting. Most of the North Atlantic’s margins and offshore basins have seen a massive reduction in IRD following the retreat and disappearance of late Quaternary ice sheets (Figures 2 and 4). In today’s world (Figure 4), the distribution of IRD-rich sediments is primarily restricted to the Greenland shelves and the eastern Canadian Arctic (Baffin Island and Labrador) margins, therefore even small traces of sand-size minerals distal to these areas may indicate intervals of iceberg rafting. However, the threshold in question is the presence of tidewater calving glaciers in the Greenland fiords. Observations from Greenland indicate that the ice sheet was well behind its present margin by 6000 years ago and probably by 7000–8000 years ago. Even on the East Greenland margin, which today is ‘well traveled’ by icebergs, sediments deposited between 6000 and 8000 years ago are largely devoid of IRD, whereas between 5000 and 6000 years ago the iceberg rafting of coarse, clastic sediments becomes a pervasive depositional process. Modern observations, however, do indicate that the production of Intermediate Atlantic Water is sensitive to modern-day atmospheric and oceanographic processes. The Great Salinity Anomaly of the late 1960s and early 1970s (depending on location) was the result of an excess freshwater output from the Arctic Ocean (as sea ice) (Figure 4). This pool of relatively fresh, surface water, caused dramatic cooling of the water column off North Iceland (by 5 1C), and as it moved into the Labrador Sea it caused a cessation in convective overturning. This resulted in a temperature drop of B2 1C on the west Greenland and Canadian margins. Because another salinity anomaly occurred in the early 1980s, this time sourced from the Hudson Bay/ Labrador Sea region, there certainly appear to be mechanisms within the present climate system that are capable of generating rather severe and abrupt oceanographic changes. The question is whether the processes responsible for multidecadal climatic

(c) 2011 Elsevier Inc. All Rights Reserved.

888

MILLENNIAL-SCALE CLIMATE VARIABILITY

Arctic Ocean Pacific Ocean

Eurasian ice sheets

Freshwater sources and transport Production of North Atlantic deep and intermediate water

Major ice sheet

Figure 4 Major sources of salinity events and location of convection areas in the North Atlantic during the present ‘interglacial’ world. Sources include the influx of freswater from the Pacific Ocean via the Bering Strait, river runoff into the Arctic Ocean, and the export of sea ice from the Arctic Ocean via Fram Strait and the Canadian High Arctic channels. Tidewater calving margins around the Greenland Ice Sheet lead to the calving of about 350 km3 of ice per year.

variability can be scaled up, so that the processes persist and produce millennial oscillations in ocean records.

Conclusions Millennial-scale changes have become an accepted reality in the climate system. Initial research

concentrated on the massive changes associated with the discharge of sediments and water into the North Atlantic Ocean during the last glacial cycle (marine oxygen isotope stages 2–5) (Figures 1–3). However, high-resolution studies of our deglacial world (Figure 4) appear to indicate that similarly spaced but subdued events persist but with very different boundary conditions. A number of publications have

(c) 2011 Elsevier Inc. All Rights Reserved.

MILLENNIAL-SCALE CLIMATE VARIABILITY

also demonstrated that millennial-scale changes in various proxy records are a feature of ocean sediments over at least the last 500 000 years. The work from tropical and subtropical sites (Figure 2) indicates that Heinrich events have manifestations in ocean reconstructions, which belie a simple association with ice sheet instability and collapse. It is far from clear how oceanographic and atmospheric changes are transmitted to the bed of large ice streams, and there is indeed disagreement as to whether the collapse of Northern Hemisphere ice sheets (Figure 2) was regionally coeval or whether the collapses are linked temporally by a mechanism, such as rapid changes in relative sea level. It has, however, been observed that the routing of fresh water (Figure 4) can have dramatic effects, even in the present world, and the key may well lie in a better understanding of the role of the ocean thermohaline circulation system in the global climate system.

See also Cenozoic Climate – Oxygen Isotope Evidence. Ocean Circulation: Meridional Overturning Circulation.

Further Reading Alley RB, Clark PU, Keigwin LD, and Webb RS (1999) Making sense of millennial-scale climate change. In: Clark PU, Webb RS, and Keigwin LD (eds.) Mechanisms of Global Climate Change at Millennial Time Scales, pp. 301--312. Washington, DC: American Geophysical Union. Anderson DM (2001) Attenuation of millennial-scale events by bioturbation in marine sediments. Paleoceanography 16: 352--357. Andrews JT (1998) Abrupt changes (Heinrich events) in late Quaternary North Atlantic marine environments: A history and review of data and concepts. Journal of Quaternary Science 13: 3--16. Andrews JT, Hardardottir J, Stoner JS, Mann ME, Kristjansdottir GB, and Koc N (2003) Decadal to millennial-scale periodicities in North Iceland shelf sediments over the last 12,000 cal yrs: Long-term North Atlantic oceanographic variability and solar forcing. Earth and Planetary Science Letters 210: 453--465. Bond G, Kromer B, Beer J, et al. (2001) Persistent solar influence on North Atlantic climate during the Holocene. Science 294: 2130--2136. Bond GC and Lotti R (1995) Iceberg discharges into the North Atlantic on millennial time scales during the last glaciation. Science 267: 1005--1009.

889

Broecker WS (1997) Thermohaline circulation, the Achilles heel of our climate system will man-made CO2 upset the current balance? Science 278: 1582--1588. Broecker WS, Bond G, McManus J, Klas M, and Clark E (1992) Origin of the northern Atlantic’s Heinrich events. Climatic Dynamics 6: 265--273. Clarke GKC, Marshall SJ, Hillaire-Marcel C, Bilodaeu G, and Veiga-Pires CA (1999) Glaciological perspective on Heinrich events. In: Clark PU, Webb RS, and Keigwin LD (eds.) Mechanisms of Global Climate Change at Millennial Time Scales, pp. 243--262. Washington, DC: American Geophysical Union. Clemens SC (2005) Millennial-band climate spectrum resolved and linked to centennial-scale solar cycles. Quaternary Science Reviews 24: 521--531. Curry WB, Marchitto TM, McManus JF, Oppo DW, and Laarkamp KL (1999) Millennial-scale changes in the ventilation of the thermocline, intermediate, and deep waters of the glacial North Atlantic. In: Clark PU, Webb RS, and Keigwin LD (eds.) Mechanisms of Global Climate Change at Millennial Time Scales, pp. 59--76. Washington, DC: American Geophysical Union. Heinrich H (1988) Origin and consequences of cyclic ice rafting in the Northeast Atlantic Ocean during the past 130 000 years. Quaternary Research 29: 143--152. Hughen KA, Overpeck JT, and Lehman SJ (1998) Deglacial changes in ocean circulation from an extended radiocarbon calibration. Nature 391: 65--68. Keigwin LD and Jones GA (1994) Western North Atlantic evidence for millennial-scale changes in ocean circulation and climate. Journal of Geophysical Research 99: 12397--12410. Lowell TV, Heusser CJ, and Andersen BG (1995) Interhemispheric correlation of late Pleistocene glacial events. Science 269: 1541--1549. McManus JR, Oppo DW, and Cullen JL (1999) A 0.5 million-year record of millennial-scale climate variability in the North Atlantic. Science 283: 971--975. Moros M, Andrews JT, Eberl DD, and Jansen E (2006) The Holocene history of drift ice in the northern North Atlantic: Evidence for different spatial and temporal modes. Palaeoceanography 21 (doi:10.1029/ 2005PA001214). Sachs JP and Lehman SJ (1999) Subtropical North Atlantic temperatures 60 000 to 30 000 years ago. Science 286: 756--759. Thomas E, Booth L, Maslin M, and Shackelton NJ (1995) Northeastern Atlantic benthic foraminifera during the last 45 000 years: Changes in productivity seen from the bottom up. Paleoceanography 10: 545--562.

Relevant Websites http://www.ncdc.noaa.gov/paleo – NOAA Paleoclimatology Program, NOAA.

(c) 2011 Elsevier Inc. All Rights Reserved.

MINERAL EXTRACTION, AUTHIGENIC MINERALS J. C. Wiltshire, University of Hawaii, Manoa, Honolulu, HA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1821–1830, & 2001, Elsevier Ltd.

Introduction The extraction of marine mineral resources represents a worldwide industry of just under two billion dollars per year. There are approximately a dozen general types of marine mineral commodities (depending on how they are classified), about half of which are presently being extracted successfully from the ocean. Those being extracted include sand, coral, gravel and shell for aggregate, cement manufacture and beach replenishment; magnesium for chemicals and metal; salt; sulfur largely for sulfuric acid; placer deposits for diamonds, tin, gold, and heavy minerals. Deposits which have generated continuing interest because of their potential economic interest but which are not presently mined include manganese nodules and crusts, polymetallic sulfides, phosphorites, and methane hydrates. Authigenic minerals are those formed in place by chemical and biochemical processes. This contrasts with detrital minerals which have been fragmented from an existing rock or geologic formation and accumulated in their present position usually by erosion and sediment transport. The detrital minerals – sand, gravel, clay, shell, diamonds, placer gold, and heavy mineral beach sands – are presently extracted commercially in shallow water. The economically interesting authigenic mineral deposits tend to be found in more than 1000 m of water and have not yet been commercially extracted. Nonetheless, between 1970 and 2000 on the order of one billion US dollars was spent collectively worldwide on studies and tests to recover five authigenic minerals. These minerals are: manganese nodules, manganese crusts, metalliferous sulfide muds, massive consolidated sulfides, and phosphorites. This article will focus on the extraction of these five mineral types. Descriptions of the formation, geology, geochemistry, and associated microbiology of these deposits are presented elsewhere in this work. As a brief generalization, manganese nodules are black, golf ball to potato sized concretions of ferromanganese oxide sitting on the deep seafloor at depths of 4000–6000 m. They contain potentially economic

890

concentrations of copper, nickel, cobalt, and manganese, and lower concentrations of titanium and molybdenum. Manganese crusts are a flat layered version of manganese oxides found on the tops and sides of seamounts with the highest metal concentrations in water depths of 800–2400 m. They are a potential source of cobalt, nickel, manganese, rare earth elements and perhaps platinum and phosphate. Polymetallic sulfides also come in two forms: metalliferous muds and massive consolidated sulfides. The sulfides contain potentially economic concentrations of gold, silver, copper, lead, zinc, and lesser amounts of cobalt and cadmium. The metalliferous muds are unconsolidated sediments (muds) found at seafloor spreading centers and volcanically active seafloor sites. The metals have been concentrated in the muds by hydrothermal processes operating at and below the seafloor. The best explored site of these metalliferous muds is in the hydrothermally active springs in the central deeps of the Red Sea. By contrast, the massive consolidated sulfides are associated with chimney and mound deposits found at the sites of ‘black smokers’, hydrothermal vents on the seafloor. The sulfides deposited at these sites have been concentrated by hot seawater percolating through the seafloor and being expelled onto the seabottom at the vents. When in contact with the cold ambient sea water, the hydrothermally heated water drops its mineral riches as sulfide-rich precipitates, forming the sulfide chimneys and mounds. The final authigenic mineral of economic interest is phosphorite, used primarily for phosphate fertilizer. The seabed phosphorites are found as nodules, crusts, irregular masses, pellets, and conglomerates on continental shelves and on the tops of seamounts and rises. Their distribution is widespread throughout the tropical and temperate oceans, although they are preferentially found in areas of oceanic upwelling and high organic productivity. Whichever of these five mineral types is the target of a mining operation, following regional exploration the process of extracting the commodity of value has seven distinct steps. These steps are: (1) survey the mine site, (2) lease, (3) pick up the mineral, (4) lift and transport, (5) process, (6) refine and sell the metal, and (7) remediate the environmental damage. These steps will be considered in turn (Figures 1–7). Naturally, there are differences for each mineral type as well as a range of possible processes that can be used. While many of these

(c) 2011 Elsevier Inc. All Rights Reserved.

MINERAL EXTRACTION, AUTHIGENIC MINERALS

891

Mining system in general

Discharge pipe

Rigid pipe

Compressed air line (option I)

Hydraulic pumps (option II) Air injection (option I) Rigid pipe Flexible pipe

Collector

Nodules

Towed

S

Collector vehicle types Self-propelled

Tracked

Archimedes screw

Sled

Figure 1 Generalized deep-sea mining system component diagram. This diagram illustrates the proposed collection and lift components which would apply to a wide variety of mineral systems. Note that a considerable variety of collector types are possible on the bottom, including tracked robotic miners, Archimedes screw-driven vehicles, or towed sleds. Both airlift and pumped systems are illustrated as possible lift mechanisms. (Reproduced with permission from Thiel et al., 1998.)

processes have been tested and many are used in traditional terrestrial minerals operations, to date there is no commercially viable full-scale deep-sea authigenic mineral extraction operation.

Survey The first step in minerals development is to find an economic mine site. This is found by surveying and mineral sampling. A great deal of mineral sampling has already been done over the last 40 years throughout the world’s ocean. These data are available for initial planning purposes. Following a detailed literature review the prospective ocean miner would send out a research vessel to sample extensively in the areas under consideration. New acoustical techniques can be calibrated to show certain kinds of bottom cover, including the density of manganese nodule cover. This is one way to rapidly survey the bottom to highlight areas with potentially economic accumulation of authigenic minerals.

Significant advances in marine electronics, navigation, and autonomous underwater vehicles (AUVs) are being brought together. New ‘chirp’ sonars which transmit a long pulse of sound in which the frequency of the transmitted pulse changes linearly with time give high resolution and long-range seafloor and sub-bottom imagery. Navigation based on the satellite global positioning system (GPS) can now give accurate underwater positions (r1m) when linked to an acoustic relay. This level of survey equipment is now available on underwater autonomous vehicles, meaning that the cost of a ship is not necessarily an impediment. Sampling for metal concentrations follows the initial surveys. Sampling may be from a ship, a remotely operated vehicle (ROV), or a submersible. Sampling is likely to begin with dredges, progress to some kind of coring, and finish with carefully oriented drilled samples giving a three-dimensional picture of the ore distribution. These data, after chemical analysis of the contained metals, will give

(c) 2011 Elsevier Inc. All Rights Reserved.

892

MINERAL EXTRACTION, AUTHIGENIC MINERALS

Continuous line bucket (CLB) system

Autonomous robot submersible shuttle (PLA) system

Figure 2 In contrast to the more conventional marine mining systems illustrated in Figure 1, two other potential systems are the continuous line bucket system and the autonomous robot submersible shuttle. The continuous line bucket is a series of dredge buckets on a line dragged over the mineral deposit. The submersible shuttle system is a theoretical system using a series of robotic transport submersibles to carry ore from a collection point on the bottom to a waiting surface ship. (Reproduced with permission from Thiel et al., 1998.)

grade and tonnage information. The grade and tonnage estimates of the deposit will be entered into a financial model to determine whether it is economically profitable to mine a given deposit. In actual fact this process is iterative. A financial model will be used to indicate the type of mineral deposit which must be found. This will narrow the search area. As new data are forthcoming, increasing levels of survey sophistication will follow, assuming that the data continue to indicate an economic mining possibility. In its simplest form the financial model looks at sales of the metals derived from the mine, compared to the total cost of all the operations required to obtain the minerals and contained metals. If the projected sales are greater than the costs by an amount sufficient to allow a profit, typically on the order of a minimum of 20% after taxes, then the mineral deposit is economic. If not, costs must be decreased, sales increased or another more attractive deposit must be found. In the end, a deposit survey will be developed of sufficient detail to allow the production of a mining plan for the development of the property. Normally such a mining plan would need to be approved by the government authority granting mineral leases before actual mining could commence. At some point during the survey and evaluation process, the mining company needs to obtain a lease on the mineral claim in

question. Usually this is done very early in the process, after the first broad area wide surveys.

Lease Ocean floor minerals are not owned by a mining company. If the minerals are within the exclusive economic zone (EEZ) of a country, they are the property of that country’s government. EEZs are normally 200 nautical miles off the coast of a given country, but in the case of a very broad continental shelf may be extended up to 350 miles offshore with recognition of the appropriate United Nations boundary commission. Beyond the EEZ, the ownership of minerals rests with the ‘common heritage of mankind’ and is administered for that purpose by the United Nations International Seabed Authority in Kingston, Jamaica. Both national regimes and the International Seabed Authority will lease seabed minerals to bone fide mining groups after the payment of fees and the arrangement for filing of mining and exploration plans, environmental impact statements, and remediation plans. The specific details of the requirements vary with the size of the proposed operation, the mineral sought, and the regime under which the application is made. In many countries, the offshore mining laws

(c) 2011 Elsevier Inc. All Rights Reserved.

MINERAL EXTRACTION, AUTHIGENIC MINERALS

893

taken. A typical deep water rental for an oil and gas lease might be $25 per acre per year with a royalty payment equal to 12.5% of the oil extracted. Hard mineral lease rentals and royalties could be expected to be lower, as the commodity is less valuable and the demand for access to offshore mineral resources is much less than for offshore oil.

Mineral Pick-up Lift pipe

Dump valve

Pick-up mechanism Figure 3 An engineering design for a proposed manganese crust mining system. (Reproduced with permission from State of Hawaii (1987). Mining Development Scenario for Cobalt-Rich Manganese Crusts in the Exclusive Economic Zones of the Hawaiian Archipelago and Johnston Island. Honolulu, Hawaii, Ocean Resources Branch.)

are modeled on the legislation which governs offshore oil development. This is not surprising as the issues are similar and worldwide the annual value of the offshore oil industry is in excess of $100 billion, whereas the offshore mining industry will not be more than a few percent of this value for the foreseeable future. Leasing arrangements vary from country to country. In the USA, offshore leasing of both hard minerals and oil and gas in the federally controlled waters of the EEZ (normally more than 3 miles offshore) falls under the Outer Continental Shelf Lands Act and is administered by the US Department of the Interior. This act requires a competitive lease sale of offshore tracts of land. The company to whom a lease is awarded is the one which offers the highest payment at a sealed bid auction. In addition to this payment, known as the bonus bid, there is also an annual per acre rental and a royalty payment which would be a percentage of the value of the mineral

Once a deposit has been characterized and leased, a mining plan is drawn up. This plan will depend on the mineral to be mined. In general, several steps must be taken to mine. The first is separating the mineral from the bottom. In the case of polymetallic sulfide, manganese crust or underlying phosphorite, a cutting operation is involved. In the case of phosphorite nodules, manganese nodules, or sulfide muds there is solely a pick-up operation. Cutting requires specialized cutting heads, often these are simply rotating drums with teeth (Figures 3–5). Such cutters have been developed for the dredging industry as well as the underground coal industry on land. The size, angle, and spacing of the teeth on a cutter are dependent on the rate of cutting desired and the size to which particles are to be broken. The overall mineral pickup rate is determined by the necessary rate of throughput at the mineral processing plant. This can be worked back to pick-up rate on the bottom. Engineering judgment then dictates whether this is best achieved with larger numbers of smaller cutters or a lower number of larger cutters. This may translate into multiple machines operating on the bottom and feeding one lift system. When the mineral is broken into sufficiently small pieces, these must then be collected for lifting. This is usually a scooping or vacuuming operation. Scooping is accomplished by blades of various shapes. Vacuuming is usually the result of a powerful airlift or pump farther up the line. One system tested by the Ocean Minerals Co. for picking up manganese nodules involved an Archimedes screw-driven robotic miner, which had two pontoons. A flange in screw-shaped spiral was welded onto each pontoon. The screws both served to drive the vehicle forward as well as pick up the nodules which were sitting in the mud it passed over. There may be a sieving or grinding step between mineral pick-up and lift. This serves two functions; the sieving gets rid of unwanted bottom sediment that may have become entrained in the ore; the grinding ensures that the particle size range going up the lift pipe is in the correct range to get optimum lift without clogging the pipe.

(c) 2011 Elsevier Inc. All Rights Reserved.

894

MINERAL EXTRACTION, AUTHIGENIC MINERALS

Separator Lift pipe

Diffusers

Power tracks

Hydraulic dredge heads

Cutter heads

Figure 4 A robotic miner designed to rip up and lift attached flat lying bottom deposits such as manganese crusts. (Reproduced with permission from State of Hawaii (1987) Mining Development Scenario for Cobalt-Rich Manganese Crusts in the Exclusive Economic Zones of the Hawaiian Archipelago and Johnston Island. Honolulu, Hawaii, Ocean Resources Branch.)

Pivot arm

Cutter teeth

Figure 5 Cutter head design for a bottom miner ripping attached ores such as polymetallic sulfides or manganese crusts. (Reproduced with permission from State of Hawaii (1987) Mining Development Scenario for Cobalt-Rich Manganese Crusts in the Exclusive Economic Zones of the Hawaiian Archipelago and Johnston Island. Honolulu, Hawaii, Ocean Resources Branch.)

Lift and Transport Once the mineral has been cut off or picked up from the bottom it must be lifted to the surface. Over the

years a number of systems have been suggested and tested for this purpose. The most successful of these tests have involved bringing the minerals to the surface in a seawater slurry in a steel pipe. This pipe would typically be 30–50 cm in diameter, hence it would be similar if not identical to pipe used in offshore oil drilling operations. Two methods have been tested for bringing the slurry up the pipe, i.e. pumps and airlift. Airlift is a commonly used technique in shallow-water dredging. Compressed air is introduced into the pipe, typically about a third of the way from the top (Figure 6). As the air expands moving upward in the pipe it draws the mineral slurry behind it. This works extremely well over lifts of a few hundred meters. It also works over lifts in the deep sea of 6000 m, for example on manganese nodules, except that it is much more difficult to control. Hydraulically driven pump systems along the pipe’s length have also been tested to full ocean depth. They also work and although easier to control may ultimately be less efficient lifting systems than an airlift. Several other lift systems are possible (Figures 2–5). The most notable of these is the continuous line bucket, which is a series of buckets on a line which is continuously dragged over the mineral deposit. This system has been successfully tested, although it suffers the disadvantages of lack of control on the bottom, potentially wider spread environmental

(c) 2011 Elsevier Inc. All Rights Reserved.

MINERAL EXTRACTION, AUTHIGENIC MINERALS

895

Air compressor 3 Phase separator ~ 6 atm

Airlift section

High pressure supply (~ 80 atm)

600 m 800 m Air supply pipe

Air injection 2400 m Lift pipe

Flexible hose Dump valve

Figure 6 Generalized design of an airlift system to lift ground minerals from the ocean bottom to the surface. (Reproduced with permission from State of Hawaii (1987) Mining Development Scenario for Cobalt-Rich Manganese Crusts in the Exclusive Economic Zones of the Hawaiian Archipelago and Johnston Island. Honolulu, Hawaii, Ocean Resources Branch.)

damage, and the possibility of the rope entangling on itself. A system proposed but not tested involves a series of ore-carrying robotic submersible shuttles. These shuttles would use waste mineral tailings as ballast to descend to the bottom, where they would exchange the ballast for a load of ore. Adjusting their buoyancy they would then rise to the surface, offload the ore onto an ore carrier, reload waste, and return to the bottom. This is potentially a very elegant system in that it handles both tailings waste disposal and mineral lift at the same time, each with the expenditure of very little energy. Once the ore is on the surface it must be taken to a processing plant. This processing plant can be an existing plant at a site on land, a purpose-built plant near a harbor to process marine minerals, a large

offshore floating platform moored near the mine site on which a plant is built, or a ship converted to mineral processing. The latter two could be right at the mine site. The minerals would be processed as soon as they were lifted to the surface. For a shore-based processing plant a fleet of transport ships or barges would be required to move the lifted ore from the mine site to the plant. This could be a tug and barge operation or a series of dedicated ore carrier vessels.

Processing Once the minerals have been mined and transported to the processing location they must then be treated to remove the metals of interest. There are two basic

(c) 2011 Elsevier Inc. All Rights Reserved.

896

MINERAL EXTRACTION, AUTHIGENIC MINERALS

Products

Mining vessels

Nodule transports

Marine terminal

Onshore transport

Processing plants Tailings

Reagents, fuels, & power Figure 7 Components involved in a marine minerals extraction operation, each of which has an environmental effect and must be considered in the context of the whole operation. (Reproduced with permission from State of Hawaii (1981) The Feasibility and Potential Impact of Manganese Nodule Processing in the Puna and Kohala Districts of Hawaii. Honolulu, Hawaii, Ocean Resources Branch.)

ways to do this: smelting and leaching (pyrometallurgy and hydrometallurgy). Once a basic scenario of either reducing the mineral with acid or melting has been decided upon, there are a number of possible steps to partially separate out unwanted fractions of the mineral. In order to be economically viable, a mineral processing plant needs to process large tonnages of material. In the case of manganese nodules this might be 3 million tonnes y 1. In order to process these large tonnages economically it is important to separate out as much of the waste mineral material as early as possible in the process. This is done most commonly by magnetic separation, froth flotation, or a density separation such as heavy media separation. In an effort to make these processes more efficient the incoming ore would normally be washed to remove salt and ground to decrease the particle size and increase surface area. Magnetic separation separates magnetic and paramagnetic fractions from nonmagnetic fractions. This technique can be done either wet or dry. Dry magnetic separation involves passing the ground ore through a strong magnetic field underlying a conveyor belt. The magnetic and paramagnetic fraction stay on the conveyor belt while the nonmagnetic fractions drop off as the belt goes over a descending roller. Manganese, cobalt, and nickel are paramagnetic, so this is a good separation technique for manganese nodules and crusts. Wet magnetic separation involves passing a slurry of ore over a series of magnetized steel balls. The magnetic and paramagnetic materials stick to the balls and the nonmagnetic fraction flows through. The level of separation depends on the strength of the magnetic field. Usually this is done with an electromagnet so that the field can be turned on and off. Froth flotation relies on differences in the surface chemistry of the ore and waste grains. Sulfide ores in particular are more readily wet with oil than water.

Grains of clay, sand, and other seabed detritus have the reverse tendency. Typically something like pine oil might be used to wet the mineral grains. The oilwet ore is then introduced into a tank with an agent that produces bubbles, a commonly used chemical for this is sodium lauryl sulfate. The bubbles tend to stick to the oil-wet grains, which float them to the surface where they can be taken off in the froth at the top of the tank. The waste material is allowed to settle in the tank and removed at the bottom. The separated ore is then washed to remove the chemicals. The chemicals are recycled if possible. Another technique commonly used in mineral beneficiation is density separation. This can be as simple as panning for gold. In a simple panning operation or more complex sluice boxes or jigging tables, the heavy mineral (the gold) stays at the bottom, whereas the lighter waste materials are washed out in the swirling motion. A more sophisticated version of the same idea is to use a very dense liquid or colloidal suspension to allow less dense material to float on top and more dense material to sink. The density of the medium is adjusted to the density of the particular ore and waste to be separated. In general, heavy media suspensions, such as extremely finely ground ferro-silicon, are superior to traditional heavy liquids such as tetrabromoethane, which have a high toxicity. Following mineral beneficiation, which removes the non-ore components, the ore itself must be broken down into its component value metals. Marine minerals can contain up to half a dozen economically desirable metals. In order to separate these metals individually, a reduction process must occur to break the chemical framework of the ore. This will either involve smelting or leaching. In a smelting operation the whole mineral structure is melted and the various metals are taken off in fractions of different densities which float on each other. Normally a series of

(c) 2011 Elsevier Inc. All Rights Reserved.

MINERAL EXTRACTION, AUTHIGENIC MINERALS

fluxing agents is used. Clean separation in the case of a multi-metal ore can be a very exacting process. For highly complex polymetallic ores such as manganese nodules or crusts many experts favor a leaching approach over smelting in order to facilitate high purity separation of the metals. However, both leaching and smelting processes have been tried on a variety of marine minerals and both types of process have been proved technically viable. The most common leach agent is sulfuric acid as it is cheap, efficient, and readily available. Other acids, hydrogen peroxide, and even ammonia are also used as leaching agents. To reduce leaching times, temperature and pressure are raised in the leaching vessel. The leaching process breaks up the mineral structure and dissolves the contained metals into an acidic solution known as a ‘pregnant liquor’. The metals become positive ions in solution. These individual metal ions must be separated out of the pregnant liquor and plated out as an elemental metal. This is done primarily in three processes: solvent extraction, ion exchange, and electrowinning. These may initially be done with the primary processing of the ore; however, the final phase of these techniques will be done in a dedicated metal refinery to achieve metal purity 499%. Solvent extraction relies on an organic solvent which is optimized to select one metal ion. This organic extractant is immiscible with the aqueous pregnant liquor. They are stirred together in a mixer forming an emulsion much like oil stirred into water. This emulsion has a very high contact area. The organic extractant takes the desired metal ion out of the aqueous phase and concentrates it in the organic phase. The aqueous and organic phases are then separated by density (usually the organic phase will float on the aqueous phase). The metal ion is then stripped out of the organic phase by an acidic solution and sent to an electrowinning step. Instead of doing this extraction in a liquid as in the case of solvent extraction, it is possible to run the pregnant liquor over a series of ion-exchange beads in a column. The ionic beads have the property of collecting or releasing a given metallic ion, e.g. nickel, at a given pH. Therefore by adjusting the pH of the liquid flowing over the column it is possible to remove the nickel, for example, from the pregnant liquor, transferring it to its own tank, then remove copper or cobalt, etc. It is possible to use both solvent extraction and ion-exchange column steps in a particularly complex metal separation process or to achieve very high metal purities. Following solvent extraction or ion exchange, the various metal ions have been separated from each other and each is in its own acidic solution. The

897

standard concentration technique used to remove these metal ions from solution is electrowinning. The metal ions are positively charged in solution. They will be attracted to the negatively charged plate in a cell. A small current is set up between two or more plates in a tank. The positive metal ions plate out on the negatively charged plate. Depending on the purity of the metal plated out and how strongly it is attached, either the metal is scraped off the plate and sold as powdered metal or the entire plate is sold as metal cathode, e.g. cathode cobalt.

Refining and Metal Sales Each of the metals produced by an offshore mining operation must be refined to meet highly exacting standards. The metals are sold on various exchanges in bulk lots. Silver, copper, nickel, lead, and zinc are largely sold on the London Metal Exchange. Platinum, silver, and gold are sold on the New York Mercantile Exchange. Both of these exchanges maintain informative internet websites with the current details and requirements for metals transactions (www.lme.co.uk and www.nymex.com). Cobalt, manganese and phosphate are sold through brokers or by direct contract between a mining company and end-users (e.g. a steel mill). Metal sales are by contract. The contracts specify the purity of the metal, its form (e.g. powder, pellets or 2 inch squares), the amount, place, and date that the metal must be delivered. Standard contracts allow for metal delivery as much as 27 months in the future (in the case of the London Metal Exchange). This allows a major futures market in metals and considerable speculation. This speculation allows metal producers to lock in a future price of which they are certain. In reality, only a few percent of the contracts written for future metal deliveries result in actual metal deliveries. The vast majority are traded among speculators over the time between the initial contract settlement and the metal delivery date. The price for a metal is highly dependent on its purity. Often metals are sold at several different grades. For example, cobalt is typically sold at a guaranteed purity of 99.8%. It may also be purchased at a discounted price for 99.3% purity and at a premium for 99.95%. In order to achieve these grades considerable refining takes place. Often this is at a facility which is removed from the original mine site or processing plant. Metal refineries may be associated with the manufacturing of the final consumer endproduct. For example, a copper refinery will often take lower grade copper metal powder or even scrap and produce copper pipe, wire, cookware, or copper plate. Most modern metal fabricators rely on electric furnaces of some form to cast the final metal product.

(c) 2011 Elsevier Inc. All Rights Reserved.

898

MINERAL EXTRACTION, AUTHIGENIC MINERALS

One of the techniques commonly used at refineries is known as ‘zone refining’ whereby a small segment of a piece of metal in a tube is progressively melted while progressing through a slowly moving electric furnace coil. The impurities are driven forward in the liquid phase with the zone of melting. The purified metal resolidifies at the trailing edge of the melting zone. This technique has been used to reduce impurities to the parts per billion level.

Remediation of the Mine Site Once mining has taken place both the mine site and any processing site must be remediated (Figure 7). Considerable scientific work took place in the 1980s and 1990s looking at the rate that the ocean bottom recovers after being scraped in a mining operation. While it is clear that recovery does take place and is slow, it is still unclear how many years are involved. A period of several years to several decades appears likely for natural recolonization of an underwater mined site. It also appears that relatively little can be done to enhance this process. Once all mining equipment is removed from the site, nature is best left to her own processes. An independent yet perhaps even more important issue is the way the waste products are handled after mineral processing. There are both liquid and solid wastes. The most advanced of a number of clean-up scenarios for the discharged liquids is to use some form of artificial ponds or wetlands, most often involving cattails (Typha) and peat moss (Sphagnum), the two species shown to be most adept at wastewater clean-up. Typically the wastewater will circulate over several limestone beds and through various artificial wetlands rich with these and related species. At the end of the circulation a certain amount of cleaned water is lost to ground water and the rest is usually sufficiently cleaned to dispose in a natural stream, lake, river, or ocean. The larger and as yet less satisfactorily engineered problem is with solid waste. In fact, recent environmental work on manganese crusts has shown that 75% of the environmental problems associated with marine ferromanganese operations will be with the processing phase of the operation, particularly tailings disposal. Traditionally, mine tailings are

dumped in a tailings pond and left. Current work with manganese tailings has shown them to be a resource of considerable value in their own right. Tailings have applications in a range of building materials as well as in agriculture. Manganese tailings have been shown to be a useful additive as a fine-grained aggregate in concrete, to which they impart higher compressive strength, greater density, and reduced porosity. These tailings serve as an excellent filler for certain classes of resin-cast solid surfaces, tiles, asphalt, rubber, and plastics, as well as having applications in coatings and ceramics. Agricultural experiments extending over 2 years have documented that tailings mixed into the soil can significantly stimulate the growth of commercial hardwood trees and at least half a dozen other plant species. Finding beneficial uses for tailings is an important new direction in the sustainable environmental management of mineral waste.

See also Authigenic Deposits. Hydrothermal Vent Deposits. Manganese Nodules. Mid-Ocean Ridge Geochemistry and Petrology. Remotely Operated Vehicles (ROVs).

Further Reading Cronan DS (1999) Handbook of Marine Mineral Deposits. Boca Raton: CRC Press. Cronan DS (1980) Underwater Minerals. London: Academic Press. Earney FCF (1990) Marine Mineral Resources: Ocean Management and Policy. London: Routledge. Glasby GP (ed.) (1977) Marine Manganese Deposits. Amsterdam: Elsevier. Nawab Z (1984) Red Sea mining: a new era. Deep-Sea Research 31A: 813--822. Thiel H, Angel M, Foell E, Rice A, and Schriever G (1998) Environmental Risks from Large-Scale Ecological Research in the Deep Sea – A Desk Study. Luxembourg: European Commission, Office for Official Publications. Wiltshire J. (2000) Marine Mineral Resouces – State of Technology Report Marine Technology Society Journal 34: no. 2, p. 56–59

(c) 2011 Elsevier Inc. All Rights Reserved.

MOLLUSKAN FISHERIES V. S. Kennedy, University of Maryland, Cambridge, MD, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 3, pp 1830–1841, & 2001, Elsevier Ltd.

Introduction The Phylum Molluska, the second largest phylum in the animal kingdom with about 100 000 named species, includes commercially important gastropods, bivalves, and cephalopods. All are soft-bodied invertebrates and most have a calcareous shell secreted by a phylum-specific sheet of tissue called the mantle. The shell is usually external (most gastropods, all bivalves), but it can be internal (most cephalopods). Gastropods live in salt water, fresh water, and on land; bivalves live in salt water and fresh water; cephalopods are marine. Gastropods include land and sea slugs (these have a reduced or no internal shell and are rarely exploited by humans) and snails (including edible periwinkles, limpets, whelks, conchs, and abalone). Bivalves include those that cement to hard substrates (oysters); that attach to hard substrates by strong, beard-like byssus threads (marine mussels, some pearl oysters); that burrow into hard or soft substrates (clams, cockles); or that live on the sea bottom but can move into the water column if disturbed (scallops). Cephalopods are mobile and include squid, cuttlefish, octopus, and the chambered nautilus (uniquely among cephalopods, the nautilus has a commercially valuable external shell). Most gastropods are either carnivores or grazers on algae, most commercial bivalves ‘filter’ suspended food particles from the surrounding water, and cephalopods are carnivores. Humans have exploited aquatic mollusks for thousands of years, as shown worldwide by shell mounds or middens produced by hunter-gatherers on sea coasts, lake margins, and riverbanks. Some mounds are enormous. In the USA, Turtle Mound in Florida’s Canaveral National Seashore occupies about a hectare of land along about 180 m of shoreline, contains nearly 27 000 m3 of oyster shell, and was once an estimated 23 m high (humans have mined such mounds worldwide for the shells, which serve as a source of agricultural and building lime, as a base for roads, and for crumbling into chicken grit). A group of oystershell middens near Damariscotta, Maine, USA stretches along 2 km of

shoreline. One mound was about 150 m long, 70 m wide, and 9 m high before it was mined; a smaller mound was estimated to contain about 340 million individual shells. Studies by anthropologists have shown that many middens were started thousands of years ago (e.g. 4000 years ago in Japan, 5000 years ago in Maine, 12 000 years ago in Chile, and perhaps up to 30 000 years ago in eastern Australia; sea level rise may have inundated even older sites). Mollusks have been exploited for uses other than food. More than 3500 years ago, Phoenicians extracted ‘royal Tyrian purple’ dye from marine whelks in the genus Murex to color fabrics reserved for royalty, and other whelks in the Family Muricidae have long been used in Central and South America to produce textile dyes. From Roman times until the early 1900s, the golden byssus threads of the Mediterranean pen shell Pinna nobilis were woven into fine, exceedingly lightweight veils, gloves, stockings, and shawls. Wealth has been displayed worldwide on clothing or objects festooned with cowry (snail) shells. Cowries, especially the money cowry Cypraea moneta, have been widely used as currency, perhaps as early as 700 BC in northern China and the first century AD in India. Their use spread to Africa and North America. Also in North America, East Coast Indians made disk-shaped beads, or wampum, from the shells of hard clams Mercenaria mercenaria for use as currency. Mollusk shells have been used as fishhooks and octopus lures (the latter often made of cowry shells), household utensils (bowls, cups, spoons, scrapers, knives, boring devices, adzes, chisels, oil lamps), weapons (knives, axes), signaling devices (trumpets made from large conch shells), and decorative objects (beads, lampshades made of windowpane oyster shell, items made from iridescent ‘mother-of-pearl’ shell of abalone). Many species of mollusks produce ‘pearls’ when they cover debris that is irritating their mantle with layers of nacre (mother-of-pearl; an aragonite form of calcium carbonate). The pearl oyster (not a true oyster) and some freshwater mussels (not true mussels) use an especially iridescent nacre that results in commercially valuable pearls. Some religious practices involve shells, including scallop shells used as symbols by pilgrims trekking to Santiago de Compostela in Spain to honor St James (the French name for the scallop is Coquille StJacques). As human populations and the demand for animal protein have grown, harvests of wild mollusks have

(c) 2011 Elsevier Inc. All Rights Reserved.

899

900

MOLLUSKAN FISHERIES

Figure 1 Post-harvest activities on shallow-draft tonging boats in Chesapeake Bay, Maryland, USA. Note the array of tongs, and the sorting or culling platform in the boat in the left foreground. (Photograph by Skip Brown, courtesy of Maryland Sea Grant.)

(c) 2011 Elsevier Inc. All Rights Reserved.

MOLLUSKAN FISHERIES

expanded. However, overharvesting and pollution of mollusk habitat have depleted many wild populations. To meet the demand for protein and to combat these losses, aquaculture has increased to supplement wild harvests. Unfortunately, some harvesting and aquaculture practices can have detrimental effects on the environment (see below).

Harvesting Natural Populations Historical exploitation of mollusks occurred worldwide in shallow coastal and freshwater systems, and artisanal fishing still takes place there. The simplest fisheries involve harvesting by hand or with simple tools. Thus marine mussels and various snails exposed on rocky shores at low tides are harvested by hand, with oysters, limpets, and abalone pried from the rocks. Low tide on soft-substrate shores allows digging by hand to capture many shallow-dwelling species of burrowing clams. Some burrowing clams live more deeply or can burrow quickly when disturbed. Harvesting these requires a shovel, or a modified rake with long tines that can penetrate sediment quickly and be rotated so the tines retain the clam as the rake is pulled to the surface. Mollusks living below low tide are usually captured by some sort of tool, the simplest being rakes

901

and tongs. For example, oyster tongs in Chesapeake Bay (Figure 1) have two wooden shafts, each with a half-cylinder, toothed, metal-rod basket bolted at one end and with a metal pin or rivet holding the shafts together like scissors. A harvester standing on the side of a shallow-draft boat lets the shafts slip through his hands (almost all harvesters are male) until the baskets reach the bottom and then moves the upper ends of the shafts back and forth to scrape oysters and shells into a pile. He closes the baskets on the pile and hoists the contents to the surface manually or by a winch, opening the baskets to dump the scraped material onto a sorting platform on the boat (see Figure 1). Harvesters use hand tongs at depths up to about 10 m. Also in Chesapeake Bay, harvesters exploiting oysters living deeper than 10 m deploy much larger, heavier, and more efficient tongs from their boat’s boom, using a hydraulic system to raise and lower the tongs and to close them on the bottom. In addition to capturing bivalves, rakes and hand tongs are used worldwide to harvest gastropods like whelks, conchs, and abalone (carnivorous gastropods like whelks and conchs can also be captured in pots baited with dead fish and other animals). Mollusks can be captured by dredges towed over the bottom by boats, some powered by sail (Figure 2) and others by engines (Figure 3). Dredges harvest attached mollusks like oysters and marine mussels, as

Figure 2 Oyster dredge coming on board a sailboat (‘skipjack’) in Chesapeake Bay, Maryland, USA. Note the small ‘push-boat’ or yawl hoisted on the stern on this sailing day. (Photograph by Michael Fincham, courtesy of Maryland Sea Grant.)

(c) 2011 Elsevier Inc. All Rights Reserved.

902

MOLLUSKAN FISHERIES

Figure 3 Harvesting vessel Mytilus with blue mussel dredges in Conwy Bay, Wales. (Photograph courtesy of Dr Eric Edwards.)

well as buried clams, scallops lying on or swimming just above the sea bottom, and some gastropods (whelks, conchs). Dredges are built of metal and usually have teeth on the leading edge that moves over the bottom. Captured material is retained in a sturdy mesh bag made of wear-resistant heavy metal rings linked together and attached to the dredge frame. Mesh size is usually regulated so that small mollusks can fall out of the bag and back onto the sea bottom. Some dredges use powerful water jets to blow buried mollusks out of soft sediment and into a metalmesh bag. Where the water is shallow enough, subtidal clams (and oysters in some regions) are harvested by such water jets, but instead of being captured in a mesh bag the clams are blown onto a wire-mesh conveyor belt that carries them to the surface alongside the boat. Harvesters pick legal-sized clams off the mesh as they move past on the belt. Mesh size is such that under-sized clams and small debris fall through the belt and return to the bottom while everything else continues up the belt and falls back into the water if not removed by the harvester. Commercial harvesting of freshwater mussels in the USA since the late 1800s has taken advantage of the propensity of bivalves to close their shells tightly when disturbed. Harvest vessels tow ‘brails’ over the bottom. Brails are long metal rods or galvanized pipe with eyebolts at regular intervals. Wire lines are attached to the eyebolts by snap-swivels. Each line holds a number of ‘crowfoot’ hooks of various sizes and numbers of prongs, depending on the species being harvested. Small balls are formed on the end of

each prong so that, when the prong tip enters between the partially opened valves of the mussel, the valves close on the prong and the ball keeps the prong from pulling free of the shell. When the brails are brought on deck the clinging mussels are removed. This method works best in river systems with few snags (tree stumps, rocks, trash) that would catch and hold the hooks. Cephalopods are captured by trawls, drift nets, seines, scoop and cast nets, pots and traps, and hook and line. A traditional gear is the squid jig, which takes various shapes but which has an array of barbless hooks attached. Jigs are moved up and down in the water to attract squid, which grab the jig and ensnare their tentacles, allowing them to be hauled into the boat. In oceanic waters, large vessels using automated systems to oscillate the jig in the water may deploy over 100 jig lines, each bearing 25 jigs. Such vessels fish at night, with lights used to attract squid to the fishing boat. A typical vessel may carry 150 metalhalide, incandescent lamps that together produce 300 kW of light. Lights from concentrations of vessels in the global light-fishing fleet off China and south-east Asia, New Zealand, the Peruvian coast, and southern Argentina can be detected by satellites. With a crew of 20, a vessel as described above may catch 25– 30 metric tons of squid per night. Some harvesters dive for mollusks, especially solitary organisms of high market value such as pearl oysters and abalone. Breath-hold diving has been used for centuries, but most divers now use SCUBA or air-delivery (hooka) systems. Diving is efficient

(c) 2011 Elsevier Inc. All Rights Reserved.

MOLLUSKAN FISHERIES

because divers can see their prey, whereas most other capture methods fish ‘blind’. Unfortunately, although diving allows for harvesting with minimal damage to the habitat, it has led to the depletion or extinction of some mollusk populations such as those of abalone. A variety of measures are in place around the world to regulate mollusk fisheries. A common regulation involves setting a minimum size for captured animals that is larger than the size at which individuals of the species become capable of reproducing. This regulation ensures that most individuals can spawn at least once before being captured. Size selection is often accomplished by use of a regulated mesh size in dredge bags or conveyor-belts as described earlier. If the animals are harvested by a method that is not size selective, such as tonging for oysters or brailing for freshwater mussels, then the harvester is usually required to cull undersized individuals from the accumulated catch (see Figure 1 for the culling platform used by oyster tongers) and return them to the water, usually onto the bed from which they were taken. Oysters are measured with a metal ruler; freshwater mussels are culled by attempting to pass them through metal rings of legal diameter and keeping those that cannot pass through. Other regulatory mechanisms include limitations as to the number of harvesters allowed to participate in the fishery, the season when harvesting can occur, the type of harvest gear that can be used, or the total catch that the fishery is allowed to harvest. There may be areas of a species’ range that are closed to harvest, perhaps when the region has many undersized juveniles or when beds of large adults are thought to be in need of protection so that they can serve as a source of spawn for the surrounding region. Restrictions may spread the capture effort over a harvest season to prevent most of the harvest from occurring at the start of the season, with a corresponding market glut that depresses prices. Finally, managers may protect a fishery by regulations mandating inefficiencies in harvest methods. Thus, in Maryland’s Chesapeake Bay, diving for oysters has been strictly regulated because of its efficiency. Similarly, the use of a small boat (Figure 2) to push sailboat dredgers (‘power dredging’) is allowed only on 2 days per week (the days chosen – usually days without wind – are at the captain’s discretion). Dredging must be done under sail on the remaining days of the week (harvesting is not allowed on Sunday). Around the world, inspectors (‘marine police’) are empowered to ensure that regulations are followed, either by boarding vessels at sea or when they dock with their catch. A great hindrance to informed management of molluskan (and other) fisheries is the lack of data on

903

the quantity of organisms taken by noncommercial (recreational) harvesters. For example, in Maryland’s Chesapeake Bay one can gather a bushel (around 45 l) of oysters per day in season without needing a license if the oysters are for personal use. Clearly it is impossible to determine how many of these bushels are harvested during a season in a sizeable body of water. If such harvests are large, the total fishing mortality for the species can be greatly underestimated, complicating efforts to use fishery models to manage the fishery. Processing mollusks involves mainly shore-based facilities, except for deep-sea cephalopod (mostly squid) fisheries where processing, including freezing, is done on board, and for some scallop species (the large muscle that holds scallop shells shut is usually cut from the shell at sea, with the shell and remaining soft body parts generally discarded overboard). Thus the catch may be landed on the same day it is taken (oysters, freshwater and marine mussels, many clams and gastropods) or within a few days (some scallop and clam fisheries). If the catch is not frozen, ice is used to prevent spoilage at sea. Suspension-feeding mollusks (mostly bivalves) can concentrate toxins from pollutants or poisonous algae and thereby become a threat to human health. If such mollusks are harvested, they have to be held in clean water for a period of time to purge themselves of the toxin (if that is possible). Thus they may be relaid on clean bottom or held in shore-based systems (‘depuration facilities’) that use ozone or UV light to sterilize the water that circulates over the mollusks. Relaying and reharvesting the mollusks or maintaining the land-based systems adds to labor, energy, and capital costs.

Aquaculture Aquaculture involves using either natural ‘seed’ (small specimens or juveniles that will be moved to suitable habitat for further growth) that is harvested from the wild and reared in specialized facilities, or producing such seed in hatcheries. Most commercial bivalves and abalone can be spawned artificially in hatcheries. The techniques involve either taking adults ripe with eggs or sperm (gametes) from nature or assisting adults to ripen by providing algal food in abundance at temperatures warm enough to support gamete production. Most ripe adults will spawn when provided with a stimulus such as an increase in temperature or food or both, or by the addition to the ambient water of gametes dissected from sacrificed adults. The spawned material is washed through a series of screens to separate the gametes

(c) 2011 Elsevier Inc. All Rights Reserved.

904

MOLLUSKAN FISHERIES

from debris. Eggs are washed into a container and their numbers are estimated by counting samples, then the appropriate density of sperm is added to fertilize the eggs. Depending on the species of mollusk, an adult female may produce millions of eggs in one spawning event, so hundreds of millions of larvae produced by a relatively small number of females may be available for subsequent rearing. Larvae are reared in specialized containers that allow culture water to be changed every few days and the growing larvae to be captured on screens, counted, and replaced into clean water until they become ready to settle. As they grow, larvae are fed cultured algae at appropriate concentrations (this requires extremely large quantities of algae to support the heavily feeding larvae). Depending on the species, after 2 or more weeks many larvae are ready to settle (‘set’) onto a solid surface (e.g., oysters, marine mussels, scallops, pearl oysters, abalone) or onto sediment into which they will burrow (e.g. clams). The solid material on which setting occurs is called ‘cultch’. Settled larvae are called ‘spat’. Spat can be reared in the hatchery if sufficient algal food is available (usually an expensive proposition given the large quantity of food required). Thus most production facilities move spat (or seed) into nature soon after they have settled. However, this exposes the seed to diverse natural predators – including flatworms, boring sponges, snails, crabs, fish, and birds. Consequently, the seed may need to be protected (e.g. by removing predators by hand, poisoning them, and providing barriers to keep them from the seed), which is labor-intensive and expensive. Mortalities of larvae, spat, and seed are high at all stages of aquaculture operations. The energy and monetary costs of maintaining hatchery water at suitable temperatures and of rearing enormous quantities of algae means that molluskan hatcheries are expensive to operate profitably. Thus only ‘high-value’ mollusks are cultured in hatcheries. One option for those who cannot afford to maintain a hatchery is to purchase larvae that are ready to settle. In the supply hatchery, larvae are screened from the culture water onto fine mesh fabric in enormous densities. The densely packed larvae withdraw their soft body parts into their shells, closing them. The larvae can then be shipped by an express delivery service in an insulated container that keeps them cool and moist until they reach the purchaser, who gently rinses the larvae off the fabric into clean water of the appropriate temperature and salinity in a setting tank. Also in the tank is the cultch that the larvae will attach to or the sediment into which they will burrow. This procedure of setting purchased larvae is called remote setting.

Among bivalves, oyster larvae settle on hard surfaces, preferably the shell of adults of the species, but also on cement materials, wood, and discarded trash. The larva cements itself in place and is immobile for the rest of its life. Thus oysters have traditionally been cultured by allowing larvae to cement to cultch like wooden, bamboo, or concrete stakes; stones and cement blocks; or shells (such as those of oysters and scallops) strung on longlines hanging from moored rafts (Figure 4C). As noted above, the settling larvae may be either those produced in a hatchery or those living in nature. The spat may be allowed to remain where they settle, or the cultch may be moved to regions where algal production is high and spat growth can accelerate. Scallop, marine mussel, and pearl oyster larvae do not cement to a substrate, but secrete byssus threads for attachment. For these species, fibrous material like hemp rope is provided in hatcheries or is deployed in nature where larvae are abundant. As these settled spat grow, they may be transferred to containers such as single- or multi-compartment nets (Figure 4A, B) or to flexible mesh tubes (used for marine mussels). Most clams are burrowers and will settle onto sediment. This may be provided in ponds, sometimes including those used to grow shrimp and other crustaceans (simultaneous culture of various species is called polyculture). Abalone settle on hard surfaces to which their algal food is attached, so they are usually cultured initially on corrugated plastic plates coated with diatoms. As they grow, they may be moved to well-flushed raceways or held suspended in nature (Figure 4C) in net cages. There are risks associated with aquaculture, just as with harvesting natural populations. A major problem is the one that affects many agricultural monocultures and that is the increased susceptibility to disease outbreaks among densely farmed organisms. Such diseases now affect cultured abalone and scallops in China. Another problem involves the reliance on a relatively few animals for spawning purposes, which can lead to genetic deficiencies and inbreeding, further endangering a culture program. The coastal location of aquaculture facilities makes them susceptible to damage by storms, and if such storms increase or intensify with global warming, aquaculturists risk losing their animals, facilities, and investment. Ice can cause damage in cold-winter regions, although this is more predictable than are intense storms and preventative steps can be taken (not always successfully). Extensive use of rafts and other systems for suspension culture (Figure 4C) may interfere with shipping and recreational uses of the water, and in some regions, laws against navigational hazards prevent water-based culture systems. If land

(c) 2011 Elsevier Inc. All Rights Reserved.

MOLLUSKAN FISHERIES

(A)

905

(B)

(C) Figure 4 (A) Single-compartment pearl oyster cages attached to intertidal longlines. (B) Multi-compartment lantern net used to culture scallops. (C) Suspended longlines for growing Pacific oyster, scallops, and abalone off Rongchen, China. (Photographs courtesy of Dr Ximing Guo; reproduced with permission from Guo et al. (1999) Journal of Shellfish Research 18: 19–31.)

contiguous to the water is too expensive because of development or other land-use practices, land-based culture systems may not be economically feasible. Finally, coastal residents may consider rafts to be an eyesore that lowers the value of their property and they may seek to have them banned from the region.

Comparisons of Wild and Cultured Production Of the 86–93 million metric tons of aquatic species harvested from wild stocks in 1996–1998 (capture

landings), fish comprised 86%, mollusks 8%, and crustaceans (shrimp, lobsters, crabs) 6% (Table 1). Among the mollusks, the average relative proportions over the 3 years were cephalopods, 44%; bivalves, 32%; gastropods, 2%; and miscellaneous, 21%. Within the bivalves, the relative proportions over the 3 years were clams, 38% of all bivalves; freshwater mussels, 25%; scallops, 22%; marine mussels, 9%; and oysters, 6%. Aquaculture had a greater effect on total landings (wild harvest plus aquaculture production) of mollusks than of fish and crustaceans. Of FAO’s estimated 27–31 million metric tons of aquatic animals

(c) 2011 Elsevier Inc. All Rights Reserved.

906

MOLLUSKAN FISHERIES

cultured worldwide from 1996 to 1998, fish comprised 64% and crustaceans comprised 5% by weight, declines from their proportions of the wild harvest (Table 1). In contrast, mollusks comprised 31% of culture production, about four times their proportion of the wild harvest. In addition, 20–40% more mollusks by weight were produced by aquaculture than were harvested from nature; by contrast, the quantities of cultured fish and crustaceans were a small fraction of quantities harvested in nature (Table 1). Of the cultured mollusks produced from 1996 to 1998, bivalves represented 87% by weight, followed by miscellaneous mollusks at 13%; Table 1 Worldwide capture landings (wild harvest) and aquaculture production and value of fish, mollusks, and crustaceans from 1996 to 1998 Category

1996 (%)

1997 (%)

1998 (%)

Capture landings (million metric tons) Fish 80.9 (87) 79.7 (86) 72.7 Mollusks 6.6 (7) 7.3 (8) 6.6 5.9 (6) 6.4 Crustaceans 5.5 (6) Total 93.0 92.9 85.7 Aquaculture production (million metric tons) Fish 17.0 (63) 18.8 (65) 20.0 Mollusks 8.6 (32) 8.7 (30) 9.2 1.4 (5) 1.6 Crustaceans 1.2 (4) Total 26.8 28.9 30.8 Aquaculture value (billion $US)Fish Fish 26.5 (62) 28.3 (62) 27.8 Mollusks 8.6 (20) 8.7 (19) 8.5 Crustaceans 7.8 (18) 8.5 (19) 9.2

Average percentage

(85) (8) (7)

86 8 6

(65) (30) (5)

64 31 5

(61) (19) (20)

62 19 19

From UN Food and Agriculture Organization statistics as at 21 March 2000.

cultured gastropods and cephalopods represented fractions of a percent of produced weight. When the FAO’s harvest values of the wild fisheries and aquaculture production from 1984 to 1998 are combined for bivalves, cephalopods, and gastropods (Figure 5), the production of bivalves is seen to have risen greatly in the 1990s, with cephalopod production having increased modestly and gastropod production not at all. The differences can be attributed to the relatively greater yield of cultured bivalves compared with the other two molluskan groups. The economic value of cultured animals ranged from 43 to 46 billion US dollars over the period 1996–1998, with fish comprising an average of 62%, mollusks 19%, and crustaceans 19% of this amount (Table 1). Thus, although production (weight) of cultured mollusks was six to seven times that of cultured crustaceans, their value just equaled that of crustaceans (this was a result of the premium value of cultured shrimps and prawns). Of the cultured mollusks, bivalves represented 93% by economic value followed by miscellaneous mollusks at 6%; cultured gastropods and cephalopods represented fractions of a percent of overall economic value. Among cultured bivalves, the relative proportions by weight and by economic value were: Oysters, 42% and 42% respectively; clams, 26% and 32%; scallops, 15% and 20%; marine mussels, 17% and 6%; and freshwater mussels, o1% in both categories. Thus, although oysters were the least important bivalve in terms of wild harvests (6% of bivalve landings, see above), they were the most important cultured bivalve. On the other hand, freshwater mussels represented 25% of wild harvests but were relatively insignificant as a cultured item.

12

Million metric tons

10

Bivalves Cephalopods Gastropods

8

6 4

2 0

1982

84

86

88

90 92 Year

94

96

98

2000

Figure 5 Worldwide production of bivalve, cephalopod, and gastropod mollusks from wild fisheries plus aquaculture activities in the period from 1984 to 1998. (From UN Food and Agriculture Organization statistics as at 21 March 2000.)

(c) 2011 Elsevier Inc. All Rights Reserved.

907

MOLLUSKAN FISHERIES

(A)

2.5

8

Oyster

2

Mariculture 6

Clam

Mollusks

Fishery

1.5

Scallop

4

Mussel 1

Year

0.5

(B)

Year

(C)

1996

1994

1988

1986

1996

1990

1978

1984

1972

1966

1960

0

1954

1996

1990

1978

1984

0

1972

0

1966

2

1960

5

Other

1992

10

1990

Aquaculture

15

1954

Production in million metric tons

20

Year

Figure 6 Production (wet, whole-body weight) in the natural fishery and in aquaculture in China. (A) Natural fishery and aquaculture. (B) Total mariculture (marine aquaculture) production and the molluskan component. (C) Comparative production among molluskan groups. (Adapted with permission from Guo et al. (1999) Journal of Shellfish Research 18: 19–31.)

China provides an excellent example of increased efforts in aquaculture, with its industry producing about 50% of the world’s aquacultural output in the early 1990s, rising to about 67% by 1997 and 1998. Oyster and clam culture has been practiced in China for about two millennia, but growth of aquaculture accelerated in the early 1950s, with production outstripping wild fishery harvests by 1988 (Figure 6A). Mollusk culture was a substantial factor in this growth (Figure 6B). The intensified culture involved new species beyond the traditional oysters and clams, beginning with marine mussels, then scallops, then abalone. Over 30 species of marine mollusks are farmed along China’s coasts, including 3 species of oysters, 14 of clams, 4 of marine mussels, 4 of scallops, 2 of abalone, 2 of snails, and 1 of pearl oyster. Marine mussel production has declined since 1992 (Figure 6C), apparently because of increased production of ‘high-value’ species (oysters, scallops, abalone).

Detrimental Effects of Molluskan Fisheries Harvesting Natural Populations

Harvesting can have direct and indirect effects on local habitats. Direct effects include damage caused by harvest activities, such as when humans trample rocky shore organisms while harvesting edible mollusks. Similarly, heavy dredges with strong teeth can damage and kill undersized bivalves as well as associated inedible species. Dredges that use highpressure water jets churn up bottom sediments and

displace undersized clams and nontarget organisms that may not be able to rebury before predators find them exposed on the surface. When commercial shellfish are concentrated within beds of aquatic grasses, dredging can uproot the plants. On the other hand, some harvesting techniques are relatively benign. For example, lures like squid jigs attract the target species with little or no ‘by-catch’ of nontarget species. Size selectivity can be attained by using different sizes of jigs. However, as with most fishing methods, overharvesting of the target species can still occur. From the perspective of population dynamics, intense harvesting pressure commonly results in smaller average sizes of individual organisms and eventually most survivors may be juveniles. Care must be taken to ensure that a suitable proportion of individuals remains to reproduce before they are harvested, thus the widespread regulations on minimal sizes of individuals that can be harvested. In terms of indirect effects, if some molluskan species are overharvested, inedible associated organisms that depend on the harvested species in some way can be affected. For example, oysters are ‘ecosystem engineers’ (like corals) and produce large structures (beds, reefs) that are exploited by numerous other organisms. Oyster shells provide attachment surfaces for large numbers of barnacles, mussels, sponges, sea squirts, and other invertebrates, as well as for the eggs of some species of fish and snails. Waterborne sediments as well as feces from all the reef organisms settle in interstices among the shells and become habitat for worms and other burrowing invertebrates. The presence of these

(c) 2011 Elsevier Inc. All Rights Reserved.

908

MOLLUSKAN FISHERIES

organisms attracts predators, including crustaceans and fish. Consequently, the diversity of species on and around oyster reefs is higher than it is for adjacent soft-bottom systems. The same is true for marine mussel beds. Harvesting these bivalves can result in lowered biomass and perhaps lowered diversity of associated organisms. Depleted populations of commercial bivalves may have other ecological consequences. These bivalves are ‘suspension feeders’ and pump water inside their shell and over their gills to extract oxygen and remove suspended food particles. When Europeans sailed into Chesapeake Bay in the 1600s, oyster reefs broke the Bay’s surface and were a navigational hazard. Today, many oyster beds have been scraped almost level with the Bay bottom and annual harvests in Maryland have dropped from an estimated 14 million bushels of oysters in the 1880s to a few hundred thousand bushels today. Knowing the pumping capacity of individual oysters and the estimated abundances of pre-exploitation and presentday populations, scientists can calculate the filtering ability of those populations. Estimates are that a quantity of water equivalent to the total amount contained in the Bay could have been filtered in about a week in the early eighteenth century; today’s depleted populations are thought to require nearly a year to filter the same amount of water. This diminished filtering capacity has implications for food web dynamics. A feeding oyster ingests and digests food particles and expels feces as packets of material larger than the original food particles. Nonfood particles are trapped in mucous strings and expelled into the surrounding water as pseudofeces. Thus, oysters take slowly sinking microscopic particles from the water and expel larger packets of feces and pseudofeces that sink more rapidly to the Bay bottom. When oyster numbers are greatly diminished, this filtering and packaging activity is also diminished and more particles remain suspended in the water column. Thus ecological changes in Chesapeake Bay over the last century may have been enormous. Many scientists believe the Bay has shifted from a ‘bottom-dominated’ system in which oyster reefs were ubiquitous and the water was clearer than now to one dominated by water-column organisms (plankton) in which light levels are diminished. Smaller plankton serve as food for larger plankton, including jellyfish, and some scientists believe that the large populations of jellyfish present in the Bay in warmer months may be a result of overharvesting of oysters over the past century. The reverse of this phenomenon is seen in the North American Great Lakes, where huge populations of zebra mussels Dreissena polymorpha and

quagga mussels Dreissena bugensis have developed from invaders inadvertently introduced from Europe in ships’ ballast water. The mussels filter the lake water so efficiently that the affected lakes are clearer than they have been for decades – they have become ‘bottom-dominated’ and plankton populations have decreased in abundance. Unintended consequences resulting from overharvesting other mollusk populations remain to be elucidated.

Aquaculture Practices

A number of environmental problems affect molluskan aquaculture. For one, heavy suspension feeding by densely farmed mollusks in coastal bays may outstrip the ability of the environment to supply algae, so the mollusks may starve or cease to grow. Another problem is that large concentrations of cultured organisms produce fecal and other wastes in quantities that can overwhelm the environment’s ability to recycle these wastes. When this happens, eutrophication occurs, inedible species of algae may appear, and abundances of algal species that support molluskan growth may be reduced. To counter these problems, countries like Australia are using Geographic Information System technology to pinpoint suitable and unsuitable locations for aquaculture. Introductions of exotic species of shellfish for aquaculture purposes have sometimes been counterproductive. A variety of diseases that have depleted native mollusk stocks have been associated with some introductions (e.g. MSX disease in the eastern oyster Crassostrea virginica in Chesapeake Bay is thought to be linked to attempts to import the Pacific oyster Crassostrea gigas to the bay). ‘Hitch-hiking’ associates that are carried to the new environment by imported mollusks have sometimes caused problems. For example, the gastropod Crepidula fornicata that accompanied oysters of the genus Crassostrea that were brought to Europe from Asia has become so abundant that it competes with the European oyster Ostrea edulis and blue mussel Mytilus edulis for food and may foul oyster and mussel beds with its wastes. Recently, the veined rapa whelk Rapana venosa has appeared in lower Chesapeake Bay, apparently arriving from the Black Sea or from Japan in some unknown fashion. It is a carnivore and may pose a threat to the indigenous oyster and hard clam fisheries (although it might prove to be a commercial species itself if a market can be found). As a result of these and other problems, many countries have developed stringent rules to govern movement of mollusks locally and worldwide.

(c) 2011 Elsevier Inc. All Rights Reserved.

MOLLUSKAN FISHERIES

See also Cephalopods. Coral Reef and Other Tropical Fisheries. Corals and Human Disturbance. Dynamics of Exploited Marine Fish Populations. Ecosystem Effects of Fishing. Exotic Species, Introduction of. Fisheries and Climate. Fishery Management, Human Dimension. Fishery Manipulation through Stock Enhancement or Restoration. Fishing Methods and Fishing Fleets. Marine Fishery Resources, Global State of. Rocky Shores.

Further Reading Andrews JD (1980) A review of introductions of exotic oysters and biological planning for new importations. Marine Fisheries Review 42(12): 1--11. Attenbrow V (1999) Archaeological research in coastal southeastern Australia: A review. In: Hall J and McNiven IJ (eds.) Australian Coastal Archaeology, pp. 195--210. ANH Publications, RSPAS. Canberra, Australia: Australian National University. Caddy JF (ed.) (1989) Marine Invertebrate Fisheries: Their Assessment and Management. New York: John Wiley Sons. Caddy JF and Rodhouse PG (1998) Cephalopod and groundfish landings: Evidence for ecological change in global fisheries. Reviews in Fish Biology and Fisheries 8: 431--444. Food and Agriculture Organization of the United Nations’ website for fishery statistics (www.fao.org). Guo X, Ford SE, and Zhang F (1999) Molluscan aquaculture in China. Journal of Shellfish Research 18: 19--31.

909

Kaiser MJ, Laing I, Utting SD, and Burnell GM (1998) Environmental impacts of bivalve mariculture. Journal of Shellfish Research 17: 59--66. Kennedy VS (1996) The ecological role of the eastern oyster, Crassostrea virginica, with remarks on disease. Journal of Shellfish Research 15: 177--183. MacKenzie CL Jr, Burrell VG Jr, Rosenfield A, and Hobart WL (eds.) (1997) The History, Present Condition, and Future of the Molluscan Fisheries of North and Central America and Europe. Volume 1, Atlantic and Gulf Coasts. (NOAA Technical Report NMFS 127); Volume 2, Pacific Coast and Supplemental Topics (NOAA Technical Report NMFS 128); Volume 3, Europe (NOAA Technical Report NMFS 129). Seattle, Washington: US Department of Commerce. Menzel W (1991) Estuarine and Marine Bivalve Mollusk Culture. Boca Raton, FL: CRC Press. Newell RIE (1988) Ecological changes in Chesapeake Bay: Are they the result of overharvesting of the American oyster, Crassostrea virginica. In: Lynch MP and Krome EC (eds.) Understanding the Estuary: Advances in Chesapeake Bay Research, pp. 536--546. Chesapeake Research Consortium Publication 129. Solomons, MD: Chesapeake Research Consortium. Rodhouse PG, Elvidge CD, and Trathan PN (2001) Remote sensing of the global light-fishing fleet: An analysis of interactions with oceanography, other fisheries and predators. Advances in Marine Biology 39: 261--303. Safer JF Gill FM (1982) Spirals from the Sea. An Anthropological Look at Shells. New York: Clarkson N Potter. Sanger D and Sanger M (1986) Boom and bust in the river. The story of the Damariscotta oyster shell heaps. Archaeology of Eastern North America 14: 65--78.

(c) 2011 Elsevier Inc. All Rights Reserved.

MONSOONS, HISTORY OF N. Niitsuma, Shizuoka University, Shizuoka, Japan P. D. Naidu, National Institute of Oceanography, Dona Paula, India & 2009 Elsevier Ltd. All rights reserved.

Introduction The difference in specific heat capacity between continents and oceans (and specifically between the large Asian continent and the Indian Ocean) induces the monsoons, strong seasonal fluctuations in wind direction, and precipitation over oceans and continents. Over the Indian Ocean, strong winds blow from the southwest during boreal summer, whereas weaker winds from the northeast blow during boreal winter. The monsoons are strongest over the western part of the Indian Ocean and the Arabian Sea (Figure 1). The high seasonal variability affects various fluctuations in the environment and its biota which are reflected in marine sediments. Marine sedimentary sequences from the continental margins thus contain a record of the history of monsoonal occurrence and intensity, which may be deciphered

to obtain insight in the history of the monsoon system. Such insight is important not only to understand the mechanisms that cause monsoons, but also to understand the influence of monsoons on the global climate system including the Walker circulation, the large-scale, west–east circulation over the tropical ocean associated with convection. The summer monsoons may influence the Southern Oscillation because of interactions between the monsoons and the Pacific trade wind systems.

Geologic Records of the Monsoons Sedimentary sequences record the effects of the monsoons. One such effect is the upwelling of deeper waters to the surface, induced by the strong southwesterly winds in boreal summer in the Arabian Sea, and the associated high productivity of planktonic organisms. Continents supply sediments to the continental margin through the discharge of rivers, as the result of coastal erosion, and carried by winds (eolian sediments). Monsoon-driven wind direction and strength and precipitation therefore control the sediment supply to the continental margin.

(a) Low

40° N

0

40° S

(b) 40° N

0

40° S

Figure 1 The domains of the monsoon system of the atmosphere during Northern Hemisphere (a) summer and (b) winter. The hatching shows the land areas with maximum surface temperature and stippling indicates the coldest land surface.

910

(c) 2011 Elsevier Inc. All Rights Reserved.

MONSOONS, HISTORY OF

Monsoonal precipitation supplies a large volume of fresh water, discharged by rivers from the continents into the oceans, with the flow directed by the topography in the regions of the river mouths. The density of fresh water is less than that of seawater, and the seasonal freshwater flow dilutes surface ocean water which then has a lower density than average seawater. The existence of strong vertical density gradients causes the development of stratification in the water column. The elevation of the continent governs the circulation of the atmosphere, and influences vegetation patterns. The vegetation cover on the continent directly affects the albedo and heat capacity of the land surface, both of which are important factors in the generation of monsoonal circulation patterns. The high Himalayan mountain range acts as a barrier for air circulation; east–west-oriented mountain ranges particularly affect the course of the jet streams. Continental topography, vegetation coverage, and elevation thus affect the monsoonal circulation and therefore the rate of sediment supply to the continental margin, as well as the composition of continental margin sedimentary sequences.

Sedimentary Indicators of Monsoons Monsoon-induced seasonal contrasts in wind direction and precipitation are recorded in the sediments by many different proxies. Most of these monsoonal indicators, however, record qualitative and/or quantitative changes of the monsoons in one, but not in both, season. For example, upwelling indicators in the Arabian Sea represent the strength of the southwest (summer) monsoon. In interpreting the sedimentary record, one must note that bottom-dwelling fauna bioturbates the sediment to depths of c. 10 cm, causing the cooccurrence of sedimentary material deposited at different times within a single sediment sample. Laminated sediments without bioturbation are only deposited in oxygen-minimum zones, where anoxia prevents activity of burrowing metazoa. The time resolution of studies of various monsoonal indicators thus is limited by this depth of bioturbation, but each proxy by itself does record information on the environment in which it formed at the time that it was produced. Single-shell measurements of oxygen and carbon isotopes in foraminiferal tests (see below) thus may show the variability within each sample, thus within the zone of sediment mixing. Oxygen isotopes. The oxygen isotopic ratio (d18O) of the calcareous tests of fossil organisms (such as foraminifera) provides information on the oxygen

911

isotopic ratio in seawater and on the temperature of formation of the test. The volume of the polar ice sheets and the local influx of fresh water controls the oxygen isotopic ratio of seawater. The d18O record of the volume of the polar ice sheets can be used for global correlation of oxygen isotopic stages, corresponding to glacial and interglacial intervals. Freshwater discharge caused by monsoonal precipitation lowers the d18O value of seawater. The temperature record of various species with different depth habits, such as benthic species at the bottom, and planktonic species at various depths below the surface, can provide a temperature profile of the water column. Temperature is the most basic physical parameter, which provides information on the stratification or mixing of the water column, as well as on changes in thermocline depth. Different species of planktonic foraminifera grow in different seasons, and temperatures derived from the oxygen isotopic composition of their tests thus delineates the seasonal monsoonal variation in sea surface temperatures. Carbon isotopes. The carbon isotopic ratio in the calcareous test of foraminifera provides information on the carbon isotopic ratio of dissolved inorganic carbon (DIC) in seawater and on biofractionation during test formation. Carbon isotopic ratios of various species of foraminifera which live at different depths provide information on the carbon isotopic profile of DIC in the water column. The carbon isotope ratio of DIC in seawater is well correlated with the nutrient concentration in the water, because algae preferentially extract both the lighter carbon isotope (12C) and nutrients to form organic matter by photosynthesis. Decomposition of organic matter releases lighter carbon as well as nutrients, and lowers the carbon isotopic composition of DIC. The carbon isotopic profile with depth thus provides information on the balance of photosynthesis and decomposition. Rate of sedimentation. Marine sediments are composed of biogenic material produced in the water column (dominantly calcium carbonate and opal), and terrigenous material supplied from the continents. The sediments are transported laterally on the seafloor and down the continental slope, and eventually settle in topographic depressions in the seafloor. Both seafloor topography and the supply of biogenic and lithogenic material control the apparent rate of sedimentation. Organic carbon content. Organisms produce organic carbon in the euphotic zone, which is strongly recycled by organisms in the upper waters, but a small percentage of the organic material eventually settles on the seafloor. Organisms (including bacteria) decompose the organic carbon on the surface

(c) 2011 Elsevier Inc. All Rights Reserved.

912

MONSOONS, HISTORY OF

of the seafloor as well as within the sediment, using dissolved oxygen in the process. The flux of organic matter and the availability of oxygen thus control the organic carbon content in the sediments. In highproductivity areas, such as regions where monsooninduced upwelling occurs, a large amount of organic matter sinks from the sea surface through the water column. The sinking organic matter decomposes and consumes dissolved oxygen, and at mid-water depths an oxygen minimum zone may develop in such high productivity zones. Oxygen concentrations may fall to zero in high-productivity environments, and in these anoxic environments, eukaryotic benthic life becomes impossible, so that there is no bioturbation. The organic carbon content in the sediment deposited below oxygen-minimum zones may become very high, and values of up to 7% organic carbon have been recorded in areas with intense upwelling, such as the Oman Margin. Calcium carbonate. Three factors control the calcium carbonate content of pelagic sediments: productivity, dilution, and dissolution. Calcium carbonate tests are produced by pelagic organisms, including photosynthesizing calcareous nannoplankton and heterotrophic planktonic foraminifera. At shallow water depths (above the calcite compensation depth, CCD) dissolution is negligible; therefore, the calcium carbonate content in continental margin sediments is regulated by a combination of biotic productivity and dilution by terrigenous sediment. Magnetic susceptibility. Magnetic susceptibility provides information about the terrigenous material supply and its source. The part of the sediment provided by biotic productivity (calcium carbonate and opaline silica) has no carriers of magnetic material. Magnetic susceptibility is thus used as an indicator of terrigenous supply to the ocean. Clay mineral composition. Weathering and erosion processes on land lead to the formation of various clay minerals. The clay mineral composition in the sediments is an indicator for the intensity of weathering processes and thus temperature and humidity in the region of origin. The clay mineral composition thus can be used as a proxy for the aridity and vegetation coverage in the continents from which the material derived. In addition, the crystallinity of the clay mineral illite depends on the moisture content of soils in the area of origin. At high moisture, illite decomposes and dehydrates, so that its crystallinity decreases. Therefore, the crystallinity of illite can be used as a proxy for the humidity in its source area. Eolian dust. Eolian dust consists of fine-grained quartz and clay minerals, such as illite. The content

and grain size of eolian dust in marine sediment, therefore, is an indicator of wind strength and direction, as well as of the aridity in the source area. Fossil abundance and diversity. The abundance and diversity of foraminifera, calcareous nannoplankton, diatoms, and radiolarians depend on the chemical and physical environmental conditions in the oceans. The diversity of nannofossil assemblages decreases with sea surface temperature and is thus generally correlated with latitude. In upwelling areas, low sea surface temperatures caused by the monsoon-driven upwelling disturb the zonal diversity patterns of nannoflora. Therefore, the diversity of nannofossils can be used as an upwelling indicator. In addition, the faunal and floral assemblages can be used to estimate sea surface temperature and salinity as well as productivity. UK37 ratio. The biomarker UK37 (an alkenone produced by calcareous nannoplankton) is used to estimate sea surface temperatures within the photic zone where the photosynthesizing algae dwell. The records are not always easy to calibrate, especially in the Tropics and at high latitudes, but global calibrations are now available.

Indicators of Present Monsoons The specific effect of the monsoons on the supply of lithogenic and biogenic material to the seafloor varies in different regions, so that specific tracers cannot be efficiently applied in all oceans. Arabian Sea

During the boreal summer, strong southwest monsoonal winds produce intense upwelling in the Arabian Sea (Figure 2). These upwelling waters are characterized by low temperatures and are highly enriched in nutrients. The process of upwelling fuels the biological productivity in June through August in the Arabian Sea. The weaker, dry, northeast winds which prevail during the boreal winter do not produce upwelling, and productivity is thus lower in the winter months. Thus, the southwest and northeast monsoonal winds produce a strong seasonal contrast in primary productivity in the Arabian Sea. Sediment trap mooring experiments have demonstrated that up to 70% of the biogenic and lithogenic flux to the seafloor occurs during the summer monsoon. Biological and terrestrial particles thus strongly reflect summer conditions, and eventually settle on the seafloor to contribute to a distinct biogeochemical record of monsoonal upwelling. Therefore, regional sediments beneath the areas affected by monsoondriven upwelling record long-term variations in the

(c) 2011 Elsevier Inc. All Rights Reserved.

MONSOONS, HISTORY OF

913

Figure 2 Vertical profiles of d13C, salinity, temperature, and oxygen values in the Arabian Sea. d13C and salinity are from GEOSECS Station 413, while water temperature and oxygen values are from ODP Site 723 in the Oman Margin.

strength and timing of the monsoonal circulation. The following proxies were used to study the upwelling strength in the Arabian Sea. Globigerina bulloides abundance. Seasonal plankton tows and sedimentary trap data document that the planktonic foraminifer species G. bulloides is abundant during the summer upwelling season in the Arabian Sea. Core top data from the upwelling zones of the Arabian Sea show that the dominance of G. bulloides in the living assemblage is preserved in the sediment. Changes in the abundance of G. bulloides in the sediments thus have been used to infer the history of upwelling intensity in the western Indian Ocean.

Oxygen and carbon isotopes. More recently, it was proposed that oxygen and carbon isotopic difference between the surface and subsurface dwelling planktonic foraminifera can be used as proxy records to reconstruct intensity of upwelling and monsoon. Lithogenic material. Lithogenic material deposited in the northwest Arabian Sea is dominantly eolian (diameter up to 18.5 mm), and is transported exclusively during the summer southwest monsoon. The lithogenic grain size in sediment cores thus provides information about the strength of the southwest monsoonal winds and associated upwelling in the Arabian Sea.

(c) 2011 Elsevier Inc. All Rights Reserved.

914

MONSOONS, HISTORY OF

China Sea

The surface circulation patterns in the China Sea are also closely associated with the large-scale seasonal reversal of the atmospheric circulation over the Asian continent. High precipitation over Asia during the summer leads to increased input of fresh water into the China Sea, lowering sea surface salinity. Therefore, the d18O record of planktonic foraminifera in sediment cores documents the magnitude of freshwater discharge, sea surface salinity, and summer monsoonal rainfall in the past. During the winter, the westerly winds lower the sea surface temperature and deliver a large amount of eolian dust to the South China Sea. The rate of eolian dust supply and sea surface temperature changes in this region thus reflect the strength of the winter monsoon. Japan Sea

The winter monsoon’s westerlies transport eolian dust from the Asian desert regions to the Japan Sea and the Japanese Islands 3–5 days after sandstorms in the source area. The thickness of the dust layer is larger in the western Japan Sea. The concentration and grain size of eolian particles in sediments of Japan Sea thus represent the strength of the winter monsoon. Continental Records

Lake levels in the monsoon-influenced regions are highly dependent on the monsoon rainfall; therefore researchers have been using lake levels to trace the monsoon history during the late Quaternary period. The layers of stalagmite (carbonate mineral) preserves the oxygen isotopic composition of monsoon rains that were falling when the stalagmite got precipitated. The oxygen isotope composition of stalagmite is proportional to the amount of rainfall: the more the rainfall, the lighter the d18O of stalagmite, and vice versa. Thus, the oxygen isotopic ratios of stalagmites from the caves of monsoon-influenced regions have been used to infer the ultra-high-resolution variability of monsoon precipitation.

Variability of Monsoons during Glacial and Interglacials Arabian Sea

Along the Oman Margin of the Arabian Sea, strong southwest summer monsoon winds induce upwelling. Detailed analyses of various monsoon tracers such as abundance of G. bulloides and Actinomma spp. (a radiolarian) and pollen reveal that the southwest monsoon winds were more

intense during interglacials (warm periods) and weaker during glacials (cold periods), recognized by the oxygen isotopic stratigraphy in the same samples. Carbon isotope differences between planktonic and benthic foraminifera show lower gradients during interglacials, and higher gradients during glacials. The lower gradients indicate that upwelling was strong (and pelagic productivity high) during interglacials due to a strong summer southwest monsoon. Similarly, the oxygen isotope difference between planktonic and benthic foraminifera along the Oman Margin reflects changes in thermocline depth, associated with summer monsoon-driven upwelling. A larger oxygen isotope difference between planktonic and benthic foraminifera during interglacials reflects the presence of a shallow thermocline, as a result of the strong summer monsoons. Oxygen and carbon isotope records from various planktonic and benthic foraminifera in several cores located within and away from the axis of the Somali Jet along the Oman Margin indicate that sea surface temperatures were lower and varied randomly during interglacials, reflecting strong upwelling induced by a strong summer monsoon (Figure 3). Studies of the oxygen and carbon isotopic composition of individual tests of planktonic foraminifera enable us to understand the seasonal temperature variability induced by monsoons in the Arabian Sea. Such studies show that the seasonality was stronger during glacials and weaker during interglacials, because during glacials the southwest summer monsoon was weaker and the northeast winter monsoon stronger. The variability in seasonal contrast during glacials and interglacials suggests that interannual and interdecadal changes in monsoonal strength were also greater during glacial periods than during interglacials (Figure 4). High-resolution monsoon records from the Arabian Sea show millennial timescale variability of SW monsoon over last 20 ky. Synchronous changes between SW monsoon intensity and variations of temperatures in North Atlantic and Greenland suggest some kind link between monsoons to the highlatitude temperature. South China Sea

High-resolution studies in the South China Sea document a high rate of delivery of eolian dust during glacial periods, as well as lower sea surface temperatures, documented by the UK37 records. These data indicate that during glacial periods the winter monsoon was more intense. During interglacials the sea surface salinity (derived from oxygen isotope records) was much lower than during

(c) 2011 Elsevier Inc. All Rights Reserved.

MONSOONS, HISTORY OF

80

16.0 15.3

12

15.2 15.4 500

450

400

Nannofossil

Δ18OB-P

15.5

13.0

12.0

11.0

9.3 10.0

15.0 15.1

8.5

18OB

11 10

300

250

200

100 m

LAD P. lacunosa 18OP

9.2

9.0

8.0 7.5 8.3

9 8

90

8.4

6.5

6.3

7

70

FAD E. huxleyi

6.4 150

6.1 6.2 100

50

%

10Ka

60

7.2 7.4

7.1

7.0 7.1 7.3

6.0 5.0 5.1 5.3a 5.3b 5.5

6

5.2 5.4

4

3.1 3.3 4.3

3

3.2

1.2

2.1

2

50

14.0

40

5 1 2

3

30

LAD H. inversa

3.0

0 1

20

1.1 2.0

−1

10

350

(a) 0 −2

915

% 3 2 1

(b) 1

13CP

0 −1 %

13CB

% −1 Δ13CB-P

−2

Figure 3 (a) Oxygen isotope profiles of planktonic foraminifera (Pulleniatina obliquiloculata, d18OP) and benthic foraminifera (Uvigerina excellens, d18OB), and the difference between oxygen isotopes of planktonic and benthic foraminifera (Dd18OB–P). (b) Carbon isotope profiles of planktonic foraminifera (d13CP) and benthic foraminifera (d13CB), and the difference between planktonic and benthic records (Dd13CB–P) over last 800 000 years at ODP Site 723 in the Arabian Sea. Large planktonic–benthic differences indicate a more vigorous monsoonal circulation during the summer monsoon.

glacials, probably as a result of increased freshwater discharge from rivers caused by the high summer monsoon precipitation. Over the South China Sea, glacials were thus characterized by an intense winter monsoon, and interglacials by a strong summer monsoon circulation. Japan Sea

The oceanographic conditions of the Japan Sea are strongly influenced by eustatic sea level changes, because shallow straits connect this sea to the Pacific Ocean. In glacial times, at low sea level, most of the straits were above sea level and the Japan Sea was connected to the ocean only by a narrow channel located on the present continental shelf. River discharge of fresh water into the semi-isolated Japan Sea caused the development of strong stratification, and the development of anoxic conditions, as documented by the occurrence of annually laminated sediments. The sedimentary sequence in the Japan Sea thus consists of mud, with laminated sections alternating

with bioturbated, homogeneous muds. During interglacials the waters at the bottom were oxygenated, although the organic carbon content of the sediment can be high (up to 5%) and diatoms abundant (up to 30% volume). During glacials, conditions on the seafloor alternated between euxinic (anoxic) and noneuxinic. Glacial and interglacial parts of the sediment section can thus be easily recognized. The sedimentary sequences contain eolian dust transported from the deserts of west China and the Chinese Loess Plateau by the westerly winter monsoon. The eolian dust content of the sediments is thus an indicator of the strength of the winter monsoon. The crystallinity of illite, a main component of the eolian dust, indicates that the source region on the Asian continent was more humid during interglacials. A high content of eolian dust and a high crystallinity of illite during glacial stages indicate that during these intervals the winter monsoon was strong and the summer monsoon weak. Summer monsoons were strong during interglacials, but winter monsoons were strong during glacials.

(c) 2011 Elsevier Inc. All Rights Reserved.

916

MONSOONS, HISTORY OF

18O −2

0

4

Nanofossil zones 0 21

2

Stage 2 7.01 m

8 7

13C

9

19

9

18

0

9

16

9

14 13

12 9

9

−2

5

Stage 4−5 15.23 m

9

12

12

Age (Ma)

11

11 10 11

9

10 7

8

Stage 6 29.81 m

10

7

12 6

15

15

5

0 Stage 22 137.47 m

50

100

Nanofossil species diversity (%) Figure 5 Calcareous nannofossil diversity for the last 15 My in the Indian Ocean. High diversity indicates weak upwelling and low diversity represents strong upwelling. The upwelling intensity is controlled by the strength of the summer monsoon winds over the Indian Ocean. The numbers on the species diversity profile represent number of species.

1.8 Ma 315.65 m

Planktic Benthic Interglacial Glacial Figure 4 Carbon and oxygen isotope ratios of individual tests of planktonic foraminifera (Pulleniatina obliquiloculata) and benthic foraminifera (Uvigerina excellens) at ODP Site 723 in the Arabian Sea. The variations in difference between planktonic and benthic values reflect the magnitude of the seasonal changes in surface and bottom waters through time.

Long-term Evolution of the Asian Monsoon

calcareous nannofossil species, and the abundance of the planktonic foraminifera G. bulloides and the radiolarian Actinomma spp. (upwelling indicators) show that the evolution of the Asian monsoon started in the late Miocene, at about 9.5 Ma. Between 9.5 and 5 Ma, the monsoon increased noticeably in strength, with smaller fluctuations in monsoonal intensity from 5 to 2 Ma (Figure 5). Upwelling indicators such as the differences in oxygen and carbon isotope between planktonic foraminifera, and the organic carbon content in sediments from the Oman Margin indicate that until about 0.8 Ma the summer monsoon was intense, as during interglacials. Oxygen and carbon isotope ratios in individual tests of planktonic and benthic foraminifera show that from 0.8 Ma onward, the strength of the summer monsoon changed with glacial and interglacial cycles, as described above.

Arabian Sea

China Sea

Long-term variations of proxies of monsoonal intensity tracers, especially the species diversity of

In sediments from the South China Sea, magnetic susceptibility and calcium carbonate percentages

(c) 2011 Elsevier Inc. All Rights Reserved.

MONSOONS, HISTORY OF

started to decrease at about 7 Ma, reflecting increased deposition of terrigenous, eolian dust. This increase in dust supply indicates that monsoons started to become noticeable at that time. Both color reflectance and magnetic susceptibility show cyclic fluctuations in monsoonal intensity from 0.8 Ma on, in response to glacial and interglacial cycles (as described above). Japan Sea

Cyclic changes with large amplitude of magnetic susceptibility and illite crystallinity in sediment cores indicate that the imprint of the monsoon signature in the Japan Sea sediment record started at about 0.8 Ma. The crystallinity of illite was much higher until 0.8 Ma, reflecting that the summer monsoon was weak until that time, as it was afterward during glacials. Asian Continent

Loess and paleosol sequences on the Chinese Plateau show cyclic fluctuations in magnetic susceptibility and illite crystallinity for the last 0.6 My. Loess and paleosol sequences in the Kathamandu Basin show cyclic fluctuations in magnetic susceptibility and illite crystallinity for the last 1.1 My. The cyclic sequences of low crystallinity, representing a humid climate in the interglacial periods in both areas, is consistent with the marine sediment records of the Japan Sea. Palearctic elements first became represented in the molluskan faunas in the Siwalik group sediments in the Himalayas after 8 Ma, suggesting the beginning of seasonal migrations of water birds crossing the Himalayas at this time. The diversity of this molluskan fauna increased around 5 Ma, a time when the summer monsoon was strong. The timing of developments in the molluskan faunas in the Himalayas is thus consistent with the long-term evolution of the Asian summer monsoon as derived from marine records.

The Asian Monsoon and the Global Climate System Uplift of the Himalayas and the Tibetan Plateau occurred coeval with the increase in strength of the Asian monsoon between 9.5 and 5 Ma, as documented by the heavy mineral composition of deposits in the Bengal Fan, derived from the weathering and erosion of the rising Himalayas. Cyclic fluctuations in the strength of the summer monsoon started at about 1.1 Ma in the southern Himalayas and Tibet,

917

whereas in the Arabian Sea, South China Sea, and Japan Sea such changes started at about 0.8 Ma. Peru Margin

As a result of strong southeasterly trade winds, nutrient-rich water upwells along the Peru coast and reaches the photic zone in a belt that is c. 10 km wide, and parallels the coastline. In this region, upwelling-induced productivity is very high. The organic carbon concentration in the sediments below this high-productivity zone has been used to trace the upwelling strength in the past. Upwelling is absent or less intense during El Nin˜o/Southern Oscillations (ENSO) events, and the Southern Oscillation is linked to the Asian monsoons in the tropical Walker circulation. Upwelling along the Peru Margin started at around 3.5 Ma, as indicated by an increase in the organic carbon content, and the decrease of sea surface temperatures (derived from UK37 records). Upwelling along the Peru Margin thus started after the Asian monsoons reached their full strength at about 5 Ma. Equatorial Upwelling

The intensity of the southeast Asian monsoon controls the easterly trade winds associated with the north and south equatorial currents, and the strength of the easterly trade winds controls the intensity of equatorial upwelling. In the equatorial Pacific, the intensity of trade winds and equatorial upwelling increased at about 5 Ma (as indicated by a high abundance of siliceous and calcareous pelagic microfossils), at the time that the Asian monsoons developed their full intensity. Teleconnections

Numerous data sets show synchroneity between changes of monsoon intensity and abrupt climate shifts in Greenland during the deglaciation (between 15 and 16 ky). In addition, monsoon proxy records from the Arabian Sea reveal that the intervals of weak summer monsoon coincide with cold periods in the North Atlantic region and vice versa. All these evidences suggest some kind of teleconnection between monsoon and global climate. However, the exact physical mechanism underlying the link between high-latitude temperature changes and SW monsoon has not been addressed yet. Cyclicity of Monsoon

Paleomonsoon records show periodicity of SW monsoon at 100, 41, and 23 ky corresponding to Earth’s

(c) 2011 Elsevier Inc. All Rights Reserved.

918

MONSOONS, HISTORY OF

orbital changes of eccentricity, obliquity, and precession, respectively. Subsequently, high-resolution monsoon variability exhibits suborbital periodicities of 2200, 1700, 1500, and 775 years.

Hemisphere, and changes in monsoonal intensity trigger global climate change.

See also Conclusions The evolution of the Asian monsoon started at around 9.5 Ma, in response to the uplift of the Himalayas. The monsoonal intensity reached its maximum at around 5 Ma, and from that time the associated easterly trade winds caused intense upwelling in the equatorial Pacific. Before 1.1 Ma, the summer monsoon was strong over the Arabian Sea, whereas the winter monsoon was strong over the Japan Sea. The glacial and interglacial cycles in intensity of the monsoons in the Arabian Sea, the South China Sea, and the Japan Sea started around 0.8 Ma, coinciding with the uplift of the Himalayas to their present-day elevation. Therefore, the chronological sequence of monsoonal events and the strength of trade winds and equatorial upwelling suggest that the Asian monsoons (linked to the development of the Himalayan mountains) were an important control on global climate and oceanic productivity. The Tropics receive by far the most radiative energy from the sun, and the energy received in these regions and the ways in which it is transported to higher latitudes controls the global climate. In the Tropics, the atmospheric circulation over the Asian continent is dominated by the area of highest elevation: the Himalayas and the Tibetan Plateau. The high heat capacity of this region causes the strong seasonality in wind directions, temperature, and rainfall, involving extensive transport of moisture and thus also latent heat from sea to land during summer. The Himalayas–Tibetan Plateau thus influence the transport of sensible and latent heat from low-latitude oceanic areas to mid- and high-latitude land areas. These high mountains act as a mechanical barrier to the air currents, and the north–south contrast across these mountains varied in magnitude with the glacial–interglacial cycles from 0.8 Ma onward, at which time the glacial–interglacial climatic fluctuations reached their largest amplitude. The Asian monsoons thus control the atmospheric heat budget in the Northern

Carbon Cycle. El Nin˜o Southern Oscillation (ENSO). Holocene Climate Variability. Oxygen Isotopes in the Ocean. Somali Current. Stable Carbon Isotope Variations in the Ocean. Upwelling Ecosystems.

Further Reading Clemens SE, Prell W, Murray D, Shimmield G, and Weedon G (1991) Forcing mechanisms of the Indian Ocean monsoon. Nature 353: 720--725. Fein JS and Stephenes PL (eds.) (1987) Monsoons. Chichester, UK: Wiley. Kutzbach JE and Guetter PJ (1986) The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years. Journal of Atmospheric Science 43: 1726--1759. Naidu PD and Malmgren BA (1995) A 2,200 years periodicity in the Asian Monsoon system. Geophysical Research Letters 22: 2361--2364. Niitsuma N, Oba T, and Okada M (1991) Oxygen and carbon isotope stratigraphy at site 723, Oman Margin. Proceedings of Ocean Drilling Program Scientific Results 117: 321--341. Prell WL, Murray DW, Clemens SC, and Anderson DM (1992) Evolution and variability of the Indian Ocean summer monsoon: Evidence from western Arabian Sea Drilling Program. Geophysical Monographs 70: 447--469. Prell WL and Niitsuma N, et al. (eds.) (1991) Proceedings of the Ocean Drilling Program Scientific Results, Vol. 117. College Station, TX: Ocean Drilling Program. Street FA and Grove AT (1979) Global maps of lake-level fluctuations since 30,000 years BP. Quaternary Research 12: 83--118. Takahashi K and Okada H (1997) Monsoon and quaternary paleoceanography in the Indian Ocean. Journal of the Geological Society of Japan 103: 304--312. Wang L, Sarnthein M, and Erlenkeuser H (1999) East Asian monsoon climate during the late Pleistocene: High-resolution sediment records from the South China Sea. Marine Geology 156: 245--284. Wang P, Prell WL, and Blum MP (2000) Leg 184 summary: Exploring the Asian monsoon through drilling in the South China Sea. Proceedings of the Ocean Drilling Program, Initial Reports 184: 1--77.

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS R. P. Trask and R. A. Weller, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The need to measure ocean currents throughout the water column for extended periods in order to better understand ocean dynamics was a driving force that led to the development of oceanographic moorings. Today’s moorings are used as ‘platforms’ from which a variety of measurements can be made. These include not only the speed and direction of currents, but also other physical parameters, such as conductivity (salinity), temperature, and sea state, as well as surface meteorology, bio-optical parameters, sedimentation rates, and chemical properties. Moorings typically have three basic components: an anchor, some type of chain or line to which instrumentation can be attached, and flotation devices that keep the line and instrumentation from falling to the seafloor. Shackles and links are typically used to connect mooring components and to secure instruments in line. The choice of hardware, line, and flotation for a particular application, as well as the size and design of the anchor, depends on the type of mooring and the environment in which it is deployed. Most moorings fall into two broad categories – surface and subsurface. The main difference between the two is that the surface mooring has a buoy floating on the ocean surface, whereas the subsurface mooring does not. Although the two mooring types have similar components, the capabilities of the two are very different. With a surface buoy, it is possible to measure surface meteorology, telemeter data, and make very near-surface measurements in the upper ocean. The surface mooring, however, is exposed to ocean storms with high wind and wave conditions and therefore must be constructed to withstand the forces associated with those environmental conditions. In addition, the wave action may transmit some unwanted motion to subsurface instruments if care is not taken. The subsurface mooring, on the other hand, is away from the surface forcing and can be fabricated from smaller, lighter components, which are less expensive and easier to handle. However, it is difficult to make near-surface measurements from a subsurface mooring.

Early attempts at mooring work in the 1960s began with surface moorings. Problems with the mooring materials and the dynamic conditions encountered at the ocean surface resulted in poor performance by these early designs, and attention turned to developing subsurface moorings. The introduction of wire rope as a material for fabricating mooring lines and the advent of a remotely triggered mechanism to release the mooring’s anchor were significant milestones that helped make the subsurface mooring a viable option. It has since proved to be a very successful oceanographic tool. Interest in the upper ocean and the air–sea interface prompted a reexamination of the surface mooring design. The evolution of the subsurface mooring as a standard platform for oceanographic observations and the more recent development of reliable surface moorings are summarized here.

Subsurface Mooring Evolution Early moorings consisted of a surface float, surplus railroad car wheels for an anchor, and lightweight synthetic line, such as polypropylene or nylon, to connect the surface float to the anchor. Several kilometers of line are required for a full-depth ocean mooring, and weight of the line itself, even in water, is not negligible. Instrumentation was connected to the synthetic line along its length. The anchor was connected to the mooring line by means of a corrosible weak link. The initial method of recovering the moorings was to connect to the surface float and pull it with the hope that the tension would break the weak link leaving the anchor behind. Unfortunately, at the time of recovery, the mooring line was often weaker than the weak link and would break, allowing the line and instrumentation below the break to fall to the seafloor. Studies showed that the synthetic ropes were being damaged by fish. Analysis of many failed lines revealed tooth fragments and bite patterns that were used to identify the type of fish responsible for the damage. Statistics concerning the number of fishbites, their depths, and their locations were collected, and it was found that the majority of fishbites occurred in the upper 1500–2000 m of the water column. Prevention of mooring failure due to fish attack required lines that could resist fishbite. Ropes made of high-strength carbon steel wires were an obvious candidate. Wire rope would not only provide protection from fish attack, but also

(c) 2011 Elsevier Inc. All Rights Reserved.

919

920

MOORINGS

would have minimal stretch, unlike the synthetic ropes, and would provide high strength with relatively low drag. Many types of wire rope construction and sizes were tested, in addition to methods for terminating the wire rope; terminations are the fittings attached to the ends of wire sections. In constructing a mooring whose components can be shipped separately and handled safely on the deck of a ship at sea, the practice is to cut the wire into sections of specific lengths (shots) that allow connection to other wire shots or to instrumentation in series (end to end). A desirable termination is one that is as strong as the wire rope itself. If the technique used to terminate a rope imposes stress concentrations, which significantly reduce the strength of the wire rope, then the whole system is weakened. Methods of terminating wire include the formation of eyes into which shackles can be attached either from swaged fittings or from zinc- or resin-poured sockets. Swaged terminations utilize a fitting that is slid onto the end of the wire and pressed or swaged onto the wire with a hydraulic press. In the case of a poured-socket termination, the wire is inserted into the socket and the individual wires are splayed outward or ‘broomed out’. Once the wires are properly cleaned and positioned, a filler material (molten zinc or uncured epoxy resin) is poured into the socket and allowed to harden. A strain relief boot is often used in conjunction with a swaged fitting termination, as well as with the poured sockets. The boots are often an injection-molded urethane material designed to extend from the fitting out over a short section of the wire to minimize the bending fatigue that can occur between the flexible wire and rigid fitting. At present, galvanized 3  19 wire rope is widely used for oceanographic applications. The designation ‘3  19’ denotes three strands or groups, each with 19 individual wires: the 19 wires are twisted together to form a strand. Three strands are then wound together to form the rope. The rotation characteristics of wire rope are critically important in certain oceanographic applications. If the rope has the tendency to spin or rotate excessively when placed under tension, there is a tendency for that wire to develop loops when the tension is reduced quickly. If the load is quickly applied again to the line, the loops are pulled tight into kinks, which can severely weaken the wire rope. Wire rope with minimal rotation characteristics is preferred for mooring applications, particularly surface mooring work. Wire ropes are available with varying degrees of rotation resistance. Swivels are sometimes placed in series with the wire to minimize the chances of kink formation. In addition to using galvanized wires when fabricating the rope to provide protection

against corrosion, some wire ropes have a plastic jacket extruded over the wire. Types of plastics used for jacketing materials include polyvinyl chloride, polypropylene, and high-density polyethylene. In the early years, mooring recoveries that were initiated by pulling on deteriorated mooring lines often resulted in line breakage and instrument loss. A preferable approach was to detach the mooring from its anchor prior to hauling on the mooring line. This would reduce the load on the mooring line since the line would never ‘feel’ the weight of the anchor nor the tension required to pull the anchor out of the bottom sediments. This approach became possible with the development of an acoustically commanded anchor release. The acoustic release is deployed inline on the mooring and is typically positioned below all instrumentation and close to the anchor. To activate the release mechanism, a coded acoustic signal is sent from the recovery vessel. The acoustic release detects the signal and disconnects from the anchor. When mooring work was in its infancy, the surface buoy was a vital, visible link to the mooring below. Without it, the exact location of the mooring was unknown. The introduction of acoustic releases not only provided a way to disconnect the line and instruments from the anchor, but also provided a way to locate the exact position of the mooring by acoustic direction finding. This eliminated the need for a surface float, which at that time, before the development of meteorological sensors for buoys, was used solely for recovery purposes. Instead of having a mooring that stretched all the way from the ocean bottom to the surface, the mooring was shortened so that the top of the mooring was positioned below the surface of the water. Sufficient buoyancy was placed at the top of the mooring to keep all of the mooring components as vertical as possible throughout the water column. With this design, which became known as the subsurface mooring, the mooring would ascend to the surface once it was acoustically released from its anchor. At the surface, it could be pulled out of the water by the waiting recovery vessel. A great advantage of the subsurface mooring is having the hardware below the ocean surface, which is the most dynamic part of the water column. As a result, there is a considerable reduction in component fatigue due to surface-wave action. In addition, the mooring is no longer visible to surface vessels and is less vulnerable to vandalism. The buoyancy used on subsurface moorings is usually in the shape of a sphere, because of its low drag coefficient. Other shapes have been used depending on the specific application. Various materials are used, including steel spheres, glass spheres with protective plastic covers, ceramic spheres, and

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

syntactic foam spheres. Syntactic foams consist of small pressure-resistant glass microspheres (2–300 mm in diameter), as well as larger glass fiber-reinforced spheres (0.15–10 cm in diameter) embedded in a thermosetting plastic binder. An advantage of the syntactic foam is that it can be molded to form custom shapes. Unlike the steel spheres, whose use is depth-limited (maximum working depth is approximately 1000 m), the syntactic foam can be engineered to withstand full ocean-depth pressures. In addition, it can be designed to provide the same buoyancy as a string of glass balls (commonly 43 cm in diameter) with considerably less drag. This makes syntactic foam spheres attractive for use in high-current regimes, since the drag on the mooring will be less; consequently, less buoyancy will be needed to keep the mooring near-vertical. Subsurface moorings with a single element of buoyancy at the top are still at some risk. Should the buoyant element be lost or damaged, the mooring would fall to the bottom, leaving no secondary means of bringing it back to the surface when the acoustic release mechanism is activated. To provide a higher degree of reliability, buoyancy is often provided in the form of glass balls attached along the length of the mooring. In addition, buoyancy is often added to the bottom of the mooring, just above the acoustic release, to provide what is sometimes referred to as the ‘backup recovery’. With this design feature (Figure 1), should a mooring component fail and the upper part of the mooring be lost, no matter where the failure occurs, there should be sufficient buoyancy below that point to bring the remaining section of the mooring back to the surface. Instrumentation that would otherwise have been lost if deployed with a single buoyant element is recoverable with this configuration. Equally important, the recovery provides the opportunity to identify the failed component and correct the problem. As a recovery aid, pressure-activated submersible satellite transmitters are frequently installed on the upper buoyancy sphere. In the event of a mooring component failure that causes the top of the mooring to surface, it can be tracked via satellite. This allows for possible recovery of whatever instrumentation hangs below. The number of measurements made from a subsurface mooring depends on several factors including the load the mooring is designed to support and the resources available for instrumenting the mooring. Depending on the resources, a mooring may have on the order of 10 instruments located at discrete depths. If one or more of the instruments malfunction during their deployment, the data from that particular depth are missing, but there may be other

95 m

921

1.52-m syntactic sphere with radio, light, satellite beacon 3 m 1.27 cm chain

100 m

Vector averaging current meter (VACM) 292 m 0.48 cm wire Glass float transponder (3) Glass flotation balls in plastic hardhats

401 m

VACM 393 m 0.48 cm wire (3) Glass flotation balls in plastic hardhats

800 m

VACM 591 m 0.48 cm wire (3) Glass flotation balls in plastic hardhats

1399 m

VACM 591 m 0.48 cm wire (3) Glass flotation balls in plastic hardhats

2000 m

VACM 491 m 0.48 cm wire (4) Glass flotation balls in plastic hardhats

2500 m

VACM 281 m 0.48 cm wire (14) Glass flotation balls in plastic hardhats

2799 m

2804 m

VACM 2 m 0.95 cm chain Acoustic release 5 m 0.95 cm chain 156 m 0.64 cm wire 20 m 1.9 cm nylon 5 m 0.64 cm wire 5 m 1.27 cm chain Anchor

3000 m deep

Figure 1 A typical subsurface mooring design. Design by S. Worrilow.

instruments possibly above and below the failed unit that continue to collect data. Such a design requires numerous lengths of terminated wire, which must be deployed in a particular order so the individual instruments end up at the desired depths. An alternative approach is to have a single instrument that uses the mooring wire as a guide along which an instrument moves up and down profiling

(c) 2011 Elsevier Inc. All Rights Reserved.

922

MOORINGS

the water column on a predetermined schedule. In this configuration (Figure 2), the bulk of the mooring consists of a single continuous length of wire, eliminating the need for multiple lengths used between discrete instruments. An advantage of such a system is that the data are collected from the entire sampling area and not from specific depths, which, due to their deployed location, may not be in an optimal position for observing an interesting phenomenon. One profiling instrument has the potential to replace many instruments deployed along the mooring. If, however, the mooring only carries one profiling instrument and that instrument malfunctions, there

Subsurface buoy

is a chance that the mooring may not collect any data. For that reason, it is not uncommon for a mooring to have a combination of instrument types, which include profiling as well as several discrete instruments. Duplicate measurements by different instruments can improve the chances of collecting a full data set. Since redundant moorings are seldom an option, a mooring failure can have a catastrophic impact on the total data return. With both instruments and mooring components, attention to detail is critical. The care taken in preparing and testing instrumentation, in the selection of quality mooring components, and in the fabrication techniques utilized is often a deciding factor. Other factors that can impact a mooring’s success include the quality of the information that went into the design of the mooring, that is, how similar were the predicted environmental conditions to the actual conditions encountered? Is the design unique or is it a variation of a design that has historically worked well? Attention to detail during the actual deployment of the mooring, as well as uncontrollable outside influences such as fishing activities in the area, can also be contributing factors. Despite all the variables, it is possible to routinely achieve success rates that are greater than 90%.

Top stop

Advances in Surface Mooring Technology

Wire rope

Profiler

Bottom stop

Backup glass ball flotation

Acoustic releases

Anchor

Figure 2 A subsurface mooring schematic with a profiling instrument that runs up and down the mooring line on a predetermined schedule. Illustration by J. Doucette/WHOI Graphics.

Growing interest in understanding interactions between the ocean and the atmosphere has rekindled interest in using surface moorings. The surface mooring is a unique structure. It extends from above the surface to the ocean bottom, providing a platform from which both meteorological and oceanographic measurements can be made in waters that range from shallow to 5 km in depth. Surface-mooring designs must consider the effects of surface waves, ocean currents, biofouling, and other factors that can vary with the time of year, location, and regional climate and weather patterns. The success of a surfacemooring deployment often depends on the abilities both to accurately estimate the range of conditions that the mooring may encounter while deployed and to design a structure that will survive those conditions. The primary goal of any mooring deployment is to keep the mooring on location and making accurate measurements. Adverse environmental conditions not only influence the longevity itself but also impact the instruments that the mooring supports. It is often very difficult to keep the instruments working under such conditions for long periods. Surface moorings are used to support submerged oceanographic instrumentation from very close to

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

the surface (sometimes floating at the surface) to near the bottom, which is typically 5 km in depth. Measurements of physical properties, such as temperature, velocity, and conductivity (salinity), as well as of biological parameters, such as photosynthetically available radiation (PAR), beam transmission, chlorophyll fluorescence, and dissolved oxygen, are routinely made from surface moorings. The surface buoy also provides a platform from which meteorological measurements can be made and a structure from which both surface- and subsurfacecollected data can be telemetered via satellite. Meteorological sensors typically deployed on a surface buoy measure wind speed, wind direction, air temperature, relative humidity, barometric pressure, precipitation, and long-wave and short-wave radiation. The meteorological data are stored in memory and telemetered via satellite to a receiving station ashore. The telemetered data often play an important part in real-time analysis and reaction to conditions on site. The data can also be passed to weather centers for forecasting purposes. There are a number of different types of surface buoys. Some shapes have been in use since the early days of mooring work, and others are relatively new. Buoy shapes include the toroid or ‘donut’, the discus, and the hemispherical hull. The toroid hull in various configurations is widely used throughout the scientific community. Where a significant amount of instrumentation must be supported, a discus-shaped buoy with as much as 6800 kg of buoyancy may be used for both deep- and shallow-water applications. Some buoys are designed with modular buoyancy elements that can be added to maximize the available buoyancy. The discus buoy design is widely used by the US National Data Buoy Center in Mississippi in coastal waters, at the Great Lakes stations, and for directional wave measurements. The 3-m discusshaped hull was also adopted by the Atmospheric Environment Service in Canada for its coastal buoys. Smaller discus-shaped hulls are used for shallowwater applications. Buoy hulls are made of aluminum, steel, fiberglass over foam, and various closed-cell foams. Several closed-cell foams are extremely resistant to wear and have low maintenance. Ionomer foam and polyethylene foam are common materials for buoy and fender applications. Depending on the material, various outer skin treatments are used to increase the hull’s resilience to wear. These include the application of heat and pressure, as well as bonding a different material, such as urethane, to the exterior of the hull. The mooring materials used on surface moorings resemble those used on the subsurface moorings.

923

Component sizes are usually increased to compensate for the larger forces and the increased wear. Materials include chain, plastic-jacketed wire rope, and synthetic line. Chain is used directly beneath the buoy for strength, ease of handling, and, because of its additional mass, stabilization of the buoy during its deployment. If the water is sufficiently deep and the design permits, the wire rope is usually extended to a depth of at least 1500 m and often as deep as 2000 m for fishbite protection. The surface mooring needs some form of built-in ‘compliance’ (ability to stretch) to compensate for large vertical excursions that the buoy may experience during the change of tides and with passing waves and swell. The compliance also compensates for the buoy being displaced laterally on the surface by the drag forces associated with ocean currents and prevents the buoy from being pulled under when such forces are applied. In deep-water applications, compliance is provided through the use of synthetic materials, such as nylon. The synthetic line acts like a large rubber band that stretches as necessary to maintain the connections between the surfacefollowing buoy and the anchor on the bottom. A challenge in the design process, particularly in shallow water, is to achieve an appropriate mix of compliant materials and fishbite-resistant materials, which tend to be unstretchable. The ‘scope’ of the mooring – the ratio of the total unstretched length of the mooring components to the water depth – can be one of the sensitive design factors. A mooring with a scope of less than 1.0 relies on the stretch of the nylon for the anchor to reach the bottom. Such a taut mooring remains fairly vertical with a relatively small watch circle (the diameter of the area on the ocean surface where the buoy can move about while still anchored to the ocean bottom), but it carries a penalty: such a vertical mooring is under considerable tension, or ‘preloaded’, at the time of deployment. Currents and waves impose additional loads beyond the initial preloaded condition. Moorings with scopes between 1.0 and c. 1.1 are generally referred to as ‘semi-taut’ designs. The mooring shown in Figure 3 is typical of a semi-taut design. Early surface moorings were designed using only a static analysis program, which used steady-state current profiles as input to predict mooring performance. However, experience has shown that it is necessary to consider the combined effects of strong currents and surface waves. An investigation of the dynamic effects of surface forcing on the performance of surface moorings found that semi-taut moorings could have a resonant response to forcing in the range of surface wave periods, causing high dynamic loads. These high tensions limit the

(c) 2011 Elsevier Inc. All Rights Reserved.

924

MOORINGS

3-m discus buoy with meteorological

Discus buoy with meteorological sensors

sensors, tension recorder, and satellite transmitter

Bridle with temperature sensor, SEACAT, tensiometer ,

10 m

3 m 1.9 cm chain 3.9 m 1.27 cm chain Vector measuring current meter (VMCM)

3.5 m 4.5 m 5m

5.5 m 0.95 cm wire rope 20 m

Vector averaging current meter (VACM)

30 m

VMCM

9.5 m 0.95 cm wire rope 5.5 m 0.95 cm wire rope 40 m

80 m

VACM 39.5 m 0.95 cm wire rope VMCM

dissolved oxygen and backup Argos transmitter 0.4 m 1.9 cm Syst. 3 chain 0.4 m 1.9 cm Syst. 3 chain

Temperature logger

Vector measuring current meter (VMCM) and temperature pod (Tpod) 2 m 1.9 cm Syst. 3 chain

15 m

Multivariable moored system (MVMS): current, temperature, dissolved oxygen, light, and other sensors 2.5 m 1.9 Syst. 3 chain VMCM

20 m

TPod

25 m

VMCM

30 m

TPod

35 m

MVMS

40 m

Acoustic instrument with TPod

45 m

VMCM

50 m

TPod

55 m

VMCM

60 m

TPod

65 m

MVMS

10 m

7.25 m 1.27 cm wire

7.25 m 1.27 cm wire

2.5 m 1.9 cm Syst. 3 chain

2.5 m 1.9 cm Syst. 3 chain

37.6 m 0.95 cm wire rope

120 m

VMCM 37.4 m 0.95 cm wire rope

160 m

VMCM

72.5 m

537.2 m 0.95 cm wire rope 700 m

2000 m

VMCM 1300 m 0.79 cm wire rope 1000 m 2.06 cm nylon 1500 m 1.9 cm nylon

80 m 90 m 100 m 125 m 150 m 175 m 200 m 225 m 250 m 300 m

7.25 m 1.27 cm wire

7.25 m 1.27 cm wire

12.1 m 1.27 cm wire

TPod MVMS

17.5 m 1.11 cm wire TPod SEACAT conductivity/temperature sensor 49 m 1.1 cm wire TPod SEACAT 49 m 1.1 cm wire TPod SEACAT 49 m 1.1 cm wire TPod SEACAT 100 m 1.1 cm wire TPod

(2) Glass flotation balls in plastic hardhats

1400 m 0.95 cm wire

Special wire/nylon termination

100 m 0.95 cm wire 500 m 2.22 cm nylon

one piece

732 m 2.22 cm nylon 500 m 2.22 cm nylon 100 m 2.54 cm nylon one piece 500 m 2.86 cm polypro 600 m 2.86 cm polypro

1500 m 1.9 cm nylon

1.27 cm trawler chain

(57) Glass flotation balls in plastic hardhats

(82) Glass flotation balls in plastic hardhats

2 m 1.27 cm chain Acoustic release 5 m 1.27 cm chain 20 m 2.54 cm nylon 5 m 1.27 cm chain Anchor 5400-m deep

Acoustic release

Anchor

5 m 1.27 cm trawler chain 5 m 1.27 cm trawler chain 20 m 2.54 cm Samson nystron 5 m 1.27 cm trawler chain

4032-m deep

Figure 3 A semi-taut surface mooring design. Design by P. Clay.

Figure 4 An inverse catenary mooring design. Design by G. Tupper.

instrument-carrying capacity of the mooring and can lead to failure of mooring components. An alternative design fashioned after the US National Data Buoy Center ‘inverse catenary’ mooring has evolved in response to difficulties encountered using taut surface mooring designs. With wire rope in the upper part of the mooring and with nylon line spliced to a buoyant synthetic line such as polypropylene below, the inverse catenary design (Figure 4)

offers larger scope (typically 1.2) for high-current periods, yet performs well in lesser currents. In low currents, the positively buoyant synthetic line keeps the slightly negatively buoyant nylon from tangling with the rest of the mooring below it. Thus, the inverse catenary design can tolerate a wider range of environmental conditions. The inverse catenary design lowers the static mooring tension, as shown in Table 1. The dynamic tension contribution to the total

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

0

Table 1 A comparison of semi-taut surface mooring and an inverse catenary design (subjected to the same ocean current forcing) Inverse catenary

1.109 2065

1.285 1602

463

2292

1783

509

1208

1735

527

Semi-taut design Tension at buoy : 1565 kg Tension at anchor : 1777 kg

Difference

500

Depth (m)

Mooring scope Tension at the buoy (kg) Anchor tension (kg) Horizontal excursion (m)

Semi-taut

925

1000 Inverse catenary design Tension at buoy : 1241 kg Tension at anchor : 1400 kg 1500

tension, however, is unchanged, and care must still be taken in the design process to prevent the mooring from having a resonant response to forcing in the range of surface wave periods. In some regions of the world’s oceans, the dynamic loading due to high wind and sea state conditions may be so severe that ultimate strength considerations are superseded by the fatigue properties of the standard hardware components. In these cases, in addition to appropriate mooring design, attention must be paid to the choice and preparation of mooring hardware. Cyclic fatigue tests revealed that, in certain applications, mooring hardware that had been used reliably in the past lost a significant part of its service life owing to fatigue and either failed or showed evidence of cracks. Where possible, different hardware components that are less susceptible to fatigue failure in the range of expected tensions are now substituted. In situations where there is no replacement hardware available, the fatigue performance is improved by shot peening. Shot peening is a process whereby a component is blasted with small spherical media, called shot, in a manner similar to the process of sand blasting. The medium used in shot peening is more rounded rather than angular and sharp, as in sandblasting. Each piece of shot acts like a small ballpeen hammer and tends to dimple the surface that it strikes. At each dimple site, the surface structure of the material is placed in tension. Immediately below the surface of each dimple, the material is highly stressed in compression so as to counteract the tensile stress at the surface. A shot-peened part with its many overlapping dimples, therefore, has a surface layer with residual compressive stress. Cracks do not tend to initiate or propagate in a compressive stress zone. Since cracks usually start at the surface, a shot-peened component will take longer to develop a crack, thereby increasing the fatigue life of the part. With both the semi-taut and the inverse catenary surface mooring designs, it is difficult to make

2000

0

500 1000 1500 Horizontal excursion (m)

2000

Figure 5 A comparison of the shape of a semi-taut design with that of an inverse catenary design when subjected to the same ocean current forcing. Note the differences in buoy and anchor tensions as well as the horizontal excursions at the surface.

deep-current measurements because the mooring line at these depths is sometimes inclined more than 151 from vertical. This is a problem for two reasons: first, some instruments fitted with compasses do not work well if the compass is inclined more than 151; and second, some velocity sensors require the instrument to be nearly vertical. An inverse catenary mooring, with its greater scope, has inclination problems at shallower depths as compared to the semi-taut design. Figure 5 compares the mooring shape of a semitaut design with that of an inverse catenary mooring subjected to the same environment conditions. In addition to the inclination problem, there is also a depth-variability problem. Compliant members on a surface mooring are usually synthetics, which must be placed below the fishbite zone (nominally 2000-m depth). The deep instruments are, therefore, in line in the synthetics; their depth can vary by several hundred meters depending on the stretch of the material. A pressure sensor on the instrument can be used to record the instrument depth, but if a particular depth is desired, it is not possible with the conventional design. Hence, the trade-off for being able to withstand a wider range of environmental conditions is a reduction in the depth range for making certain kinds of measurements. A partial solution to the problem of deep measurements on a surface mooring is illustrated in the mooring design shown in Figure 6, which combines features of both the subsurface- and inverse

(c) 2011 Elsevier Inc. All Rights Reserved.

926

MOORINGS

0 3-m discus buoy with meteorological sensors and satellite telemetry

Combination inverse catenary and subsurface design Buoy tension = 1438 kg 3500 m VMCM angle = 8°

1000 Bridle with T-pod, IMET and VAWR temperature sensors, and tensiometer 6.7 m 1.9 cm chain Vector measuring current meter (VMCM)

30 m

VMCM

17 m 0.95 cm wire

17 m 0.95 cm wire 50 m

2000 Depth (m)

10 m

3000 Inverse catenary design Buoy tension = 1471 kg 3500 m VMCM angle = 25°

VMCM 17 m 0.95 cm wire

70 m

VMCM

90 m

VMCM

110 m

VMCM

17 m 0.95 cm wire

4000

5000

17 m 0.95 cm wire

6000 37.3 m 0.95 cm wire

150 m

VMCM

200 m

VMCM

300 m

VMCM

310 m

VMCM

47.3 m 0.95 cm wire

97 m 0.95 cm wire

0

1000

2000 3000 4000 5000 Horizontal excursion (m)

6000

Figure 7 A comparison of the shapes of two mooring designs subjected to the same ocean current forcing, showing the differences in mooring inclination at 3500-m depth.

7.3 m 0.95 cm wire 200 m 0.95 cm wire 237 m 0.79 cm wire 750 m

VMCM

1500 m

VMCM

500 m 0.79 cm wire 252 m 0.79 cm wire

3500 m

400 m 0.79 cm wire 100 m 0.79 cm wire one piece 50 m 2.06 cm nylon Special wire/nylon termination 450 m nylon one piece 500 m 2.54 cm polypro 1.27 cm trawler chain (47) Glass flotation balls Tensiometer with safety chain 10 m 0.95 cm wire VMCM

2078 m 0.95 cm wire

1.27 trawler chain (52) Glass flotation balls 1.27 cm trawler chain Acoustic release 5 m 1.27 cm trawler chain 20 m 2.54 cm nylon 5 m 1.27 cm trawler chain Anchor 5670-m deep

Figure 6 A mooring design that combines the features of an inverse catenary mooring with those of a semi-taut mooring in order to improve the quality of deep (3500 m) current measurements made from a surface mooring. Design by G. Tupper.

catenary-type moorings. The upper 2000 m of the mooring is similar to any surface mooring, with the instrumentation at the appropriate depths and wire rope in between. The lower part of the mooring from the bottom up to the 3500 m instrument is all wire with a cluster of glass ball flotation just above the

release near the anchor and another immediately above the 3500 m instrument. The compliance of the mooring consists of 1500 m of nylon and polypropylene inserted between the 3500 m instrument and the base of the wire at 2000 m. The combination of nylon and polypropylene gives the mooring enough stretch (from the nylon) and built-in buoyancy (from the polypropylene) to handle the range of expected current conditions. The polypropylene actually performs a double duty in that during low current periods the buoyant polypropylene keeps the excess nylon from tangling with the lower part of the mooring; when the currents increase, that buoyant member becomes available in the form of extra scope. The shape of the combination mooring design is compared with the shape of an inverse catenary design in Figure 7. Surface moorings can also be used as a communications link between the ocean surface and points along the mooring all the way to the seafloor. With the appropriate mooring components, information can be passed in both directions. Data collected by instrumentation deployed on the mooring line or in close proximity to the mooring can be sent to the buoy, where it is transmitted via satellite to a receiving station. Two-way communications between a shore station and the buoy via satellite permit buoy systems to be reprogrammed so as to modify instrument sampling schemes, as well as to repair system malfunctions. Being able to diagnose and repair a buoy data collection system without having to dispatch a vessel to the site is of great value.

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

Telemetry of data from subsurface instruments on surface moorings is possible through various techniques. One approach is to utilize electromechanical (EM) cable for the transmission of an electrical signal. EM cables typically have two elements, a strength member and conductors. One type of EM cable has electrical conductors in the center with an outer armor of steel wire that provides strength, as well as fishbite protection of the conductors. Another EM cable design that has been used successfully utilizes 3  19 oceanographic cable with three conductors laid in the valleys that are formed by the three strands. The plastic jacket is then extruded over the wire and conductors. A disadvantage of this design is that the conductors are on the outside of the strength member and are more susceptible to fishbite damage than in a cable with the outer armor. Mooring cables are also being designed with optical fibers to take advantage of their capability to transmit over long distances with minimal losses, and its inherently high data-carrying capacity. Not all communications rely solely on special purpose cables. Through the use of high-speed acoustic modems, signals can be sent from instrumentation deployed on the ocean floor to an acoustic transducer

927

located on the mooring. Those signals are then transmitted up the mooring line to the surface buoy and then to a receiving station via satellite. Data from ocean-bottom seismometers that record undersea earthquakes have been monitored in this manner as part of a prototype tsunami-warning network. Figure 8 depicts ocean-bottom seismographs communicating acoustically with a surface mooring that is linked with a satellite network. The interface where a mooring transitions from just below the ocean surface to a buoy on the surface is a challenging area in the design of a mooring because the components used in that section are subject to excessive wear and fatigue failure from the constant motion of the surface buoy. To reduce the wear at this juncture, a universal joint is employed between the buoy and the mooring. It is extremely challenging to pass cables with electrical conductors through this interface. The universal joint is configured with a central hole, which provides an unbending pathway for conductors that must pass through the universal. Getting the vulnerable conductors through this dynamic near-surface region and into the buoy is not a trivial part of the design. Several approaches have been taken with reasonable success.

Link to satellite network

Acoustically linked surface buoy

Acoustic transducer

Hydrothermal vent sensors, ADCP, and modem

Glass floats Acoustic release

Ocean-bottom seismograph and modem

Figure 8 A moored buoy system configured to receive acoustically transmitted data from seafloor instruments in near-real time and able to communicate data and commands between shore-based labs and the observatory via a satellite network. Illustration by J. Doucette/WHOI Graphics.

(c) 2011 Elsevier Inc. All Rights Reserved.

928

MOORINGS

One technique is to utilize a special chain assembly directly below the buoy consisting of chain that is wrapped with a spiral shaped multiconductor cable and completely encapsulated with urethane. This component has the required strength and flexibility from the chain while the spiraled cable configuration and urethane protects the conductors from bending strain produced by the buoy motion. Another approach that has been used, particularly in shallow-water applications, is an ultrastretchy rubber hose that is reinforced with nylon and is capable of stretching to twice its unstretched length. Electrical conductors are embedded in the wall of the hose. The angle that the conductors make with respect to the axis of the hose is a critical part of the hose design in order to allow the hose to stretch and not damage the conductors. The system also allows power generated by solar panels on the surface buoy to flow to the bottom instruments. It is possible to use conventional plastic-jacketed wire rope, without copper electrical conductors, to send a signal up the mooring cable through the use of inductively coupled modems. In such a system, the signal is applied to the primary winding of a toroidal transformer. The mooring wire only has to pass through the toroid, to form a single-turn secondary that conveys the data along the mooring cable. The ends of the mooring cable are grounded to the seawater, which permits a current to flow through the mooring wire and seawater. So as not to have to thread the mooring cable through the toroid, it is split and clamped around the cable. It is not necessary to break the cable at the instrument position or to provide any electrical connection between the sensor and the cable. Another advantage of this system is the flexibility it offers for sensor placement. Since it does not require discrete cable lengths, sensors can be clamped along the wire at any location and easily repositioned if necessary. Surface moorings have capabilities which make them good candidates for use with ocean observatories. The moorings have the flexibility to be deployed nearly anywhere in the world’s oceans and offer near-real-time access due to the satellite communications capability from the surface buoy. Moored observatories can vary in complexity depending on their intended application. If data must be collected from high-bandwidth sensors, such as those used in acoustic arrays or for continuous seismic monitoring, then high-speed satellite links may be required in conjunction with fiber optic cables connected to junction boxes on the seafloor. Such systems typically require large quantities of power to operate and could require active power generation. The buoy size increases to accommodate such

capabilities as does the mooring hardware necessary to keep the structure on station. Other systems, which do not require high-speed satellite communications or have only a moderate number of sensors, do not have the same power requirements as a larger system and can make use of smaller moored platforms with smaller hardware.

Mooring Deployments Deep-ocean surface and subsurface moorings are typically deployed using an anchor-last technique. As the name implies, the anchor is the last component to be deployed. The entire mooring, starting at the top, is put over the side and strung out behind the deployment vessel and towed into position. At the appropriate location, the anchor is dropped. If the current and the wind are from the same direction, the deployment begins by positioning the ship down-current of the desired anchor-drop position. By doing this, the ship can maintain steerage as it slowly steams against the current while the mooring components are deployed and are carried away from and behind the ship by the wind and current. When the wind and current are opposing each other, it becomes necessary to alter the deployment plan. In such cases, the important factor is the relative speed of the ship, which is typically 50–100 cm s 1 through the water. Depending on the length of the mooring, its complexity, and the wind and current conditions, the start position could be as much as 10 km from the anchor-drop position. The goal is to put the mooring line over the side at a rate that is slightly less than the ship’s speed through the water and, thus, have the entire mooring stretched out without kinks and loops behind the ship by the time it arrives at the anchordrop site. With the ship at the position for the start of the deployment sequence, the upper buoyancy of the mooring is lowered into the water. Figures 9(a)–9(g) illustrate the deployment sequence of a deep-water surface mooring. The mooring components are attached in series and paid out with the assistance of a winch. Instruments are attached to the mooring at the appropriate locations between premeasured mooring line shots. The last component put in line is the anchor. The ship tows the mooring into position with the anchor still on deck and actually steams past the desired anchor position by a distance equal to approximately 7–10% of the water depth. The anchor is deployed by either sliding it off the deck of the ship and into the water by means of a steel tip plate or it is placed into the water with a crane and

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

(b)

(a)

Stopped off on deck

From mooring winch on deck

(c)

Deploying additional upper instruments Stopped off on deck

Initial instrument deployment

From mooring winch on deck

From trawl winch

Start of buoy deployment Quick-release hook

From mooring winch on deck

(d) From trawl winch

929

(e) Buoy outboard of ship ready to be lowered

Buoy released and in tow

Quick-release hook

From mooring winch on deck

From mooring winch on deck

(g)

(f)

Glass ball buoyancy attached

Anchor away

Figure 9 Surface mooring deployment sequence. (a) The first instrument is lowered into the water. (b) Instrumentation in the upper part of the mooring is lowered into the water before deploying the buoy. (c) The upper part of the mooring is attached to the surface buoy. (d) The surface buoy is placed into the water. (e) The ship steams forward slowly as additional mooring line and instrumentation are deployed. (f) The entire mooring is in tow behind the ship as the glass ball buoyancy is deployed. (g) The anchor free-falls to the ocean bottom, pulling the buoy along the surface.

mechanically released once it is just below the surface. As the anchor falls to the bottom, the mooring is pulled under with it. The mooring line takes the path of least resistance, following the anchor as it descends, resulting in the top of the mooring moving toward the anchor-drop position as the anchor falls to the bottom. The normal drag on the mooring line is greater than the tangential drag; therefore, a watersheave effect takes place as the anchor falls to the bottom. The anchor does not, however, fall straight down but rather falls back a distance equal to a small percentage of the water depth, hence the reason for steaming past the desired anchor position before deploying the anchor. Depending on the design of the

mooring, the anchor can fall at a rate of approximately 100 m min 1. Some situations make anchor-last deployments difficult or near impossible. An application requiring the deployment of moorings through the ice in high latitudes is one example where an anchor-first deployment could be used. In these cases, there is not enough open water to completely stretch the mooring out behind the vessel. Instead, the vessel breaks the ice where the mooring is to be deployed, creating a small pool. Starting with the anchor, the mooring is deployed vertically through the opening in the ice. The upper buoyancy is the last component to go into the water and the mooring is allowed to drop

(c) 2011 Elsevier Inc. All Rights Reserved.

930

MOORINGS

straight down to the bottom. Such moorings have to be designed so that each component can support the total weight of all the components below it, including the anchor. Larger wire sizes with greater breaking strengths are typically needed for anchorfirst deployments. Another application requiring an anchor-first deployment is a mooring that must be deployed in a specific depth or at a precise location. With the mooring hanging below the deployment ship and the anchor close to the bottom, the vessel can be maneuvered to the desired location, at which time the mooring is allowed to free-fall the short distance to the bottom.

Discussion and Summary All moorings have similar components, but each design is unique. Factors such as the mooring’s intended use, the environment in which it will be deployed, the water depth, the payload it must support, and the deployment period greatly affect the design. Although we have discussed vertical arrays, moorings can assume a variety of orientations. For some applications, a U-shaped array (Figure 10) is required, or a mooring

may require multiple legs to provide stability and to minimize mooring motion. A horizontal mooring (Figure 11) may be needed to investigate spatial variability. Moorings provide one means of collecting temporal and spatial data. Oceanographic gliders, which are able to freely move both vertically and horizontally through adjustments in buoyancy and attitude control, offer another approach to collecting spatial data. Gliders can be configured to travel long distances (B3000 km) for extended periods (B200 days) while making approximately 600 vertical excursions. With the proper configuration, gliders could make the vertical and horizontal measurements typical of a moored array. Each approach has its limitations (power constraints, deployment duration, and spatial variability) and, depending on the application, the most appropriate technique should be chosen. The ability to model mooring performance both statically and dynamically now permits extensive design studies before the mooring is taken into the field. As a result, it is possible to explore new designs and have greater confidence in how they will perform prior to cutting any wire or splicing any line. It is important to point out, though, that regardless of the

Radar reflector/marker

Radar reflector light/Argos Surlyn telemetry buoy

Urethane chain 0.63 cm polypropelene 25.3 m

0.95-m steel sphere 2 m 0.95 cm chain/swivel WHOI navigator

S-tether EM cable

30.5 m 5 m 0.95 cm chain in Tygon tubing

16-element hydrophone array Panther Plast floats 150 m 1.9 cm nylon

150 m 1.9 cm nylon

84 m

Yale grip and flounder plate Acoustic release/pinger Anchor sled

Yale grip and flounder plate Acoustic release 2 Glass flotation balls Anchor

Figure 10 A U-shaped moored hydrophone array. Design by J. Kemp.

(c) 2011 Elsevier Inc. All Rights Reserved.

MOORINGS

MicroCAT TCP

Glass flotation balls

160 m 158 m 0.95 cm wire rope

931

20 m

Panther Plast floats

1.22 m steel sphere 2 m 1.27 cm trawler chain

Ax pack 2 m 1.27 cm trawler chain

3D ACM

SBE 39

44 m 0.95 cm wire rope

62 m 0.95 cm 64 m wire rope

Brancker TP recorder 15 m 1.27 cm trawler chain 2000 lb. 1000 lb. 909 kg 454 kg Dor-Mor Dor-Mor anchor anchor

Weight

MicroCAT TC

BACS acoustic release 5 m 1.27 cm trawler chain

Swivel 15 m 1.27 cm trawler chain 15.8 m steamer chain

10 m 1.9 cm chain

Figure 11 A two-dimensional moored array. TCP, temperature, conductivity, and pressure; BACS, binary acoustic command system; SBE, Sea-Bird Electronics; ACM, acoustic current meter; Ax pack, acceleration package. Design by R. Trask.

amount of time spent designing, modeling, and fabricating a mooring, the success of a deployment will often come down to the ability of trained personnel to pay close attention to all the details and to get the mooring safely in and out of the water while working under extremely adverse conditions at sea.

See also Air–Sea Gas Exchange. Drifters and Floats. Gliders. Ocean Circulation. Seismology Sensors. Sensors for Mean Meteorology. Tsunami. Wave Energy.

Further Reading Berteaux HO (1991) Coastal and Oceanic Buoy Engineering. Woods Hole, MA: Henri Berteaux.

Berteaux HO and Prindle B (1987) Deep sea moorings: Fishbite handbook. Woods Hole Oceanographic Institution Technical Report WHOI 87-8. Woods Hole, MA: WHOI. Dobson F, Hasse L, and Davis R (eds.) (1980) Air–Sea Interaction: Instruments and Methods. New York: Plenum. Frye D, Ware J, Grund M, et al. (2005) An acousticallylinked deep-ocean observatory. Proceedings of Oceans 2005 – Europe, vol. 2, pp. 969--974. Woods Hole, MA: WHOI. Morrison AT, III, Billings JD, and Doherty KW (2000) The McLane Moored Profiler: An autonomous platform for oceanographic measurements. In: OCEANS 2000 MTS/ IEEE Conference and Exhibition, Providence, RI, 11 Sep.–14 Sep. 2000, vol. 1, pp. 353–358. East Falmouth, MA: McLane Research Laboratories Inc. Warren BA and Wunsch C (eds.) (1981) Evolution of Physical Oceanography, Scientific Surveys in Honor of Henry Stommel. Cambridge, MA: Massachusetts Institute of Technology Press.

(c) 2011 Elsevier Inc. All Rights Reserved.