The Community Ecology of Sea Otters

Ecological Studies, Vol. 65 Analysis and Synthesis

Edited by

W D. Billings, Durham, USA F. Golley, Athens, USA 0. L. Lange, Wiirzburg, FRG J. S. Olson, Oak Ridge, USA H. Remmert, Marburg, FRG

Ecological Studies

Volume 52 Air Pollution by Photochemical Oxidants Formation, Transport, Control and Effects on Plants Edited by Robert Guderian 1985. XI, 346p., 54 figures. cloth ISBN 3-540-13966-4

Volume 53 The Gavish Sabkha A Model of a Hypersaline Ecosystem Edited by G.M. Friedman and W.E. Krumbein 1985. X, 484p., 246 figures. cloth ISBN 3-540-15245-8

Volume 54 Tidal Flat Ecology An Experimental Approach to Species Interactions By Karsten Reise 1985. X, 198p., 69 figures. cloth ISBN 3-540-15447-7

Volume 55 A Eutrophic Lake Lake Mendota, Wisconsin By Thomas D. Brock 1985. XII, 308p., 82 figures. cloth ISBN 3-540-96184-4

Volume 56 Resources and Society A Systems Ecology Study of the Island of Gotland, Sweden By James J. Zucchetto and Ann-Mari Jansson 1985. X, 248p., 70 figures. cloth ISBN 3-540-96151-8

Volume 57 Forest Ecosystems in the Alaskan Taiga A Synthesis of Structure and Function Edited by K. Van Cleve, F.S. Chapin III, L.A. Viereck, C. T. Dyrness and P.W. Flanagan 1986. X, 240p., 81 figures. cloth ISBN 3-540-96251-4 Volume 58 Ecology of Biological Invasions of North America and Hawaii Edited by H.A. Mooney and J .A. Drake 1986. X, 320p., 25 figures. cloth ISBN 3-540-96289-1

Volume 59 Acid Deposition and the Acidification of Soils and Waters By J.O Reuss and D.W. Johnson 1986. VIII, 120p., 37 figures. cloth ISBN 3-540-96290-5

Volume 60 Amazonian Rain Forests Edited by C.F. Jordan 1987. X, 133p., 55 figures. cloth ISBN 3-540-96397-9

Volume 61 Potentials and Limitations of Ecosystem Analysis Edited by E.-D. Schulze and H. Zw6lfer 1987. XII, 435p., 141 figures. cloth ISBN 3-540-17138-X

Volume 62 Frost Survival of Plants By A. Sakai and W. Larcher 1987. XI, 321 p., 200 figures. cloth ISBN 3-540-17332-3

Volume 63 Long-Term Forest Dynamics of the Temperate Zone By Paul A. Delcourt and Hazel R. Delcourt 1987. XIV, 450 p., 90 figures. 333 maps. cloth ISBN 3-540-96495-9

Volume 64 Landscape Heterogeneity and Disturbance Edited by Monica Goigel Turner 1987. XI, 241 p., 56 figures. cloth ISBN 3-540-96497-5

Volume 65 The Community Ecology of Sea Otters Edited by G.R. VanBlaricom and J.A. Estes 1987. XVI, 247 p., 71 figures. cloth ISBN 3-540-18090-7

The Community Ecology of Sea Otters Edited by G.R.VanBlaricom and lA. Estes

With 71 Figures

Springer-Verlag Berlin Heidelberg New York London Paris Tokyo

Professor Dr. GLENN R. VANBLARICOM Professor Dr. JAMES A. ESTES

U.S. Department of the Interior Fish and Wildlife Service Institute of Marine Sciences 272 Applied Sciences Building University of California Santa Cruz, CA 95064 USA

Legendfor cover motif: A sea otter carries a common prey species, the red sea urchin Strongylocentrotusfranciscanus, to the ocean surface. Illustration by Jenny Wardrip, based on an underwater photograph by Richard Mattison.

ISBN-13:978-3-642-72847-1 e-ISBN-13:978-3-642-72845-7 DOl: 10.1007/978-3-642-72845-7

Library of Congress Cataloging in Publication Data. The Community ecology of sea otters. (Ecological studies; v. 65). Bibliography: p. Includes index. 1. Sea otter - California - Pacific Coast - Ecology. 2. Animal populations - California - Pacific Coast. 3. Mammals - CaliforniaPacific Coast- Ecology. 1. VanBlaricom, G. R. (Glenn R.), 1949- . II. Estes, J. A. (James A.), 1945- . III. Series. QL737.C25C66 1987 599.74'47 87-20553.

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. Duplication of this publication or parts thereof is only permitted under the provisions of the German Copyright Law of September 9,1965, in its version of June 24, 1985, and a copyright fee must always be paid. Violations fall under the prosecution act of the German Copyright Law.

© Springer-Verlag Berlin Heidelberg 1988. Softcover reprint of the hardcover 1st edition 1988

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

2131/3130-543210

Preface

The impetus for this volume comes from two sources. The first is scientific: by virtue of a preference for certain large benthic invertebrates as food, sea otters have interesting and significant effects on the structure and dynamics of nearshore communities in the North Pacific. The second is political: because of the precarious status of the sea otter population in coastal California, the U.S. Fish and Wildlife Service (USFWS) announced, in June 1984, a proposal to establish a new population of sea otters at San Nicolas Island, off southern California. The proposal is based on the premise that risks of catastrophic losses of sea otters, due to large oil spills, are greatly reduced by distributing the population among two geographically separate locations.

The federal laws of the U.S. require that USFWS publish an Environmental Impact Statement (ElS) regarding the proposed translocation of sea otters to San Nicolas Island. The EIS is intended to be an assessment of likely biological, social, and economic effects of the proposal. In final form, the EIS has an important role in the decision of federal management authority (in this case, the Secretary of the Interior of the U.S.) to accept or reject the proposal.

As a result of our positions with USFWS and our experience in the study of sea otter-community interactions, the authors of the draft EIS for the above proposal (USFWS 1986) solicited our views on predicted effects of sea otters on the nearshore benthic communities of San Nicolas Island. Many of our views were incorporated into an early draft of the EIS, which was then made available for review by experts in various relevant disciplines. Several of the reviewers were intensively critical of portions of the early draft ElS. Discussions concerning the relationships of sea otters to kelp forest communities were perceived as particularly contentious, stimulating vociferous accusations of bias and even ulterior motives.

Our response to the above events was to seek a forum wherein the various views of sea otter-community interactions could be objectively stated, compared, and discussed, in an atmosphere free of political overtones. Thus, we organized and moderated a half-day symposium on community effects of foraging by sea otters. The symposium was part of the 66th annual meeting of the Western Society of Naturalists, held in December 1985 at Monterey, California. The gathering included presentations by M. Foster, D. Laur, A. Ebeling, D. Duggins, and S. Levin, in addition to ourselves (see Chaps. 4 through 10, this Vol.). Although the symposium was well attended, well

VI Preface

presented, and well received, the participants agreed that there was much more to be said, and more topics to be covered. In response to this need, we solicited the manuscripts which comprise this volume. Selection of contributors and subject matter was intended to cover all categories of habitat in which sea otters are known to have biological effects. Thus, Kvitek and Oliver review interactions of sea otters and soft-sediment communities (Chap. 3), VanBlaricom presents data on effects of sea otters in rocky intertidal systems (Chap. 4), and several authors contribute views on relationships of sea otters and kelp forests (Chaps. 5 through 9). We emphasize kelp forests in this Volume because more work has been done with sea otters and kelp forests than with other habitat types influenced by sea otters. In addition (probably as a consequence of the greater research effort), interactions of sea otters and kelp forests are quite controversial, especially in California. Our selection of contributors reflects our desire to have the full range of views in the controversy represented in this Volume. Finally, we sought the views of S.A. Levin as a widely recognized theoretical ecologist and applied mathematician, and as one experienced in environmental policy disputes. Dr. Levin's role was to assess available data on sea otter-community interactions, and to determine if existing evidence and lines of inquiry are appropriate for the kinds of answers needed for ecological generalizations and environmental policy decisions.

With the exception of Chap. 1, all chapters in this volume were subjected to peer review. We established a panel of 20 scientists to provide the reviews. Panel members are listed below, with the exception of one member, who requested anonymity:

Ames, J .A., California Department of Fish and Game, Monterey, California, USA

Beddington, J .R., Center for Environmental Technology, Imperial College of Science and Technology, London, England

Cailliet, G.M., Moss Landing Marine Laboratories, Moss Landing, California, USA

Dayton, P K., Scripps Institution of Oceanography, University of California, La Jolla, California, USA

Gaines, S D., Hopkins Marine Station, Stanford University, Pacific Grove, California, USA

Goodman, D., Department of Biology, Montana State University, Bozeman, Montana, USA

Hines, A.H., Chesapeake Bay Center for Environmental Studies, Smithsonian Institution, Edgewater, Maryland, USA

Hixon, M.A., Department of Zoology, Oregon State University, Corvallis, Oregon, USA

Jackson, G.A., Scripps Institution of Oceanography, University of California, La Jolla, California, USA

Jameson, RJ ., US Fish and Wildlife Service, San Simeon, California, USA Johnson, A.M., US Fish and Wildlife Service (retired), Anchorage, Alaska,

USA

Preface VII

Kenyon, K.W., US Fish and Wildlife Service (retired), Seattle, Washington, USA

Menge, B.A., Department of Zoology, Oregon State University, Corvallis, Oregon, USA

North, W J., Kerckhoff Marine Laboratory, California Institute of Technology, Corona del Mar, California, USA

Pearse, J .S., Institute of Marine Sciences, University of California, Santa Cruz, California, USA

Peterson, C.H., Institute of Marine Sciences, University of North Carolina, Morehead City, North Carolina, USA

Simenstad, C.A., Fisheries Research Institute, University of Washington, Seattle, Washington, USA

Sousa, W.P., Department of Zoology, University of California, Berkeley, California, USA

Suchanek, T .H." Division of Environmental Studies, University of California, Davis, California, USA

Chapters 2 through 10 were each reviewed by at least two members of the panel. No panel member reviewed more than one chapter. Chapter 11 (our concluding summary) was reviewed by all of the principal authors of the contributed chapters. In addition, several authors of contributed chapters sought additional review comments from colleagues. These reviewers are acknowledged in the individual chapters as appropriate.

Because of our joint roles as editors and contributors to this Volume, we established a procedure to avoid conflict of interest in the review process. For all chapters that we authored or co-authored, Dr. A.W. Ebeling (University of California, Santa Barbara, California, USA) assumed the role of editor, and held final authority on any matters of disagreement between us and our reviewers.

The review process outlined above was intended to insure that this Volume would be as informative, constructive, and objective as is humanly possible. We believe that the process has been reasonably successful, recognizing that such goals are, in practice, difficult to obtain. We sincerely hope that this Volume will contribute to improved knowledge of the ways in which sea otters influence natural communities. In addition, we hope the Volume will provide a model for the discussion of sea otter ecology in an atmosphere relatively free of the heavy constraints of political intrigue.

Santa Cruz, California, USA G.R. VanBlaricom J.A. Estes

Acknowledgments

We thank the contributors to this Volume for their efforts toward improvement of our knowledge of sea otters and natural communities. We are grateful for the responsiveness and tolerance of the contributors to the review process, and to deadlines and various other impositions. Likewise, we thank the members of the review panel, whose prompt and constructively critical evaluations led to substantial improvements in the quality, clarity, objectivity, and credibility of the volume.

We are particularly grateful to Dr. A.W. Ebeling for his efforts as an editor of manuscripts that we contributed. Dr. Ebeling's time and efforts had a significant positive influence on the quality and objectivity of our chapters.

We thank Dr. David Montgomery for his support and assistance in the presentation of our symposium at the 1985 annual meeting of the Western Society of Naturalists, in Monterey, California. The completion of this volume was, to a large degree, a consequence of the successful presentation of our symposium.

We thank Dr. Dieter Czeschlik and the editorial staff of Springer-Verlag for their remarkable support, enthusiasm, and tolerance of this project. We also thank the editors of Ecological Studies for supporting the inclusion of this Volume in the Series.

We are grateful for the administrative support and patience of the U.S. Fish and Wildlife Service and the University of California, San ta Cruz (UCSC), especially Drs. Robert 1. Brownell, Jr., and William Doyle. The National Ecology Center of the Fish and Wildlife Service was most helpful in the preparation of graphics for the volume. The office staff of the Institute of Marine Sciences, UCSC, provided extensive assistance with manuscript preparation, and David Carlson provided valued help with computers.

Our own research activities on the community ecology of sea otters probably would not have developed without the support and encouragement of Dr. Clyde Jones during the time he served as Director of the National Fish and Wildlife Laboratory and the Denver Wildlife Research Center. We gratefully acknowledge his wisdom, friendship and commitment to scientific excellence.

It is a particular pleasure to express our thanks to Mrs. Patti Himlan for her contributions to the completion of this volume. In the face of an endless stream of manuscripts, letters, unreasonable deadlines, and self-destructing hard discs, she has maintained her patience, good humor, and a steady commitment to the completion of this volume.

Finally, we thank our families for their continuous support and patience throughout the project.

G.R. VanBlaricom and J.A. Estes

Contents

1 Introduction

G.R. VanBlaricom and J.A. Estes ......................... .

2 A Review of the History, Distribution and Foraging Ecology of Sea Otters

M.L. Riedman and J .A. Estes (With 12 Figures)

2.1 Introduction ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 4 2.2 Evolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 4 2.3 History, Distribution and Present Status of Populations . . . . . . .. 7 2.4 Diet and Foraging Behavior .......................... 12

2.4.1 Habitat .................................... 12 2.4.2 Diet ...................................... 13 2.4.3 Foraging Behavior. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 15 2.4.4 Activity ................................... 16 2.4.5 Class and Individual Variation ..................... 19

2.5 Summary ...................................... 20

3 Sea Otter Foraging Habits and Effects on Prey Populations and Communities in Soft-Bottom Environments

R.G. Kvitek and 1.S. Oliver (With 5 Figures)

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 22 3.2 Soft-Bottom Foraging Habits ......................... 23

3.2.1 Epifaunal Prey Communities. . . . . . . . . . . . . . . . . . . . .. 28 3.2.1.1 Prey Populations ....................... 28 3.2.1.2 Prey Communities ...................... 29

3.2.2 Shallow-Burrowing Infaunal Prey. . . . . . . . . . . . . . . . . .. 30 3.2.2.1 Prey Populations ....................... 30 3.2.2.2 Prey Communities ...................... 31

3.2.3 Deep-Burrowing Infaunal Prey .................... 32 3.2.3.1 Prey Populations ....................... 32 3.2.3.2 Prey Communities ...................... 37

3.3 Prey Vulnerability and Patch Dynamics .................. 38

XII Contents

3.3.1 Prey Resistance .............................. 38 3.3.2 Prey Resilience ............................... 41

3.4 Substrate Disturbance .............................. 41 3.5 Rocky and Soft Substrata ........................... 43 3.6 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 45 3.7 Summary ...................................... 45

4 Effects of Foraging by Sea Otters on Mussel-Dominated Intertidal Communities

G.R. VanBlaricom (With 19 Figures)

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 48 4.2 Models of Mussel-Dominated Communities ................ 48 4.3 Sea Otters as Predators of Mussels .................... .. 50 4.4 Case 1: Sea Otters and Mussels on the Coast of Central California . 51

4.4.1 Study Location .............................. 52 4.4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 54 4.4.3 Consumption of Mytilus californumus by Sea Otters:

The Basic Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 55 4.4.4 Creation of Gaps in Mussel Cover by Sea Otters: Spatial and

Temporal Aggregation. . . . . . . . . . . . . . . . . . . . . . . . .. 57 4.4.5 Size Distribution of Gaps Created by Sea Otters. . . . . . . .. 61 4.4.6 Mussel Size and Vulnerability to Foraging Sea Otters. . . . .. 63

4.5 Case 2: Sea Otters and Mussels in Prince William Sound, Alaska. .. 65 4.5.1 Study Location .............................. 66 4.5.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 69 4.5.3 Consumption of Mytilus edulis by Sea Otters:

The Basic Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 70 4.5.4 Size Distribution of Intertidal Mussels and the Population

Status of Sea Otters ........................... 73 4.5.5 Mussel Size and Vulnerability to Foraging Sea Otters. . . . .. 82

4.6 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 84 4.7 Summary ...................................... 88

5 Kelp Communities and Sea Otters: Keystone Species or Just Another Brick in the WaD?

M.S. Foster and D.R. Schiel (With 3 Figures)

5.1 Introduction ............. . . . . . . . . . . . . . . . . . . . . . .. 92 5.2 Kelp Community Structure .......................... 94 5.3 The Otter as a Keystone Species in California: Local or General? .. 95

5.3.1 Approach and Methods .... . . . . . . . . . . . . . . . . . . . .. 95 5.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 98 5.3.3 Possible Bias ................................ 99

5.4 Otter Effects: Geographic and Historical . . . . . . . . . . . . . . . . . . 102

Contents XIII

5.5 Beyond Otters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 5.6 A Model for Structure and Organization .. . . . . . . . . . . . . . . . . 103 5.7 Conclusions ..................................... 106 5.8 Summary ...................................... 107 Appendix ......................................... 108

6 Sea Otters, Sea Urchins, and Kelp Beds: Some Questions of Scale

J .A. Estes and C. Harrold (With 13 Figures)

6.1 Introduction .................................... 116 6.2 The Questions ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 6.3 The Evidence .................................... 119 6.4 Variation in Space and Time .......................... 123

6.4.1 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 6.4.2 Variation in Space ............................ 125

6.4.2.1 Quadrats ............................. 125 6.4.2.2 Swath Counts ......................... 127 6.4.2.3 Regional Variation ...................... 129

6.4.3 Variation in Time ............................. 130 6.4.3.1 Temporal Variation in Kelp Canopies .......... 135 6.4.3.2 Variation Between Algal Assemblages and Sea Urchin

Barrens . . . . . . . . . . . . . . . . . . . . .......... 136 6.4.3.3 Long-Term Changes ..................... 137

6.5 Directions for Future Research ........................ 138 6.6 Summary .......................... ............ 140 Appendix ......................................... 142

7 Effects of Sea Otter Foraging on Subtidal Reef Communities off Central California

D.R. Lauf, AW. Ebeling, and D.A. Coon (With 7 Figures)

7.1 Introduction .................................... 151 7.2 Study Sites ..................................... 152 7.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

7.3.1 Sampling Schedule ............................ 154 7.3.2 Macroin verte brate Sampling . . . . . . . . . . . . . . . . . . . . . . 154 7.3.3 Algae and Sessile Invertebrates .................... 155

7.3.3.1 Central Californian Sites .................. 155 7.3.3.2 Urchin Exclusion Experiment ............... 155

7.3.4 Fish Sampling ............................... 156 7.3.5 Data Analysis ................................ 157

7.4 Results and Discussion .............................. 158 7.4.1 Macroinvertebrates ............................ 158 7.4.2 Algae and Sessile Invertebrates .................... 159

XIV Contents

7.4.2.1 Central Californian Sites .................. 159 7.4.2.2 Urchin Exclusion Experiment ............... 162

7.4.3 Fish ...................................... 162 7.5 General Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . 163

7.5.1 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 7.6 Summary ...................................... 167

8 Fish Populations in Kelp Forests Without Sea Otters: Effects of Severe Storm Damage and Destructive Sea Urchin Grazing

A.W. Ebeling and D.R. Laur (With 9 Figures)

8.1 Introduction .................................... 169 8.2 The System ..................................... 170 8.3 Methods ....................................... 171 8.4 Results and Discussion .............................. 175

8.4.1 PhYSical Variables ............................. 175 8.4.2 Total Fish Density ............................ 176 8.4.3 Species Densities ............................. 176

8.4.3.1 Surfperch Adults ....................... 176 8.4.3.2 Surfperch Young, Subadults, and Summer

Transients ............................ 177 8.4.3.3 Midwater Planktivores .................... 179 8.4.3.4 Switch-Feeding Predators and Plant-Cropping

Omnivores . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 8.4.4 Biogeographic Species Groups ..................... 181 8.4.5 Fish Assemblage Structure ....................... 183

8.5 General Discussion and Conclusions . . . . . . . . . . . . . . . . . . . .. 183 8.5.1 Predicted Behavior of Our System in the Presence of Sea

Otters ..................................... 188 8.5.2 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190

8.6 Summary ...................................... 191

9 The Effects of Kelp Forests on Nearshore Environments: Biomass, Detritus, and Altered Flow

D.O. Duggins (With 3 Figures)

9.1 Introduction .................................... 192 9.2 HabitatModel ................................... 193 9.3 Trophic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 194 9.4 Hydrodynamic Model .............................. 197 9.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 200 9.6 Summary ...................................... 201

Contents xv

10 Sea Otters and Nearshore Benthic Communities: A Theoretical Perspective

SA Levin

10.1 Introduction: Regulatory Issues ...................... 202 10.2 Ecological and Regulatory Parallels .................... 203 10.3 Measures of Ecosystem Health ....................... 205 10.4 Risk Assessment: Predicting Fate, Transport, and Effects of

Otters ........................................ 206 10.5 Summary ..................................... 208

11 Concluding Remarks

J.A. Estes and G.R. VanBlaricom

11.1 Introduction ................................... 210 11.2 Patterns, Processes and Paradigms in Communities Occupied by

Sea Otters - A View Among Systems, and Through Space and Time ........................................ 210

11.3 Variation in Community Structure ..................... 213 11.4 Future Research Needs ............................ 215 11.5 The Approach to Variation - A Philosophical Perspective ..... 217

References ........................................ 219

Subject Index ...................................... 239

Contributors

Coon, David A., Environment Health and Safety, University of California, Santa Barbara, CA 93106, USA

Duggins, David 0., Friday Harbor Laboratories, University of Washington, 620 University Road, Friday Harbor, WA 98250, USA

Ebeling, Alfred W., Department of Biological Sciences, University of California, Santa Barbara, CA 93106, USA

Estes, James A., U.S. Fish and Wildlife Service, Institute of Marine Sciences, University of California, Santa Cruz, CA 95064, USA

Foster, Michael S., Moss Landing Marine Laboratories, P.O. Box 450, Moss Landing, CA 95039-0450, USA

Harrold, Christopher, Division of Research, Monterey Bay Aquarium, 886 Cannery Row, Monterey, CA 93940, USA

Kvitek, Rikk G., Department of Zoology NJ-15, University of Washington, Seattle, WA 98195, USA

Laur, David R., Marine Science Institute, University of California, Santa Barbara, CA 93106, USA

Levin, Simon A., Section of Ecology and Systematics, 347 Corson Hall, Cornell University, Ithaca, NY 14853, USA

Oliver, John S., Moss Landing Marine Laboratories, P.O. Box 450, Moss Landing, CA 95039-0450, USA

Riedman, Marianne L., Division of Research, Monterey Bay Aquarium, 886 Cannery Row, Monterey, CA 93940, USA

Schiel, David R., Ministry of Agriculture and Fisheries, Fisheries Research Centre, P.O. Box 297, Wellington, New Zealand

VanBlaricom, Glenn R., U.S. Fish and Wildlife Service, Institute of Marine Sciences, University of California, Santa Cruz, CA 95064, USA

1 Introduction

G. R. VANBLARICOM and J. A. ESTES

Some species in natural communities seem to be have disproportionately significant effects on populations of co-occurring species. Improved knowledge of "important" species is seen by many ecologists as vital to understanding natural selection within communities. Protection of such species may be regarded as essential to the conservation of ecosystems. However, the identification and characterization of "important" species may be criticized by ecologists who argue that natural ecosystems will be oversimplified. If large tracts of habitat must be set aside as reserves for protection of "important" species, harvesting interests and resource managers may object to the placement of valued resources beyond the reach of human utilization.

A well-known "important" species is the sea otter (En hydra lutris), which lives in cold-temperate nearshore waters of the North Pacific Ocean. Two views of sea ottercommunity interactions are widely held and frequently discussed, often critically and emotionally. The first is that sea otters, by controlling herbivore densities, are of immense functional significance in nearshore benthic ecosystems, and therefore are worthy of protection or enhancement. The second is that sea otters consume large quantities of commercially valuable shellfish, thereby conflicting with fisheries' interests and justifying limitation of the range and numbers of sea otters. Controversies surrounding these two views are among the principal causes of substained interest in sea otters by marine ecologists, fisheries' interests, conservation organizations, and regulatory agencies. Resolution of these controversies probably will require many further ecological, social, economic, and philosophical investigations, and certainly will require compromise among interested parties.

Current controversies involving sea otters first emerged in California in the early 1960's. McLean (1962) published the first report on the community -level consequences of sea otter foraging. McLean presented qualitative observations that removal of sea urchins by the expanding sea otter population in California was causing increased abundance of kelps. North (1965) attributed changes in California kelp forests at additionallocations to sea otter foraging on urchins. Miller and Geibel (1973) disputed North's findings, arguing that other factors must have caused the changes that North associated with sea otters. To this point, however, none of the listed authors had offered any data in support of their arguments. Estes and Palmisano (1974), Dayton (1975), Palmisano and Estes (1977) and Estes et al. (1978) broadened the geographic scope of the issue to Alaska, and published the first quantitative accounts. Their data from the western Aleutian Islands indicated that sea otters limit sea urchin size and abundance, with a consequent proliferation of kelps and kelp-aSSOciated fauna. Duggins (1980) offered the first experimental tests of the developing sea otter-kelp paradigm,

2 Introd uction

showing that removal of sea urchins could indeed cause a significant increase in kelp abundance. Breen et al. (1982) presented qualitative observations that extended the paradigm to nearshore Canadian waters.

Recent studies in California provide conflicting results. Pearse and Hines (1979) correlated disease-induced urchin mortality (arguably analogous to sea otter effects) with increased kelp abundance, and VanBlaricom (1984) presented historical observations consistent with McLean's original report. In contrast, Cowen et al. (1982) showed experimentally that some kelp forests in California are influenced more by storm disturbance and turbidity than by effects of sea otters on grazing urchins.

Interactions of sea otters and shellfisheries (reviewed by Estes and VanBlaricom 1985) first became a public issue in the early 1960's. The return of sea otters to northern San Luis ObiSpo County, California, correlated with a sharp decline in commercial catches of subtidal red abalone (Haliotis rufescens). Additional expansion of the California sea otter population was later correlated with loss of a commercial fishery for red sea urchins (Strongylocentrotus franciscanus) and recreational fisheries for Pismo clams (Tivela stultorum). In the early 1980's, expansion of sea otter populations into eastern Prince William Sound, Alaska, was associated with a crash in the subsistence harvest of Dungeness Crabs (Cancer magister). However, in reviewing sea ottershellfishery conflicts, Estes and VanBlaricom (1985) observed that several recreational invertebrate fisheries (including clams, mussels, and crabs) survive within the range of sea otters, even though otters continue to feed on the harvested species. Moreover, Estes and VanBlaricom argue that quantification of fishery conflicts with sea otters often is inadequate and ambiguous.

Our purpose here is to identify and discuss unanswered questions pertaining to the effect of sea otters on nearshore communities. Through the nine contributed chapters to this Volume, we seek to present the range of views on the sea otters' role in community dynamics, and to identify needs and directions for future research on relationships of sea otters and the communities they occupy.

In assembling this volume, we did not expect a consensus view on sea otter-community interactions to emerge. One finds in the chapters a range of views on answers to certain contentious questions. In some cases, the contributors disagree as well on how such questions might be answered by additional research. Such disagreements exacerbate the "uncertainty associated with prediction" discussed in Levin (this Vol.), complicating the search for ecological generalizations, and making the decisions of regulatory agencies more difficult and controversial.

Although effects of sea otters on kelp forests have been discussed frequently in the literature, a principal message of this Volume is the general paucity of information on interactions of sea otters and nearshore benthic communities. The problem is most acute for intertidal and soft-sediment systems, both of which are important sources of food for sea otters and may, in some areas, support very large otter populations. For example, Cimberg and Costa (1985) recently found preliminary evidence that sand dollars and flatfish may be important prey for a large sea otter population living in Bristol Bay, Alaska. Both sand dollars and flatfish are known to be functionally significant in some communities, but the ecological roles of sea otters, sand dollars, and flatfish are entirely unknown for Bristol Bay. Thus, any generalizations about effects of sea otters on soft-sediment ecosystems must be tempered by the expectation that

Introduction 3

new data will almost certainly alter the generalizations. The same is true, of course, for intertidal and kelp forest systems. The likely consequences of new data add to the "uncertainty associated with prediction."

Most of the contributed chapters in this Volume emphasize, in one form or another, the substantial variation across space and time in the dynamics of nearshore benthic systems and in the observed or predicted effects of sea otters. Factors contributing to variation include differing life histories of prey species, spatial and temporal changes in physical factors that affect nearshore communities, fluctuations in recruitment success, and the differing ecological needs of various age and sex categories of sea otters. Thus, even if the difficulties presented by data-gaps and disagreements among investigators can somehow be overcome, a very important component of the "uncertainty associated with prediction" concept remains: the natural temporal and spatial variation of nearshore benthic communities.

This Volume is organized both by habitat type and by the "distance" between the foraging activities of sea otters and the affected community. In Chapter 2, the history, distribution, and foraging patterns of sea otters are reviewed, providing a background for the ensuing chapters on sea otter-community interactions. Chapters 3 and 4 discuss direct effects of sea otters in soft-sediment and intertidal communities, providing the first reviews of these subjects in the literature on sea otters. Chapters 5 through 8 focus on sea otters in kelp forests. Each chapter begins at a point of common agreement -that sea otters often eliminate sea urchins as functionally significant herbivores in kelp forests - then explores various aspects of the importance of urchin grazing, such as the consequences (or lack thereof) for algae, other invertebrates, and associated populations of fishes. Higher-order effects are considered in Chapter 9. Beginning with the premise that sea otters cause increased size and persistence of kelp forests by removing sea urchins, models are outlined for the consequent effects of kelp as habitat for other species, as a source of organic carbon for nearshore food webs, and as a modifier of coastal hydrodynamic regimes. Chapter 10 develops the "uncertainty associated with prediction", assessing the consequences of uncertainty for those most concerned with sea otters: ecologists, fishermen, conservation organizations, and regulatory agencies. The concluding chapter is a brief overview of the volume, with an effort to identify threads of consistency, both in the effects of sea otters and in needs for future research.

These introductory words may suggest to some that we are skeptical for the of tencelebrated importance of sea otters as predators in the nearshore benthic communities of the North Pacific. To the contrary, we have little doubt either that sea otters are capable of producing Significant effects, or that such effects occur in a variety of habitats. Indeed, there are few species for which such effects are known or suspected across comparable ranges of geography and habitat type. Even so, we urge that generalizations be made with caution. As the contents of this volume amply demonstrate, numerous sources of natural variation can complicate or even obscure the effects of predation by sea otters.

2 A Review of the History, Distribution and Foraging Ecology of Sea Otters M. L. RIEDMAN and J. A. ESTES

2.1 Introduction

The purpose of this Volume is to summarize and discuss what is known or suspected about the community ecology of sea otters. Each of the following contributions focuses on a specific aspect of sea otter predation and its influence on coastal marine ecosystems in the North Pacific Ocean. Here we provide background information for readers who may be unfamiliar with the species.

This chapter is not intended to be a comprehensive review of the biology of the sea otter. Such a review is available in Riedman and Estes (in prep.). Rather, we concentrate on topics related to their effects as predators. We begin by briefly reviewing paleontology and evolution. Next we discuss the history, distribution and current status of sea otter populations. Finally, we present in more detail what is presently known of foraging biology, since this is the crucial link between the natural history of sea otters and their potential effects as predators in coastal marine communities.

2.2 Evolution

Sea otters (Enhydra lutris) are the largest mustelid, although next to the South American marine otter (Lutra fe/ina) they are the smallest species of marine mammal (Fig. 2.1). Otters comprise the subfamily Lutrinae (Carnivora; Mustelidae). All of the 13 extant species (Corbet and Hill 1980) forage largely or exclusively in aquatic environments and can be divided into two distinct feeding categories (Chanin 1985). One contains species that feed predominately on fish. Typically these species capture prey with their mouths and have especially well-developed sensory-motor development of the facial region. Like most other carnivores, their carnassials are well developed for shearing rather than crushing. The other category includes those species that feed mainly on invertebrates. These species typically capture prey with their forelimbs, and thus have well-developed sensory-motor function of that body region.

Based on dentition (Van Zyll de Jong 1972) and cerebral morphology (Radinsky 1968), the fish-eating forms (represented by the extant genera Lutra and Pteronura) are thought to be primitive, with invertebrate feeders having been twice derived from the fish-eating lineage. In one group, represented by the extant Aonyx (clawless and small-clawed otters), shearing function by the carnassials was retained. In the other, represented by the extant genus Enhydra (modern sea otters), shearing carnassials were

Evolution 5

Fig. 2.1. Sea otter eating squid in California (photograph by Fred Bavendam)

Fig. 2.2. Dentition of the sea otter. Note bunodont molars (photograph by Steve Webster)

6 Review of Sea Otter Biology

Fig. 2.3. Cladogram of the lutrinae. (Berta and Morgan 1985)

progressively modified to flattened and more massive premolars and molars, and thus improved crushing ability (Fig. 2.2).

Berta and Morgan (1985) recognized two lineages of extinct and living sea otters (Fig. 2.3). One led to the extinct Enhydriodon, the other to the extinct Enhydritherium and Enhydra. Enhydriodon is known only from Eurasia and Africa, with three welldescribed species. In addition, there are several more poorly known specimens from Greece, England, and east Africa that have provisionally been assigned to the genus. All material is of late Miocene/Pliocene age. It is not known if Enhydriodon lived in marine or freshwater habitats, or both. However, they were as large or larger in size than modern sea otters and had Similarly well-developed molariform dentition (Repenning 1976a).

Enhydritherium is known from the late Miocene of Europe and the late Miocene/ middle Pliocene of North America (Berta and Morgan 1985). Two species have been described: E. lluecai from Spain (Villalta and Crusafont-Pairo 1945; Crusafont-Pairo and Golpe 1962) and E. terraenovae from Florida and California (Berta and Morgan 1985). Based on the locations of fossil material thus far discovered and described, Enhydritherium apparently lived exclusively in coastal marine habitats, like modern sea otters.

The EnhydritheriumjEnhydra lineage apparently originated in the Old World, dispersed via continental margins to the New World, and ultimately reached the North Pacific Ocean. Routes of dispersal are still uncertain, although several have been suggested. One is via Asia and the Bering Land Bridge, and then secondarily to the western Atlantic via the Central American Seaway. A second possible route is from Europe to North America via the North Atlantic, and then across the New World arctic to the North Pacific. A third possibility, favored by Berta and Morgan (1985), is that Enhydritherium dispersed from Europe to eastern North America via the North Atlantic, but then entered the Pacific from the south via the Central American Seaway.

Regardless of exact lineage relationships and routes of dispersal, Enhydra is thought to have arisen in the North Pacific and never to have dispersed to other regions of the World (Repenning 1976a). The earliest occurrence of Enhydra is uncertain, although the only undisputed fossil records are early Pleistocene or later. One extinct species,

History, Distribution and Present Status of Populations 7

Enhydra macrodonta (Kilmer 1972), which was slightly larger than the extantE. lutris, has been described from the late Pleistocene.

The evolution of sea otters is related to, and may have been in part affected by, two important events that occurred in the late Cenozoic. One was a global cooling trend at high latitudes that began during the late Miocene (Durham 1950). Modern sea otters live exclusively in temperate regions. Available evidence indicates that their recent southward distribution in the eastern North Pacific extended to about central Baja California (Kenyon 1969), where cool, upwelled water from the California current turns seaward and disperses. Prior to the late Miocene, tropical and subtropical conditions prevailed throughout the North Pacific (Hopkins 1967) and it seems unlikely that modern sea otters could have survived in such environments. Perhaps the highly productive coastal environment that accompanied development of temperate conditions in the North Pacific was a prerequisite to the evolution of sea otters.

A second event of possible evolutionary significance to sea otters was the widespread extinction of odobenid pinnipeds in the Pliocene (Repenning 1976b). Some of these odobenids, especially species in the Dusignathinae, were adapted to shallow-water, benthic foraging. Repenning (1976b) suggested that extinction of these odobenids may have opened an environment suitable for radiation of the sea otters.

2.3 History, Distribution and Present Status of Populations

The exploitation and near extinction of sea otters has been well documented by Ogden (1941), Lensink (1962) and Kenyon (1969). Earliest human exploitation is unknown, although hunting of sea otters could have occurred from the mid- to late Pleistocene along the Asian Coast of the North Pacific. However, we are unaware of evidence for pre-Recent human activity from that area. Early humans are thought to have crossed the Bering Land Bridge and spread southward in the New World late in the Pleistocene (Martin 1973). However, little is known of the earliest coastal inhabitants of the New World because 15,000 years ago sea level was some 150 m lower than it is at present. Thus, coastal village sites or other cultural remains from those early times are now well submerged (Bickel 1978). With melting of the continental ice sheets, sea level advanced rapidly for the next several thousand years, reaching a plateau near the present level by about six or seven thousand years ago. Records of human activity along the Pacific Coast from that time onward are much better.

Midden remains indicate that sea otters were exploited by coastal inhabitants of the North Pacific region. Records from Amchitka Island suggest that aboriginal Aleuts limited sea otter popUlations (Simenstad et al. 1978), although it is likely that such reductions were restricted to areas near village sites, or at least were not so widespread as subsequent population reductions that occurred throughout the otter's range.

The earliest recorded history of sea otter exploitation began in the 1740's with discovery of Alaska and the Aleutian Islands by the Bering Expedition. Upon returning to Russia, these explorers told of vast numbers of sea otters from lands to the east. The decline of the sea otter, and the abuses suffered by aboriginal people in these lands, are tragic testimony to human greed, and represent historical events of broad and lasting


consequence (Hone 1984). Sea otters quickly declined from overhunting, causing the Russians to press further east, establishing outposts at Kodiak, Sitka, and Fort Ross. A trade with China was established for the valuable furs, resulting in extensive influence by the fur traders in the Hawaiian Islands. Anticipated revenue from further fur harvests (of otter and fur seals, Callorhinus ursinus) was largely responsible for the purchase of Alaska by the United States in 1867, and it was the realization by Russia that these resources were mostly spent that prompted the willingness to sell for a mere 7.2 million dollars (Fig. 2.4).

The harvesting of sea otters intensified under United States jurisdiction (Lensink 1962), so that by the beginning of the twentieth century the species was nearly extinct. It is impossible to know how many animals remained in 1911 when unregulated killing was finally stopped, but it was a tiny fraction of the original number. The few remaining animals survived in 13 known locations (Kenyon 1969). Several of these surviving colonies subsequently became extinct because of either continued illegal hunting or their precariously small size (Estes 1980) (Fig. 2.5). However, those that survived increased with the cessation of human exploitation, although there are few records detailing the patterns of growth.

Sea otter populations apparently grew at a rate of about 15% per year during the early phases of recovery (Kenyon 1969). At several locations, such as Amchitka Island (Kenyon 1969) and Medny Island in the Commander Islands (A. Zorin, unpubl.), declines have been noted following peaks in the recovery phase, presumably because of food limitation. Kenyon (1969) estimated that the population at Amchitka Island peaked at about 4500 individuals in the early 1940's and then declined to about

Fig. 2.4_ Young sea otter pup with light-colored natant pelage (photograph by Steve Webster)

History, Distribution and Present Status of Populations

Aleutian Islands

I Rat Is. 0eIar0f Is.

PACIFIC OCEAN

1111111 Original Distribution (1740)

tmft Present Distribution (1984)

Remnant Colonies (1911) ..

Surviving Translocated Populations * , .' o.~ Hawaiian Islands

I I:

9

: CANADA

Fig. 2.S. The North Pacific ocean showing past/present sea otter range, location of remnant populations, and locations of translocated populations

1500 individuals by the late 1940's. Zorin (pers. commun.) documented a decline at Medny Island from 2500 individuals at the peak of the growth phase, to about 900 to 1200 individuals several years latter.

Sea otters currently are distributed throughout most of their historical range west of about eastern Prince William Sound (Fig. 2 .5). Populations in the Kuril Islands presently are estimated at about six to seven thousand individuals, which is probably near maximum size (M. Maminov, pers. commun.). Similarly, most available habitat along the east coast of the Kamchatka Peninsula has now been recolonized, except for the extreme northern area. Populations in this area are estimated to contain about 2500 individuals, which is probably near maximum size (B.V. Khromovskikh, pers. commun.). The Commander Islands contain an estimated 2700 individuals, of which about 900 to 1200 are at Medny Island, with the remainder at Bering Island. The population at Medny Island is probably resource-limited, whereas the population at Bering Island is still increasing (A. Zorin, pers. commun.).

Populations in Alaska are more poorly known, except in areas where specific studies have been done. Sea otters now occur throughout most of the Aleutian Islands, and probably are at or near carrying capacity except at their western end (Near Islands) and perhaps at several areas near their eastern end. The Aleutian Islands have not been surveyed thoroughly for several decades, but they undoubtedly contain tens of thousands of individuals. Sea otters are also abundant along much of the Alaska Peninsula. On its north coast, they occur from Cold Bay to about Port Moller. In this area they appear to forage over the broad shelf of the eastern Bering Sea, and may migrate seasonally through Unimak Pass into the Pacific Ocean (Cimberg and Costa 1985) in

". ".

1984

1975

1981

CALIFORNIA

··':iJ~:~·. ~ari~ Cruz. ., .... "':~U81 Pt.

1938

1947

1959

1972

PACIFIC OCEAN

Longitude (OW)

Fig. 2.6. Range expansion of the sea otter population in California

38°

30'

37"

30'

Z L Q) '0 ::J

36' ~

30'

35°

History, Distribution and Present Status of Populations 11

response to the formation of winter sea ice in Bristol Bay (Schneider and Faro 1975). Distribution and population numbers are largely unknown on the southern coast of the Alaska Peninsula. A total of 2947 animals were counted in 1984 during a survey of the Kodiak archipelago, where the population apparently is continuing to increase to the southeast into areas of unoccupied habitat (Simon-Jackson et al. 1985). In a recent survey of Prince William Sound, about 5000 sea otters were counted (Irons 1984, unpublished data). Populations in the southwestern part of Prince William Sound probably are near maximum size, whereas those to the north and east appear to be still increasing.

Except for a small remnant population in central California, sea otters were exterminated from the area between Prince William Sound, Alaska, and the southernmost extent of their range in central Baja California. However, there were efforts to translocate sea otters from Alaska to Oregon, Washington, British Columbia, and southeast Alaska during the 1960's and early 1970's (Jameson et ai. 1982). For unknown reasons, the Oregon population became extinct in the mid-1970's. The population in Washington remained very small for a number of years following the initial translocation in 1969. However, by 1980 it began to increase and now contains about 65 animals (Jameson et aI., in press). Growth of the sea otter population in British Columbia is poorly known due to its remote location and infrequent surveys of the area. In 1977,70 animals were counted by Bigg and MacAskie (1978), and the population was thought to have changed little in size until 1984, when 345 animals were counted (MacAskie 1984). Thus, the population in British Columbia appears to be well established and increasing at a high rate. From 1965 to 1969, a total of 412 sea otters were translocated to southeast Alaska (Jameson et al.1982). The population was most recently surveyed in 1983 when 1124 animals were counted (Johnson et al. 1983).

The sea otter population in central California is comparatively well known, due to its accessibility and because various management problems have drawn attention to it. Although few survey data are available from before the mid-1960's the population apparently increased in size and range from early in this century until the late 1960's or early 1970's (Riedman and Estes, unpubl.). Geibel and Miller (1984) estimated that the population contained about 1800 animals in 1976. However, the population has not increased since that time, and may have declined slightly. It now appears that mortality from entanglement in the coastal set-net fishery has been a major cause of this lack of growth (Wendell et al. 1985). Presently, the sea otter population in California probably contains about 1400 animals, excluding pups (RJ. Jameson and J .A. Estes, unpubI.) (Fig. 2.6).

To summarize, sea otters now occupy most of their historical range from the northeastern Gulf of Alaska, westward across the Pacific rim, to the southern end of the Kuril Islands. Throughout much of this area populations appear to be at or near carrying capacity. In contrast, most of their historical range southeastward from Prince William Sound along the west coast of North America remains unoccupied, and for this reason, where populations do occur, they probably are below carrying capacity.

One last point concerning the near extinction and recovery of sea otters should be mentioned, and that is that these changes undoubtedly had substantial effects on certain invertebrate populations and other related components of coastal communities in the North Pacific Ocean. Many shellfish species that are so highly prized today were


most likely held at low levels by sea otter predation prior to intensive commercial hunting of otters (Estes and VanBlaricom 1985). The increased standing stocks of many invertebrate species that followed regional extinctions of the sea otter were of great historical significance, for these changes were prerequisite to the development of many North Pacific shellfisheries. Most sea otter population increases in past years have occurred in areas sparsely populated by humans, and therefore have not conflicted severely with shellfisheries. However,future increases will be into more heavily populated areas, and the developing fishery conflicts will no doubt be a central issue in the management of this unique and interesting marine mammal.

2.4 Diet and Foraging Behavior

2.4.1 Habitat

Sea otters forage in the benthos of rocky and soft-sediment communities, as well as within the algal understory and canopy. Prey are carried to the surface where they are eaten (Fig. 2.7). Foraging activity generally takes place in subtidal zones, although otters also forage intertidally at times (Vandevere 1969; Kovnat 1982; Harrold and Hardin 1986; VanBlaricom this Vol.; Jameson, unpublished data). In California, sea otters typically forage close to shore in waters less than 25 m in depth (Wild and Ames 1974), beyond which canopy-forming kelp and many types of prey become scarce (Abbott

Fig. 2.7. Sea otter carrying sea urchin to the surface in California (photograph by Richard Mattison)

Habitat 13

and Hollenberg 1976). Occasionally, California otters are observed feeding in depths of up to 36 m (Hardy unpubl.). Siniff and Ralls (1986) have recently observed some otters feeding up to 3 km offshore in the central and southern portions of the California range. Territorial males in the Monterey area sometimes forage about 1 km from shore (Deutsch, pers. commun.). However, sea otters in the Aleutian Islands commonly feed at depths of 40 m or more (Estes 1980). The deepest dive recorded for a sea otter is 97 m and was observed in Alaska (Newby 1975).

In California, sea otters are primarily associated with subtidal habitats characterized by rocky substrata, although they also occur in sandy areas. Rocky substrata support diverse assemblages of plants and animals, including prey frequently consumed by sea otters. Sea otter density within most of the range, with the exception of some areas at the northern and southern population fronts, is related to substrate type. Rocky bottom habitats support an average density of five otters per km2 , while sandy bottom areas support an average of 0.8 otters per km2 (California Department of Fish and Game 1976).

2.4.2 Diet

The diet of California sea otters consists almost exclusively of macroinvertebrates (Ebert 1968; Wild and Ames 1974; Estes et al. 1981), in contrast to that of Alaskan and Russian sea otters, which also feed on epibenthic fish in some areas where their populations are at high levels (Estes et al. 1981; Maminov and Shitikov 1970). The availability of prey species, which varies in relation to geographic location and the length of time an area has been occupied by sea otters, in part determines diet throughout the sea otter's range.

In recently reoccupied areas of central California that are characterized by rocky substrata, the diet consists principally of abalones (Haliotis spp.), rock crabs (Cancer spp.), and sea urchins (Strongylocentrotus spp.) (Ebert 1968; Vandevere 1969; Wild and Ames 1974; Wade 1975; Stephenson 1977; Benech 1981; Estes et al. 1981). These foods are higher in caloric value and therefore more rewarding than other prey species (Costa 1978a, b). As populations of preferred prey are reduced with continued occupancy of a particular area, the diet diversifies to include a larger proportion of kelp crabs (Pugettia spp.), clams (various spp.), turban snails (Tegula spp.), mussels (Mytilus spp.), octopus (Octopus spp.), barnacles (Balanus spp.), scallops (Hinnites spp.), sea stars (Pisaster spp.), chitons (Cryptochiton stelleri) and echiuroid worms (Urechis caupo) (Boolootian 1961; Limbaugh 1961; Ebert 1968; Hennessey 1972; Wild and Ames 1974; Estes 1980; Estes et al. 1981; Benech 1981; Ostfeld 1982)(Fig. 2.8).

Sea otter predation on seabirds has been reported to occur occasionally (VanWagenen et al. 1981; Baldridge, pers. commun.; Jameson, pers. commun.; Vandevere, pers. commun.; Riedman and Estes, in review). Predation on fish in California is extremely rare (Hall and Schaller 1964; Miller 1974; Estes et al. 1986).

Diet is also related to habitat type and time of year. In soft-sediment communities, sea otters feed primarily on bivalve molluscs. For example, Pismo clams (Tivela stultorum) make up a Significant proportion of the diet of sea otters foraging in sandy areas near Monterey and Morro Bays (Miller et al.1975; Wade 1975; Shimek 1977b; Stephen-


Fig. 2.8. Sea otter eating crab on the ocean surface in California

son 1977; Hines and Loughlin 1980). The diet of sea otters foraging in the Elkhorn Slough estuary of Monterey Bay consists principally of deep-burrowing bivalves (Tresus nuttalii and Saxidomus nuttali) (Kvitek et al., in press; Kvitek and Oliver, this Vol.).

Alaskan sea otters in the Montague Strait area of Prince William Sound feed primarily on clams (especially Saxidomus gigantea), which are the most abundant food resource (Calkins 1978). The remainder of the diet consists largely of crab (Telmessus cheiragonus). Estes et al. (1981) found that otters consume mainly clams and mussels at Green Island and Sheep Bay in Prince William Sound.

In recently reoccupied areas of the Aleutian Islands where sea otter populations exist below equilibrium densities (such as Attu Island), otters feed primarily on sea urchins (Strongylocentrotus polyacanthus), in addition to some molluscs and crustaceans. Fish are rarely eaten. In areas such as Arnchitka Island where otter populations are near equilibrium density, fish form an important part of the diet. At Amchitka Island, populations of herbivorous invertebrates have been reduced by sea otter predation, while the abundance of kelp beds and nearshore fishes (especially rock greenling, Hexagrammos lagocephalus) associated with kelp beds have correspondingly increased. The importance of fish in the diet of Amchitka Island sea otters is therefore apparently related to the comparative scarcity of sea urchins and the increased availability of nearshore fish (Estes et al. 1982).

Russian studies in the Kuril Islands indicate a similar relationship between dietary composition, prey availability, and length of time a particular area has been occupied by sea otters (see Estes et al. 1981 for review). At the Kuril Islands ofParamushir and

Foraging Behavior 15

Urup, where otter populations are near eqUilibrium density , the otter's diet has broadened to include a substantial amount of fish (Hexagrammos, Sebastichus [a subgenus of Sebastes) , and Cyclopterichtys [= Aptocyclus)) in addition to sea urchins and bivalve molluscs (Myti/Us, Modiolus, and Tellina). As at Amchitka Island, the sea urchins at Urup Island are small and sparsely distributed (Maminov and Shitikov 1970; Shitikov et al. 1973). In contrast, at Simushir Island where the sea otter population was well below eqUilibrium density in the late 1960's, sea urchins were large and abundant, and otters consumed urchins almost exclusively (Shitikov 1973).

Diet also varies seasonally among sea otters in the Kuril Islands. Sea urchins, molluscs, fish,andfisheggs are the most common prey during summer (Barabash-Nikiforov et al. 1947 ; Shitikov 1971). In the Aleutian Islands, rock greenling spawn and defend their eggs during summer, at which time they are probably most vulnerable to sea otter predation. Sea urchins reach maximum gonadal development during winter, at which time they are presumably of greatest nutritional value to sea otters. Thus, seasonal variation in diet in the Kuril Islands may reflect a shift in foraging strategies as otters take advantage of seasonal changes in availability and quality of food sources (Fig. 2 .9).

2.4.3 Foraging Behavior

Sea otters capture prey with their forepaws, often storing food items within loose pockets of skin beneath the axilla of each forelimb until the prey can be consumed

Fig. 2.9. Sea otter eating a fish on the sea surface in Alaska (photograph by Jane Watson)


Fig. 2.10. Sea otter on surface breaking prey winter rock tool in California (photograph by James Mattison Jr.)

at the surface (Barabash-Nikiforov et al. 1947; Kirkpatrick et al. 1955). Stealing offood by conspecifics, usually adult territorial males, sometimes occurs (Fisher 1939; Miller 1980). The use of tools such as rocks to break open or dislodge hard-shelled molluscs is common among California sea otters (Fisher 1939; Umbaugh 1961; Hall and Schaller 1964; Kenyon 1969; Houk and Geibel 1974; Miller 1974). Tool-use is less frequently observed in the Aleutian Islands (Kenyon 1969;1.A. Estes, unpubl.)(Fig. 2.10). However, Calkins (1978) and Garshelis (1983) often observed otters in Prince William Sound breaking open prey with rock tools. In soft-substrate habitats, bivalve molluscs are captured by vigorous and repetitive digging with the forelimbs (Shimek 1977a; Hines and Loughlin 1980). Sea otters are capable of learning new capture techniques to facilitate effective foraging. For example, otters in California have been observed biting into aluminum beverage cans and extracting octopuses which had taken refuge inside (McCleneghan and Ames 1976).

2.4.4 Activity

Because of their small body size and lack of blubber (which provides insulation as well as a reserve of stored energy in other marine mammals), sea otters compensate for the problem of thermal stress not only by means of their insulative fur, but by maintaining a high level of internal heat production (Iverson and Krog 1973; Morrison et al. 1974; Costa and Kooyman 1982). High energetic requirements are necessary for maintenance of the otter's elevated standard metabolic rate. Costa (1978a, 1985) estimates that free-

Activity 17

ranging adults consume an amount of food equivalent to 23%- 33% of their body weight each day .

Sea otters therefore invest a substantial proportion of their time in foraging activities to satisfy their high energy requirements. Foraging takes place during both day and night (Shimek and Monk 1977 ; Loughlin 1979; Ribic 1982). Diurnal activity cycles of sea otters are generally characterized by crepuscular peaks in foraging activity in California (Fisher 1939; Hall and Schaller 1964; Sandegren et al. 1973; Miller et al. 1975; Loughlin 1977; Shimek and Monk 1977; Benech 1981; Ribic 1982; Ralls et al. 1985; Estes et al. 1986), in the Aleutian Islands (Lensink 1962; Kenyon 1969; Estes 1977; Estes et al. 1982), and in some areas of Prince William Sound (Garshelis 1983). Observations made in California suggest that a third peak in foraging activity takes place between approximately 2300 and 0200 h in some areas (Shimek and Monk 1977; Benech 1981; Ribic 1982; Payne and Jameson 1984; Ralls et al. 1985). According to Loughlin (1977), the average duration of a foraging bout is 2 .5 h in Monterey, and at least three foraging bouts take place within a 24-h period. In the San Simeon area, Ribic (1982) showed that, on average, an otter is active for three or four periods, each of which lasts about 3 h and is followed by an inactive period of about 4 h . Despite these general activity patterns observed among groups of otters, the variables influencing activity cycles and time allocated to various activities in sea otters are complex, and appear to be related to a number of differences in the environment and in the individual (Fig. 2.11).

Diurnal activity time-budgets at the population level throughout much of the otter's range in California are characterized by a high degree of similarity with respect to geographic location and time of year. Observations in a number of locations along the California coast from Point Piedras Blancas to Santa Cruz show that otters allocate about 21% to 28% of their time to foraging activity during daylight hours (King 1976; Harris 1977; Shimek and Monk 1977; Yellin et al. 1977; Estes et al. 1986). In comparison, sea otter populations in Alaska and Oregon below equilibrium density, and at

Fig. 2.11. A group of resting sea otters in California (photograph by Larry Minden)


equilibrium density, allocate 15-17% and 50- 55% of their daylight time, respectively, to foraging activities (Estes et al. 1982). The greater amount of time spent foraging during the day in areas where populations are well established may relate to the relative scarcity of preferred prey, and the fact that otters in these areas feed on fish which require more time to capture than sessile invertebrates. Unfortunately there are no data on nocturnal activity from the Aleutian Islands, although Garshelis (1983), using radio telemetry, found that otters from long-established (and presumably food-limited) populations in Prince William Sound were more active on a 24-h cycle than were otters from areas that were more recently recolonized.

A considerable amount of individual variation in 24-h activity rhythms and the amount of time allocated to foraging appears to exist among California sea otters. Foraging activity may vary with the individual's sex, age and reproductive status (whether or not a female has a pup), in California (Loughlin 1977, 1979; Ribic 1982; Ralls et al. 1985) and in Alaska (Garshelis 1983). Throughout a 24-h period, several individual radio-telemetered otters in Monterey Bay spent an average of 34% of their time foraging. Approximately 45% of the feeding activity took place at night, although there was substantial individual variation (22-73%) in the proportion of nocturnal foraging (Loughlin 1977).

Siniff and Ralls (1986) radiotagged California otters in the central and southern portion of their range, and found variation in 24-h activity budgets among five different age/sex groups, as well as among individuals within each group. Adult females and males spent about 37% of their time foraging, while adult males, juvenile males, and females with pups allocated about 40% of their time to foraging. Juvenile females spent 48% of their time foraging, suggesting that it is more difficult for young females to obtain adequate food. With the exception of the juvenile females, other California otters spent slightly more time feeding than adult otters in Nelson Bay, an area recently reoccupied by sea otters (discussed below).

During daylight hours in California, females with large pups appear to spend as much or more time feeding than other animals, while females with very young pups may spend less time feeding than other otters (Sandegren et al. 1973; Lyons, pers. commun.; Estes and Riedman, unpubl.). According to Siniff and Ralls (1986), California females with pups spent about one more hour feeding during a 24-h period than pupless females. In Prince William Sound,females with pups spent more time feeding over a 24-h period than independent adults (Garshelis 1983), probably because mothers must obtain additional food for their pups. Recently weaned pups in Prince William Sound also spend more time feeding than solitary adults or older juveniles, apparently because they are less adept at obtaining prey (Garshells 1983) (Fig. 2.12).

Garshelis (1983) found that foraging activity patterns in Prince William Sound differed between two locations: Nelson Bay, an area recently occupied by male as well as some female sea otters, and Green Island, an area inhabited by otters for over 25 years. Green Island was occupied primarily by females and seasonally breeding males. Sea otters in one area of Nelson Bay rested during the day and foraged at night, while Green Island otters rested at night and fed during the day. At Green Island, both males and females shared similar average activity patterns, although females with large pups spent significantly more time feeding. The extensive nocturnal foraging in one area of Nelson Bay appears to be related to the fact that Nelson Bay otters fed heavily on Dungeness

Class and Individual Variation 19

Fig. 2.12. Female sea otter with her pup (photograph by Fred Bavendam)

crab (Cancer magister) which is nocturnally active (Wild and Tasto 1983). In another area of Nelson Bay where otters fed primarily on clams, the activity cycle was crepuscular. The activity pattern of males that traveled between each area shifted to correspond with the general activity cycle characteristic of a particular area (diurnal vs. nocturnal feeding).

Nelson Bay adult males and females spent 37% of their time foraging, while Green Island adults spent 47% of the time feeding (Garshelis 1983). Females with pups and juveniles at Green Island spent 53% and 51 % of their time foraging. Male otters in Nelson Bay allocated 23% less of their time to foraging activity but obtained 38% more calories per day than otters at Green Island, where food resources were apparently less abundant. Large Dungeness crabs provided 70% of the caloric intake of some otters, yet were captured on only 9% of the foraging dives at one male area in Nelson Bay (Garshelis 1983).

2.4.5 Class and Individual Variation

In California, an average of 70% to 73% of all diurnal feeding dives result in the successful capture of prey (Loughlin 1977; Estes et a1. 1981), although a complex array of variables may affect the proportion of successful dives, type of prey obtained, dive times, and foraging tactics (Estes et a1. 1981). Adult otters often have more unsuccessful dives than juveniles, although adults also obtain more rewarding but less easily


captured prey. Often, several successive dives of long duration are required to capture prey items are necessary to obtain are less accessible but energetically more rewarding (such as abalone) than less valuable prey, such as turban snails (Loughlin 1977; Costa 1978a,b;Esteseta1.1981).

There appears to be substantial variation among individual California sea otters with respect to diet, as well as the amount of time spent diving underwater and on the surface when foraging (Estes et al. 1981). Observations of tagged sea otters in Monterey Bay indicate that diet and foraging strategies differ significantly among individuals (Lyons and Estes 1985). Individual females tend to specialize in one to three types of prey. For instance, one female prefers to eat kelp crabs, turban snails and purple urchins, while another specializes on abalone and rock crab. Individual dietary patterns are maintained for at least 3 years (Lyons and Estes 1985). Studies focusing on how individual foraging strategies are acquired, and if dietary patterns change or persist throughout an individual's lifetime are currently in progress in Monterey (Riedman, Staedler, and Estes, unpubl.).

Prey choice and foraging tactics also vary Significantly according to a female's reproductive status. Several females, following the birth of their pups, shifted foraging locations and changed their diets, thus specializing in different types of prey, requiring less effort to obtain or to break open than those captured when they were not accompanied by pups (Lyons and Estes 1985). Females with large pups, however, may specialize in food that is difficult to capture, such as abalone (Riedman, Staedler, and Estes, unpubl.). In Prince William Sound, the diet of females with pups was often of relatively poor quality because the mothers frequently foraged on prey items of low nutritive value that were easily captured by pups (Garshelis 1983; VanBlaricom, this Vol.). Riedman, Staedler, and Estes (unpubl.) also found individuals use tools in different ways to break open prey. Lyons and Estes (1985) speculate that specialization in diet and the resultant potential partitioning of resources could influence survival and reproductive success of individuals in well-established sea otter popUlations. An otter may also be able to feed more efficiently during a foraging bout by repeatedly searching for the same type of prey. In addition, it is possible that the selection of preferred prey by certain individuals in limited area over time could have a subtle and complex localized impact on kelp forest communities.

2.5 Summary

In this chapter we review the evolution, recent history, present distribution and abundance, and foraging behavior of sea otters. Modern sea otters (genus Enhydra) apparently arose from a lineage containing the extinct Enhydriodon and Enhydritherium. All known members of this lineage had strongly molariform dentition, and thus were apparently predators on marine invertebrates. Enhydriodon and Enhydritherium are known from the late Miocene and Pliocene. Enhydritherium arose in the Old World, dispersed to the New World, and probably gave rise to Enhydra in the North Pacific Ocean in the early Pleistocene. Enhydra apparently never occurred elsewhere.

Summary 21

The modern sea otter (En hydra lutris) was distributed across the Pacific rim from the northern Japanese archipelago to central Baja California. The species was hunted extensively during the 18th and 19th centuries, and was nearly extinct by the beginning of the 20th century. Following protection by an international treaty, sea otter populations recovered in many areas. Presently the species is abundant throughout the Kuril and Aleutian Islands, and along the coast of North America westward of about Prince William Sound. Smaller isolated populations occur in southeast Alaska, British Columbia, Washington and California.

Sea otters live in shallow coastal areas with rocky and soft-sediment substrata. Most foraging occurs in water depths less than about 25 m in California and 40 m in Alaska, although much deeper dives have been recorded. Diet consists of macroinvertebrates, and in some areas of Alaska and the Soviet Union, fishes. In many areas, dietary diversity increases with increased time of occupancy. Bivalve molluscs are the principal prey in soft-sediment habitats. Seasonal and longer-term dietary shifts in response to changes in prey availability have been reported.

Prey are captured with the forelimbs and consumed on the ocean's surface. Rock tools or other hard objects often are used to dislodge or break open prey. The rate of tool use appears to be higher in California and in some areas of Prince William Sound than in the Aleutian Islands.

Production rates of internal body heat are high, due to the cold, aquatic environment in which sea otters live and the absence of a blubber layer. Free-ranging adults may consume food equivalent to 23% -33% of their body weight daily. Foraging occurs during the day and night, with crepuscular peaks in activity level in most areas. Recently established populations allocate less time to feeding than do longer-established populations, apparently because of prey reduction in the latter situation.

Individuals in the same general area may vary considerably in activity and diet. Females with small pups spend comparatively little time foraging during the day, whereas females with large pups often spend more time feeding than other animals. Individuals appear to specialize on one to three or more of numerous available prey types and these patterns of specialization may be maintained for 3 or more years. Dietary specialization and foraging strategy also varies with respect to a female's reproductive status, and may be related to her reproductive success, as well as her pup's subsequent survival and breeding success.

3 Sea Otter Foraging Habits and Effects on Prey Populations and Communities in Soft-Bottom Environments R. G. KVITEK and J. S. OLIVER

3.1 Introduction

Sea otter populations have made dramatic recoveries during the last 75 years and today playa major functional role in some nearshore communities (Estes and Palmisano 1974; Estes et al. 1978; Duggins 1980; Breen et al. 1982; Estes et al. 1982). Sea otters have considerable influence on the distribution, abundance, and demography of their benthic invertebrate prey (Ebert 1968; Lowry and Pearse 1973; Wild and Ames 1974; Miller et al. 1975; Stephenson 1977; Hines and Pearse 1982; Garshelis 1983; Estes and VanBlaricom 1985; Wendell et al. 1986). In general, otters are thought to invade a new area, rapidly consume the largest individuals of the most calorically rich prey, and then switch to smaller and less desirable prey as the preferred species become scarce (Ebert 1968; Wild and Ames 1974; Estes and Palmisano 1974; Ostfeld 1982; Riedman and Estes, this Vol.) or inaccessible (Lowry and Pearse 1973; Hines and Pearse 1982). Along rocky shores, the depletion of herbivorous prey like sea urchins and abalone (Estes and Palmisano 1974; Wild and Ames 1974; Estes et al. 1978) is followed by an increase in the cover of kelp forests and in the species richness of fishes (Estes and Palmisano 1974; Dayton 1975; Simenstad et al. 1978; Duggins 1980). Although sea otters may not play this role in all kelp forests (see Foster and Schiel, and Estes and Harrold, this Vol.), they clearly have a dramatic effect on some kelp forest communities (Estes and Palmisano 1974; Estes et al. 1978; Duggins 1980). In addition to rocky habitats, sea otters also forage in large areas of soft sediment (Calkins 1978; Kenyon 1969; Wild and Ames 1974; Miller et al. 1975; Stephenson 1977; Estes et al. 1981; Garshelis 1983; Estes and VanBlaricom 1985; VanBlaricom, this Vol.), where their effects on prey populations and community structure are quite different. Compared to the rocky shore, there is little information on the foraging habits and community role of sea otters in soft-sediment environments.

Sea otters will probably increase their use of estuaries, sandy beaches, and deeper soft-sediment habitats as their populations continue to grow and spread. Therefore, this is an excellent juncture to summarize what is known about the foraging habits of sea otters in soft-sediment environments including their influence on prey populations and communities. This chapter describes the diets of sea otters and the different prey communities in soft-bottom habitats. It considers the vulnerability of prey populations, the community effects of feeding disturbance, and some major differences between foraging on soft and rocky substrates. Whenever possible we suggest additional research needed to further our understanding of the ecology of sea otters in soft-sediment communities.

Soft-Bottom Foraging Habits 23

3.2 Soft-Bottom Foraging Habits

The major soft-sediment prey of sea otters can be divided into three functional groups: epifauna, shallow-burrowing infauna, and deep-burrowing infauna. These groups represent a general hierarchy of vulnerability to sea otter predation. Epifauna are the most easily captured and deep-burrowing in fauna are the most difficult to capture. The major epifaunal prey usually live at the sediment surface and include crabs, shrimp, snails, mussels, sea stars, and sand dollars. The major shallow-burrowing infauna usually live in the surface 5 -10 cm of sediment and primarily include certain species of clams. The deep-burrowing in fauna often live deeper than 20 em into the sediment and include long-siphoned clams, echiuroid worms, and polychaete worms.

Sea otters consume over 30 species of soft-sediment prey throughout their range (Table 3.1, see Fig. 3.1 for locations). However, there is no general geographic pattern related to prey-living depths in the diets of sea otters feeding in soft-bottom habitats. Clams are the most important prey, accounting for 68% and 99% of the volume of prey

Table 3.1. Species of soft-sediment prey present in the diets of sea otters. (Revised from U.S. Fish and Wildlife Service, 1982. Southern Sea Otter Recovery Plan)

Prey category California Oregon Sitka Prince Alaskan Kurile and Food organism William Peninsula and Commander

Sound Aleutian Islands Islands

Epifaunal prey Crustacea

Decapoda Blepharipoda occidentalis x Cancer magister x x x Cancer productus x x Chionecetes bairdi x Chionecetes opillio X Hyas coarctatus x Lopholithodes form ina-

tus x Sclerocrangon boreas x Telemessus cheiragonus x x

Mollusca Bivalvia

Modiolus difficilis x Mytilus edulis x Pecten beringia nus x Pecten islandica x

Gastropoda Natica clausa x x Polin ices lewisi x

Echinodermata: Asteroidea

Evasterias troschelii x Echinoidea

Dendraster sp. x

24

Table 3.1. (continued)

Prey category Food organism

California Oregon

Shallow-burrowing infaunal prey Mollusca

Bivalvia Clinocardium cilia tum Clinocardium facanum x Macoma inquinata Macoma incongrua Macoma sp. Protothaca staminea Protothaca sp. x x Serripes groenlandicus Tellina sp. Tivela stultorum x Venericardia paucicostatus

Deep-burrowing infaunal prey Echiura

Echiurus echiurus Urechis caupo x

Annelida Polychaeta

Arenicola sp. Mollusca

Bivalvia Mya truncata Sax idomus gigantes Saxidomus nuttalli x x Siliqua patula x Spisula alascana Spisula hempelli x Tresus nu tta llii x

Sources:

Sea Otters and Soft-Bottom Communities

Sitka Prince William Sound

x x

x

x

x X x x

Alaskan Kurile and Peninsula and Commander Aleutian Islands Islands

x

x

x x

x

x

x

x x x

x

California. Fisher 1939; Limbaugh 1961; Hall and Shaller 1974; Ebert 1967 and 1968; Vandevere 1969; Hennessey 1972; Wild and Ames 1974; Miller et al. 1975; McCleneghan and Ames 1976; Shimek 1977a; Stephenson 1977; Hines and Loughlin 1980; Kvitek et al. in press: Authors' unpublished data Oregon. Jameson 1975 Sitka, Alaska. Rosenthal and Barilotti 1973. Prince William Sound. Calkins 1978; Garshelis 1983; VanBlaricom this Vol. Aleutian Islands and Alaskan Peninsula. Williams 1938; Murie 1940; Lensink 1962; Kenyon 1969; Cimberg et al. 1984; Authors' unpublished data Kurile Islands, USSR. Estes et al. 1981. Commander Islands, USSR. Barabash-Nikiforov et al. 1947.

Soft-Bottom Foraging Habits

, Commander

Islands

Bering Sea

~Amchltka Is.

Prlbilof

OlabloCyn.

25

GUll of ",IUke

CANADA

Sound

Vancouver

BAY

Fig. 3.1. Map showing locations of Alaska and California sea otter soft-sediment feeding grounds referred to in text

26 Sea Otters and Soft-Bottom Communities

in sea otter stomachs from the Bering Sea and along the Alaskan Peninsula, respectively (Table 3.2). Clams are also consumed on over 75% of the successful foraging dives of otters in Prince William Sound (with the exception of Green Is.) and California (Tables 3.3 and 3.4). With the exception of the Pismo clam (Tivela stU[forum), the largest number of clam prey in most studies are deep-burrowing species (Table 3.1; Wild and Ames 1974; Calkins 1978; Hines and Loughlin 1980; Estes et al. 1981; Kvitek et al. in press).

Although there are some good data on the diets of sea otters in several soft-sediment environments, there are very few quantitative data on the size structure and abundance of prey populations, and none on the effects of sea otters on entire soft-bottom communities. In some cases, we have only anecdotal observations and catch information from sport fisheries to describe how sea otters utilize local prey populations. When fishery information is carefully interpreted, it can suggest changes in prey and help direct future investigation.

Table 3.2. The diet of sea otters in soft-sediment habitats from the Bering Sea and the Alaskan Peninsula. Percent frequency or volume of prey in scat or stomach samples

Prey type Bering Sea and Alaskan Peninsula

Glazenap Is. a Unimak Island b Shumagin Islandsc 1982 1960 1960

Scats Stomach Scats Stomach (%F) (%V) (%F) (% V)

Cnidaria 11 Nemertina 11 Mollusca

Clam 33 68+ 8 99 Mussels 33 <1 77 1 Gastropod 4

Arthopoda Amphipod 22 Crabs

Brachyura 66 7 16 Hermit 24

Shrimp 33 Echinodermata

Echinoids Sea urchins 4 Sand dollars 44

Chordata Fish 22 60 10

Sample Size 9 2 75 2 Reference (1) (2) (2) (2)

a Near Izembeck Lagoon, Alaskan Peninsula. b Bering Sea. C Alaskan Peninsula. % F = Percent of scats containing item. % V = Percent volume. Reference: (1) Cimberg et al. 1984. (2) Kenyon 1969.

Tab

le 3

.3.

Th

e d

iet

of

sea

ott

ers

in s

oft-

sedi

men

t ha

bita

ts f

rom

Pri

nce

Will

iam

Sou

nd, A

lask

a. D

ata

are

give

n as

per

cent

ages

of

succ

essf

ul f

orag

ing

dive

s pr

o-du

cing

a p

rey

item

and

sca

ts c

onta

inin

g pr

ey t

ypes

Pre

y ty

pe

Pri

nce

Will

iam

Sou

nd

Mon

tagu

e S

trai

t G

reen

Is.

S

heep

Bay

N

elso

n B

ay

Orc

a In

let

1971

19

75

19

80

-81

19

74

19

80

-81

1

98

0-8

1

Div

es

Div

es

Div

es

Sca

ts

Div

es

Div

es

Div

es

(%F

) (%

F)

(% F

) (%

F)

(%F

) (%

F)

(% F

)

Ann

elid

a E

chiu

roid

a 3.

3 2

5 M

ollu

sca

Cla

m

81

43.6

76

76

70

.3

85

97

Mus

sel

0.3

39.7

6

34

21.8

O

ctop

us

Art

ho

po

da

Cra

bs

Bra

chyu

ra

7 1.

6 13

4

6

1.5

9 3

Ech

inod

erm

ata

Ech

inoi

ds

Sea

urc

hins

0.

3 <

1

<1

H

olot

huro

ids

0.3

<1

A

ster

oids

0.

8 <

1

3 U

nid

en ti

fied

10

11

.6

6.4

10

Sam

ple

size

59

7 42

0 61

6 15

8 25

1 26

5 89

R

efer

ence

(1

) (2

) (3

) (3

) (2

) (3

) (3

)

Dur

atio

n o

f o

tter

occ

upan

cy p

rior

to

sam

plin

g(2,

3)

Pre

-fur

tra

de

Pre

-fur

tra

de

<1

yr

1-2

yrs

<

1 y

r

% F

= P

erce

ntag

e o

f su

cces

sful

div

es o

r sc

ats.

R

efer

ence

: (1

) C

alki

ns 1

978.

(2)

Est

es e

t a1

.198

1. (

3) G

arsh

elis

198

3.

U'l

0 ::;>

e:,

~

S 3 'TI

0 ..., ~ !:f

()q

:I: '" s: ~ tv

--.J


Table 3.4. The diet of sea otters in soft-sediment habitats from California. Percent frequency of dives producing prey handled by feeding otters

Prey type Southern California Monterey Bay Area

Pismo Atascadero Stillwater Monterey Moss Elkhorn Beach State Beach Cove Harbor Landing Slough 1979-81 1973 1974 1976 1973 1985 Dives Dives Dives Dives Dives Dives (%F) (% F) (% F) (%F) (%F) (% F)

Echiuroida P <1 Mollusca

Clam Pismo clam 96 93 49

Tresus nuttallii 19.7 Saxidomus nuttalli 88.6 21.7 Tresus or Saxidomus 22.7

Rock jingle 1.4 Mussel 9.4

Gastropod Moonsnail 0.5

Arthopoda Crabs <1 50

Spiny mole crab 3.3 2 Cancer sp. 4.2 Pugettia 2.4 24.1

U niden tified 3 2

Sample Size 505 29 211 77 203 Reference (1) (2) (3) (4) (5) (6)

Duration of otter occupancy prior to sampling 1-2 yrs <1 yr >10 yrs 10 yrs <1 yr <1 yr

P = present. % F = percent frequency of dives. Reference: (1) Wendel eta1.1986 ;(2) Wild and Ames 1974;(3) Shimek 1977a; (4) Hines and Lough-lin 1980; (5) Stephenson 1977; (6) Kvitek et a1. 1986.

3.2.1 Epifaunal Prey Communities

3.2.1.1 Prey Populations

Epifaunal prey are most active at the surface of the sediment. While at times they may also burrow into surface sediments, they are much less cryptic than the infaunal prey. The most important epifauna are the crabs, particularly the Dungeness and red crabs (Cancer magister and Cancer producta) which are also fished commercially.

Fishery observations in both Alaska and California suggest that sea otters reduced the numbers of large cancer crabs in local populations. The 85% decline in Dungeness crabs caught per pot from 1978-1981 in northeastern Prince William Sound, Alaska, is linked to the arrival of sea otters in 1979 (Kimker 1982; Garshelis 1983). The decline of the red crab fishery in Monterey Harbor, California, was also coincident with the arrival of

Epifaunal Prey Communities 29

sea otters. In Monterey Bay, the red crab fishery declined where otters were present, but did not decline at Capitola or Santa Cruz pier where otters were absent (California Department of Fish and Game 1976; Hines and Loughlin 1980).

Unfortunately, there is no quantitative information on the abundance or size structure of cancer crab populations in areas where sea otters presently feed or are likely to feed in the future. Since sea otter populations are expanding their range along the coasts of Alaska, Vancouver Island, and California, there are a number of opportunities to document the population structure of crabs before and after otters migrate into a new feeding ground. For example, there are areas in southeastern Alaska where crab populations can be accurately monitored by scuba divers or video cameras. In locations such as the wave-exposed Californian coast this is a very difficult task.

The bay mussel, Mytilus edulis, is a frequent otter prey only in Alaska (Tables 3.2 and 3.4) and often lives on soft substrates. Sea otter interactions with mussels are considered in Chapter 4 of this volume.

Otters also consume a number of other epifaunal invertebrates (Table 3.1), but these species are generally not frequent prey (Tables 3.2,3.3, and 3.4) and, with the exception of the sand dollar, Dendraster excentricus, are not abundant. Sand dollars occur in relatively dense beds along the Alaskan Peninsula (authors, pers. observ.) and especially along wave-exposed beaches in California (Merrill and Hobson 1970; Kastendiek 1982; Highsmith 1982; Morin et al. 1985). Along the peninsula, sand dollars occur in densities as high as lO-50 m-2 in water depths of 15-20 m. Along the Californian coast, dense beds (often> 100 m-2 ) occur in narrow bands on the seaward edge of the surf zone. Despite their high abundance, sand dollar tests contain very little tissue (4% by weight, Moss and Lawrence 1972) and with the exception of Galznap Is. (Table 3.2), this probably accounts for their low frequency of occurrence in sea otter diets (Tables 3.2,3.3 and 3.4).

3.2.1.2 Prey Communities

The reduction of cancer crab populations by sea otters may have an important effect on infaunal in vertebrate communities. Other species of epifaunal crabs (Virnstein 1977; Woodin 1981) and even smaller shrimp (Bell and Coull 1978) strongly influence the populations of their small infaunal prey. Since cancer crabs also prey on infaunal invertebrates (Gotshall 1977; Morris et al. 1980; Peterson 1983; Ambrose 1984), they may playa similar community role. Other than Peterson's (1983) work demonstrating that Protothaca staminea (Conrad) had a much lower annual survival than the thickershelled Chione undatella (Sowerby) in the presence of Cancer anthonyi, there is little information available on the community effects of the foraging activities of cancer crabs. Furthermore, no data are available on the abundance of these crabs in different soft-bottom communities. Although cancer crabs are not as active or aggressive as the blue crabs that feed on U.S. east coast sand flats (Virnstein 1977), they may also limit the numbers of bivalves in some communities (Peterson 1983).

There are environments where cancer crabs can be excluded from large bottom areas by corrals to evaluate their effects on infaunal communities. One area is the Boca de Quadra Fjord in Southeastern Alaska, where the water is extremely clear and there is


no disruptive wave action. Large corrals can be made to both exclude and include crabs, thereby avoiding cage-related artifacts (Virnstein 1978; Hulberg and Oliver 1980). The water clarity and high number of cancer crabs also permits scuba divers to quantify their abundance and population structure. We predict that cancer crabs have little measurable effect on the numerous polychaete worms in this community, but may influence the species composition, abundance, and size structure of bivalve prey as Virnstein (1977) documented for the blue crab and Peterson (1983) showed for Cancer anthonyi. As a result, otter predation on cancer crabs may have an indirect effect on in faunal communities.

3.2.2 Shallow-Burrowing Infaunal Prey


The shallow-burrowing infaunal prey are primarily suspenSion-feeding clams that live in the surface layer (top 5-10 cm) of the sediment (Table 3.1). The most important shallow-burrowing prey is the Pismo clam, Tivela stu[torum, which lives in the lower intertidal and shallow subtidal sandy beaches along the wave-exposed coast of California (Morris et a1. 1980). Since there is a major sport fishery for Pismo clams, information on catch was available along several beaches prior to the arrival of sea otters.

The number of Pismo clams caught by clammers decreased markedly after foraging sea otters first arrived at several California beaches (Miller et a1. 1975; Stephenson 1977;Wendelleta1.1986).AtPismoBeach the catch (mean catch per clammer) dropped to zero within 1 year after sea otters arrived (Fig. 3.2). Similar patterns were documented in Monterey Bay (Miller et a1. 1975; Stephenson 1977). Again, catch information gives no insight into the effects on smaller sizes that are not collected in the fishery. However, at Moss Landing Beach, Stephenson (1977) documented the abundance and size distribution of Pismo clams before and after the otters arrived. His data show a significant decrease only in the largest clams. Otters may use visual or tactile cues to locate Pismo clams, but in either case large shallow-burrowing clams are probably easier to locate than small ones.

Pismo clam populations fluctuate widely in time and space even in the absence of sea otters. Mass mortality has been attributed to changes in environmental conditions and human factors (Coe and Fitch 1950; Fitch 1950), as well as otter predation. Settlement and recruitment success are also extremely variable, infrequent and unpredictable, with 10 to 20 years often separating significant recruitment events in California (Coe and Fitch 1950; Fitch 1950; Baxter 1962; Fitch 1965; Carlisle 1973). The ability to predict the fate of a population is further complicated by the apparent lack of relationship between stock size and recruitment success (Tomlinson 1968). As a result of this natural variability, it may often be very difficult to track population changes and link them to sea otters.

A number of other species of clams are relatively shallow-burrowing sea otter prey (Table 3.1). Cockles such as Protothaca can be very abundant in Alaska (Paul and Feder 1976) and consequently may be more important prey than the existing data suggest. The tellinid clams, such as Macoma, live at intermediate depths in the sedi-

Shallow-Burrowing Infaunal Prey

M e a n

C a t c h

p e

C I a m m e

9

7

5

3

9

7

9

7

5

3

9

7

5

3

Clam Catch • Otters

Beach 1

Beach 2

Jan Jan Jan Jan Jan Jan Jan Jan Jan 76 78 8D 82 84

Month

45

35

25

15

5

31

s 45 e 35

a

25 0 t

15 t 5 e

45

35

25

15

5

45

35

25

15

5

r

c o u n t s

Fig. 3.2. The change in the sport-fishery catch of Pismo clams (Tivela stultorum) in relation to the arrival of sea otters at four beaches along Pismo Beach, Ca. High counts of sea otters given by month. (Wendell et al. 1986)

ment and do not easily fit the classification scheme of epifauna, shallow and deep burrowers. Nevertheless, the effects of sea otters on other shallow-burrowing in faunal prey are not known nearly as well as their effects on Pismo clams (the largest of the shallowburrowing clams) and on the deep-burrowing clams discussed below.


Among the shallow-burrowing sea otter prey, only the sand dollar appears to play an important role in structuring benthic communities. Highsmith (1982) argues that sand dollars exclude tube-dwelling tanaid crustaceans from their beds and Morin et al. (1985) argue that sand dollars limit the movements and other activities of several epifaunal invertebrates along wave-exposed beaches. Although Woodin (1976) predicts that sus-


pension feeding bivalves such as the Pismo clam are age-class dominated because they eat their own larvae as well as those of other benthic infauna, there is no evidence that this process effects Pismo clams or the beach community. The seasonal and episodic movements of the beach and the fauna probably reduce the probability of adults consuming larvae (Oliver et aI. 1980; Morin et aI. 1985).

3.2.3 Deep-Burrowing Infaunal Prey


The deep-burrowing infauna are primarily species of long-siphoned clams that often live below 30 cm and sometimes deeper than 1 m into the sediment (Table 3.1). Deepburrowingworrns are much less frequent prey than the clams (Tables 3.2, 3.3, and 3.4). Unlike the epifauna and shallow-burrowing in fauna , sea otters must excavate considerable quantities of sediment to obtain the deep-burrowing species. In the process they create distinct feeding excavations or pits on the sea floor (Table 35) (Shimek 1977a; Calkins 1978; Hines and Loughlin 1980; Kvitek et aI., in press).

Unlike some of the epifauna and shallow-burrowing infauna, there is no evidence that sea otters cause rapid declines in populations of any deep-burrowing prey, although otters have foraged intensively and for extended periods on the deep burrowers. For example, up to 24 otters invaded the Elkhorn Slough and fed primarily on the deepburrowing gaper (Tresus nuttallii) and Washington clams (Saxidomus nuttalli) for 7 months (Table 3.4). Actual field counts of prey abundance (not indirect estimates of abundance from fishery catch data) showed no Significant change during the period of otter occupancy (Fig. 3.3, Table 3.6) (Kvitek et aI. in press). While otters did show a

Table 3.5. Dimensions of sea otter foraging pits

Pit dimensions Prince William Sound, Alaska Monterey Bay Area, California

Montague Green Stillwater Monterey Elkhorn Strait Island Cove Harbor Slough

Width (m) 0.50 X 0.30 1.5 X 0.5 mean 0.34 range 0.15-0.45 0.10-0.70 0.5-3.0

Area (m') mean 0.10 0.15 0.75 1.0 range 0.01-0.38 0.25-6.0

Depth (m) 0.50 0.20 0.5 0.3-0.5 Prey Clams Echiurus Urechis Sax idomus Saxidomus

echiurus caupo nuttalli nuttalli Tresus Tresus nuttallii nuttallii

N na 94 1 2 22 References 1 2 3 4 5

References: (1) Calkins 1978. (2) VanBlaricom, unpublished data. (3) Shimek 1977a. (4) Hines and Loughlin 1980. (5) Kvitek et a!., in press.

Deep-Burrowing Infaunal Prey

Fig. 3.3. The change in abundance of deep-burrowing clams, Tresus nuttallii and Saxidomus nuttalli, in relation to the num-ber of sea otters in Elkhorn Slough, Ca. (After Kvitek et aI., ... in press) E

i ~ II

U

25

20

15

10

5

0 Mar Apr May Jun

1984

33

I Otters ..... Deep-Burrowing Clams

25 en CD

20 III

0 15 ; .. 10 0

0 c

5 3.

0 Jul Aug Sep Oct Nov Dec Jan

1985 Month

Table 3.6. Prey abundance in soft-sediment feeding grounds. Mean number of prey m-2 (SO except Moss Landing which is 95% CL)

Prey type

Bivalves Pismo clam (Tivela stultorum) Gaper clam (Tresus nuttalli) Washington clam (Saxidomus

nuttallii) Gaper or

Washington clam Geoduck (Panopea generosa)

Echiuroida

Sample size Reference

Maximum number of otters observed feeding

Duration of otter occupancy prior to sampling

P = present.

Morro Bay Area Monterey Bay Area

Morro Bay Harbor Monterey Harbor

1986 1976 1985

3.5 0.7 0 (3.18) ~1 12.8 <1

13.5 (8.9)

<I

P P

13 19 (1) (2) (3)

50-60(7) 4 69

18 mo(7) 10 yrs 16 yrs

Moss Landing

1972 1973

0_65 0.05 (0.49) (0.10)

20 20 (4) (5)

0 10

pre-<>tter 7 mo

Elkhorn Slough

1984 1985

8.8 14.4 (9.24) (14.80) 10.4 4.7

(11.63) (5.86)

19.2 19.1 (15.77) (18.12) ~I

05 1.7 (1.62) (2.66)

30 30 (6) (7)

24 0

1 mo 5 mo

References: (1) Kvitek and Anderson, unpublished data; (2) Hines and Loughlin 1980 (area A); (3) Author's unpublished data; (4) Stephenson 1977 (pre-<>tter); (5) Kvitek et aI., in press (middle transect); (6) Brian Hatfield, Univ_ of Minnesota, pers. comm.; (7) Marianne Riedman, Monterey Bay Aquarium Sea Otter Census.

F r e q U e n e y


preference for the largest prey species available in the slough, samples from the prey population and from the record of discarded prey shells indicated no preference for the largest individuals of the preferred prey, and no change in the size distribution of living clams during the period of sea otter occupancy (Fig. 3.4).

Although otters did not show a preference for larger individuals or for local areas with the highest prey abundance, they did feed preferentially on deep-burrowing clams with the shallowest burrow depths (Kvitek et aI., in press). This was a patch with smaller and less abundant individuals which was only a short distance « 25 m) from patches with larger and more abundant clams living deeper in the sediment (Fig. 3.5). The clams in the preferred patch had a mean burrow depth 30% shallower than those in the patches of larger and more abundant clams (Fig. 3.5). As a result, otters needed to excavate 65% less sediment to capture the deep-burrowing prey living in the preferred prey patch (Table 3.7). Foraging efficiency in soft-sediment habitats may be defined by the biomass gained/volume of excavated sediment (Kvitek et aI., in press).

Deep-burrowing clams also persisted for over 10 years of sea otter presence in Monterey Harbor. Although otters first arrived in the harbor in 1966 (California Depart-

120

80

40

0

40

30

20

10

0

40

30

20

10

o 6

Living Shells

(Middle and OUier Transect)

N . 153

Discarded Shells

N . 466

Living Shells

(Inner Transect)

N . 177

8 10 12 14 16 18 20

Shell Length (em)

Fig. 3.4. Shell lengths of gaper clams (Tresus nuttallii) discarded by feeding sea otters and living along three transects in Elkhorn Slough, Ca., during 1984-85. (After Kvitek et aI., in press)

Deep-Burrowing Infaunal Prey

STUDY SITE CROSS SECTION

I t

~ ~ >40cn

i 39cm 28cm >40cm f

I.~. '.'.1. cIa la af

~()~·T S

OUTER TRANSECT MIDDLE TRANSECT

I I I

I f'

I

I

.//

T-Tresus nuttallii

S-SaxldomU8 nuttalill

35

INNER TRANSECT

f-------25m------~')I""' ... ----25m------!

Fig. 3.5. A schematic of changes in sediment layers, clam depth in the sediment, and foraging habits of sea otters in Elkhorn Slough, Ca. (Kvitek et aJ., in press)

ment of Fish and Game (1976), considerable numbers of Tresus nuttallii and Saxidomus nuttalli (13.5 clams m -2) were still present in 1976 (Table 3.6) (Hines and Loughlin 1980). During the period of Hines and Loughlin's (1980) study there was no change in the abundance of Washington and gaper clams for 13 months at local sites used by feeding otters. However, before 1966, the gaper clam was apparently the numerically dominant clam, but by 1976 gaper clams accounted for only 5% of the two species (Table 3.6) (Hines and Loughlin 1980). In December 1985, almost 20 years after the sea otters first returned to Monterey Harbor, the authors re-surveyed the major study sites of Hines and Loughlin (1980) and found no gaper clams and very low numbers « 1 m-2 ) of Washington clams in only a few restricted patches. Since the authors surveyed large bottom areas in good water clarity (several meters) and the clam siphons and siphon burrows are conspicuous, the deep-burrowing clams in Monterey Harbor are doubtless much less abundant today. However, their decline took many years and may also be related to harbor dredging and other local human activities (Hines and Loughlin 1980).

Populations of deep-burrowing clams also persisted for a period of heavy sea otter feeding in Morro Bay, California. Here 50- 60 otters fed on gaper clams for 18 months beginning in early 1984 (B. Hatfield, pers. commun.). In January 1986, 2 years after the initial arrival of sea otters, and 6 months after the otters had left Hatfield's study area, large gaper clams (mean shell length = 150 mm, SD = 20.2, n = 12) were common (3.5 m-2 ) at the site (Table 3.6; Kvitek and Anderson, unpubl.). While there are


Table 3.7. Sea otter foraging efficiencies calculated for populations of shallow- and deep-burrowing infaunal prey at Moss Landing, Elkhorn Slough and Morro Bay, California

Prey type Mean preya burrow depth (cm)

Shallow-burrowing prey

Mean prey shell length (cm)

Mean prey wet-meat weight (g)

Sedimentb excavated prey item (m 3 )

Moss Landing Beach Pismo clam (Tivela stultorum)

5 125 (1) 0.01

Deep-burrowing prey Elkhorn Slough

Gaper and Washington clams (Tresus nuttallii and Saxidomus nuttalli) (5) Inner

transect 21.1 12.2 120 0.08 Outer

transects 34.0 13.5 187 0.24

Fat Innkeeper worm (Urechis caupo) 36(3) 17.7(4) 0.09

Gaper clam (6) Morro Bay

51.4 15.0 254 0.63

Geoduck clam (Panopea generosa) 100(6) 158(7) 715 (7) 3.36

Foraging efficiency

Biomass gained per sediment volume excavated (g IIC3)

16,400

1,590

800

178

403

213

a For clams this is the distance from the sediment surface to the upper most edge of the shell. Thus the excavation depth used to calculate the volume of sediment excavated is the burrow depth plus the shell length. b A 35° angle of sediment repose was used to calculate the volume of the cone of excavation. Reference: (1) Stephenson 1977; (2) Tomlinson 1968; (3) MacGinitie and MacGinitie 1968; (4) Moss Landing Marine Labs Museum specimens (n = 6); (5) Kvitek et ai., in press; (6) Kvitek and Anderson, unpublished data; (7) Goodwin's (1976) means from Puget Sound populations (n = 1692).

no data on clam numbers before the otters arrived, they clearly did not eat all the clams after they arrived. Although the very large geoduck clam (Panopea generosa) was also present « 1 m2 ) where sea otters were feeding in Morro Bay (Table 3.6), this species did not appear to be eaten by otters. All the prey shells discarded at the site were gaper clams (N = 30). Geoducks are highly prized by clammers and are more conspicuous and much larger than gaper clams, weighing up to 6 kg (Morris et al. 1980). They have a colorful (bright orange) and large siphon (diameter> 4 em and height > 10 cm). However, the geoduck is a much deeper burrower than either the gaper or Washington clam (Morris et al. 1980). In Morro Bay, the mean burrow depth of the gaper clam was 51.4 cm (SD = 11.8, N = 24), while all the geoduck burrows (N = 10) exceeded 100 cm in depth (Kvitek and Anderson, unpubl.).

Dense populations of the butter clam (Saxidomus giganteus) have also perSisted for many years at Green Island, Prince William Sound, where clams are obtained on 40% to 75% of successful foraging dives made by sea otters (Table 3.3) (Estes et al. 1981;

Deep-Burrowing Infaunal Prey 37

Garshelis 1983). While local reduction of butter clam populations is related to sea otters in Prince William Sound (Paul and Feder 1976; Johnson 1982), there is evidence that otters and a recreational butter clam fishery can coexist (Estes and VanBlaricom 1985). The 1964 earthquake uplifted dense beds of butter clams (Baxter 1971) that had been exposed to sea otter predation for at least 10 years before the earthquake (Lensink 1962; Estes et al. 1981; Garshelis 1983). Despite sea otter predation, the uplifted beds contained 25-75 large clams (> 50 mm shell length) m-2 (VanBlaricom, unpubl. data; Estes and VanBlaricom 1985). In Orca Inlet, Prince William Sound, where otters did not arrive until 1980 (Garshelis 1983), there were similar densities of butter clams in the locally uplifted sediment.

Most of the deep-burrowing clams are relatively sessile and have long siphons (Morris et al. 1980). The one major exception is the razor clam, Siliqua patuia, which is consumed by sea otters along the north side of the Alaskan Peninsula (Table 3.1). This species is quite large (length> 15 cm) and is an active burrower (e.g., Ricketts et al. 1985).

The most important worm prey may be echiuroid worms. They are consumed in Alaska and California (Table 3.1). In Monterey Harbor, where the density of clams is very low compared to 20 years ago, echiuroid worms are now a common prey (Marianne Reidman, Monterey Bay Aquarium Sea Otter Census; authors' pers. ODS.). In the Elkhorn Slough, where the densities of both clams and echiuroids are high, the worms are rarely eaten (Kvitek et al., in press). Echiuroid worms are probably less preferred than clams by sea otters, although there are exceptions. At Green Island, Prince William Sound, Alaksa, some territorial males specialized on in tertidal echiuroids (Echiurus echiurus), almost to the exclusion of all other prey (VanBlaricom, unpubl. data).


The deep-burrowing bivalve prey are also suspension feeders that could influence the recruitment of their larvae and those of other infauna. As we discussed earlier (Sect. 3.2.2.2), this idea has been tested in only a few settings and does not appear to be a major factor affecting recruitment success (Williams 1980; Peterson 1982). In addition, our unpublished observations indicate that large numbers of small infauna successfully recruit into dense beds of gaper and Washington clams.

There is some evidence of competition for space among deep-burrowing bivalves and with other infauna (Peterson 1979b; Peterson and Andre 1980). Since otters commonly prefer one species of prey over another (Ostfeld 1982), they could reduce this competition and influence the population dynamics of the prey and their competitors. This process may already be occurring in the Elkhorn Slough (K vitek and Anderson, unpubl. data). In 1985 and 1986, sea otters foraged almost exclusively in an extensive bed of large (mean shell length = 118 mm, SD = 10.2, n = 27) and abundant (12 m-2 ,

SD = 4.8, n = 61) Washington clams. This area is directly adjacent to another extensive patch of deep-burrowing bivalves dominated by very high numbers (36 m-2 , SD = 22.8, n = 19) of good-sized (mean shell length = 77 mm, SD = 8.4, n = 99) rough piddock clams (Zirphaea pilsbryi) in which no evidence of otter feeding has been found or observed. While these two patches do differ in substrate type, both species can be found


in each area. If there is a competitive interaction between these two species, such preferential foraging over time may well lead to changes in distribution and community structure.

The positive effects of the burrows and activities of the deep burrowers may be more important than the negative effects from predation or competition. Numerous commensals live in burrows (MacGinitie and MacGinitie 1968; Schembri and Jaccarini 1978; Morris et a1. 1980) as well as within the shell cavities of deep-burrowing bivalves (Morris et a1. 1980). In addition, the siphon plates of Tresus spp. provide a hard substratum in regions otherwise dominated by soft sediments. Some 50 species have been found growing on the siphon plates of Tresus nuttallii (Stout 1970). Not only do the deep burrowers provide substrata and refugia for a variety of species, but their burrows also affect the geometry of the porewater system. By increasing the extent and depth of the sediment - water interface, the burrows facilitate the diffusion of solutes across this boundary, creating a complex mosaic of microenvironments (Aller 1982; Aller et a1. 1983; Boudreau 1984). These factors may account for the differences in distributions and activities of microbes (Aller et a1. 1983; Aller 1983), meiofauna (Bell et al. 1978; Reise 1981) and macrofauna (Ricketts et al. 1985; MacGinitie and MacGinitie 1968; Schembri and Jaccarini 1978; Morris et a1. 1980) within and around burrows. Otter predation on deep burrowers removes these structures and, we predict, the associated communities. Field manipulations of deep-burrowing prey and burrow densities are needed to test this prediction.

Sea otters, like walrus (Oliver et a1. 1985), discard the shells of bivalve prey in softsediment habitats. In some areas along the Kodiak archipelago, Alaska, foraging otters have significantly increased the availability of hard substrata for colonization, not only by discarding preyed upon shells but also by re-exposing buried shells in the process of digging. Over 75% of the laminarian species and Metridium anemones in some of these soft-sediment areas are attached to shells (authors unpub1.). Shells discarded by otters may provide shelter for additional species as documented for walruses (Oliver et a1. 1985). The community effects of shells discarded by otters could be evaluated by scattering empty shells over an area unexploited by otters and tracking the resultant changes in species composition.

Sediment must be excavated to remove the deep-burrowing prey. Therefore, the physical and biological implications of substrate disturbance are much more pronounced in this group than in the epifauna and shallow-burrowing infauna. This disturbance is considered below in Sect. 3.4.

3.3 Prey Vulnerability and Patch Dynamics

3.3.1 Prey Resistance

Sea otter prey were divided into three major functional groups that partially reflect their vulnerability to predation. Epifauna have the poorest refuge in the sediment. They are more active on the surface of the sediment and are more visually conspicuous than the two other functional groups. Cancer crabs are the most important epifaunal

Prey Resistance 39

prey and are probably most susceptible to sea otter predation. Sea otter predation was apparently thorough enough to eliminate local crab fisheries in both Alaska and California (California Department of Fish and Game 1976; Hines and Loughlin 1980; Garshelis 1983). Among the species of shallow-burrowing infauna, the effects of sea otter predation is well documented only for the Pismo clam. Although this clam has a better refuge in the sediment than the cancer crabs, it is more susceptible to sea otter predation than any other soft-bottom prey. While it took at least several years for otters to reduce the crab fisheries (ifthey were the cause), the Pismo clam fisheries were eliminated within 1 year on every beach the otters utilized (e.g., Fig. 3.2). Even preferred rocky prey species such as abalone, urchins, and rock crab persisted longer in the face of otter predation than Pismo clam populations (Wild and Ames 1974; Carroll 1982; Estes and VanBlaricom 1985). Although crabs have better vision, mobility, and active defenses than clams, these factors probably do not deter a hungry sea otter. However, despite a better sediment refuge, the shallow-burrowing Pismo clams are apparently more vulnerable to otter predation than the epifaunal crabs.

The high vulnerability of Pismo clams is probably related to the spatial dispersion of the population and the absence of other suitable prey along the beach habitat. Pismo clams inhabit a narrow band in and just behind the surf zone along wave-exposed sandy beaches (Ricketts et al. 1985). The surf zone provides excellent visual and tactile cues for locating and relocating dense clam beds. In a similar manner, groups of skates (Fager 1964) and gray whales (Kvitek and Oliver 1986) probably eliminated dense mats of polychaete prey living in this same zone. In contrast to Pismo clams, mussels and other clams such as Protothaca live in narrow well-defined bands where there are a number of other suitable sea otter prey (Morris et al. 1980). Perhaps the presence of several abundant prey accounts for their persistence compared to Pismo clams.

The deep-burrowing infauna are the least vulnerable to sea otter predation. Otters have no significant effect on deep-burrowing clam populations during the first and probably several years after they arrive at a feeding area (Hines and Loughlin 1980; Estes and VanBlaricom 1985; Kvitek et al., in press; Table 3.6). Sea otters apparently take many years to deplete these prey populations. Long-term observations exist from only one location. Monterey Harbor, where the numbers of deep-burrowing clams were high after more than 10 years of sea otter foraging, but were quite low 20 years after the first otters arrived in the harbor (Table 3.6). Qualitative observations from the harbor also suggest that gaper clams are the preferred deep-burrowing clam and are consumed before the Washington clam (Hines and Loughlin 1980). Although both species burrow to similar sediment depths, gaper clams have a more conspicuous siphon, both visually and tactily, and therefore may be located more easily than the Washington clam.

The foraging effiCiency of sea otters in soft sediments is probably influenced mostly by the time spent searching and capturing prey. Search time must be considerably less for prey with the density and predictable patchiness of the Pismo clam. Since capture time can be estimated as the biomass obtained per volume of sediment excavated, Pismo clams are easily captured (Table 3.7) as well as located. Apparently, capture time is so important in the deep-burrowing prey that sea otters do not select the largest clams and do not forage in the densest patches, but prefer patches and individuals that require less excavation (Kvitek et al., in press; Table 3.7). Since little or no excavation

40 Sea Otter and Soft-Bottom Communities

is required to obtain the epifaunal prey, capture time is probably minimal for this group.

Perhaps as a result, the dive times for deep-burrowing prey are longer than those for shallow-burrowing and epifaunal prey (Table 3.8). For example, in Sheep Bay (prince William Sound), the mean time of sea otter foraging dives was 55 s (n = 247) and the otter's diet was 70% deep-burrowing butter clams and 21 % epifaunal mussels (Mytilus edulis). In contrast, the mean time offoraging dives was much shorter at Green Island (35 s, n = 399), where the deep-burrowers (44%) and mussels (40%) occurred in nearly equal proportions in otter diets. Mean foraging dive times were also longer for otters feeding on the deep-burrowing bivalves Tresus nuttallii and Saxidomus nuttalli in the Elkhorn Slough, California (75 s, range at 35-125 s) (Kvitek et aI., in press) than for those feeding on Pismo clams in Monterey Bay, California (37.2 s, range 16- 69 s, Miller et ai. 1975) (Table 3.8). Estes et al. (1981) found a significant correlation between dive times and certain prey types in both Prince William Sound and Piedras Blancas, California. Dive times were significantly longer for otters capturing deep-burrowing clams and echiuroid worms than those feeding on epifaunal prey (Table 3.8). While the above studies are from a variety of sites which no doubt differ in many ways (water

Table 3.8. Mean dive times in seconds by prey type for otters foraging in Prince William Sound, Alaska and central California. Sample size in parentheses

Prey type Prince William Sound, Alaska Central California

Green Island (1) Sheep Bay(l) Piedras Monterey Elkhorn (N) Dive (N) Dive Blancas(1) Bay(2) Slough (3)

times times (N) Dive (N) Dive (N) Dive (s) (s) times times times

(s) (s) (s)

Epifaunal prey Mussels

Mytilus edulis (123) 18.2 (41) 40.8 Mytilus sp. (12) 38.8

Crabs Telemessus sp. (13) 41.0 (3) 46.0 Cancer antenarius (61) 48.6 Pugettia spp. (50) 39.8 uniden tified (5) 36.2 (3) 46.0 (69) 44.2

Shallow-burrowing prey Clams

Tivela stultorum (17) 37.2

Deep-burrowing prey Echiuroid worms

Urechis caupo (11) 48.2 (13) 49.3 Clams

Saxidomus nuttalli and Tresus nuttalli (15) 75.0 uniden tified (128) 50.2 (142) 58.6 (11) 62.0

Reference: (1) Estes et al. 1981; (2) Miller et al. 1975; (3) Kvitek et aI., in press.

Substrate Disturbance 41

depth, currents, sediment type, etc.) and thus present uncontrolled contrasts, the consistent results suggest a positive relationship between dive times and depths of prey in the sediment.

3.3.2 Prey Resilience

In addition to prey resistance, the other component of prey vulnerability is the resilience of prey populations after exploitation by sea otters. There is essentially no information on how different species of soft-sediment prey recover from sea otter predation. The Pismo clam typifies this lack of knowledge. This species has pelagic larvae that live in the plankton several weeks before settlement (Fitch 1950). We do not know how they are released by local populations, how they disperse in local and regional currents, or how they reach the narrow zone where most adults live. What is known is that recruitment is highly variable in both space and time, with major recruitment events often separated by many years. These are dramatically variable among different beaches, and are not correlated with the size of the breeding population (Fitch 1950).

There may be a number of predictable patterns in the resilience of prey populations related to life histories or even among the major functional groups. Perhaps recruitment in deep-burrowing prey is higher than recruitment in the shallov] burrowers and epifauna, because the deep spatial refuge insures a larger breeding population. If so, in addition to a higher resistance to predation, the deep burrowers may also be more resilient to otter predation. When patterns of recruitment and survival are better documented in sea otter prey (e.g., Muus 1973), this and other trends can be evaluated. Today, our understanding of prey vulnerability depends on information about spatial refuges that help prey to resist predation, not on prey resilience.

3.4 Substrate Disturbance

In addition to foraging by sight (Kenyon 1969), sea otters probably winnow surface sediments with their forelimbs and sift through superficial sediment with their paws in searching and capturing epifaunal and shallow-burrowing prey. Divers routinely search and collect relatively shallow-burrowing infauna, especially clams, by winnowing sediment with back-and-forth movements of the hands and by dragging fingers through the surface sediment until a clam is contacted. These methods are excellent for locating shallow burrowers such as Protothaca, Macoma, and Tivekl (authors, pers. observ.). Sea otters also dig deeper pits and furrows in the bottom with their forelimbs (Shimek 1977a; Hines and Loughlin 1980). Both visual and tactile cues may also be used to locate deep-burrowing clams, which often have visually conspicuous burrows and Siphons that protrude above the bottom and retract relatively slowly after contact.

In general, the capture of soft-sediment prey is very different from the capture of rocky prey. Sea otters must dig considerable quantities of sediment to capture many infaunal prey (Shimek 1977a; Calkins 1978; Hines and Loughlin 1980; Kvitek et a1., in press), and probably winnow and sift through even larger volumes of surface sedi-


ment to capture others. A sea otter was observed enlarging an existing excavation to 50 X 30 X 20 cm to capture fat innkeeper worms (Urechis caupo), by digging with the forepaws and rooting with the head while maintaining position by kicking the hind legs (Shimek 1977a). In Monterey Harbor, a sea otter was observed excavating a deep trench (1.5 X 0.5 X 0.5 m), where it obtained six deep-burrowing clams (gaper and Washington clams) on six successive dives. It dug in a dog-like fashion (Hines and Loughlin 1980), which produces characteristic features with sediment deposited just outside the feeding pits (Oliver et al. 1983; Kvitek et al., in press).

These and other feeding excavations of sea otters produce a record of their foraging activity in the form of pits and trenches in soft sediments (Table 3.5). Changes in foraging behavior are reflected in this bottom-feeding record. When prey are dense and live deep in the sediment, the most efficient strategy (Table 3.7) may be to enlarge an existing excavation on successive dives rather than dig a new pit for each prey item or on each dive. This is the type of trenching behavior observed in a dense (mean = 13.5 m 2 ,

Table 3.6) patch of deep-burrowing (to 50 cm) gaper and Washington clams in Monterey Harbor (Hines and Loughlin 1980). In Elkhorn Slough, Kvitek et aI. (in press) measuring only fresh pits, found that sea otters dug the largest pits (4 and 6 m2 ) in dense patches of the deepest burrowing gaper and Washington clams (mean burrow depth = 34 cm, siphon density = 18.2 m2 ). In contrast, otters dug much smaller pits (0.25-1.0 m2 ) in a patch of shallower-burrowing individuals at similar density (mean burrow depth = 21.1 em, siphon density = 17.6 m2 ) only 25 m away from the first patch. Apparently otters vary their foraging behavior when prey occur at different depths in the sediment (K vitek et aI., in press). Clearly, sea otters can disturb much of the substrate, the primary benthic habitat, in a local soft-sediment feeding ground. VanBlaricom (unpubI. data) found between 7 and 12% of the area disturbed in an in tertidal feeding ground at Green Island, Prince William Sound, Alaska. Virtually all the pits were made by the same otter feeding on echiuroid worms (Echiurus echiurus).

Although sea otters mainly consume large invertebrates (> 3 cm in length) that rarely occur in densities greater than several dozen per square meter, searching and capturing prey in soft sediments probably disrupts very large numbers of smaller infauna « 1 mm in length). This is well documented for feeding rays, gray whales and walruses (VanBlaricom 1982; Oliver and Slattery 1985; Oliver et aI. 1985) but not for sea otters. Considerable areas of the sea floor may be highly disturbed when large numbers of otters aggregate and feed on infaunal prey. In June 1982, over 2000 sea otters foraged between Izembech Lagoon and Amak Island (Fig. 3.1), producing a dense record of feeding pits on the sea floor in water depths of 10-20 m, where there were as many as ten razor clams and one feeding pit m-2 (E.F. O'Connor, pers. commun.). The pits were about 30 cmin diameter and 10-20 cm deep in a fme sand bottom. Like feeding pits from other regions (Table 3.5), they were probably excavated with the otter's forelimbs, as small excavation piles were often next to pits.

Winnowing, sifting, and excavating sediment probably displaces, buries, or damages many thousands of small infauna per square meter of sea floor. Others are forced into the water, where fishes quickly consume them (e.g., VanBlaricom 1982). The relatively defaunated feeding depressions trap suspended material and larger items moving along the sediment surface. This organic-rich deposit can be colonized by a number of relatively opportunistic benthic animals, especially crustaceans. While there are no publish-

Rocky and Soft Substrata 43

ed data on the infauna in the feeding depressions of sea otters, the feeding depressions of other bottom-feeding marine mammals and rays have been sampled. These data show that feeding disturbances cause dramatic declines in the density and biomass of infaunal communities (VanBlaricom 1982; Nerini and Oliver 1983; Oliver et al. 1983; Oliver and Slattery 1985; K vitek and Oliver, in press), and except for the walrus (Oliver et al. 1985), many infauna colonize the fresh feeding excavations.

Very dense tube-mats of sabellid polychaetes occur in the feeding grounds of sea otters around Kodiak Island (authors' unpubl.). Since several species of clam prey are abundant in the mats, otters are likely to destroy this community when the sites are located and exploited. Since many historical feeding grounds have not been recolonized by the expanding sea otter populations in Alaska, there are unique opportunities to document the short and long-term changes in prey communities caused by the feeding disturbances of otters.

The foraging activities of sea otters probably feed a number of scavengers both inside and outside excavations. Excavating otters may expose numerous infaunal species to predation by fishes (e .g., VanBlaricom 1982). Scavengers probably consume remnant soft tissue on discarded clam shells, while other species use the shells for habitat, attaching to the surface or nestling under the shell for protection (e.g., Oliver et al. 1985). Since feeding excavations and discarded shells can be easily reproduced in an experimental setting (VanBlaricom 1982; Nerini and Oliver 1983; Oliver et al. 1985;Oliver and Slattery 1985), the response of the benthic community to sea otter feeding can be explored with considerable success.

3.5 Rocky and Soft Substrata

Sea otters apparently have a similar effect on the major invertebrate prey on the rocky shore and the epifaunal and shallow-burrowinginfauna from soft-bottom habitats. They generally select the largest individuals and cause major declines in population size. This is true for sea urchins (Estes and Palmisano 1974; Lowry and Pearse 1973) and abalone (Lowry and Pearse 1973; Hines and Pearse 1982) and probably for cancer crabs (California Department of Fish and Game 1976; Garshelis 1983) and Pismo clams (Stephenson 1977).

Although refuges from sea otter predation occur in both rocky and soft substrate communities, the nature and consequences of the refuges are very different. In rocky areas, the only protection available is for smaller individuals in crevices too small for otters to enter (Lowry and Pearse 1973; Hines and Pearse 1982). Since the abalone fishery depends on larger individuals that do not benefit from the crevice refuge, sea otters and commercial exploitation of rocky epifauna are often not compatible.

Soft-sediment refugia are fundamentally different. Unlike a rock crevice, which physically prevents the entry of an otter and is size-selective for smaller individuals, a sediment refuge is not absolute and may not be size-selective, at least among the size classes of prey consumed by otters. Infaunal refuges exist for deep-burrowers like the gaper, Washington, and geoduck clams and for some shallow-burrowers embedded in the tough rhizome mats of eelgrass beds such as the littleneck clam (Protothaca staminea)


(Estes and VanBlaricom 1985). These infaunal refuges are not absolute because sea otters can no doubt excavate even the deepest burrowers. However, the existing evidence indicates that otters maximize the biomass gained/volume of sediment excavated (K vitek et aI., in press). As a result, the larger individuals living at greater burrow depths are probably consumed less than smaller clams living at shallower depths (Table 3.7). It is important to emphasize that within the size range of clams eaten by sea otters, there is not always a simple relationship between individual size and depth into the sediment. A large range of adult sizes can be found deep in the sediment in several deep-burrowing species (e.g., Kvitek et aI., in press).

There is no obvious mechanism for assessing the size of a clam before it is excavated by a sea otter. This is not a problem when feeding on shallow burrowers like the Pismo clam which can be sorted to size by pawing through a small volume of unconsolidated surface sediment typical of its habitat (Morris et ai. 1980). As a result, otters preyed primarily on the larger size classes of Pismo clams at Moss Landing beach (Stephenson 1977). On the other hand, it is costly and inefficient for otters to discard smaller individuals of deep-burrowing prey during excavation. As a result, sea otters preyed on all the size classes of gaper and Washington clams in Elkhorn Slough (Kvitek et aI., in press).

Many large individuals of the deep-burrowing prey remain uncaptured for many years in spite of persistent foraging activities of sea otters. These large individuals probably have a high reproductive output, which may help to maintain population numbers even during heavy sea otter predation. This pattern of otter predation and prey resistance could allow sea otters and a competing clam fishery to coexist. In Prince William Sound, although butter and littleneck clams have been an important part of the sea otters' diet for many years, the same area still supports a productive sport fishery (Estes and VanBlaricom 1985).

There are no major invertebrate prey on the rocky shore like the deep-burrowing infauna. In addition to their depth refuge, these species often live in large geographical areas (Morris et ai. 1980). Therefore, it is much more difficult for sea otters to locate and consume most of a population, except locally where the prey live in well-defined patches such as Monterey Harbor. Nevertheless, it took between 10 and 20 years for sea otters to remove most of the deep-burrowing clams from the harbor, if in fact the otters were the primary cause. With the exception of the Pismo clam, mussels, and littleneck clams, most soft-sediment prey live in much larger habitat areas than the prey from rocky shores. They have little or no distinguishing features on the bottom and no visual markers such as the kelp canopy. As a result, most soft-sediment prey may enjoy better spatial refuges than rocky prey by virtue of the habitat area and absence offeatures for effective searching and orientation.

The community role of sea otters in soft-bottom habitats must be very different from the role described along the rocky shore. Here, sea otters eat herbivores that eat algae. When herbivore numbers are reduced, the algal cover increases and the kelp canopy expands (Estes and Palmisano 1974; Dayton 1975; Estes et al. 1978; Pearse and Hines 1979; Duggins 1980). Fishes are attracted to the new habitats provided by algal structures (Estes and Palmisano 1974; Estes et al. 1978; Estes et ai. 1982). No similar process occurs in soft sediments, where macroalgae are very rare. Moreover, there is no convincing evidence from the numerous studies of predation in soft-bot-

Summary 45

tom communities that any soft-sediment predator acts like the keystone predator described from the rocky shore (Peterson 1979b).

3.6 Future Research

The greatest future research opportunities involve the movement of sea otters into new soft-sediment feeding grounds. Much of the historical sea otter range has not been recolonized by otters. Otters commonly move in dense aggregations or foraging fronts as they exploit new feeding grounds. As a result, relatively unexploited feeding grounds can be located, sampled, and experimentally manipulated before the foraging front arrives and can be monitored during its presence and after it moves. It is also possible to locate unexploited areas to control for changes in prey communities that are not caused by sea otter feeding and to find several unexploited areas to serve as replicate treatments. Although the control and replication of natural experiments such as the exploitation of a new feeding ground are much more variable than the experimental setting in the laboratory or in field manipulations, these experiments have tremendous potential for revealing the short and long-term effects of sea otters on prey populations, communities and habitats.

This chapter considers the vulnerability, resistance and resilience, of various prey to otter predation, based on a handful of incomplete and usually highly qualitative field studies. Only one study quantifies prey communities and sea otter activities during the colonization of a new feeding ground, and this involves only the initial colonization period. There is no information on the resilience of prey populations and on the disturbance of associated nonprey communities and habitats. These problems can be addressed best in the natural experiments described above. In addition, within these newly exploited feeding grounds, many manipulative field experiments can be established to evaluate specific hypotheses about prey preference, size selection, excavation behavior, and other foraging tactics, and the numerous potential responses of the benthic community to sea otter disturbance.

3.7 Summary

As sea otter populations expand back to their aboriginal range, they are also increasing their use of estuaries, sandy beaches, and other soft-sediment areas. While it has long been known that otters forage in such soft-sediment habitats, we know very little about their role in structuring these communities as compared to hard-substrate areas.

In this chapter we review what is known about the foraging habits of sea otters in soft-sediment environments, including their influence on prey populations and communities. The diets of sea otters and the different prey communities in soft-bottom habitats are also described. We consider the vulnerability of prey populations, the community effects of feeding disturbance, and some major differences between foraging on soft and rocky substrates. Whenever possible, we suggest where future research is


needed to increase our understanding of the community ecology of sea otters in softsediment environments.

Sea otters consume over 30 species (primarily clams) of soft-sediment prey. These can be divided into three functional groups, which represent a hierarchy of potential vulnerability to sea otter predation: epifauna, shallow-burrowing in fauna and deepburrowing infauna. The most important epifaunal prey are crabs, especially the Dungeness and red crabs (Cancer magister and Cancer productus) , which are also fished commercially. Fishery observations in both Alaska and California suggest that sea otters reduce the numbers oflarge cancer crabs in local populations. This reduction may have an important effect on infaunal invertebrate communities, since some species of epifaunal crabs are known to influence their prey populations.

The shallow-burrowing infaunal prey are primarily suspension-feeding clams that live in the surface layer of the sediment. The best-known shallow-burrowing prey species is the Pismo clam, Tivela stultorum, which lives in the lower intertidal and shallow subtidal sandy beaches along the wave-exposed coast of central California. Since there is a major sport fishery for Pismo clams, information on catch prior to the arrival of sea otters is available along several beaches. These data reveal that the number of Pismo clams caught by clammers decreased markedly after foraging sea otters first arrived at several California beaches. The decline in these fisheries may be a result of several factors in addition to sea otter predation, including extremely variable recruitment success, physical, and human factors. Other shallow-burrowing prey species have been shown to be less vulnerable (e.g., mussels and littleneck clams). However, these are also found in areas that present otters with a greater variety of alternate prey, unlike the Pismo clam, which is often the only large prey species available in its habitat. While suspension feeders may block recruitment into an area by ingesting settling larvae, including their own, this theory is not well supported by the evidence. Furthermore, recruitment success does not appear to be related to stock size. As a result, the removal of local populations of these prey species by otters may have little effect on subsequent colonization. The disturbance caused by the otters digging a few centimeters into the sediment to obtain shallow-burrowing prey is probably of little consequence, given the highly dynamic nature of most of these habitats.

The deep-burrowing in fauna are primarily species of long-siphoned clams that often live below 30 cm and sometimes deeper than 1 m into the sediment. Several popUlations of these deeply buried species have persisted for longer periods of time than the Pismo clam or crabs in spite of otter predation, suggesting an effective depth refuge. This refuge may be important in maintaining a larger brood stock and as a consequence deep-dwelling infaunal prey may be more resistant (not rapidly depleted) and resilient (prompt recovery following depletion) to otter predation. Unlike the epifauna and shallow-burrowing infauna, sea otters must excavate considerable quantities of sediment to obtain the deep-burrowing species. In the process they create distinct feeding excavations or pits on the sea floor which greatly modify the primary substrate, the structure of the porewater system and the burrow-associated microhabitats. The discarded shells of bivalve prey also provide hard substratum for colonization by fouling organisms in habitats where this substrate is otherwise scarce. Although shells are also exposed for shallow-burrowing prey, these shells are probably moved and buried more rapidly in the dynamic shallow-water and wave-swept environments where Pismo clams

Summary 47

live. Thus, sea otter predation on deep-burrowing infauna may have the broadest ecological effects despite this prey's reduced vulnerability to otter predation.

In rocky habitats otters select the largest individuals of prey and cause major and rapid declines in preferred prey populations. In soft-sediment areas this is only documented for the epifaunal and shallow-burrowing in faunal prey. Prey refuges in rocky areas are size-selective for smaller individuals, whereas the sediment depth refuges available to deep burrowers are not absolute and benefit a range of larger size classes at a given depth. Many soft-sediment prey enjoy better refuges than rocky prey due to greater habitat expanse and a more cryptic life habit. Since rocky habitats are like small islands in a sea of soft sediment, sea otters may depend on soft-bottom habitats as much or even more than on the rocky shore.

Acknowledgments. We would like to thank both Charles Peterson and Anson Hines for their patient and thorough reviews of this chapter. We are also indebted to the U.S. Fish and Wildlife Service and Tony DeGange for their excellent support on Kodiak Island, Alaska. Additional help in the field has been provided by Brian Anderson, Andrew deVogelaere, and Heath Foxlee. Logistical and manuscript preparation support was provided in part by Mark Silberstein and Elkhorn Slough Foundation along with a grant from the National Estuarine Sanctuary branch of NOAA.

4 Effects of Foraging by Sea Otters on Mussel-Dominated Intertidal Communities G. R. VANBLARICOM

4.1 Introduction

My purpose here is to examine the role of predation by sea otters in two mussel-dominated intertidal benthic communities in the North Pacific. I defme mussel-dominated communities as those in which mussels (Pelecypoda, family Mytilidae, especially Mytilus spp.) are a significant part of community biomass and secondary productivity, and appear to be capable of competitively excluding many other sessile species in the absence of interfering factors. I will present data from concurrent long-term field studies on the coasts of central California and south central Alaska, and will list issues which are, in my view, unresolved but essential areas for additional research on the relationships of sea otters and mussel-dominated communities. This chapter is limited to consideration of mussel populations containing significant numbers of adults. Consideration of factors influencing the establishment of new mussel populations is beyond the scope of this chapter, although such factors clearly are important in mussel-dominated communities.

4.2 Models of Mussel-Dominated Communities

Ecologists continue to seek an understanding of mechanisms which permit co-occurring species to resolve competing needs for a common resource in short supply. Such mechanisms have a direct bearing on patterns of life history, relative abundance, local diversity, and spatial and temporal patchiness in natural communities.

Studies of mussel-dominated communities have contributed substantially to our knowledge of processes which influence the provision and utilization of vital resources. Such studies have had a prominent influence for several reasons (see also Paine 1977, 1980): (1) The predominant organisms are often sessile or slow-moving, and are easily counted, measured, and followed over time; (2) adult populations of intertidal organisms are often distributed on relatively small, operationally manageable spatial scales; (3) mussel-dominated communities often contain guilds of species for which primary space on rock surfaces is an essential, easily quantifiable resource (Paine 1966, 1974). These species often form a secondary matrix of physical structure on rock surfaces, providing extensive surface area and interstitial habitat for a number of additional species (Suchanek 1979). Thus, much of the biological structure of the community emanates from patterns of space utilization by a relatively small number of species whose

Models of Mussel-Dominated Communities 49

abundances are easily mani pula ted for experimen tal purposes; (4) life histories of many species in such communities operate on time scales sufficiently brief that meaningful results can be obtained from field work months or years in length (Suchanek 1981), rather than decades or centuries; (5) mussel-dominated intertidal communities are often characterized by strong interactions (sensu MacArthur 1972; Paine 1980). In many cases, competition, predation, and disturbance are easily recognized and quantified, and often have significant consequences for the biological organization of the community (Suchanek 1985).

Two processes are of substantial importance in maintaining the structure of musseldominated communities: consumption of mussels by predators, and creation of discrete gaps in mussel stands by physical disturbance. It has been shown that predation by sea stars or snails prevents mussel populations from dominating primary space in certain well-defined portions of rocky intertidal habitats (Paine 1966,1971,1974; Menge 1976; Lubchenco and Menge 1978; Suchanek 1978; Peterson 1979a). On the outer coast of the northwestern U.s., absence of sea stars from a given area for sufficient time allows mussels to occupy the area through larval recruitment or immigration of adults. Mussels eventually overgrow and exclude other species of the space-using guild (Paine 1966, 1974). Effects of the return of sea stars depend on the duration of their absence, and on growth characteristics of the resident mussel populations. If sea stars are absent for a relatively short time, on their return they most likely will consume most mussels which have colonized the site, releasing primary space for exploitation by other species. If sea stars are absent for a sufficiently long interval, colonizing mussels can grow to a size large enough to resist consumption by sea stars (Paine 1976). In other regions, however, mussels may not grow large enough to resist consumption by sea stars (e.g., Paine et al. 1985). In such cases, the return of sea stars leads to removal of colonizing mussels regardless of the duration of the absence.

Local physical disturbances are important to mussel-dominated communities. Principal sources are direct impact and shear stress from the large breaking waves which typically sweep open-coast rocky intertidal habitats (Lubchenco and Menge 1978; Paine and Levin 1981 ; Witman and Suchanek 1984), and crushing by heavy wave-borne objects such as drifting logs (Dayton 1971). As a result of persistent long-term study of mussel populations at several open-coast sites, there is an emerging understanding of the complex relationships among disturbances of varying frequency, intensity, and size, and the dynamics of mussel-dominated communities (Levin and Paine 1974, 1975; Levin 1976, 1978, 1981; Paine and Levin 1981, Sousa 1984a,b, 1985). Several points need emphasis here: (1) The primary effect of local physical disturbance to an established mussel population is functionally similar to the effect of removal of mussels by sea stars (Sousa 1984b); (2) rates of physical disturbance vary significantly over time. There are recognizable rate differences among seasons within years and within seasons among years (Paine and Levin 1981); (3) patches of mussels dominated by large individuals are relatively immune to depredation by sea stars, but are highly vulnerable to shear stresses generated by large waves (Paine 1976; Paine and Levin 1981). Relative risk levels are reversed in patches of mussels dominated by small individuals; (4) the biological recovery sequence in a gap (sensu Paine and Levin 1981) in a stand of mussels depends on the size of the gap (Suchanek 1978, 1979; Paine and Levin 1981; Sousa 1984a, 1985).

50 Sea Otters and Mussel-Dominated Communities

4.3 Sea Otters as Predators of Mussels

There are a number of published accounts of the foraging habits of sea otters at various sites in the North Pacific. Many of these indicate that sea otters eat intertidal mussels (Table 4.1). In some cases, mussels are a major part of the diet of sea otters (e.g., Hall

Table 4.1. Summary of available data on consumption of intertidal mussels by sea otters

Location Date Method Mussels N Reference in diet (% freq.)

Kurile Islands, V.S.S.R.a: Vrup Island 1969-1970 Spraint analysis 11-31 617 Shitikov 1971 Chirpoi Island 1969-1970 Spraint analysis 11 308 Shitikov 1971 Simushir Island 1969-1970 Spraint analysis 10 40 Shitikov 1971 Onekotan Island 1969-1970 Spraint analysis 28 18 Shitikov 1971 Pararnushir Island 1968 Spraint analysis 62 116 Maminovand

Shitikov 1970 Paramushir Island 1969-1970 S praint analysis 57 219 Shitikov 1971

Aleutian Islands, V.S.b : Attu Island 1976-1977 Direct observation 3.0 580 Estes et al. 1981 Amchitka Island 1959 Spraint analysis 0 422 Kenyon 1969 Amchitka Island 1970's Direct observation 0.2 563 Estes et al. 1981

Shumagin Islands, V.S. b : Simeonof, Nagai Islands 1960 Spraint analysis 77 75 Kenyon 1969

Prince William Sound, V.S. b : Green Island late 1970's Direct observation 39.7 420 Estes et al. 1981 Green Island 1980-1981 Direct observation 6.0 780 Garshelis 1983 Green Island 1980-1981 Spraint analysis 34.0 158 Garshelis 1983 Montague Island 1971 Direct observation 0.3 c 597 Calkins 1978 Sheep Bay late 1970's Direct observation 21.8 251 Estes et al. 1981 Orca Inlet 1980-1981 Direct observation 0 115 Garshelis 1983 Nelson Bay 1980-1981 Direct observation 0 358 Garshelis 1983

California, V .S. d: Pt. Santa Cruz 1977 -1979 Direct observation 0.4 748 Ostfeld 1982 Pt. Lobos 1963 Direct observation 35.7 510 Hall and Schaller

1964 Rocky Point 1956 Direct observation 5.0 5961 McLean 1962 Pt. Piedras Blancas 1976 Direct observation 2.1 820 Estes et al. 1981 Pico Creek 1966 Direct observation 0.8 243 Ebert 1968 Pt. Buchon 1973 Direct observation 7.1 266 Wild and Ames

1974 Pt. Buchon 1975 Direct observation 0.3 567 Estes et al. 1981

a Mussel category includes lumped frequencies of Mytilus edulis and Modiolus spp. In the Kurile Islands, especially near Paramushir Island, combined densities of mussels can reach 10 kg m-2 at depths of 20-50 m (Kuznetsov 1963), within the diving range of sea otters. b Mussels are Mytilus edulis, and are abundant only in the intertidal zone. c Calkins noted difficulty in enumerating mussels captured by sea otters when observation distance exceeded 100 m. d Mussels are Mytilus californianus, and are abundant only in the intertidal zone.

Case 1: Sea Otters and Mussels on the Coast of Central California 51

and Schaller 1964; Estes et al. 1981). Sea otters are well known for their ability to influence density, demography, and distribution of other marine benthic invertebrates (e.g., Lowry and Pearse 1973; Estes and Palmisano 1974; Kvitek and Oliver, this Vol.; Laur et aI., this VoL), and for the consequent effects on the dynamics of some sublittoral benthic communities (e.g., Estes et al. 1978; Simenstad et al. 1978; Duggins 1980, this Vol.; Breen et al. 1982; Laur et aI., this VoL). Given the broadly recognized significance of the empirical models outlined above, it is a matter of concern that the role of sea otters in mussel-dominated communities is largely unknown, with the exception of indirect effects resulting from sea otter-kelp forest interactions (Palmisano 1975, 1983; Palmisano and Estes 1977). I now address that concern through presentation of the following case studies.

4.4 Case 1: Sea Otters and Mussels on the Coast of Central California

From 1978 through 1984 I studied the relationship of sea otter foraging and the dynamics of populations of mussels (Mytilus californianus Conrad) at Pt. Piedras Blancas (35 0 40'N, 121 0 17'W), an exposed rocky headland in northern San Luis Obispo County, California (Figs. 4.1,4.2 and 4.3). Here I focus on the distribution of mussel consumption by sea otters with regard to time, space, and the size range of individuals present in mussel populations.

Q N

Pt. Piedras/ Blancas

PACIFIC

OCEAN

I- 'I 4000km

Fig. 4.1. Map of the Pacific Ocean, showing the location of the two study areas and the original coastal distribution of sea otters (hatched areas; see also Riedman and Estes, this Vol.)

52

Tatoosh Island

48°24'N,124°44'W

Mukkaw Bay

48°19'N,124°40'W

Z « IJ.J ()

o ()

u. () « a.

Sea Otter Range

~ Aboriginal

• Present

4.4.1 Study Location

OREGON

Sea Otters and Mussel-Dominated Communities

USA

MEXICO

Fig. 4.2. Coastline map of Washington, Oregon, and California (USA), showing aboriginal and present distributions of sea otters, location of the study site at Pt. Piedras Blancas, and locations of two sites in Washington used by R.T. Paine for long-term study of mussel-dominated communities

Study sites are located on a wave-cut expanse of metamorphosed Jurassic pillow basalts (Hall 1976) on the western shore of the headland (Fig. 4.3). The study sites are fully exposed to oceanic sea and swell. Wave heights typically are 1-2 m, increasing to 4-10 m during winter storms (National Data Buoy Service, unpubl.; U.S. Fish and Wildlife Service, unpubl.). Salinities in the area typically are 32- 33 g kg-1 (Fleminger and Klein 1963). Sea surface temperatures ranged from 90 to 18°C during the study, but are most often in the range 10°-14°C (Fish and Wildlife Service, unpubl.). Air tempera-

Fig. 4.3. Study sites at Pt. Piedras Blancas, California. Top photograph Aerial view of the Point, showing typical exposure to oceanic swell coming from the northwest (left). The three permanent study areas are located on an exposed bench of weathered basalt shown by the arrow. Center photo· graph Study areas on the northwest side of the point, as viewed from the blufftop overlook used to survey the area for foraging sea otters (see Sect. 4.4.2).Arrows indicate permanent study sites (from left Channel Bank, Mussel Island, and Pisaster Point). Bottom photograph Mussel Island study area, showing dense mussel cover (appears light gray) and the "lee area" (lower right, see also Sect. 4.4.4) used by sea otters to sort and consume mussels. Although partially obscured in this photograph, a deep surge channel surrounds the site

Study Location 53

Fig. 4.3


tures rarely exceed 20°C and rarely fall below 5°C (National Weather Service, unpubl.). Fog is common during summer.

At the descriptive level, biological communities at the study sites bear a strong resemblance to those described on the outer coast of the Olympic Penninsula, Washington. There are large areas in which the mussel Mytilus californianus is the predominant user of primary substratum. Other common species include Alaria marginata, Postelsia palmaeformis, Phyllospadix spp., Pollicipes polymerus, and Pisaster ochraceus.

Sea otters were removed from the area by fur hunters sometime during the nineteenth century (Odgen 1941; Kenyon 1969; Riedman and Estes, this Vol.), but returned to the area in 1959 and have been present without interruption since that time (Wild and Ames 1974; Estes and Jameson 1983b; Riedman and Estes, this Vol.). Typical densities in the area are five to eight individuals per linear km of shoreline (R. Jameson, personal commun.). All age and sex categories of sea otters have been seen in the vicinity of the study area in recent years.

4.4.2 Methods

Three study sites (Pisaster Point, Mussel Island, Channel Bank) were established on the exposed western shore of Pt. Piedras Blancas during spring 1978 (Figs. 4.2 and 4.3). Each site is characterized by an upper zone (elevations of approximately + 1 m MLLW and above) in which primary space is dominated by Mytilus californianus and Pollicipes polymerus, and a lower zone (below +1 m MLLW) in which Pisaster ochraceus is abundant and algae and surfgrasses are the dominant cover. Sites were chosen to be as similar as possible with regard to exposure to incoming oceanic waves. Each study site was visited at least once during each series of minus tides (effectively, about twice per month). Sustained rough seas occasionally prevented site visits. Over the entire duration of the study, the average interval between site visits was 20 days. During each visit, sites were carefully surveyed for the appearance of new gaps in mussel-dominated zones. When a gap was found, the source (sea otter foraging or wave shear) was identified on the basis of wave activity and otter surveys (see below) since the previous visit. Gaps were permanently tagged by attaching a numbered plastic disc to the bare substratum with marine epoxy compound. Dimensions of gaps were recorded by placing a meter bar or surveyor's tape across the long dimension of the gap. At 10-cm intervals along the bar or tape, perpendicular distances to the upper and lower edges of the gap were recorded. This allowed the reconstruction of details of gap dimensions without the need for assumptions regarding gap shape. Gaps monitored in this study frequently had irregular shapes.

In central California, sea otters typically forage during early morning hours, enter a period of grooming and resting during an interval from approximately 0830 to 1230 h local standard time, then resume feeding during the early afternoon (Hall and Schaller 1964; Loughlin 1977, 1979; Ribic 1982). Although there can be substantialindividual variation in this pattern, most observations of foraging by sea otters in the study areas were in fact made during early morning or early afternoon. Study sites were scanned an average of 3 days per week, most often between the hours of 0800 and 0830, but frequently in early afternoon as well, to look for foraging sea otters in the area. Searches

Consumption of Mytilus californianus by Sea Otters: The Basic Pattern 55

for otters were made from the top of a nearby bluff which overlooks the study site (Fig. 4.3). These efforts were, on occasion, disrupted by fog, heavy rainfall, or rough seas which obscured the study sites.

On several occasions when sea otters were observed foraging in the study areas, data were collected by watching the otters with a spotting telescope or binoculars from the bluff top overlook. When conditions permitted, prey species were identified and recorded after each foraging dive, along with qualitative notes on the nature of the foraging behavior. When prey items included mussels, rough estimates were made of maximum shell length by comparing mussel size with size of known features of sea otter morphology, such as paw width or head diameter.

Samples of live mussels were collected from the intertidal by blind, haphazard drop of an indicator in an area of interest, and removal (down to the primary substratum) of all mussels in a O.l-m2 area about the site where the indicator fell. Mussels were subsequently washed over a screen with O.5-mm mesh, separated from debris and from one another, and me~sured to the nearest mm with Vernier calipers. A special effort was made to search screening residues and the byssal threads of large mussels in order to locate minute juvenile mussels.

4.4.3 Consumption of Mytilus californianus by Sea Otters: The Basic Pattern

Sea otters of all age and sex categories were seen foraging on mussels on a number of occasions in the study sites. Mussel predation by otters coincided with the appearance of gaps in mussel cover, as described below and plotted in Figs. 4.5 and 4.7. As outlined below, foraging activity was aggregated in both space and time, on several scales. The general features of foraging activity were, however, relatively consistent, and are comparable to other accounts of mussel consumption by sea otters (Hall and Schaller 1964; Kovnat 1982). Sea otters typically obtain mussels by diving on submerged musseldominated areas during high water. A clump of several mussels is removed and brought to the surface during each successful dive. After surfacing the foraging sea otter usually swims a distance of 10-20 m into deeper water away from areas of breaking waves. During the surface interval, the sea otter rapidly separates and sorts mussels. Small mussels are immediately dropped into the water. Typically, larger mussels are broken open by rapid pounding on a flat stone balanced on the upper chest area. Hall and Schaller (1964) provide a detailed description of the use of stones as anvils by sea otters for breaking mussel shells. When a stone is absent, the otters break mussel shells by striking one mussel against another. Once the shells are opened, soft tissues are removed by scraping with the tongue, lower incisors, and canines. Shell fragments are discarded, the otter rolls in the water to clean the pelage, and then repeats the procedure for the next mussel. When all mussels from a clump have been consumed or discarded, the sea otter typically swims toward the site of the previous dive to make the next foraging dive.

When foraging on mussels, otters dive repeatedly at the same location to remove clumps of mussels. The result is the creation of discrete gaps in the mussel-dominated zone. Removal of mussels typically is complete within the gaps; edges of gaps are sharply defined and substrata within gaps are primarily bare rock, to which remnant clumps


Fig. 4.4. Photographs of gaps made in stands of Mytilus californianus by foraging sea otters. Top photograph Large gap, several months after formation. Width of gap is approximately 1.5 m. Surviving mussels and barnacles are seen across the upper third and left center of photograph. Cleared area is seen in center and right center. Ephemeral algae have begun to appear in the gap. Bottom photograph Close-up photograph of the rock surface of a newly formed gap in mussel cover. Remaining byssal threads from mussels are clearly seen. Barnacles are Semiba/anus cariosus. Numbers on tape indicate cm; individual lines on tape are at intervals of 2 mm

Creation of Gaps in Mussel Cover by Sea Otters: Spatial and Temporal Aggregation 57

of byssal threads from recent removed mussels are still attached (Fig. 4.4). Formation of a gap by a sea otter sometimes involves several foraging episodes spread over more than 1 day.

On occasion, sea otters are known to leave the water and climb into the intertidal to collect mussels, abalones, and sea urchins (Harrold and Hardin 1986; unpublished observations of R. Jameson, G. VanBlaricom, and J. Vandevere). Harrold and Hardin (1986) report a brief observation of a sea otter consuming mussels on an emergent mussel bed at low tide, at a site 7.2 km north of Pt. Piedras Blancas in May 1984. Mussels were removed in two discrete patches, 60 and 3S cm2 in area. The sea otter did not enter the water to handle or consume mussels. I saw no foraging of this kind in the study sites at Pt. Piedras Blancas, but its occurrence cannot be ruled out.

4.4.4 Creation of Gaps in Mussel Cover by Sea Otters: Spatial and Temporal Aggregation

Sea otters were seen foraging on mussels in only two of the three study sites (Mussel Island and Pisaster Point) during the course of field work. Forty gaps in mussel patches were found in study sites during the period 1978-1984. Of these, 31 were made by sea otters at Mussel Island (see below). Four of these gaps subsequently were merged into a single larger gap by wave shear during the major wave events of winter 1983. Eight gaps appeared at Pisaster Point, and were formed by sea otters, Pisaster ochraceus, or wave shear. The single gap seen at Channel Bank appeared after a localized influx of Pisaster ochraceus during calm weather.

Gaps were determined to be a result of sea otter foraging only if the following criteria were met: (1) sea otters were seen removing mussels from the site since the previous visit: (2) no episodes of waves large enough to make gaps occurred since the previous visit; (3) large numbers of Pisaster ochraceus were not present in or near the new gap. Because Pisaster were systematically removed from Mussel Island from 1978 through 1984 (VanBlaricom, in prep.), it is unlikely that predation by Pisaster produced any of the gaps found on Mussel Island during the study.

Aggregation of gap creation in space by sea otters correlates with the local topography of study sites, and may also be influenced by size distribution of mussels at each site. Most gaps made by sea otters were at the Mussel Island site, a reef nearly surrounded by relatively deep (> 2 m at spring low tides) surge channels. Topography is such that most mussel patches on the reef are within 2 m of the edges of the surge channels (Fig. 4.3). Shoreward of the reef is an area of relatively deep water (hereafter termed the "lee area") protected by the reef from wave-generated surge and turbulence (Fig. 4.3). Thus, patches of mussels at the Mussel Island site can be reached by foraging sea otters without an extended swim across areas of shallow, highly turbulent water with breaking surf. Once mussel clumps are removed, the lee area is readily available to sea otters as a site for sorting and consuming mussels, without the risk of disturbance from breaking waves. Comparable lee areas were not present at other study sites.

From this point forward I limit discussion of gaps in mussel cover to those created at Mussel Island, because all were made by sea otters (see above criteria for this determination). Gap creation by sea otters was aggregated among seasons (Fig. 4.5). Two environmental variables correlate with the seasonal pattern of gap creation. First, periods

58

(/') a. < Cl :;::r: wI-ZZ u. 0 a::!:

a: (/')w a: a. w ID ::!: ::::l Z

12

9

6

3

Gaps made by sea otters 1978-1984

(N=31)

MONTH


Fig_ 4.5 _ Seasonal distribution of gap creation by sea otters at the Mussel Island study area, Pt. Piedras Blancas. These data coincide closely with the frequency of occurrence of foraging sea otters in the general vicinity of the three study areas at Pt. Piedras Blancas

of calm frequently occur during late fall, winter, and early spring between passages of weather systems. Indeed, such intervals, typically 2 to 3 days long, are among the calmest periods of the year along the coast of central California. Both locally generated wind waves and distantly generated oceanic swell are at minima during such periods. Locally generated rough seas are much more frequent during late spring and summer as a result of persistently strong northwesterly winds, producing moderate to heavy surf and turbulence at the study sites.

Extended calm periods often occur during late summer and early fall. Intertidal foraging by sea otters during this period may, however, be influenced by the local distributions of sublittoral kelps. Floating kelp canopies (Macrocystis pyrifera and Nereocystis luetkeana) are principal locations for resting and grooming by sea otters near the study sites, and are an important feeding area as well (R. Jameson, pers. commun.; Foster and Schiel 1985). Kelp canopy size and distribution are markedly seasonal off central California (Fig. 4.6; Gerard 1976; Foster 1982; Foster and Schiel 1985). During winter, canopies are minimal, restricted to small patches in shallow, nearshore sites protected from oceanic waves by reefs or headlands. During spring and summer, kelp canopy area expands rapidly, peaking in August and September. In late fall (usually in November), north Pacific frontal systems begin to generate waves of sufficient size to tear out much of the kelp canopy, reducing it to the fragmented state typical of winter. The distribution of sea otters over distance from shoreline closely tracks seasonal fluctuations in kelp canopy size; greater numbers of sea otters are found relatively far from shore during late spring, summer, and autumn, when offshore kelp canopies are most fully developed (R. Jameson and G. VanBlaricom, unpubl.). In addition, the expanded presence of kelp canopies offshore during summer and fall may improve access to some species of alternative prey; kelp crabs (Pugettia producta) and turban snails (Tegula spp.; Calliostoma spp.) are important prey species for sea otters, and may prefer kelp stipes and canopies as substrata when they are available (Riedman et al. 1981; Hines 1982; Watanabe 1984a,b). A seasonal shift in sea otter diet toward greater use of canopy-associated prey would most likely reduce the probability that

Creation of Gaps in Mussel Cover by Sea Otters; Spatial and Temporal Aggregation 59

Fig. 4.6. Seasonal fluctuations in surface canopies of kelps near Pt. Piedras Blancas (see Sect. 4.4.4). Top photograph Typical mid-winter condition, with canopies almost completely absent. Note two large emergent rocks at right center. Bottom photograph Same area, showing typical late-summer condition, with floating canopies (primarily Macrocystis pyrifera) at maximum development. Kelp is growing primarily in water 4-18 m in depth. Note emergent rocks at lower center. Width of area shown is approximately 1 km. (Photographs taken from an aircraft with infrared film, by R.F. Van Wagenen.)


sea otters would make gaps in intertidal mussel beds. However, actual documentation of such patterns is not available.

Seasonal changes in food qUality could influence foraging on mussels by sea otters. Seasonal cycles in gonadal development of sea urchins (Strongylocentrotus poiyacanthus) correlate with consumption rates by sea otters in the Kurile Islands of the Soviet Union (Barabash-Nikiforov et al. 1947; Shitikov 1971; Riedman and Estes, this Vol.). Apparent seasonal cycles in spawning of Mytilus califomianus have been described for some locations (Coe 1932; Whedon 1936; Young 1942, 1946), but such cycles are absent elsewhere (Suchanek 1981). Marked seasonal cycles in food quality are more likely to occur in M. edulis (see Sect. 4.5) than inM. califomianus (Suchanek 1981 and pers. commun.). As data on seasonal reproductive cycles are lacking for mussels at Pt. Piedras Blancas, a possible association with foraging patterns of sea otters remains entirely speculative.

The rate of gap formation by sea otters at Mussel Island also varied substantially between years (Fig. 4.7). During the second half of the study (1981-1984), rate of gap formation was associated inversely with the frequency of moderate and major wave events, as defined by Seymour et al. 1984 (Fig. 4.7).

The 1982-83 winter was noteworthy because no gaps were made by sea otters. Large wave events were frequent, and there were extended periods during which there was no break in heavy sea and swell striking the study sites.

No association between winter calm periods and gap formation by sea otters was apparent from 1978 through 1980 (Fig. 4_7). Qualitative observations and wave height calculations (no major events, three moderate events during the 3-year period, as defined by Seymour et al. 1984) indicate that many periods of mid-winter calm occurred during this interval. However, no gaps were formed in mussel cover in the study sites by foraging sea otters during the period.

It is possible that the patterns outlined above are generated in part by variation in foraging at the level of individual otters. Estes et al. (1981) examined patterns of foraging among sea otter populations at six locations in the North Pacific. A principal conclusion of this work was identification of the importance of individual variation in foraging patterns. Lyons and Estes (unpubl.) have observed the foraging habits of tagged, individually recognizable sea otters along the north shore of the Monterey Penninsula, in central California. Their principal result bears out the predictions of Estes et al. (1981): Individual sea otters have strikingly different foraging patterns, even when one compares animals within the same categories of age and sex. Within a given category, each animal observed has a diet dominated by one or two prey types, and these prey types vary markedly among individuals. Although there were a number of tagged, individually recognizable sea otters in the vicinity of Pt. Piedras Blancas during 1978 to 1984 (Jameson, in prep.), none was seen making gaps in mussel cover in the study sites. The possible importance of individual variation of foraging habits of sea otters to variable rates of gap formation during the study cannot, therefore, be directly evaluated.

Size Distribution of Gaps Created by Sea Otters

==a:: W< Z W 12 u.> ° a:: enW a:: a.. 6 Wen ala.. ~< ~C!)

en I-a:: Z< Ww i;j> wa:: >W <a.. ==

8

4

Gaps made by sea otters

(N - 31)

Moderate wave events

Wave ht. ~ 3.0m Period ~ 12s

Major wave events

en Wave ht. ~ 6.0m I-a:: ffi < 4 Period ~ 12s >w w> wa:: > W 2 <a.. ==

78 79 80 81 82 83 84

YEAR

61

Source: This study

Source: Seymour et al.

( 1984)

Source: Seymour et al.

( 1984)

Fig. 4.7. Interannual distribution of gap creation by sea otters at the Mussel Island study area, Pt. Piedras Blancas. As in Fig. 4.5, data coincide closely with frequency of occurrence of foraging sea otters in the area. Also shown are the interannual distributions of moderate and major wave events, as defined by Seymour et al. (1984), along the California coastline during the period of this study

4.4.5 Size Distribution of Gaps Created by Sea Otters

During the course of the study, about 20% of the mussel cover at the Mussel Island study site was removed by sea otters. The 31 gaps created by sea otters ranged in size from 0.03 to 1.34 m2 . The mean gap size was 0.25 m2 , the median gap size 0.08 m2 •

The central tendency in size of gaps formed by sea otters is similar to central values for areas of gaps formed by wave shear along the outer coast of the Olympic Peninsula, Washington state (Fig. 4.8). Inspection of the data suggests that there are differences in the tails of the distributions. It appears that wave shear tends to produce more very large and very small gaps in mussel beds at Tatoosh Island (Olympic Peninsula) than sea otters do in central Calfomia. At more protected locations in Washington (illustrated

62

0.3

0.2

...J « 0.1 I-o I-


Gaps caused by sea otters 1981-1984

N=31

Mean area: 0.25 m2

Median area: 0.08 m2

~ O.O~~"~~-,---.--~~~~,--o Z o i= ()

« c: ~

0.3 Gaps caused by wave shear

0.2

0.1

O.O+-~~--.--,---.--~~~~~

o 2 4 6 8 10 12 14

NATURAL LOG, GAP AREA (cm2 )

Fig. 4.8. Distribution of areas of gaps made by sea otters at Mussel Island study area (no gaps were made by sea otters during 1978-1980). Also shown are area distributions of gaps caused by wave shear at two sites on the coast of the Olympic Peninsula, Washington state (see Fig. 4.2; lower plots after Paine and Levin 1981)

by data from Mukkaw Bay), frequencies of large gaps are lower than the frequencies of large gaps made by sea otters, but very small gaps are more common in Washington. The differences may be a simple artifact of the much larger sample size gathered by Paine and Levin (1981). Alternatively, the differences may be real, reflecting practical limitations on the foraging tendencies of sea otters. The minimum gap size attributed to sea otters (about 30 cm2 ) was most likely created by only one or two foraging dives; it is probably difficult for a sea otter to remove one or two clumps of mussels without clearing an area of at least 30 cm2 • Likewise, formation of gaps larger than 1.34 m2 by sea otters probably would require sequential foraging episodes at the same location over a number of days. Given that sea otter foraging patterns may be sensitive to environmental variables subject to rapid change (e.g., wave height), such an occurrence may be unlikely.

Eight (26%) of the 31 gaps formed by sea otters were subsequently enlarged by additional sea otter foraging activity (examples shown in Fig. 4.9). In some cases, substantial enlargement occurred up to 34 months after initial gap formation. Such patterns differ from those described for gap formation by wave shear along the shore of the Olympic Penninsula (Paine and Levin 1981). In the latter case it is argued that enlargement of an existing gap is uncommon because of a leaning response of the mussels

Mussel Size and Vulnerability to Foraging Sea Otters 63

MUSSEL ISLAND, EXPOSED SIDE MUSSEL ISLAND, PROTECTED SIDE

GAP +048 GAP +043

A;e;~IYo.~~~12--~ CO I I

21 April 1983---+-~ &<\j Area: 0.27m2 18/ ~

1m

I I I I I

CTITIill Gap

D Unbroken Mussel Cover

7 March 1982 Area: 0.07 m2

O.Sm

6 May 1984 Area: 0.62 m2

Fig. 4.9. Two examples of enlargement of existing gaps by additional sea otter foraging. Both gaps were originally made by sea otters. Dashed line in left example indicates position of the original gap through time

on the edges of the gap; this response is said to "cure" the edge (language of Paine and Levin), stabilizing the margins of the gap and reducing the probability of its subsequent enlargement. Available data indicate that enlargement of existing gaps is in fact un· common along the Washington coast, and tends to occur only as a result of major storms during winter. When enlargements do occur, they typically happen soon (i.e., within a few months) after the time of initial gap formation [Paine and Levin 1981; but see Dayton 1971; Sousa 1985 for discussion of an alternative view; T. Suchanek (pers. commun.) has observed occasional enlargements occurring 25-30 months after gap formation in Washington]. In the case of gaps formed by sea otters in California, a similar "leaning response" was observed in mussels along the edges of gaps. This response apparently stabilized gap edges to subsequent enlargement by wave shear (the only exceptions occurred during the major wave events described above), but did not prevent subsequent enlargement by foraging sea otters.

4.4.6 Mussel Size and Vulnerability to Foraging Sea Otters

Observations of sea otters foraging on mussels of various sizes in the study sites indicate two patterns of interest. First, when viewed from the perspective of mussel populations, sea otters are entirely nonselective with regard to individual size of captured prey. By removing mussels in clumps, and by removing all mussels from discrete pat-


ches, sea otters remove all size and age categories of the local mussel population, exactly in proportion to the relative importance of each category in the population. Although only selected individuals are actually consumed by sea otters, all categories are removed from the primary substratum. Those mussels not eaten are dropped into deeper water (see Harrold and Hardin 1986 for an exception), where mortality is almost certainly 100% due to other predators, burial in sand, or eventual stranding on the shore at a height above normal physiological tolerances. Effectively, then, sea ot-

0.4

0.2

0.4

...J < 0.2 I-0 l-ll.. 0 Z 0 I-0 0.4 < ~ u..

0.2

0.4

0.2

40

Sample of mussels, Pisaster Point.

N = 691

Sample of mussels, Mussel Island.

N = 675

Sample of mussels, Channel Bank.

N =531

80

Mussels eaten by sea otters

N. 116

120 160

MAXIMUM SHELL LENGTH (mm)

Fig. 4.10. Representative samples (upper three plots) of size frequencies of Mytilus californianus in the three study areas at Pt. Piedras Blancas. Bottom plot shows estimates of sizes of mussels actually consumed by sea otters in the area during the study

Case 2: Sea Otters and Mussels in Prince William Sound, Alaska 65

ters are nonselective with respect to mussel size, within the patches in which otters actually forage. Viewed from a larger perspective, a similar pattern emerges. Based on samples of mussel size frequency distributions collected in the study sites (Fig. 4.10), it appear that sea otters are able to clear gaps among mussel patches with a variety of size distributions. Mussel stands dominated by large individuals are not immune to sea otter predation.

In contrast to the above, sea otters are indeed selective for mussel size, when viewed from the otter's perspective. Small individuals are discarded, while larger ones are consumed, often with the help of tools used to break the shell (Fig. 4.10). During this study I watched sea otters eat about 300 individual mussels, encompassing most of the larger portion of the size range of mussels present at Pt. Piedras Blancas. In only one case did a sea otter appear to be deterred from eating a large mussel; a mussel about 14 cm in length was discarded (effectively killed) after a juvenile sea otter was unable to open the shell with its teeth. The otter in question was not carrying a tool rock. On a number of other occasions I saw sea otters successfully handle and consume mussels of comparably large size, in all cases with the use of a stone as an anvil for breaking the shell.

The largest mussels occurring in samples collected from the study sites are 15 cm in length. Mussels of this size are readily removed from the substratum and, with the help of tools, consumed by sea otters; the data and observations further suggest that larger mussels are preferentially selected out of clumps of mussels removed by otters. Sea otters are known to seek out large, difficult-to-capture prey such as abalone, presumably because the energy return is worth the considerable cost in effort and time (e.g., Estes et al.1981). Sea otters are also known to be efficient and innovative in their use of tools as aids in capturing and handling prey (e.g., Hall and Schaller 1964; Houk and Geibel 1974; McCleneghan and Ames 1976; Riedman and Estes, this Vol.). It is therefore unlikely that mussels can attain a refuge in size from consumption by sea otters in the vicinity of the study areas.

4.5 Case 2: Sea Otters and Mussels in Prince William Sound, Alaska

In January 1978 I began a long-term study of effects of predation by sea otters on populations of mussels (Mytilus edulis L.) in Prince William Sound, Alaska, in collaboration with A.M. Johnson and J .A. Estes. We began with the knowledge that mussels are consistently an important component of the diet of the sea otter population in the Sound (Calkins 1978; Estes et al. 1981; Johnson 1982; Garshelis 1983), and an awareness that important variations occur in the Sound across space and time, both with regard to the age and sex structure of the resident sea otter population (Garshelis et al. 1984), and to the density and demography of populations of mussels in the intertidal zone. The purpose of our work was to determine if foraging by sea otters was important to the structure of mussel populations in the Sound, and to evaluate the extent to which sex- and age-specific differences in the distribution and foraging habits of sea otters were important in observed spatial variations in mussel population structure.

Our work on mussels in Prince William Sound included collection of extensive data on density and size distribution of mussel populations. Data were collected from a


nwnber of locations, but repetitive sampling over a number of years was done primarily at three principal study sites, described below. Data on mussel demography were compared to known patterns of distribution and foraging activity of sea otters in the Sound. A key element of this work is the consistent spatial segregation of sea otters in Prince William Sound by age and sex.

Previously abundant populations of otters in Prince William Sound (Cook 1784) were drastically reduced during the fur-hunting era, but a remnant population survived in the southwestern part of the Sound (Lensink 1960, 1962; Kenyon 1969; Pitcher 1975; Calkins 1978; Estes et al. 1981). Since the cessation of commercial hunting, sea otters have increased substantially in numbers and range (Pitcher 1975; Johnson 1982). In recent decades numbers of otters have increased greatly in the northeastern and northwestern parts of the Sound (Johnson 1982; Garshelis 1983; Garshelis et al. 1984; Irons 1984; Garshelis and Garshelis 1984; VanBlaricom and Estes 1986; additional detail provided below). The most recent census (1984-85) of sea otters in Prince William Sound produced a count of 5000 animals, including 800 pups (D. Irons, pers. commun.).

4.5.1 Study Location

Prince William Sound is a complicated array of interconnecting fjords, embayments, and stretches of relatively open water located on the northern edge of the Gulf of Alaska (Figs.4.1 and 4.11). Physical characteristics of the Sound are presented in detail

Prince William Sound

I· >, 20km

Alaska

Fig. 4.11. Location map of study areas in Prince William Sound, Alaska (see also Fig. 4.1)

Study Location 67

elsewhere (lslieb and Kessel 1973 ; Calkins 1978; Garshelis 1983; National Ocean Service 1985; VanBlaricom and Estes 1986) and are briefly reviewed here. Shorelines in the Sound are variable and can change abruptly over small distances. Precipitous rocky shores, gently sloping flats of mud and sand, and moderately sloping shores of mixed mud, cobble, and shell litter are all common. Rocky substrata in the study areas are metamorphosed sandstone, siltstone, and argillite (Moffit 1954). Climate is maritime, with heavy precipitation, especially in fall and winter. Surface salinities are 18-20 gm kg -1, fluctuating with rainfall and runoff. Sea surface temperatures range from 11 ° to 12°C in late summer to 3° to 4°C in winter. The complex local topography protects most shores in the Sound (including all study areas discussed here) from the direct impact of oceanic swell. The maximum tidal range in the Sound is approximately 6 m.

Many kinds of substrata are utilized by mussels in Prince William Sound (Fig. 4.12). In areas of solid rock small-scale cracks and crevices are common. Other common substrata for mussels include fields of large, angular bolders usually near formations of solid rock (termed bolder fields herein), extensive, gently sloping shores composed of a mixture of mud, sand, gravel, angular cobble, and bivalve shell fragments (termed mixed substrata), and nearly horizontal bars and flats of sand and mud with little cobble or shell litter (termed mudbars).

Two species, the mussel Mytilus edulis and the brown alga Fucus distichus, were among the most abundant large sessile organisms in most intertidal habitats we examined in the Sound, occurring primarily at middle tidal levels (+0.5 to +3 m MLLW). Mytilus and Fucus were prominant species in all primary study areas. Other abundant sessile organisms in the region are listed by Nybakken (1969); Hubbard (1971) and Johanson (1971). On mixed substrata in the intertidal, most algae and invertebrates are found only on or among mussel shells or the exposed portions of cobbles. On mudbars, mussels form the primary substratum for most other nonburrowing organisms. In such habitats mussels form discrete patches of high denSity, and are attached by byssal fibers primarily to one another (see also Reise 1985).

Field work on mussels was concentrated in three areas of the Sound: Green Island, Simpson Bay and other nearby embayments, and Orca Inlet. Green Island (60° 17'N, 147°25'W) is located in the southwestern part of the Sound (Fig. 4.11). Mussel sampling was done primarily in and near Gibbon Anchorage, on the northwestern part of the Island. Gibbon Anchorage is a complex array of coves, channels, and rocky reefs and islets largely protected from heavy seas. Samples were gathered at least once per year from 1977 through 1984. Sea otters have been continuously present at Green Island since at least the early 19S0's (Lensink 1962), and possibly much longer (Estes et al. 1981). Since 197 S the sea otter population at Green Island has consisted primarily of reproductively active females and (seasonally) their dependent pups, seasonally territorial adult males, and independent juveniles of both sexes (Garshelis et al. 1984). Hereinafter Green Island is termed a breeding area for sea otters. Sea otter population size may have peaked in the middle to late 1970's at Green Island (Estes et al. 1981; A. Johnson, pers. commun.); since that time there has been a gradual decline in total numbers of sea otters in the area (Irons 1984; A. Johnson, pers. commun .).

Simpson Bay (600 38'N, 14SoS0'W) is located in the northeastern part of the Sound and consists of northwestern and southeastern arms (Fig. 4.11). Mussels were sampled from solid rocky substrata, mixed substrata, and mud bars at the upper end of the


Fig. 4.12

Methods 69

southeastern arm in 1978, 1983, and 1984, and a boulder field in the same area was sampled in 1984. Mixed substrata in the northwestern arm were sampled in 1979. All sites sampled in Simpson Bay are relatively protected from wave action. Sea otters returned to the northwestern arm of Simpson Bay in 1977 and the southeastern arm in 1978. The Bay became a breeding area during 1983 -1985. However, during the period of mussel sampling (1978-1984), the study sites in the Bay were primarily male areas.

A number of sites were sampled in Orca Inlet, in the northeastern part of the Sound (Fig. 4.11), during the period 1979-1984. The principal study area was on the southeast shore, 5 km northeast of the town of Cordova and near Odiak Channel. This site is unofficially termed Odiak Point (600 34'N, 145°42'W). Mussel populations on solid rock and mixed substrata at Odiak Point were examined at least annually from 1979 through 1984. I also sampled mussel popUlations at Spike Island (600 34'N, 145° 45'W); samples from solid rock, mixed substrata, and mudbars taken in 1980, 1983, and 1984) and Hartney Bay (600 30'N, 145°50'W; mudbar mussel population sampled in 1979 and 1980). All sites in Orca Inlet are somewhat more exposed to locally generated wind waves than study sites in Gibbon Anchorage (Green Island) and Simpson Bay. Large numbers of male sea otters returned to Orca Inlet in winter 1980, and Orca Inlet has remained a "male area" since that time. Spatial distribution of sea otters in Orca Inlet has varied predictably with season, at least through 1982 (Garshelis and Garshelis 1984). During late fall and winter, sea otters occur primarily in the central part of the Inlet, which includes the study sites at Odiak Pt., Spike Island, and Hartney Bay. During spring, summer, and early fall sea otters congregate in the northwestern and southeastern ends of the Inlet, apparently avoiding seasonally intensive traffic in the central Inlet by fishing vessels using Cordova as home port.

4.5.2 Methods

Because of constraints on field schedules, detailed, direct, systematic observations of consumption of mussels by sea otters were not attempted. Instead, I rely largely on accounts of foraging behavior provided by other investigators (Calkins 1978; Estes et al. 1981; Johnson 1982 and pers. commun.; Garshelis 1983 and pers. commun.; Garshelis et al. 1984; Garshelis and Garshelis 1984; C. Monnett, pers. commun.; D. Siniff, pers. commun.; F. Sorensen, pers. commun.) all of whom collected data at one or more of the same study sites during the same time interval as my studies on mussel populations. I made opportunistic observations of sea otter foraging activity at irregular intervals; in all cases, my observations were consistent with the findings of the observers listed above. The assertion that mussels are important to sea otter diets in the Sound is based principally on direct observation of foraging animals in the field (e.g., Estes et al. 1981) .

.. Fig. 4.12. Photographs of populations of Mytilus eduIis on different substrata in Prince William Sound. Top Solid rock substrata; Center Mixed substrata (see Sect. 4.5.1; this is the substratum type most commonly used by mussels in the Sound); Bottom Mussels on a mudbar (in this case, mussels attach only to one another, with no active attachment to the substratum). All photographs were taken in Orca Inlet. Sampling frames in the upper two photographs are 0.25 m2 in area


I compared mussel demography among sites that differ in density, age, and sex composition of resident sea otters. At each site, mussel populations were monitored repeatedly over time, and temporal patterns were compared to known changes in the demography and behavior of sea otters in the area.

Sites for sampling of live mussel populations were selected by blind, haphazard drop of a 0.25 m2 quadrat frame in an area of relatively dense mussel cover at a predetermined tidal level. Beginning at a predetermined corner of the frame, mussels were cleared (down to the primary substratum) toward the center of the frame until approximately 500 individuals had been collected. Dimensions of the cleared area were measured, and the tidal level of the sampled site was determined. All mussels in each sample were washed over a screen with 0.5 mm mesh, separated from debris and from one another, and measured (maximum shell length) with Vernier calipers. A special effort was made to search screening residues and the byssal threads of large mussels in order to locate minute juvenile mussels. Most samples contained 500 to 700 individuals (range: 300 to 1200 individuals). The relationship of shell length to dry tissue weight was determined by measuring shell length as noted above, removing tissues and drying to constant weight at 60-80 °C,and weighing dried tissues on a precision balance. The calculated relationship is:

Y = 0.02 x2.46 (r2 = 0.82, n = 388) ,

where x is the maximum shell length in mm, and Y is the dried tissue weight in mg. At Gibbon Anchorage sea otters frequently haul out on beaches or on patches of

terrestrial vegetation just above the high tide line. Hauling out is particularly common during winter, and otters often defecate while hauled out. Bivalves and crabs are the predominant prey of sea otters in Prince William Sound, and are often consumed whole, without prior removal of the shell or carapace. Thus, spraints (feces) often consist largely of shell fragments which are easily identified to species (Fig. 4.13). During springtime visits to Gibbon Anchorage in 1979 and 1980, I collected all spraints left by sea otters above the high tide line along a predetermined segment of shoreline. Each spraint was later examined with a microscope, allowing listing of component species. Because of characteristically colored periostraca and characteristically structured hinges, mussel shells are easily recognizable in spraints. Sea otters rarely haul out at Simpson Bay or Orca Inlet, and spraint collections were not attempted.

4.5.3 Consumption of Mytilus edulis by Sea Otters: The Basic Pattern

The frequency of mussel consumption by sea otters in Prince William Sound varies markedly with age and sex of sea otters. Solitary adult sea otters rarely eat mussels. Mussels are a major dietary item of young otters and females with large dependent pups

Fig. 4.13. Photographs of spraints left by sea otters above the high tide line at Green Island, Prince William Sound. In all cases, prey species are readily identified because otters ingest and pass shells as well as soft tissues. Top Spraint consisting entirely of shell fragments from Mytilus edulis; Center Spraint containing shell fragments of several clam species, including Macoma inquinita and Protothaca staminea; Bottom Spraint containing remains of hard parts of the crab Telmessus cheiragonus. Lens cap diameter is 52 mm in all photographs

Consumption of Mytilus edulis by Sea Otters: The Basic Pattern 71

Fig. 4.13


(Garshelis 1983; A. Johnson, pers. commun.). Exceptions to these patterns are noted below. Sea otters gather mussels by diving when mussel beds are submerged at high tide. To my knowledge, emergent consumption of mussels by sea otters (sensu Harrold and Hardin 1986) has not been observed in Prince William Sound. Captured mussels are brought to the surface in clumps, together with attached substrata such as cobble, gravel, shell fragments, and algae. Mussels and debris are quickly sorted at the surface; small mussels and debris are dropped into the water, while larger mussels are crushed, shell included, with the molars and ingested whole. Unlike otters observed in California, otters in Prince William Sound apparently do not use stones as anvils to break mussel shells. In some cases otters swim a few meters away from shore while handling and ingesting mussels. Frequently, however, otters remain at the point of surfacing during the prey-handling interval. Otters typically roll in the water to rinse the fur after each mussel is ingested. When all mussels in a clump have been consumed or discarded, the otter rolls in the water and grooms briefly, then makes another dive.

On occasion I observed sea otters handle and consume relatively large mussels (shell length of about 4 cm or more) in a somewhat different manner. Rather than crushing the shell with the molars, the valves are pried apart with a canine. One of the valves is then broken off with the teeth, and the soft tissues are scraped from the shell with the tongue and incisors. The shell is then discarded, rather than ingested as above.

In contrast to observations at Pt. Piedras Blancas, California, sea otters in Prince William Sound typically do not remove discrete, continuous patches of mussels by diving repeatedly at precisely the same location. Observations during foraging episodes, and subsequent examination of exposed mussel beds at low tide, are consistent with the notion that mussel clumps are removed independently of one another. On a larger scale, however, foraging activity is often concentrated in space, in areas meters to tens of meters in diameter. Sea otters frequently return to the same stand of mussels to forage for many successive days, although dense, accessible, largely unexploited stands of mussels may be available nearby (A. Johnson, pers. commun., Garshelis 1983). The effect of such effort is the creation of a relatively large (tens of meters) area in which mussel density is substantially reduced.

Mussels comprise up to 40% (by frequency of occurrence) of sea otter diet in breeding areas such as Green Island (Estes et al. 1981). Mussels are rarely taken in male areas and typically comprise less than 5% of the diet in such areas. Exceptions (up to 20% mussel frequency) have been noted in male areas containing many young animals (Estes et al. 1981; Garshelis 1983). On rare occasions, newly arrived male groups may forage intensively on mussels, with Significant effects on mussel populations (see below).

Estes et al. (1981) presented data on dive time and success rate for otters feeding on mussels at Green Island and at Sheep Bay (near Simpson Bay; a male area when sampled). Dive time averaged 18 s (N = 123) at Green Island and 41 s (N = 41) at Sheep Bay; both figures are substantially less than mean dive times for all other prey taken within each site. Based on data for all sites sampled in the Sound, otters feeding in mussel beds captured prey successfully on 99% of foraging dives (N = 156). The success rate was 75% for otters foraging in other patch types. These patterns are a predictable correlate to the relative accessibility of mussels as prey. Whereas other bivalve prey are infaunal and must be excavated and withdrawn from the sediments, mussels are simply lifted from the surface of the substratum, in most cases after very little search time.

Size Distribution of Intertidal Mussels and the Population Status of Sea Otters 73

It is not clear how sea state influences the consumption of mussels by sea otters in Prince William Sound. Sea otters frequently are seen removing mussels from locations exposed to large waves, but observational effort is almost certainly biased in favor of calm conditions (A. Johnson, pers. commun.). Rough seas often are associated with stormy weather which produces poor visibility and reduced opportunities for observation of foraging sea otters. Storm activity does, however, have a decided influence on the local distribution of sea otters in the vicinity of Gibbon Anchorage. During periods of high winds and large seas, sea otters from a relatively large area (possibly most of the shorelines of Green, Little Green, and Channel Islands, and the shallow-water habitat near Applegate Rocks) converge on Gibbon Anchorage, apparently seeking shelter (A. Johnson, pers. commun.). Density of sea otters in the Anchorage may increase by an order of magnitude during such episodes. Because most categories of preferred prey of sea otters occur at low density within the Anchorage, sea otters may forage intensively on mussels.

4.5.4 Size Distribution of Intertidal Mussels and the Population Status of Sea Otters

Data collected at the three study sites indicate consistent differences among areas in average mussel size, and in the proportion of large individuals (herein arbitrarily defined as mussels with a maximum shell length of 40 mm or more) present in mussel populations.

At Green Island, large mussels were rare in samples (Figs. 4.14,4.15). This pattern was consistent across time for all habitat types and levels of exposure at which samples were collected. The only exceptions were patches of solid rock substrata that included relatively narrow, deep crevice habitat. Most of the large mussels found at Green Island were located in crevices where, presumably, they found refuge from foraging sea otters (Fig. 4.16). The utility of such refugia in reducing predation by sea otters has been recognized in California (e.g., Lowry and Pearse 1973). During the latter part of the study, I made extensive searches of the intertidal zone at Green Island in order to locate areas of mussels dominated by large individuals. I was unable to locate such areas. In contrast to Simpson Bay and Orca Inlet (see below), mussel cover was typically low at Green Island, especially on solid rock substrata where it was unusual to find more than a single layer of small mussels attached to the rock surface.

Fragments of mussel shells were consistently important (both volumetrically and by frequency of occurrence) components in the spraints of sea otters collected at Gibbon Anchorage (Table 4.2).

Large mussels were common on all dates in all sites sampled in Simpson Bay (Figs. 4.14, 4.15). Large mussels were particularly prominent on solid rocky substrata, where they grew in large, multi-layered clumps covering much of the exposed rock surface (Fig. 4.16).

The patterns of mussel size distribution at the sites sampled in Orca Inlet were similar to those seen in Simpson Bay, with an important, albeit transient, exception noted below. Large mussels were common (Figs. 4.14, 4.15) regardless of type of substratum or relative exposure to wave action. On all types of substrata, mussel cover was typically high; mussels were commonly found in dense, multi-layered clumps on solid rock surfaces (Fig. 4.16).

74

20

10

30

20

10 en a: w III :?: ::J Z I-Z W 30 () a:

20 W a..

10

30

20

10

Green Island,

Mixed Substrata

20 40

April 1978

N=600

April 1980

N=517

October 1982

N=521

August 1984

N=523

60


30

20

10

30

20

10

30

20

10

30

20

10

Green Island,

Solid Rock

20 40

April

1979

N=584

April

1980

N=692

October 1982

N=539

August 1984

60


Fig. 4.14. Representative samples of shell length frequency of Mytilus edulis at the three study areas in Prince William Sound. Data are partitioned by substratum type, time, and location. Frequency bars for mussels greater than 40 mm in shell length are arbitrarily shaded black for emphasis


Simpson Bay, Simpson Bay,

Mixed Substrata Solid Rock

30 May 30 May 1978 1978

20 N=166 20 N=80

10 10

30 September 1979

20

10 en a: w !Xl ~ ::> z I-Z June June W 30 30 0 1983 1983 a: W 20 N=519 20 N=900 a.

10 10

30 August 30 August 1984 1984

20 20

10 10

20 40 60 20 40 60


Fig. 4.14


Orca Inlet.

Mixed Substrata

30 April 1980

20

10 en a: w ID ~ ::l Z I-Z

October W 30 0 1982 a:

20 w a..

10

August

30 1984

N=552 20

10

20 40 60

30

20

10

30

20

10

30

20

10

Orca Inlet.

Solid Rock

April 1980

20 40

October

1982

N=623

August

1984

N=563

60


Fig. 4.14

Table 4.2. Composition of sea otter spraint samples collected at Gibbon Anchorage, Green Island, Prince William Sound

Category a

Mussels Clams Crabs

Frequency of Occurrence (%) April 1979 (n = 49) April 1980 (n = 63)

31 98 39

41 68 63

a Categories shown include only hard-shelled prey species which are easily recognized in spraints. Other important prey (e.g., Echiurus echiurus) are seldom recognizable in spraints and are, therefore, under-represented in spraint sampling.


Green Island, Green Island,


30 April 30 April 1978 1979

20 20 Total weight: Total weight:

10 32.9gms 10 17.S gms

30 April 30 April

en 1980 1980 en 20

Total 20

Total weight: <C weight:

~ 10 0

37.0gms 10 14.8gms

m >-a:: Cl

I-Z W 30 October 30 October () a:: 1982 1982 W 20 20 a.. Total weight: Total weight:

10 18.Sgms 10 12.Sgms

30 August 30 August 1984 1984

20 Total weight:

20 Total weight:

10 2S.2 gms 10 20.5gms

20 40 SO 20 40 SO


Fig. 4.15. Representative samples of the distribution of dry tissue biomass as a function of maximum shell length in Mytilus edulis collected at the three study areas in Prince William Sound. Data partitioning and bar shading are as in Fig. 4.14


Simpson Bay. Simpson Bay.


30 May 1978

Total weight: 29.4gms 30 May 1978

Total weight: 27.1gms

20 20

10 10

September 1979 30 Total weight: 92.0gms

en 20 en c:( 10 ~ 0 CD

>-c: 0

I- June 1983 June 1983 Z W

30 Total weight: 58.6gms 30 Total weight: ()

c: 20 20 W 0.. 10 10

August 1984 August 1984 30 Total weight: 55.2gms 30 Total weight: 55.2gms

20 20

10 10

20 40 60 20 40 60

MAXIMUM SHELL LENGTH (mm) Fig. 4.15


Orca Inlet, Orca Inlet,


April 1980 April 1980 30 Total weight: 62.2gms

30 Total weight: 73.5gms

20 20

10 10

C/) C/)

« ~

October 1982 0 October 1982

ell 30 Total weight: 89.1gms 30 Total weight: 62.3gms

>-c: 20 20 0

I- 10 10 Z W t)

c: w a..

August 1984 August 1984



20 20

10 10

20 40 60 20 40 60

MAXIMUM SHELL LENGTH (mm) Fig. 4.15

As noted above, large numbers of male sea otters entered Orca Inlet in late 1979 and early 1980. During winter 1979-80 a significant, large-scale mortality event occurred among mussel populations at many locations in Orca Inlet, including all the study sites discussed here. I did not observe the mortality event itself, nor could I locate residents who directly observed the widespread mussel mortality. Below a tidal level of about + 1.7 m, mussels were removed selectively from large areas of the intertidal zone, as indicated by remnant byssal threads still attached to the rock surfaces. Most barnacles were left undisturbed (Fig. 4.17). Mortality was 100% at Hartney Bay. In many areas (but especially Spike Island) there were large piles of relatively fresh (unweathered, periostraca intact, valves still articulated), empty mussel shells in the lower intertidal. Shell pairs in such piles typically consisted of one intact valve and one valve broken in half. Most mussels in the shell piles were large (Fig. 4.18).


Fig. 4.16. Representative examples of utilization of solid rock substrata by Mytilus edulis in Prince William Sound. Top Typical mussel cover at Green Island. Mussels are almost entirely confined to crevices that apparently provide refuge from foraging sea otters. Bottom Typical mussel cover at Orca Inlet. Mussels occur as dense, multi-layered patches on open as well as cryptic substrata. Lens cap diameter is 52 mm; scale is similar in both photographs


Fig. 4.17. Close-up photograph of an area at Spike Island, Orca Inlet, apparently stripped of mussels by sea otters during the winter of 1979-1980. Remnant byssal threads from mussels are clearly seen in the center. Note that most barnacles in the area apparently survived the mussel mortality event. Maximum shell length of largest surviving mussel is approximately 4 cm

Removal of mussels by male sea otters is the most likely explanation for the mortality event described above, based on the following circumstantial evidence: (1) The event coincided in time with the return of large numbers of male sea otters to Orca Inlet during the winter of 1979-80 (Garshelis and Garshelis 1984). (2) Large-scale mortality was restricted to mussels. Most other sessile invertebrates survived in the same locations from which mussels were removed. (3) Condition of mussel shell litter in areas of mortality is consistent with the manner of consumption of large mussels by sea otters (see Sect. 4.5 .3). (4) Episodes of unusually severe storms, large waves, or ice scour did not occur in Orca Inlet during the winter of 1980 (A. Kimker, Alaska Dept. of Fish and Game, Cordova, pers. commun.). (5) There were no indications of "outbreaks" or unusual behavior by other mussel predators in Orca Inlet. Although quantitative data are lacking, I observed no unusual fluctuations in numbers of foraging habits of predatory sea stars (Evasterias, Pisaster) and whelks (Nucella) during 1979 or 1980, within the limits defined by the field schedule. (6) A similar event occurred on the north shore of the Monterey Penninsula, California, in the late 1960's. This event was likewise coincident with the appearance of a large group of male sea otters in a previously vacant area (D. Abbott, pers. commun.; Hines and Loughlin 1980). D.P. Abbott (pers. commun.) asserted without reservation that sea otters caused the decline of mussels at Monterey.

82

en !La: ow

ID Z~ 0:> j:Z O...J «« a: I!LO

I-

0.15

0.10

0.05

20 40 60


Mussel shell litter. Spike Island. Orca Inlet. April.1980. N=149

80


Fig. 4.18. Size frequency distribution of a sample of empty mussel shells found in the intertidal at Spike Island, Orca Inlet, in April 1980. Condition of shells indicated that they had recently been killed by foraging sea otters

Mussel populations in Orca Inlet recovered quickly from the 1980 mortality event. By 1984 dense patches oflarge mussels were again common in all study locations from which mussels had been stripped in 1980 (Fig.4.l9). Although male sea otters remained abundant near the Orca Inlet study sites during winter months, additional intensive foraging on local mussel populations apparently had not occurred through 1984, based on mussel demographic data and observations of sea otters.

Spatial variations of sea otter populations clearly are not a unique potential explanation for observed variations in mussel size distribution in Prince William Sound. Spatial variations in foraging habits of other predators and spatial differences in other environmental variables (e.g., nutrient supply) are plausible alternative explanations. Experiments were conducted to evaluate these alternatives (presented in detail in VanBlaricom and Estes 1986; VanBlaricom, in press; VanBlaricom and Johnson, in prep.). Exclosure experiments showed that mussels can grow to large size at Green Island if protected from foraging sea otters. Tagging/transplantation experiments showed that mussels at Green Island grow as rapidly as mussels at other sites, indicating that site-specific environmental factors (other than predation) probably do not account for observed variation in mussel size.

4.5.5 Mussel Size and Vulnerability to Foraging Sea Otters

In Prince William Sound, it appears unlikely that Mytilus edulis can attain a refuge in size (sensu Paine 1976) from sea otter predation. Most mussels appear to present little difficulty in handling or consumption by sea otters. The largest mussels which occur in the Sound (shell length of 9 cm) are easily handled without the need for tools (shells of Mytilus edulis generally are thinner and more easily broken than those of M. californianus [Harger 1972]). Although juvenile sea otters are frequently unsuccessful when foraging on other prey types in Prince William Sound, they typically are highly

Mussel Size and Vulnerability to Foraging Sea Otters 83

Fig. 4.19. Photograph of dense intertidal cover of Mytilus edulis at Spike Island, near Cordova in Orca Inlet, Prince William Sound, in autumn 1984. This site had been completely stripped of mussels, apparently by foraging sea otters, during winter 1979-1980. Mussel cover at this site was nearly zero in April 1980

successful when foraging on mussels, and are able to capture, handle, and consume the largest mussels available in local populations (Estes et al. 1981).

Selection of mussels by size appears to be a similar process in Prince William Sound and in central California. Mussel mortality is largely nonselective because mussels are removed as clumps by sea otters. Mussels not selected for consumption from clumps are discarded, and most probably fall in unsuitable habitat. However, in cases where foraging otters sort mussel clumps without swimming away from the collection site, it is likely that many rejected mussels fall in or near the mussel patches from which they were removed. In such cases sea otter foraging may effectively become size selective.


4.6 Discussion

Data presented from central California indicate that size distribution of gaps, variation among season and year in gap creation, and vulnerability of large mussels to gap formation are markedly similar whether the forcing mechanism is wave shear or foraging sea otters. In view of such similarities I argue that the general structure of existing models of the gap creation process in Mytilus californianus assemblages (e.g., Paine and Levin 1981) is largely adequate to deal with foraging by sea otters. However, it is apparent that some adjustments in the details of such models are needed. Based on the data from Pt. Piedras Blancas, sea otters make gaps primarily during calm periods in winter and spring. Thus the addition of sea otters potentially reduces the level of interannual variation in combined birth rate of gaps. Storm waves will produce gaps during rough winters and sea otters may produce gaps during mild winters. The presence of sea otters appears to increase the probability that existing gaps will be enlarged before they return to a mussel-dominated "background" community. As a result, different parts of a single gap may be quite different in successional "age".

Gap formation by sea otters may have important implications for species whose survival depends on gaps in Mytilus californianus cover. If sea otters improve the odds that some gaps will be formed each year, then effort invested in seasonal production of propagules by gap users is more likely to produce a significant return, in the form of successful recruitment. If the presence of sea otters improves the odds that a small gap will be enlarged before it is closed by mussel cover, then recruitment by gap users into small gaps is more likely to succeed in producing individuals which survive to reproductive age.

By virtue of their ability to make gaps in Mytilus californianus cover, sea otters are able to function in a manner similar to the predatory sea star Pisaster ochraceus. Although Pisaster was removed from the Mussel Island study site at Pt. Piedras Blancas for 6.5 years, the predicted downward extension of mussel cover did not occur (VanBlaricom, in prep.). On 9 of 16 transects established at the site to sample the vertical distribution of mussel cover, the lower limit of mussel cover was actually displaced vertically upward because sea otters made gaps at or near the lower limit of mussel cover. Interannual variation of mussel recruitment was an additional important contributing factor in the failure of the mussel population to extend dominance of space (VanBlaricom, in prep.). These results raise the issue of completeness in models that attribute "keystone" ecological roles to populations of Pisaster in the Pacific Northwest (Paine 1966, 1974, 1984). Paine's experiments were done within the aboriginal range of sea otters, but at locations at which sea otters have been locally extinct for at least 80 years. I suggest that the recognition of the potential effects of sea otters requires no change in the underlying conceptual framework of Paine's models. Effects of the removal of mussels are comparable regardless of the agent of removal. Likewise, the more episodic nature of mussel removal by sea otters, as compared to sea stars, does not violate the conceptual integrity of existing models (e.g., Paine and Levin 1981) because, as noted above, sea otter foraging is remarkably similar to wave disturbance. However, models based on the assertion that mussels can attain a refuge in size from predators must be revised in deference to the foraging capabilities of sea otters. My data suggest that mussels along the shores of the North Pacific cannot become too large for

Discussion 85

consumption by sea otters. In the presence of sea otters, survival of large mussels will be affected by factors which influence the frequency and spatial distribution of intertidal foraging by sea otters.

There is a distinct possibility that study sites used by Paine and colleagues will be reoccupied by sea otters within the next few decades. Sea otters were reintroduced to the outer coast of the Olympic Peninsula by translocation from Alaska in 1969 and 1971 (Jameson et al. 1982). At present the population numbers about 65 animals, ranges from Destruction Island (47°41'N, 124°29'W) to Cape Alava (48°10'N, 124° 44'W) and is growing in numbers at a rate of 10 to 15% per year (Jameson et aI., in press). There has been little change in the spatial distribution of the population in recent years. Most of the population is concentrated near Cape Alava, less than 30 km south of Paine's study locations. The eventual return of sea otters to Tatoosh Island and Mukkaw Bay seems likely, and will permit direct examination of the integrity of existing models of gap dynamics.

Competitive relationships among common intertidal species in Prince William Sound have not been worked out, and it is unclear how similar this system (possibly dominated by Mytilus edulis) is to outer coast systems dominated by Mytilus californianus, where removal of mussels may provide essential opportunities for the survival of competitively inferior species. At the simplest descriptive level, rocky intertidal communities in the Sound appear to resemble those in protected locations of the northwestern Atlantic coast, in which space utilization by barnacles and fucoid algae is sensitive to factors that limit the abundance of Mytilus edulis (Menge 1976). Qualitative observations suggest that some species of barnacles and fucoid algae benefit from the removal of mussels in Prince William Sound, but these notions have not been tested.

For many marine benthic invertebrates that are broadcast spawners, gonad biomass and potential for gamete production are proportional to individual body size. As a result, populations dominated by large individuals have a much greater total fecundity than populations in which large individuals are rare. Recognizing that Mytilus edulis is a broadcast spawner, and assuming that senescence is unimportant in mussels, it is likely that mussel populations in sea otter breeding areas have reproductive potentials much below potentials for other parts of Prince William Sound. Ifthis hypothesis is accurate, a substantial selective advantage may accrue to individual mussels that occupy cryptic habitats and, as a result, escape from sea otter predation and grow to large size. The advantages of cryptic microhabitat selection are emphasized if one considers that the amount of sea otter breeding area in Prince William Sound continues to increase.

The physical differences in habitat occupied by sea otters in central California and Prince William Sound, Alaska, are obvious and striking. There are, nevertheless, common features in the relationship of sea otters and mussel-dominated intertidal communities in the two regions. The most important of these is the substantial spatial and temporal variation in the effect of sea otters on mussel populations. Such patterns add a dimension of difficulty to the understanding of effects of sea otters on mussel-dominated assemblages: One cannot predict patterns of response in a mussel community simply by knOwing that sea otters are present or absent.

For Prince William Sound, I offer two alternative hypotheses for observed variation across space in mussel size distribution. In the first (termed the refuge hypothesis), nearshore habitats at Green Island are used by sea otters as refugia from severe weather


during fall and winter. When large numbers of otters are thus constrained to forage in the finite area provided by refugia, available prey, including mussels, are heavily exploited. The result is the observed size distribution of mussels at Green Island. Observed mussel size distributions require that Green Island is used by otters as a refuge from weather to a significantly greater degree than Simpson Bay or Orca Inlet.

The second hypothesis (termed the maternal care hypothesis) relates more directly to the segregation of breeding areas and male areas. Description of this hypothesis requires consideration of foraging techniques and patterns of parental care in sea otters. Sea otters forage in a manner indicating the ability to innovate and adjust on the basis of available resources. The advantage of this approach is that changes in prey availability can be exploited by altering the foraging strategy. The disadvantage is that young animals must learn how to forage if they are to succeed in obtaining prey which are elusive, cryptic, or require extensive handling. There is evidence from Prince William Sound that young animals are less skilled than adults in foraging (Estes et al. 1981), and that among independent animals, age-specific mortality rates are highest in year 1, with symptoms often suggesting starvation (A. Johnson, pers. commun.; D. Siniff, pers. commun.).

Post-partum parental care in sea otters is provided entirely by the mother (Kenyon 1969; Garshelis et a1.1984; Payne and Jameson 1984), a pattern that is common among mammals (Kleiman and Malcolm 1981; Trivers 1985). In the maternal care hypothesis, females with pups should locate and utilize areas in which prey are easily captured and handled, where pups can therefore learn to forage on their own. In such areas, females must have access to sufficient food which, combined with existing somatic energy reserves, allows at least marginal survival during the pups' period of dependency. In terms of individual fitness, such a strategy provides important benefits to the pup (a greater probability of surviving to reproductive age) and to the mother (improved inclusive fitness through pup survival, and a less frequent need to share or relinquish captured prey to the pup, thus improving chances for her own survival). Among the major prey types used by sea otters in the Sound (crabs, clams, echiurids, and mussels; Estes et al. 1981), I propose that mussels provide the best solution to the problem of maternal care (see also Sect. 4.5.3 and 4.5 .4). Thus, in breeding areas, sea otters eat mussels with sufficient frequency that large mussels are rare and contribute little to the biomass of the local mussel population. In male areas, sea otters focus primarily on other more energetically rewarding prey and rarely consume mussels in numbers. As a result, large mussels are locally common, contributing the bulk of the biomass present in the population. However, male groups of sea otters have the potential to cause episodically intensive local mortality in mussel populations in response to unknown factors.

A key element in the maternal care hypothesis is the trade-off between ease of capture and nutritional value among prey species in the Sound. Garshelis (1983) suggests that recently weaned young otters which feed exclusively on mussels are not likely to survive. Garshelis argues that mussels are poor-quality prey because the relative ease of capture and handling is overridden by the inadequate return in quantity of consumable biomass per captured individual. However, utilization of more energetically desirable prey (crabs or clams) requires that young animals overcome difficult learning threshholds. Until the threshholds are passed, the hypothesis favors continued use of mussels, possibly with supplementary food provided by the mother. Given Garshelis' observa-

Discussion 87

tions, it appears to be essential for newly independent animals to expand foraging skills quickly beyond mussels, if the young otters are to survive.

The refuge and maternal care hypotheses cannot be rigidly set apart as mutually exclusive alternatives. For example, mussels may be consumed preferentially by females and pups during good weather at Green Island, but by all otters when foul weather forces many other individuals into protected waters. I suggest that available data are not yet sufficient to evaluate the hypotheses, but that the system is sufficiently accessible to allow testing of both scenarios.

Because my work in California was confined to one location, examination of effects of sea otter sex and age on mussel consumption rates is not possible. Based on data gathered at Pt. Piedras Blancas, however, I suggest that variation across space and time in mussel predation by otters is so large in California that differences relating to otter age or sex are unimportant. Additional work in male areas would be helpful in this regard.

I conclude by listing some issues which involve the relationship between sea otters and mussel-dominated intertidal communities, which are unresolved, and which should be resolved if the subject relationships are to be understood. The issues are as follows:

1. Prediction of the effects of sea otters on intertidal communities on most scales will remain difficult until we can attain an understanding of the causes of variation in mussel consumption rates among various subdivisions of sea otter populations. It is essential to understand what drives foraging behavior at the level of individuals and various categories of age and sex. To what extent do genetic variability, learning, and ingestive conditioning (sensu West 1986) contribute to foraging differences among individual sea otters? What do sea otters gain from foraging on mussels in California, given that such foraging is strikingly sporadic? With regard to age and sex, which of the two hypotheses I have proposed for Prince William Sound (refuge hypothesis, maternal care hypothesis) is most accurate? Are there other, better descriptions of this system?

2. Little is known about competitive relationships, responses to disturbances, and life history parameters of species in mussel-dominated communities of the North Pacific north and west of the Olympic Penninsula. Sea otters and mussel-dominated communities range along the shorelines of British Columbia, the southeastern, southern, and southwestern coasts of Alaska, the Aleutian, Kommandorskii, and Kurile Islands, and the northeastern coast of the Soviet Union. There is a rich potential for interaction in all these sites, and they should be investigated. Such information will be particularly useful in evaluating the patterns I have described for Prince William Sound.

3. Mobile, carnivorous invertebrates (e.g., crabs and lobsters) which may consume significant numbers of mussels (e.g., Robles, in prep.) may also be important sea otter prey (e.g., Antonelis et al. 1981; Estes and VanBlaricom 1985; Kvitek and Oliver, this Vol.). The abundance and distribution of sea otters, crabs, and lobsters have all been influenced significantly by human exploitation. What, then, are the food web relationships of mobile carnivores and mussels under circumstances free of human intervention? Will sea otters compete with crabs or lobsters for mussels? Or will sea otters consume most lobsters and crabs, reducing their importance as mussel predators to insignificant levels? Evaluation of these issues is straightforward in concept, but can only be achieved where the subject species are relatively free from human intervention, and are allowed to reach something approaching natural levels of abundance.


4. I have suggested that recently weaned sea otters may need to change their diets from mussels to higher-quality prey if they are to survive their first year. If this suggestion is accurate, how will social interactions of sea otters influence the diet of young, independent animals? How will survival of young males, for example, be affected if territorial males aggressively deny younger animals access to food-rich areas? What are the survival rates of young otters that are socially constrained to feed primarily on mussels? Such interactions could have significant effects on the dynamics of otter populations, and on the relationships of otters with prey populations.

S. Native human populations exploited both sea otter and mussel populations along much of the coastline of the North Pacific prior to the intrusion of European man and the consequent cultural or biological extinction of many aboriginal societies (e.g., Simenstad et al. 1978). Aboriginal man was ecologically significant in the coastal ecosystems of certain localities, such as parts of the Aleutian Islands. However, it is not known if the ecological effects of aboriginal man were important on a large scale. If we are to argue that community models are incomplete without consideration of effects of sea otters or other predators, we must make the same argument for abori~nal man, given that the latter was a known consumer of both sea otters and mussels. While I have not presented data on possible effects of aboriginal man, such effects are, potentially, quite important. Thus, there is a need for aquisition of data sufficient to model the impacts of harvest by native humans on sea otter-mussel interactions.

The listed issues are broad in scope and will require much work if they are to be resolved. I suggest that two sites are particularly suitable for study of sea otter-mussel interactions in the near future. Destruction Island, Washington, is within the current range of the translocated sea otter population along the Olympic Penninsula. It is part of a federal wildlife refuge, is geographically isolated, and is therefore largely free of the intrusive effects of contemporary human activity. Intertidal communities at the Island appear, based on qualitative observations, to be structured in a manner consistent with the models of Paine and Levin (G. VanBlaricom, unpubl.). Kelp forests are present at the Island and probably fluctuate on a seasonal basis. The wave climate.at the Island is probably quite similar to the climate at Tatoosh Island. Thus, it is an excellent site for testing the ideas discussed herein regarding effects of sea otters on communities dominated by Mytilus californianus. Simpson Bay, Prince William Sound, Alaska, is likewise a good site for additional work on effects of sea otters on communities in which Mytilus edulis may be a dominant species. The area has recently undergone the transition from a male (sea otter) area to a breeding area, experiences little human disturbance, and is relatively close to research support facilities available in Cordova. Most importantly, there is a relatively good record available for the demography of mussel and sea otter populations over the past 7 years.

4.7 Summary

Sea otters commonly forage on mussels in rocky intertidal habitats throughout the north Pacific. Mussel-dominated intertidal communities have been the basis for important generalizations regarding competition, predation, disturbance, and habitat com-

Summary 89

plexity in natural ecosystems. Many key studies of these processes have been done at locations within the aboriginal range of sea otters, but at which sea otters now are locally extinct. Here I describe patterns of intertidal mussel consumption by sea otters, and discuss some community-level effects of such foraging, based on concurrent studies done during 1978-1984 at Pt. Piedras Blancas, California, and Prince William Sound, Alaska.

Pt. Piedras Blancas is an established breeding area for sea otters, and is fully exposed to oceanic wave activity. The predominant mussel species in the rocky intertidal is Mytilus californianus. Sea otters forage on mussels by diving repeatedly at the same location during high tide, removing clumps of mussels during each dive. During the surface interval, small mussels and debris are discarded, while large mussels are broken open by pounding against a rock or another mussel balanced on the chest. Soft tissues are ingested and the shells discarded.

As a consequence offoraging technique, sea otters create areas of bare space ("gaps") within mussel cover. Gaps range in size from 0.03 to 1.34 m2 , and the size distribution of gaps is remarkably similar to that of gaps caused by wave shear in other locations. During the study 26% of gaps made by sea otters were subsequently enlarged by additional otter foraging, in some cases over 2 years after gaps were created. Within the study sites, sea otters made more gaps than did wave shear during the period of field work, despite the occurrence of several major wave events.

Formation of gaps was strongly aggregated in space and time. Spatial aggregation correlated with the proximity to sheltered sites for prey handling and consumption. Temporal aggregation was both seasonal and interannual. Most gaps were formed in winter and spring. Among years, gap formation correlated inversely with storm frequency during the second half of the study, but not during the first half. Other possible environmental correlates to gap creation by sea otters include seasonal variation in nearby kelp canopies (which provide resting sites for otters and habitat for alternative prey), seasonal and interannual variation in the quality of mussels as food, and variation among individual sea otters in the inclination to feed on mussels.

Mussels clearly are unable to attain a refuge in size from foraging sea otters at Pt. Piedras Blancas. Sea otters seem to prefer large mussels, which are readily shelled and eaten with the aid of tools.

Prince William Sound is a complex system of bays, fjords, and passages. Most shores in the Sound, and all study sites considered here, are protected from the ef';)cts of oceanic swell. Mussels (Mytilus edulis) are common in patches on several kinds of substrata, including solid rock, mixed sand/gravel/cobble, and mudflats. Sea otters forage on mussels by diving during high tide, bringing to the surface clumps of mussels and associated debris. Small mussels and debris are discarded, while larger mussels are crushed with the teeth and ingested whole, including the shell. On occasion, the shells of the largest mussels are opened with the canines, the soft tissues ingested, and the shell discarded.

Sea otter-mussel interactions were examined at three permanent study sites in Prince William Sound. Green Island was a breeding area and Simpson Bay a male area throughout the study. Orca Inlet was initially vacant of sea otters, but after 1979 fluctuated seasonally from male area to relatively vacant. Based on data from other studies, it is apparent that most mussel predation by sea otters in the Sound is done by adult females,


their dependent pups, and recently independent juveniles. Thus, mussel consumption by sea otters occurred much more frequently at Green Island than at Simpson Bay or Orca Inlet.

Patterns of size distribution in mussel populations appear to correlate with patterns of use by sea otters. Most mussel biomass occurs as large individuals at high density in Simpson Bay and Orca Inlet. In contrast, large mussels were rare at Green Island, regardless of exposure or substratum type, and most mussel biomass occurred as small individuals. The above patterns were strikingly consistent through time. The onlyexception was in Orca Inlet during winter 1980, when male otters stripped large areas of mussels down to bare substratum. Affected mussel populations recovered and were again dominated by large individuals by 1984.

As in California, mussels clearly do not attain a refuge in size from sea otter predation in Prince William Sound, Alaska. Sea otters readily consume the largest available Mytilus edulis without using tools for assistance.

Two alternative models are proposed to explain variation in mussel demography across study sites in Prince William Sound. The "refuge" hypothesis is based on observations that large numbers of sea otters concentrate in protected sites at Green Island during foul weather. Under such circumstances prey populations, including mussels, are heavily exploited, with consequent demographic effects. The "maternal care" hypothesis proposes that females with dependent young and newly weaned juveniles forage preferentially on mussels because they are much easier to capture than other, energetically more rewarding prey. Females choose to forage in mussel patches because dependent young can learn foraging skills, and because females will have to share or relinquish food to the pups less often, both of which ultimately improve the survival rate and inclusive fitness of females and young.

I conclude that sea otters can have significant effects on the structure and dynamics of mussel-dominated communities in both locations studied. Effects include direct modification of population structure and life history options for mussels themselves, and provision of space for species that are competitively subordinate to mussels. Effects of sea otters vary markedly across space and time, and additionally vary with local sex and age composition, and possibly individual inclination, of the sea otter population.

There are a number of important issues which can only be resolved with additional research. It is essential to understand what drives variation in consumption of mussels by sea otters at the level of age, sex, and individual. Consequences of mussel removal for the intertidal community are not well understood in Prince William Sound. Effects of other mobile carnivores (crabs, lobsters) on mussels can be significant, but it is unclear to what extent otters will control populations of other carnivores, thus influencing effects of the other species on mussels. Young sea otters may need to aquire the skill to obtain prey species more energetically favorable than mussels, if the otters are to survive to maturity. How, then, do young otters survive if social interactions prevent them from using food-rich areas? Aboriginal man harvested both sea otters and mussels, and the possible effects of such harvests on sea otter-mussel interactions is entirely unknown.

Summary 91

Acknowledgments. Jim Bodkin, Jerry Busch, Jim Estes, Jean Gravning, Jon Gravning, David Irons, Ancel Johnson, Michael Kenner, Mark Rauzon, and Fred Sorensen assisted with field and laboratory work. Jim Bodkin, Mike Bogan, Al Ebeling, Ancel Johnson, Carlos Robles, Wayne Sousa, and Tom Suchanek provided constructively critical reviews of an earlier version of the chapter. David Carlson, Patti Himlan, Lynn Rathbun, Jennifer Shoemaker, and Susan Strawn assisted with preparation of the chapter in final form. This work was supported financially by the Fish and Wildlife Service, U.S. Department of the Interior, and by the National Geographic Society. My sincere thanks to all of the above.

5 Kelp Communities and Sea Otters: Keystone Species or Just Another Brick in the Wall? M. S. FOSTER and D. R. SCHIEL

5.1 Introduction

Recent reviews have discussed the effects of many biotic and abiotic factors on the abundances of kelp species and how these vary in space and time (Dayton 1985a; Foster and Schiel 1985; Schiel and Foster 1986). Giant kelp (Macrocystis pyrifera) communities along the California coast have been of particular interest because of the diversity of organisms they harbor, their importance to nearshore productivity, and their economic and recreational value (reviewed in Foster and Schiel 1985). The need to develop management directives for these nearshore communities has fostered much debate about the relative importance of various factors affecting community structure, and has engendered what we see as essentially opposing views about the development and maintenance of community structure. The focus for this debate has been the sea otter, Enhydra lutris, which presently occurs along the coasts of central California, Alaska, Washington, and British Columbia (Riedman and Estes, this Vol.). Sea otters consume a variety of invertebrates, particularly sea urchins and abalone, which can be important consumers of kelp.

In discussing other marine systems, Paine (1969) suggested that the "integrity" and "stability" of tropical coral reefs and the rocky intertidal zone of the northeast Pacific were the result of the activities of keystone species. These species occurred high in the food web and, by controlling the abundance of particular prey, kept these prey from greatly altering the community. More recent observations in kelp forests, made as sea otters have increased their abundance and range in Alaska and California, have shown that these mammals can Significantly reduce densities of large sea urchins and alter sea urchin size frequency and spatial distribution (Lowry and Pearse 1973; Estes and Palmisano 1974; Estes et al. 1978; Laur et al., this Vol.). Because changes in sea urchin populations may result in dramatic changes in the distribution and abundance of algae and perhaps other groups, sea otters have also been called keystone species (Estes and Palmisano 1974; Palmisano and Estes 1977; Estes et al.1978; Duggins 1980; Palmisano 1983; Estes and Harrold, this Vol.). Moreover, experiments and observations at a few sites have been generalized to suggest that the hierarchical interaction of sea otter-sea urchin-algae is "extremely" or "most" important in structuring nearshore communities in the ancestral range of the sea otter (Estes and Palmisano 1974; Dayton 1975; Duggins 1980).

Observations of changes in subtidal community composition ranging from high densities of large, exposed sea urchins and low algal abundance, to low densities of these sea urchins and high algal abundance have also led to proposals that alternate

Introduction 93

stable states exist for these communities, with abundant kelp representing one state and deforested areas, dominated by echinoids, the other state (sensu Lewontin 1969; Sutherland 1974). In kelp stands within their range, sea otters are said to mediate these states through their predatory activities (Simenstad et al. 1978). Similarly, Duggins (1980) suggested that sea otters, by feeding on sea urchins, "transform one biological community to another".

These generalization about keystone species and alternate stable states in kelp communities have been incorporated into textbooks (e.g., Carefoot 1977; Nybakken 1982; Valiela 1984), become popularized (e.g., Davis 1977; Woolfenden 1979; Scheffer 1981), used in the context of various management issues (e.g., Armstrong 1979; Cicin-Sain 1982; Estes and VanBlaricom 1985; Levin, this Vol.) and used as partial justification for research related to management alternatives (Estes and Harrold, this VoL).

In this paper we examine whether, as a general hypothesis, this hierarchical view of the cause of kelp community structure - that is, the removal of grazers by otters which in turn allows the persistence of kelp stands - is supported by the available evidence from kelp communities in California. If this hypothesis is generally true for these communities within the ancestral range of sea otters in California, one would predict that sea urchins are the most important controlling factor of kelp community structure within this range in the absence of sea otters. That is, the absence of sea otter predation within this range should allow increased numbers of sea urchins and consequent general deforestation (i.e., a high percentage of deforested sites). Alternatively, a more complex view, with sea otters being one of many factors affecting these communities may be necessary, both to understand community structure and to attempt to "manage" it.

Many of the early studies that generalized the importance of the hierarchical view were based on observations of natural experiments that assessed its predictions at selected sites. Because experimental removals of otters have not been done, and there are few published accounts comparing sites before and after sea otter foraging (e.g., Laur et aI., this Vol.) , these studies relied on information about comparative community structure within vs. outside the range of the sea otter. Examples of this are Estes and Palmisano (1974) and Estes et al. (1978), who compared sites with and without otters in Alaska and argued for the pervasive effect of Enhydra lutris on kelp communities.

We accept the fact that otters can have a great impact on the invertebrate fauna of nearshore areas, and that this can lead to great changes in macroalgal abundance. Our surveys (below) indicate, however, that while the keystone species hypothesis can be locally true, it is not generally true in California, and that even on a local scale many processes other than predation by sea otters can greatly affect both echinoids and kelp. Multi-site information reviewed here and by Foster and Schiel (1985) further indicates that kelp communities in California can vary substantially in species composition and abundance at time scales shorter than one complete turnover of the common macroorganisms associated with them, indicating a lack of stability (as discussed by Connell and Sousa 1983). If one recognizes this variability as a natural property of these communities (see also Estes and Harrold, this VoL), we argue that it is difficult to support the concept of alternate stable states for them. In fact, we suggest that the stable state view tends to obscure rather than clarify an understanding of natural variability by

94 Kelp Communities and Sea Otters

focusing on the extreme cases of either densely forested or echinoid-dominated communities.

Finally, we propose that a better understanding of the influence of sea otters on kelp communities in the North Pacific will only result from the explicit recognition of spatial and temporal variation and the diversity of factors that contribute to it. To aid in this understanding, we propose that generalizations about processes might be developed in the context of a "type" approach that associates particular types of community structure at various spatial and temporal scales with particular sets of physical factors.

5.2 Kelp Community Structure

There are several recent reviews that include a discussion of kelp community structure both worldwide (Dayton 1985a; Schiel and Foster 1986; Harrold and Pearse, in press) and in California (Foster et al. 1983; Foster and Schiel 1985). Although this structure can be highly variable in space and time, we have summarized it in composite form for a hypothetical, well-forested site in California (Fig. 5.1) to provide a context for the discussion that follows. Animals, and especially plants, generally occur in broad zones along a depth gradient. As originally suggested by Neushul (1965), three broad benthic zones may be recognized (ZI ,Z2, and Z3; Fig. 5.1), with the surface canopy kelps that typify kelp forests usually most abundant in the mid-depth zone (5-20 m). Vegetation layering from multiple canopies can result in plant cover exceeding 300% over a point

'" a: 4 UJ

20cm

within patch

( Macrocystis

~ereOCystis P Phylloapadlx

'Egregia

!cystoseira

I- Lamlnarla setchellii I Pterygophora

t Articulated corailines

~'" FO~O::o=~dAlgae .jjo Strongylocentrotus

/""" __ Encrusting Coraliines

.A'\ Sessile Invertbrates

,.p Kelp Gametophytes

,00 :: Sed I men t

Fig. 5.1. Kelp forest community structure at different depths and spatial scales

>< ... z o .... < .... UJ CI UJ >

Approach and Methods 95

on the bottom; cover would no doubt be much higher if multiple blade layers within a canopy were included. Plant cover is often highest at mid-depths. At any particular depth, organisms are commonly distributed in patches that are often associated with subhabitats such as flat or vertical substrates, holdfasts, bolders, gravel, etc. Relatively flat surfaces have been most frequently studied, and plants also commonly occur in patches on these (e.g., Dayton et a1. 1984).

Large (> 3 cm test diameter) sea urchins, especially Strongylocentrotus franciscanus, are also known to vary in abundance, distribution, and behavior both within and outside the range of sea otters in California. We have summarized this variation in space and time into five general categories representing patterns observed in the field (Fig. 5.2). All of these patterns have been observed at sites in California outside the range of the sea otter (e.g., Fig. 5.2B, Foster 1975; Cowen 1983; on isolated bolders in Rosenthal et al. 1974; see Appendix 5.1). The abundance of sea urchins and their effects may also vary with time at a particular site (e.g., Pearse and Hines 1979; Dean et a1.1984; Ebeling et a1. 1985; Harrold and Reed 1985), and with depth (e.g., Mattison et al. 1977).

The smaller, white sea urchin, Lytechinus anamesus, may also deforest areas in southern California (Clarke and Neushul1967; J. Dixon, S. Schroeter, A. Ebeling, pers. commun.). The effects of sea otter predation on this species are not known.

5.3 The Otter as a Keystone Species in California: Local or General?

5.3.l Approach and Methods

Some of the earliest observations of high sea urchin densities and associated deforestation were from southern California (e.g., North 1965; North and Pearse 1970; Pearse et a1. 1970), and these observations have been used to support generalization about the overall importance of keystone predators to kelp community structure (e.g., Mann

A COMPLETEL Y

FORESTED NO LARGE

~ :::fAf~, u. 100 a. o II: w

~:~~~ u 100 b. 11-

o 200

B

URCHINS IN CREVICES

C URCHINS IN

SMALL PATCHES ( I-3m)

D URCHINS IN

LARGE PATCHES (20-50m)

E DEFORESTED

LARGE URCHINS

ALL OVER

brtrn:hillb o 200

DISTANCE (m)

Fig. 5.2. Graphical representation of categories of sea urchin abundance and distribution, and associated algal cover that have been observed in kelp forests in California. Graphs are of variation in algal cover that might be measured along a 200-m transect in the mid-depth zone (Fig. 5.1) with varying abundances and distributions of sea urchins. Cover can exceed 100% because of vegetation layering. Numbers in () are patch diameters. Subcategory A.b. represents sites with high sessile invertebrate cover. This subcategory could occur in any of the other categories except E


1973; Estes and Palmisano 1974). However, while these early observations provided information about how widespread this deforestation was along portions of the southern California mainland (Pearse et al. 1970), many of them suggested that the large numbers of sea urchins were associated with sewage discharge (Pearse et al. 1970; North 1983). A variety of recent evidence also suggests that oceanographic conditions and sewage discharge, not increases in sea urchins, may have been directly responsible for much of the initial deforestation in this region (reviewed by Foster and Schiel 1985).

Additional evidence for possible keystone effects of sea otters in California is based on observations of kelp surface canopy extent and composition at particular sites in central California before and after occupation by sea otters (cf. VanBlaricom 1984). However, historical observations made prior to occupation are often available for only a very limited period of time. Moreover, as discussed by VanBlaricom (1984) and others (Miller and Geibel 1973;Foster 1982; Kimura and Foster 1984), surface canopies can vary substantially in time regardless of the presence of sea otters. Therefore, the lack of necessary information on temporal variation prior to occupation by sea otters makes the significance of any before vs. after differences difficult to determine given the available data. The best information comes from sites in the vicinity of Monterey. High densites of echinoids and red abalone (Haliotis rufescens) were reported in the late 1950's before otters reached this area (Mclean 1962). Ten years after otters became established, these invertebrates were confined mostly to crevices, and giant kelp was abundant (Lowry and Pearse 1973). Given the otters' diet and food consumption rates (Woodhouse et al. 1977), and that sites deforested by sea urchins (Fig. 5 .2 E) presently exist outside the sea otters' range, there is no doubt that otters can have a great impact on the abundance of large sea urchins that, in turn, can lead to great changes in algal assemblages at particular sites in California (see Laur et al., this Vol.). As posed in the Introduction, however, the question we address is: how prevalent is this effect?

Rigorous evaluations of the generality of hypotheses are rarely done in ecology, and this is probably one reason for much of the controversy surrounding the "importance" of factors (Underwood and Denley 1984). The commonest evidence for generality, even if the hypothesis has been tested by well-designed experiments, is usually presented in the discussion sections of papers in the form of natural history information from other sites, citations of supporting papers, and a review of supporting theory. This evidence can be quite subjective. The best evidence for generality would be an unbiased estimate of how frequently the hypothesis is true. For the hypothesis under consideration here, obtaining this evidence would require a random selection with subsequent sampling of rocky subtidal habitats that are, or could be, forested. These habitats must be within the ancestral range, but outside the present range, of the sea otter. One would then look for the frequency .)f sites in category E (deforested with numerous large urchins; Fig. 5.2E) in these habitats; if the keystone effect of sea otters is general, one would predict a high frequency of this category. This approach assumes that deforestation (Fig. 5.2 E) is caused by the sea urchins, and accepts that sea otters would remove these sea urchins (but see below). It also assumes that the frequency of deforestation does not vary significantly in time. This assumption could be tested by sampling over some appropriate time interval.

Estes and Harrold (this Vol.) suggest another approach to determining generality; sampling similar to that above but both within and outside the sea otters' range to test

Approach and Methods 97

"the null hypothesis that sea otters have no measurable influence on plant/herbivore interactions." This would not test the generality of the keystone effect of sea otters, only whether or not there is a difference in extent of deforestation within and outside their range. As far as we are aware, that sea otters have a "measurable influence on plant/herbivore interactions" has never been questioned.

The random surveys discussed above have not been done. However, using recent reviews as models, we reviewed and summarized information on the abundances of sea urchins and macro algae contained in all the available surveys of shallow subtidal reefs done outside the present range, but inside the ancestral range, of the sea otter in California. Recent reviews have tabulated the results of experimental studies to determine the general importance of competition (Connell 1983; Schoener 1983) and predation (Sill et al. 1985) to the organization of communities. In these reviews, one measure of the general extent or importance of each process to the structure of natural communities was the frequency of studies that demonstrated significant effects of competition or predation. This approach may be flawed by bias in the intent of the original investigator, by bias in the habitats and organisms investigated (Sih et al. 1985; Conner and Simberloff 1986) and by the criteria and interpretation of the reviewer (Ferson et al. 1986). However, it provides better evidence than typical discussion sections because all available studies are examined (there was some selection according to journal and year).

The information for our review was obtained from several sources (Appendix 5.1). In addition to the published scientific literature and available reports, we requested unpublished information from investigators who have been doing multi-site surveys for many years. The surveys were tabulated into the five categories shown in Fig. 5.2 to determine what range of variability in macroalgal and echinoid abundances is found where otters do not yet forage. Areas in the vicinity of Los Angeles were not included because these probably reflect significant impacts of humans, especially due to sewage discharge (e.g., Grigg and Kiwala 1970; Pearse et al. 1970; Meistrell and Montagne 1983; North 1983). We also excluded sites or studies where commercial fishing for sea urchins (discussed in Foster and Schiel 1985) may have influenced community structure.

One difficulty in categorizing the results of these surveys was that the abundances of macroalgae and echinoids were described over many spatial scales, from localized patches (ca. 100 m2 ) to broader surveys encompassing several depth zones and locations. Because of this, our designation of "sites" was often subjective and highlights the problems associated with inferring general patterns from non-standard and often unstructured field surveys (Schiel and Foster 1986). We arbitrarily defined "site" as an area within a depth zone (i.e., shallow, mid, deep: cf. Fig. 5.1) that could be observed during a period of a dive. For example, the observations of Rosenthal et al. (1974) were made over an area of several hundred square meters at a depth of 14-20 m. We deSignated this as a site. Engle (unpubl.; see Appendix 5.1) often listed data and observations within two depth zones (3-6 m and 7-20 m) in several locations. For our purposes, each depth zone at a location was considered a site. At many sites, particularly in northern California, sea urchins were often described as inhabiting the bases and sometimes the sides of large boulders while algae covered the tops. We placed such sites in category C or D depending on the abundance of sea urchins. A site was placed in category A only if the investigators stated that large sea urchins were absent. If sea


urchins were not mentioned at all, the site was placed in category B on the assumption that some were probably present. Some sites have been repeatedly visited, and some of these changed categories during the course of observations. In these cases each different category was counted as if it were a single site. We reviewed our categorizations twice, and Engle (pers. commun.) independently reviewed our categorizations of his data.

5.3.2 Results

Considering first the spatial variation in kelp communities, only 19 of the 224 sites covered by the review fell into category E, deforested with abundant sea urchins (Table 5.1). This is the "urchin deforestation" (urchin-dominated barren grounds) state described in many areas (see reviews by Lawrence 1975; Harrold and Pearse, in press) and predicted by the keystone species hypothesis in the absence of sea otters. Because of the overall scarcity of sites dominated by sea urchins, this tabulation indicates that

Table 5.1. Summary of site classifications based on a review of surveys of kelp/reef habitats in California outside the range of the sea otter

General location of sites Total number of sites Number of sites in each category ABC D E

California Mainland: Oregon Border to Santa Cruz 69 2 43 15 6 3

California Mainland: San Luis Obispo to Mexican Border 41 1 20 14 2 4

Southern California Islands 72 0 29 30 9 4

Total 182 3 92 59 17 11

Sites that changed No. of sites that Categories of change b categories changed a

One change 36 5 5

5 5 2 2

5 5 Two changes 6 1 1

1 1 1

Total 42 2 10 17 5 8

Grand total 224 5 102 76 22 19

% of total 2.2 45.5 33.9 9.8 8.5

a Counting each change as a separate site. b One site changed between categories A and B, five sites between Band C, etc. Specific survey locations and references are listed in Appendix 5.1, and general locations are dis-cussed in Riedman and Estes (this Vol.). Categories range from A (completely forested - no large sea urchins) to E (deforested - abundant sea urchins, encrusting algae only; see Fig. 5.2).

Possible Bias 99

otters would have only local, not general, "keystone" effects if they were to occupy all sites currently outside their range in California. The removal of echinoids from sites in categories D, C, and B would not cause an overall change in the character of the site, but rather would potentially change the relative abundance of deforested patches within the site, shifting it toward category A. Sites in category B would be virtually unchanged in terms of macroalgal abundance after the removal of sea urchins. Sites in category A would obviously not change, as large sea urchins are absent. However, even within the range of sea otters, this category may not be common (cf. Lowry and Pearse 1973; Hines and Pearse 1982) except perhaps as a result of disease (Pearse et al. 1977).

Twenty sites were observed to change categories with time (42 changes; Table 5.1). The Naples Reef site was the only one that changed across the nearly complete range of possible categories twice (sea urchins in small patches and abundant algae to numerous sea urchins and deforested to sea urchins in crevices and abundant algae; Ebeling et al. 1985 in Appendix 5.1). These changes occurred in less than 5 years. Recent high recruitment and survival of sea urchins at this site appear to be causing another change to deforestation (D. Reed, pers. commun.). These and other relatively long-term observations (Dean et al. 1984; Harrold and Reed 1985) indicate that variability across a range of composition, rather than stability in one of two possible states over at least one turnover of kelp and sea urchin populations, is characteristic of kelp forest communities in California. Sea otters could reduce this temporal variability caused by changes in sea urchin abundance and behavior at a particular site (VanBlaricom 1984). They would not necessarily eliminate it, as indicated by the recent observation of a category E site within the sea otters'range near Monterey (S. Webster and C. Harrold, pers. commun.; M. Foster, pers. observ.). Assuming the distribution of site categories in space (Table 5.1) is indicative of the distribution in time, then the temporal effects of sea otter foraging would be similar to the spatial effects discussed above; the overall character of sites presently outside the range of the sea otter would generally not change (less than 10% are deforested), but the time spent in categories A, and especially B, would increase.

Although these data indicate that kelp communities in California are not generally structured around interactions between sea otters, sea urchins, and macro algae , it could be argued that sea otters and sea urchins are at least relatively more important "bricks in the wall" (Fig. 5.3) than other factors. This issue is probably much more complex because of regional differences in oceanographic conditions and frequency of extreme water motion (Foster and Schiel 1985), in other sea urchin predators (Tegner and Dayton 1981; Cowen 1983) and no doubt other factors. In addition, there are factor interactions such as those between storms and sea urchin abundance and behavior (Cowen et al. 1982; Ebeling et al. 1985). As these differences and interactions are just now being described, it seems premature to speculate in general about relative importance.

5.3.3 Possible Bias

As previously mentioned, the data used in these analyses were from sites not chosen at random. Sites were selected by the original investigators for a variety of reasons,

SUBSTRATUM

MULTIFACTOR

INTERACTIVE VIEW

RECRUITMENT

WATER MOTION ETC.

SEA OTIERS OTHER

PREDATORS

LIGHT

NUTRIENTS

DISPERSAL SEA URCHINS

(/) w

100 -

!:: 60 (/)

=II:

20

01HER

~r I r ,.

ABC 0

, SITE CATEGORIES IN AREAS

WITHOUT OTIERS

~ +otters -otte rs , ..

SEA

OTTERS

SEA URCHINS

MACRO AlGAE

SINGLE FACTOR

HIERARCHICAL VIEW

, r

o

~ ~ ll.; W o

ORGANISMS

Fig. 5.3. Alternative views of the causes of kelp forest community structure. Graph is of data in Table 5.1

Possible Bias 101

from structured sampling of coastal areas for environmental surveys, to research focused at sites of particular interest. At least three problems could result from this nonrandom selection. First, if sites were consistently chosen in areas where kelp was abundant, then the frequency of sites in categories A through D would be overestimated. Second, the sites are unequally distributed in space, with a large number around islands in southern California (Table 5.1). This region could naturally have more sites in a particular category than other regions of equivalent size. Finally, the frequency distribution of categories may vary over time due to some spatially large episodic event.

None of these problems can be eliminated with the available data, but these data and other observations do not suggest a strong bias toward underestimating the prevalance of deforestation. Many of the surveys covered large sections of the coast (e.g., Areas of Special Biological Significance; see Appendix 5.1 and associated citations) that would presumably find sites in each of our five categories if they occurred. Moreover, one of the objectives of the multi-site surveys by Davis (unpubl.) and Engle (unpubl.) is to provide general assessments of large areas. Other investigators chose sites for a variety of reasons and, in some cases, because sites had high abundances of sea urchins (consult citations in Appendix 5.1). Thus, deforested sites could be overestimated as suggested by Harrold and Pearse (in press) for reports of deforestation by sea urchins world-wide. Moreover, the Area of Special Biological Significance surveys (ASBS) represent the largest set of published surveys that are likely to be unbiased in their choice of sites. These areas were not necessarily chosen because they contained kelp forests, and surveys in those that did were not restricted to locations with a surface canopy. The frequency distribution of site categories from the ASBS surveys suggests that the frequency of sites with a high abundance of sea urchins may be overestimated by the grand total in Table 5.1 (% ASBS/% Other: category A -0/3; B -70/ 39; C -26/36; D -2/12; E -2/10). A comparison of the frequency distribution of southern California island sites (SCI), including those that changed categc ,·es, with all others suggests that this regional bias may also contribute to an overestimate of sea urchin abundance (% SCI/% Other: category A -0/4; B - 33/56; C -44/26; D -14/16; E -9/8). The subtidal surveys used were not intentionally biased in time, and cover nearly the entire period of subtidal research in California. More studies have been done in recent years, but the possible bias of this is unknown.

On the other hand, few surveys specifically examined shallow reefs without kelp (e.g., Pequegnat 1964). In addition, the surveys by Yellin et al. (1977) did not include areas beyond the seaward edge of the surface kelp canopy. If sea urchins produced deforested areas seaward of the canopy as noted by Mattison et al. (1977) at a site in the same general area, then the frequency of deforestation in this area would be underestimated by these surveys. Seaward portions of four of the sites surveyed by Yellin et al. (1977) were later investigated by Foster and Reed (1980); the seaward edge of one (Greyhound Rock; Appendix 5.1) was deforested, while those of the other three were not.


5.4 Otter Effects: Geographic and Historical

Would a review of similar surveys in Alaska or the western coast of Canada also reveal a general lack of deforestation in the absence of sea otters? We cannot answer that question, but our results show that the null hypothesis of an absence of a general keystone effect will have to be examined more rigorously in evaluating extrapolations from site specific observations. Moreover, it is far from clear how pervasive the effects of otters were prior to their near extinction by fur traders in the 1800's. Estes and VanBlaricom (1985) stated that the near extinction of sea otters allowed the development of many present-day fisheries. However, Simenstad et al. (1978) indicated that aboriginal humans had a considerable impact on otter distribution near village sites in Alaska, and Dayton and Tegner (1984a) have pointed out the possibility of similar impacts in California. Much of the variability in sea urchin distribution and effects that we presently see outside the sea otters' range, therefore, may have occurred in at least localized regions of the coast when aboriginal humans were present prior to the fur trade. Sea otters were present in the northeast Pacific prior to substantial occupation by aboriginal humans (Simenstad etal.1978). However, we are unaware of any evidence for or against pervasive effects of these animals during this time other than suggestions based on present-day observations and generalizations. Moreover, the recent deforestation of a site near Monterey by sea urchins ( discussed above) suggests that assumptions concerning the uniformity of prey reduction within the range of the sea otter (Estes and VanBlaricom 1985) may be questionable.

5.5 Beyond Otters

Recent discussions of kelp communities in other parts of the world where sea otters do not occur indicate that, as in California, sea urchin abundances and effects are spatially and temporally variable (Andrew and Choat 1982; Dayton 1985a,b; Miller and Colodey 1983; Scheibling 1986; Pringle 1986; Schiel and Foster 1986; Harrold and Pearse, in press). As the results of more long-term studies become available, it appears that variability, not stability, is characteristic of kelp forests and perhaps most other communities (Connell and Sousa 1983).

This variability, particularly in sea urchin abundances and effects, suggests to us that these communities do not exhibit alternate stable states but vary continuously over a dynamic range of composition where deforestation and dense macro algal cover -whether caused by sea urchins, storms, nutrients, or other factors - are only the extremes. Generalizing from these extremes poses many problems of interpretation. Most sites in our review fell into categories with abundant macro algae , with few sites being dominated by echinoids in large patches (Fig. 5.3). About 44% of sites (categories C and D) had significant amounts of both macroalgae and echinoids. An understanding of how kelp grows in dense stands, or of echinoid grazing effects in widespread, dense aggregations, may not be particularly useful in these intermediate situations. For example, sea urchins may alter their foraging behavior when high density aggregations break up, when drift algae become available, or when algal recruitment is very high

A Model for Structure and Organization 103

(Mattison et al. 1977; Dean et al. 1984; Harrold and Reed 1985; Vadas et al. 1986). The different densities of the major organisms, therefore, may have significant effects on the dynamics of kelp communities, and on any predictive scheme we might devise to "manage" these communities. A more realistic consideration on the entire spectrum of patterns observed should form a basis for better understanding kelp forest structure and dynamics.

5.6 A Model for Structure and Organization

Kelp forest communities in California differ in structure, and many of these differences have been associated with particular sets of biotic and physical factors (reviewed by Foster and Schiel 1985). It is apparent from that review and the surveys considered in this paper that no single factor, such as grazing by sea urchins, explains a high percentage of community structure over many locations. The effects of foraging by sea otters also occur in the complex context of other factors affecting structure seen in kelp communities, and these effects are seen over many temporal and spatial scales. There are many observations showing that in localized patches or sites, grazing by dense aggregations of echinoids can remove large macroalgae and produce deforested areas. Removal of sea urchins, however, by storms, disease, humans, sea otters, other predators or other processes may not produce such predictable results. Many factors come into play such as the availability of algal propagules, the nature of the substratum, nutrient and light levels, etc. (Table 5.2; Fig. 5.3; reviewed in Foster and Schiel 1985). Various species interactions among algae may occur that change relative abundances, and hence the character of the community through time, quite independent of grazers (e.g., Pearse and Hines 1979; Duggins 1980; VanBlaricom 1984; Ebeling et al. 1985). At present, prediction of these changes is site-specific at best.

This highlights two problems in designing investigations of kelp forests. The first of these is the question of context. If information from one or a few sites is to be generalized to a larger area of the coast, sufficient sampling is required so that the full range of kelp forest structure along a coast is identified. The survey data summarized in this paper and the review by Foster and Schiel (1985) indicate that such sampling is likely to show that sites over a larger area vary in structure. This leads to a second consideration. Can this variability be associated with particular physical characteristics at particular spatial scales and, if so, can these associations be partitioned into a limited number of natural divisions or types for more detailed investigations? Our experience and review of the literature suggest that these associations do occur (see also Dayton et al. 1984) and that natural divisions can be made. At the scale of sites, the resulting "types" are conceptually similar to the divisions of "wave-exposed" and "protected" and, on a smaller scale, the different vertical zones in rocky intertidal communities. These divisions of rocky shores have proven to be a natural way to partition variation in structure (e.g., Lewis 1964; Stephenson and Stephenson 1972) and reflect differences in the relative importance of factors affecting structure (Foster et al., in press, b). For kelp communities, there may be a relatively small number of characteristic types at a given spatial scale in many locations. These types are unrelated to the categories in


Table 5.2. Examples of a "type" approach for sampling kelp communities. Factors are selected a priori for the spatial scale of interest a and random sampling is done within sites of particular types to quantify the abundances and sizes of selected species or groups. These factors can be used both within and between areas of coastline (only a few possible factors and variables are listed)

Question: How can sampling be partitioned to assess the community structure of kelp forests?

Factors

Wave exposure Depth Substratum

Surface canopy type

Type Relief

Shallow (0-5 m) Mid (5-20 m) Deep (> 20 m)

Low Moderate High

Hard rock Soft rock

Variables measured

A. Abundances of kelp species (surface and understory) B. Abundances of other understory algae

Low Moderate High

Macrocystis Nereocystis No canopy

C. Abundances of selected invertebrates (e.g., sea urchins, abalone) D. Size-frequencies

Site

1.

2.

Type

- mid-depth high exposure soft rock low relief Macrocystis canopy

Two·structure types

Physical characteristics

mudstone reef; sand abundant water motion high but varies with season

mid depth - hard, massive rock, - moderate exposure some cobble, little sand

hard rock - water motion moderate moderate relief but varies with season Macrocystis canopy

Community structure

Macrocystis surface canopy with maximum cover in late summerearly fall few understory kelps bottom cover of annual foliose red algae

- low abundance of sessile invertebrates

- moderate abundance of large sea urchins

Macrocystis surface canopy with some seasonal variation in cover

- abundant understory kelp bottom cover of perennial articulated and encrusting coralline algae low abundance of sessile invertebrates

- moderate abundance of small sea urchins

a Mid-depth, between sites in the examples given; 1. Sandhill Bluff; 2. Stillwater Cove (see text for further discussion of these sites)

Fig. 5.2; types are a higher level of classification that indicate an association between the physical environment and community structure.

To illustrate how this approach would be used to partition sampling for community structure at mid-depth between sites within a region, we have chosen two examples (Table 5.2) based on sites in central California described in detail by Cowen et al.

A Model for Structure and Organization 105

(1982), Foster (1982) and Reed and Foster (1984). We suggest that Sandhill Bluff and Stillwater Cove are "types" because their physical characteristics of depth, exposure, substratum, and surface canopy appear to be typical of many sites in central California influenced by similar factors. The low abundance of understory kelps at Sandhill Bluff, the foliose red algal understory composed of apparently annual species, and the large seasonal and annual changes in the cover of surface and bottom canopies are characteristic of many sites examined between Point Santa Cruz and Ano Nuevo Island, a distance of over 35 km (Foster 1982). Recent, unpublished surveys from the Big Sur coast (VanBlaricom and Foster, in prep.) suggest that the mid-depth structure of many kelp forest sites in this region are similar to Stillwater Cove. Whether the causal factors are also similar is unknown. However, many of these forests are so similar in structure and relative abundances of species to Stillwater Cove, which was examined experimentally by Reed and Foster (1984), that similar factors probably affect their structure. Observations suggest that other structural types at mid-depth occur in this region of central California: (1) Point Cabrillo, inside Monterey Bay, characterized by low exposure, moderate relief, and a Macrocystis canopy (Lowry and Pearse 1973; Hines and Pearse 1982; Watanabe 1984b; Breda and Foster 1985; Foster and Schiel 1985); and (2) more open coast, high relief sites with Macrocystis or Nereocystis canopies, such as around the entrance of Carmel Bay and at Grimes Point along the Big Sur coast (M. Foster, pers. observ.; VanBlaricom and Foster, in prep.). Point Cabrillo is characterized by a generally thick, persistent canopy of Macrocystis, high abundance of the fucoid Cystoseira osmundacea, a paucity of understory kelps, a foliose red algal understory of mainly perennial species, and a moderate cover of sessile invertebrates. It is similar to other sites at the southern end of Monterey Bay. The causes of its structure are largely unknown, but may be related to relatively low wave exposure combined with conditions favoring sessile invertebrate growth (Breda and Foster 1985). The open coast sites may reflect more oceanic conditions, particularly low sedimentation.

Clearly, further work needs to be done to determine if these types are truly representative of the majority of sites in the region, and to investigate further the factors thought to be responsible for the observed structures. We emphasize that this "type" scheme only provides a contextual partitioning of sampling that forms a basis for hypothesis testing through experiments within and between types. Dayton et al. (1984) and Laur et al. (this Vol.) suggest similar associations between the physical properties of reefs and their biological characteristics. Many more types are suggested by the combinations of factors in Table 5.2, and not all factors are included (e.g., nutrients). It is highly likely that such types may change or be abandoned altogether with increased understanding of kelp forest communities; we are imposing discrete groupings on what are surely more continuous variables. However, this approach seems to us to provide a heuristic middle ground between generalizing from the single-factor, two-state approach evaluated above, and a more rigorous understanding based on multi-site demographic analyses.


5.7 Conclusions

A hierarchical view of kelp forest organization, with sea otters ameliorating the pervasive grazing effects of echinoids, may be appealing in its simplicity, but it has the disadvantage of depreciating the importance of many other factors known to influence the presence of macroalgae and the structure of nearshore communities (Fig. 5.3). The evidence presented in this and other reviews strongly indicates that the concept of sea otters as keystone species is applicable to only a relatively small number of sites and thus does not constitute a general explanation of kelp community structure in California. Moreover, rather than existing in alternate stable states, kelp communities exhibit a dynamic range of composition that appears to be driven by a complex of interacting factors. The effects of these other factors are not trival. Our tabulations show that only about 9% of the 224 sites examined were dominated by echinoids (Table 5.1). In addition, many studies have shown that any of a number of factors such as water motion, light, nutrient levels, and substratum type and availability can strongly influence community structure at particular sites (see reviews by Kain 1979; Foster and Schiel 1985; Dayton 1985a; Schiel and Foster 1986). Some of these "bricks" may be bigger than others but, in the absence of information from many sites and about processes such as dispersal, recruitment and demographics, this remains speculation.

The principal task we face as ecologists trying to understand the processes that determine the distribution and abundance of organisms is to recognize the variability that exists, and to cope with it by partitioning it into useful categories for experimentation. The available field observations suggest that a limited set of community structures or types exist at least at mid-depths within kelp forests, and these types and the factors that affect them may constitute models such as those proposed by Foster (1982) from which testable hypotheses may be generated. They might also provide the context for determining under what conditions a particular factor may become important. General "types" may be found at other spatial scales such as within patches.

Sea otters will probably increase in numbers and range in California, and this should provide additional opportunities to examine their effects on kelp communities. Enhydra lutris is a legally protected, threatened species. Experimental reductions in the density of existing populations is impracticable, and any manipulations fall well into the political realm. However, since their main community effects are initiated via the removal of sea urchins, an appropriate way of examining the potential effects of sea otters is with experiments involving manipulations of echinoid densities (additions and removals). These could be done at different spatial scales in various "categories" of sites, so that experimental factors such as site effects, seasonality of propagule availability, different "types" of kelp forest structure, and stochastic processes can be accounted for. Such experiments have rarely been done in northeast Pacific kelp communities (e.g., Palmisano and Estes 1977; Duggins 1980; Cowen et ai. 1982; Laur et aI., this VoL).

This review also indicates there are many interesting questions about the distribution and abundance of sea urchins that remain to be investigated. Do sea urchins maintain extensive deforested areas in Alaska for very long periods? If so, how does the ecology of sea urchin populations differ in these areas compared with sites in California? What accounts for the different distribution, abundance, and behavior patterns of sea

Summary 107

urchins in California? What are sea urchin recruitment patterns and how do they vary at different spatial scales? What is the ecology of juvenile sea urchins? In addition, there have been few investigations of the recruitment, growth, reproduction, and survival of even the prominent species of macroalgae. How do these change at different densities and between different forest types? The focus of investigations leading to a clearer understanding of kelp forest structure must incorporate these important demographic questions. We believe that these questions are tractable, but will require thoughtful and detailed experiments. These will probably de-emphaSize generalities in the first instance, but at least have the advantage that they offer some future hope that predictive models needed for management will be realistic.

5.8 Summary

Observations at selected sites in the northeast Pacific have led to two generalizations: (1) kelp communities on nearshore, subtidal reefs exist in one of two stable states, forested with few large sea urchins or deforested with abundant large sea urchins, and (2) changes of state are controlled by a keystone predator, the sea otter. In contrast, many observations made both within and outside the sea otters' range in California indicate that these subtidal communities are spatially and temporally variable. Our review of over 220 descriptive surveys of such communities in California that occur outside the range of the sea otter shows that sea urchin grazing effects can be highly variable in the absence of sea otters and that deforestation by sea urchins is the exception « 10% of the sites surveyed). In addition, the communities do not exist in two states controlled by otters, but rather exhibit a dynamic range of composition where the above "states" are the uncommon extremes.

A number of different factors affect community structure and the relative importance of these factors can vary over small distances and short time intervals. We therefore reject the keystone species hypothesis as a geographic generality for the control of kelp community structure in California and perhaps elsewhere. We propose that more realistic general hypotheses be developed and tested using a "type" approach that associates particular kinds or types of community structure with particular sets of physical factors.

Acknowledgments. We thank G. Davis and 1. Engle for allowing us to use their unpublished surveys, and N. Andrew, A. Ebeling, J. Estes, S. Gaines, 1. Pearse, D. Reed, L. Stocker and G. VanBlaricom for reviewing the manuscript.

Ap

pen

dix

5.1

. S

ite

cate

gori

zati

ons.

Cat

egor

ies

are

show

n in

Fig

. 5.

2 an

d di

scus

sed

in t

ext.

If

a si

te w

as s

urve

yed

mor

e th

an o

nce,

it

was

onl

y co

unte

d on

ce

unle

ss a

cha

nge

in c

ateg

ory,

no

t as

soci

ated

wit

h se

a o

tter

for

agin

g, o

ccur

red

wit

h ti

me.

If

chan

ge o

ccur

red,

eac

h ch

ange

was

co

un

ted

as

a "s

ite"

(th

ese

site

s li

sted

und

er "

Cha

nge

obse

rved

" be

low

). I

f di

ffer

ent

site

s w

ere

surv

eyed

at

the

sam

e ge

nera

l lo

cati

on b

y d

iffe

rent

inv

esti

gato

rs,

the

gene

ral

loca

tion

is l

iste

d m

ore

than

onc

e

Gen

eral

A

rea

Tri

nida

d H

ead(

i)

Kla

mat

h R

iver

(i)

Poi

nt A

rena

Sau

nder

s R

eerC

i )

Ger

stle

Cov

e(i)

Pyg

my

For

est(

i)

Poi

nt R

eyes

(i)

No.

of

site

s S

urve

yed

15 2 3(b

)

2 8 4 2

Sit

es s

urve

yed

once

or

whe

re n

o ch

ange

in

cate

gori

es o

ccur

red

1. O

rego

n B

orde

r to

San

ta C

ruz

No.

of

site

s in

Urc

hin

cate

gori

es

Not

es

Ref

eren

ce

Dep

th (

m)

A

B

C

D

E

0-7

7

Mos

tly

sand

B

oyd

19

80

6

-9

1 be

low

12

m;

7-1

4

7(a)

li

ght

low

be

low

7 m

0

-6

1 L

ight

low

6

-21

1 (

a)

belo

w 6

m

Boy

d et

al.

1981

2-8

G

otsh

all

et a

l. 19

74

8-1

5

15

-23

1

0-8

M

ostl

y sa

nd a

nd g

rave

l, F

oste

r et

al.

1980

8

-12

1

and

turb

id,

belo

w 7

m

0-6

3

2 M

ostl

y sa

nd

DeM

arti

ni 1

979

6-9

1

belo

w 9

m

7 -

12

1 9

-14

1

0-2

U

rchi

ns a

t m

ore

expo

sed

Pen

ning

ton

et a

l. 19

81

0-4

1

site

s 2

-7

1 (a)

4-1

2

1

0-6

1

San

d be

low

12

m

Cha

n 19

80a

6-9

1 (

a)

.... o 00

:;>;:

co .a n o 3 3 s:: a riO

b Q-

uo

~ S?

M

e.

~

Bod

ega(

i)

2 0

-6

1 S

and

belo

w 7

m

Ris

tau

et a

l. 19

79

CI'l

c

6-1

2

S S D

oubl

e P

oint

(i)

2 0

-8

1 C

han

19

80

b

<0 ...

1 (a)

'<

8

-12

D

ux

bu

ry R

eef(

i)

0-6

1

San

dy a

nd t

urbi

d be

low

3 m

B

lake

197

9

Pil

lar

Poi

nt(i

) 2

0-1

1

Low

lig

ht b

elow

6 m

V

and

erw

ilte

tal.

19

79

>

11

1 (a)

Pig

eon

Poi

nt

2 ( c

) 2

Alg

ae o

n s

hall

ow r

eef

top

s V

anB

lari

com

198

4 A

flo

Nue

vo

1 9

-10

L

ow r

elie

f Y

elli

n et

al.

1977

E

llio

tt C

reek

1

9-1

1

Low

rel

ief

Yel

lin

et a

l. 19

77

Wad

dell

Cre

ek

1 1

0-1

1

1 L

ow r

elie

f Y

elli

n et

al.

1977

G

rey

ho

un

d R

ock

1 6

-10

1

Yel

lin

et a

l. 19

77

Gre

yh

ou

nd

Roc

k 3

4-9

F

oste

r an

d R

eed

19

80

9

-12

1

2-1

8

1 (d)

Sco

tt C

liff

s 1

6-1

0

1 Y

elli

n et

al.

1977

Sco

tt C

reek

2

6-9

1

Yel

lin

et a

l. 19

77

9-1

2

Dav

enpo

rt

3 3

-9

2 U

rchi

ns a

ppea

r to

lim

it a

lgae

C

ayo

t and

No

rth

109

1 9

-13

be

low

12

m(d

)

Dav

enpo

rt

2 4

-8

1 Y

elli

n et

al.

1977

5

-9

1

Lid

dell

Cre

ek

5-9

1

Yel

lin

et a

l. 19

77

Yel

low

Ban

k 7

-9

1 Y

elli

n et

al.

1977

S

andh

ill B

luff

4

-6

1 F

oste

r an

d R

eed

1980

T

able

Roc

k 7

-10

1

Yel

lin

et a

l. 19

77

Thr

ee M

ile

1 5

-9

1 Y

elli

n et

al.

1977

N

eedl

e R

ock

1 6

-9

Yel

lin

et a

l. 19

77

Ter

race

Poi

nt

1 6

-8

1 Y

elli

n et

al.

1977

S

tock

ton

Ave

nue

1 6

-9

1 Y

elli

n et

al.

1977

S

oque

l P

oin

t 1

(c)

Van

Bla

rico

m 1

984

To

tal:

6

9

2 43

15

6

3 .....

. 0 \0

App

endi

x 5.

1. (

cont

inue

d)

.... 0

Sit

es s

urve

yed

once

or

whe

re n

o ch

ange

in

cate

gori

es o

ccur

red

2. M

orro

Bay

to

Mex

ican

Bor

der

(Mai

nlan

d)

Gen

eral

N

o. o

f si

tes

No.

of

site

s in

Urc

hin

cate

gori

es

Not

es

Ref

eren

ce

Are

a S

urve

yed

Dep

th (

m)

A

B

C

D

E

Dia

blo

Cov

e 2

3-5

N

orth

et

al.

1979

>

5

1

San

ta R

osa

Ree

f 1

(c)

1 L

aur

et a

I.,

this

Vol

.

Lon

e B

lack

Roc

k 2

4-6

L

aur

et a

I.,

this

Vol

. 6

-10

1

Isle

Vis

ta

2 0

-6

Fos

ter,

per

s. o

bser

v.

6-1

2

Gol

eta

2 0

-6

Fos

ter,

per

s. o

bser

v.

6-1

8

San

ta B

arba

ra

1 2

-12

1

Fos

ter,

per

s. o

bser

v.

Car

pint

eria

1

10

-12

1

Fos

ter,

per

s. o

bser

v.

Lat

igo

Poi

nt(i

) 3

0-5

U

rchi

ns a

roun

d re

ef b

ases

M

orin

and

Har

ring

ton

5-1

0

1979

1

0-2

0

New

port

Bea

ch(i

) 1

0-1

5

:;>::

Bru

sca

and

Win

n 19

79

(1)

Hei

sler

Par

ke i)

9

-18

1

Bru

sca

and

Wic

kste

n oS

n

1979

0 :3

Cor

ona

Del

Mar

(i)

1 0·

-15

Bru

sca

and

:3 Z

imm

erm

ann

1979

s:: s..

Car

ona

Del

Mar

3

9-1

2

1 (a)

P

eque

gnat

196

4 ~.

12

-14

1 (

a)

., ::l

14

-16

"'" U>

San

Mat

eo F

ores

t 1

2-1

5

Fos

ter

et a

l. 19

83

(1) .,

San

Ono

fre

For

est

1 1

2-1

5

Urc

hins

in

fron

ts

Dea

n et

al.

1984

0 .... ....

Del

Mar

1

14

-20

R

osen

thal

et a

l. 19

74

(1) ;;l

La

Joll

a C

anyo

n(i)

3

6-8

K

obay

ashi

197

9 8

-10

>

10

La

Joll

a 3

< 1

2 N

eush

ul 1

965

en

~

12

-14

1

3 >

14

1 3 el

Poi

nt L

oma(

e)

11

0-6

3

Tur

ner

et a

l. 1

96

8

'<

12

-18

1

2 >

18

3 1

Impe

rial

Bea

ch

(c)

No

rth

19

74

Tot

al:

41

20

14

2 4

So

uth

ern

Cal

ifor

nia

Isla

nds

3. S

an M

igue

l Is

land

Tyl

er B

ight

8

-20

E

ngle

(f)

Cro

ok P

oint

1

0-2

0

Eng

le(f

) W

ycko

ff L

edge

1

2-1

5

Dav

ilg )

H

are

Roc

k 6

-9

Dav

is(g

)

4. S

anta

Ros

a Is

land

For

d P

oint

1

6-2

0

1 E

ngle

(f)

Eas

t P

oint

2

3-1

2

Eng

le(f

) 1

2-2

0

Eas

t B

eche

rs B

ay

1 3

-12

1

Eng

le(f

) N

W B

eche

rs B

ay

1 6

-20

1

Eng

le(f

) B

eaco

n R

eef

1 3

-20

1

Eng

le(f

) R

odes

Ree

f 1

7-2

0

1 E

ngle

(f)

Rod

es R

eef

1 1

2-1

5

Dav

is(g

) T

alco

tt S

hoal

1

10

-20

1

Eng

le(f

) B

ee R

ock

1 3

-20

1

Eng

le(f

) C

hina

Cam

p 1

8-2

0

Eng

le(f

) N

Joh

nson

's L

ee

1 8

-11

1

Dav

is(g

) S

John

son'

s L

ee

1 1

4-1

7

Dav

is(g

)

.....

.....

.....

App

endi

x 5.

1. (

cont

inue

d)

.- .- N

Sou

ther

n C

alif

orni

a Is

land

s

5. S

anta

Cru

z Is

land

Gen

eral

N

o. o

f si

tes

No.

of

site

s in

Urc

hin

cate

gori

es

Not

es

Ref

eren

ce

Are

a S

urve

yed

Dep

th (

m)

A

B

C

D

E

Gul

l Is.

3

0~8

Neu

shul

1965

8~1O

10~18

Coc

hes

Pri

etos

0~9

1 F

oste

r 19

75

Fry

's H

arbo

r 1

6~9

1 H

ixon

198

0 F

ry's

Har

bor

1 11~14

Dav

is(g

) S

corp

ion

1 5~6

Dav

is(g

) S

an P

edro

Poi

nt

1 3~20

1 E

ngle

(f)

Cav

ern

Poi

nt

6~20

1 E

ngle

(f)

Pot

atoe

Har

bor

1 8~20

Eng

le(f

) P

riso

ner's

Har

bor

1 3~20

1 E

ngle

(f)

Cue

va V

alde

z 1

0~20

1 E

ngle

(f)

11 F

atho

m C

ove

1 0~20

Eng

le(f

) F

rase

r C

ove

1 0~20

1 E

ngle

(f)

Mor

se P

oint

0~20

1 E

ngle

(f)

~

Will

ows

3~20

1 E

ngle

(f)

'" .;-A

lber

t's

0~20

Eng

le(f

) n 0

San

dsto

ne P

oint

1

0~20

Eng

le(f

) 3

Yel

low

Ban

ks

1 0~20

Eng

le(f

) 3 s::

Hun

gry

Man

Gul

ch

0~20

Eng

le(f

) s.: ~

6. A

naca

pa I

slan

d $»

::l

"'

- en

Sur

vey

Roc

k 3

0~8

Cla

rke

and

Neu

shul

'" $»

8~34

1967

0 ::::

> 3

4 '" ... '"

Ana

capa

1

6-9

H

ixon

19

80

en

D

avis

(g)

:= A

dmir

al's

Ree

f 1

15

-18

:3

Cat

hedr

al C

ove

1 6

-11

D

avis

(g)

:3

1 0

-20

E

ngle

(f)

~

Cat

Roc

k '<

Fre

nch

y's

Cov

e 1

0-2

0

Eng

le(f

)

Fis

h C

amp

1

0-2

0

Eng

le(f

)

7. S

anta

Bar

bara

Isl

and

Arc

h P

oin

t 1

8-1

1

Dav

is(g

)

Arc

h P

oin

t 2

3-6

E

ngle

(f)

6-2

0

Arc

h R

eef

1 1

0-2

0

Eng

le(f

) S

hag

Ro

ck

1 3

-20

E

ngle

(f)

Sut

il I

s.

3-2

0

Eng

le(f

)

8. S

an N

icol

as I

slan

d

SW

En

d

11

Ove

rall

sit

e; c

hang

e be

low

H

arro

ld a

nd

Ree

d 19

85

Du

tch

Har

bor

6-1

8

1 C

owen

19

83

D

utc

h H

arb

or

8-2

0

Eng

le(f

) S

E S

ands

pit

1 1

0-2

0

Eng

le(f

) N

W B

oile

r 1

10

-20

E

ngle

(f)

9. S

anta

Cat

alin

a Is

land

Far

nsw

ort

h B

ank(

i)

17

-24

E

ngle

an

d C

oy

er 1

981

Wes

t o

f A

rrow

Po

int

2 0

-10

E

ngle

(f)

10

-20

1

Eas

t o

f A

rrow

Po

int

1 0

-20

E

ngle

(f)

Wes

t E

nd

to

Cat

H

arbo

r 1

0-2

0

Eng

le(f

) E

ast

En

d t

o C

at

Har

bor

0-2

0

Eng

le(f

) .....

. .....

. v.

>

App

endi

x 5

.1. (

cont

inue

d)

- -.j:o. S

outh

ern

Cal

ifor

nia

Isla

nds

10. S

an C

lem

ente

Isl

and

No.

of g

ener

al s

ites

N

o. o

f si

tes

in U

rchi

n ca

tego

ries

N

otes

R

efer

ence

A

rea

Sur

veye

d D

epth

(m

) A

B

C

D

E

Cas

tle

Roc

k 1

6-2

0

Eng

le(f

) W

est C

ove

1 9

-20

1

Eng

le(f

) C

him

ney

Ree

f 1

9-1

5

Eng

le(f

) Se

al C

ove

1 3

-12

1

Eng

le(f

) C

ove

Poi

nt

1 6

-11

1

Eng

le(f

) Sm

uggl

ers

Cov

e 1

6-1

5

1 E

ngle

(f)

Pyr

amid

Cov

e 1

3-2

0

Eng

le(f

) P

yram

id H

ead

1 3

-20

E

ngle

(f)

Wils

on C

ove

1 3

-20

1

Eng

le(f

)

Tot

al:

72

0

29

30

9 4

Sit

es w

here

cha

nge

obse

rved

(h)

(1.,

2.,

etc

. ref

er t

o r

egio

ns o

r is

land

s ab

ove;

fir

st o

bser

vati

on =

A,

seco

nd =

B, t

hird

= C)

One

Cha

nge

;;>::

(1)

1. C

hina

Lad

der

2 9

1B

1A

Aft

er s

torm

Y

elli

n et

al.

1977

>S

" n 1.

Poi

nt S

anta

Cru

z 8

6 1B

1A

C

hang

e af

ter

urch

in d

isea

se

Mat

tiso

n et

al.

1977

; 0 3

7 IB

1A

P

ears

e an

d H

ines

3

8 1B

1A

19

79

§ 10

1B

lA

~:

0-2

0

1A

1B

Eng

le(f

) 0:>

::s

0

-12

1A

IB

E

ngle

(f)

c:>.

5-9

1A

1B

D

avis

(g)

en

(1)

0:>

3-6

1A

IB

E

ngle

(f)

9 3

-8

1A

1B

Eng

le(f

) .... (1

) ~

3. C

uyle

r's

Har

bor

2 4.

Joh

nson

's L

ee

2 5.

Pel

ican

Bay

2

5. C

aver

n P

oint

2

5. P

otat

oe H

arbo

r 2

5.

Pel

ican

Bay

5

. O

riza

ba C

ove

5.

Fo

rney

Cov

e 5

. G

ull

Isla

nd

7. S

eali

on

8. S

WE

nd

8. S

E L

ight

10

. N

W H

arbo

r

1. S

andh

ill

Blu

ff

2. N

aple

s R

eef

Tot

al:

Gra

nd t

ota

l:

2 ,.., "' 2 2 2 2 2 2 3 2

42

224

3-8

3

-8

0-2

0

0-2

0

14

-17

11

9-2

0

3-2

0

8-1

2

1C

5-1

5

2 5

1B

1B

1B

1C

10

102

1A

1A

1A

1A

1A

1B

1A

1A

1B

1B

1B

Tw

o C

hang

es

1A

1A

17

76

5 22

1B

1B

1A

IB

8 19

One

lar

ge p

atch

; hi

gh a

lgal

re

crui

tmen

t

Eng

le(f

) E

ngle

(f)

Eng

le(f

) E

ngle

(f)

Dav

is(g

) H

arro

ld a

nd

Ree

d 19

85

Eng

le(f

) E

ngle

(f)

1st

chan

ge a

fter

sto

rm,

Cow

en e

t al

. 1

98

2;

2nd

afte

r st

orm

s an

d pe

rhap

s F

oste

r, p

ers.

obs

erv.

o

tter

s C

hang

es a

fter

sto

rms

Ebe

ling

et

al.

1985

(a) H

igh

abun

danc

e o

f se

ssile

inv

erte

brat

es.

(b)T

his

stud

y su

rvey

ed a

nu

mb

er o

f si

tes

aro

un

d P

t. A

rena

bu

t in

suff

icie

nt d

etai

ls a

nd d

iscu

ssio

n to

cat

egor

ize

mos

t o

f th

ese

site

s -

thre

e ge

nera

l d

epth

cat

egor

ies

assi

gned

bas

ed o

n d

iscu

ssio

n, p

rese

nce

of

brow

n al

gae,

and

sea

urc

hin

abun

danc

e. (

c) D

epth

no

t gi

ven.

(d

) Few

fol

iose

alg

ae a

t th

is d

epth

in

this

reg

ion

may

ind

icat

e ef

fect

s o

f lo

w l

ight

, n

ot

sea

urch

in g

razi

ng.

(e) N

umer

ous

surv

eys

have

bee

n d

on

e in

th

is la

rge

kelp

for

est

-w

e us

ed t

he

sour

ce c

ited

bec

ause

the

y su

rvey

ed m

ulti

ple

tran

sect

s pr

ior

to s

ea u

rchi

n ha

rves

ting

in

th

e ar

ea.

(f) U

npub

lish

ed d

ata

from

mui

tisi

te,

mul

tiye

ar s

urve

ys b

y D

r. J

ack

Eng

le,

The

Tat

man

Fou

ndat

ion,

Cat

alin

a M

arin

e S

cien

ce C

ente

r, P

.O.

Box

398

, A

valo

n, C

a. 9

0704

. (g

) Unp

ubli

shed

inf

orm

ati

on

fro

m m

uiti

site

, m

ulti

year

sur

veys

by

Mr.

Gar

y D

avis

, Cha

nnel

Isl

ands

Nat

iona

l P

ark,

190

1 S

pinn

aker

Dr.

Ven

tura

, C

A.

9300

1. (

h)C

ount

ing

each

cha

nge

as a

sep

arat

e si

te.

(i) A

rea

of

Spe

cial

Bio

logi

cal

Sig

nifi

canc

e (A

SB

S)

surv

ey.

Vl S 3 ~ -lJ\

6 Sea Otters, Sea Urchins, and Kelp Beds: Some Questions of Scale

J. A. ESTES and C. HARROLD

6.1 Introduction

For more than a century ecologists have been interested in population and communitylevel consequences of consumer foraging. More recently, papers such as those by Connell (1961), Brooks and Dodson (1965), Paine (1966), Janzen (1970); Harper (1969) and Dayton (1971) have been influential in focusing attention on the topic.

It was this intellectual development, coupled with opportunities provided by a history of near extinction and subsequent recovery of sea otters (En hydra lutris), that led to the following scenario of how sea otters influence kelp forest communities in the North Pacific Ocean (Fig. 6.1): Grazing by sea urchins can reduce or eliminate fleshy macroalgae. Sea otter predation reduces the size and density of sea urchins. Communities in which sea otters are abundant are thus typified by well-developed assemblages of macroalgae, which, having escaped herbivory, derive much of their structure from competitive interactions. The plant assemblage produces detritus which faUs to the substratum and is fed upon by herbivores. In this setting the remaining urchins are largely sessile, employing what can be thought of as a "sit and wait" foraging strategy,

Sea Otter Predation

$/ ,\8

Plant Competition Intense

"

Sea Urchins Low Population Density

of Small Individuals

/ Grazing Intensity Low on

Kelp & Other jI Macrophytes ,

\ Sea Urchin

Foraging Strategy Sit & Wait

Detrita, "',,-Our) High

Sea Urchins High Population Density

of Large Individuals

'\ Plant Competition Weak or Absent

If Grazing Intensity High on

Kelp & Other

! I Macrophytes,...

Sea Urchin Foraging Strategy

Mobile-Active Searching

~D"'" ',,,,-o,r Low

Fig. 6.1. A dichotomous model for sea otter-sea urchin-kelp interactions. The presence or absence of sea otter predation is indicated by + or -. Arrows indicate cause and effect relationships in positive feedback loops. See text for further explanation

The Questions 117

one consequence of which is the further reduction of destructive grazing. In the absence of sea otters, similar physical settings take on a different structure and organization of plants and their herbivores. Left unchecked by sea otter predation, sea urchin populations increase and grazing intensifies, ultimately resulting in what have been called "urchin barrens" (Pearse et al. 1970; Lawrence 1975; recently reviewed by Dayton 1985a and Harrold and Pearse, in press). Such areas are characterized by a preponderance of encrusting coralline algae, with upright species and large fleshy algae being rare or absent. Detritus production is reduced. Thus faced with a diminished supply of detritus, the "sit and wait" foraging strategy by urchins switches to one of active searching and extensive movement, resulting in destructive grazing upon living plants. As described above, both settings are positive feedback systems because once achieved, the processes in one case tend to maintain high plant biomass, whereas those in the other tend to maintain sea urchin barrens (Fig. 6.\).

The preceding scenario is based largely on work by Lowry and Pearse (1973), Estes and Palmisano (1974), Dayton (1975), Estes et al. (1978), Duggins (1980), Breen et al. (1982), VanBlaricom (1984), Reed and Foster (1984), and Harrold and Reed (1985). Supporting evidence is now available from several studies over a broad geographical range (see Sect. 6.3). However, in the minds of some (e.g., Tegner 1980; Foster and Schiel 1985 and this Vol.) this scenario has, unjustly, been given the status of a paradigm. The objection seems not to be with the proposed effect of sea otters, but that its importance must be judged relative to the importance of other structuring processes. We concur with this view.

In our opinion the questions are these: (1) Do kelp-dominated habitats and sea urchin barrens exist as alternate stable states of community organization, or is there a continuum of change between them (see Foster and Schiel, this Vol.)? (2) In regions, areas, or habitats where sea otters enhance kelp beds, what is the "breadth" of the interaction? In other words, what are the effects on other species? If coastal food webs are organized around strong linkages among species (sensu Paine 1980), then the influence of sea otter predation may be far ranging indeed. (3) How general is the scenario?

This chapter considers the question of generality. First, we state the question formally. Second, we critically examine available evidence. Third, we discuss variation in kelp forest communities over space and time. Finally, we suggest two directions for future study to help resolve the contentious question of generality in the sea otter/sea urchin/kelp scenario.

6.2 The Questions

We begin by posing the question of generality in a heuristically useful way. Consider some measure of species abundance, such as density or percent cover. We define community structure as the vector of such measures for all species. We then allow the vector be to affected by three variables - space, time, and the presence or absence of sea otters, which we portray by the linear model

(1)

118

where Ai Tj Ok Yijk

= = = =

Questions of Scale

the effect of area i, the effect of time j, and the effect of k sea otters. (k ranges from 0 to equilibrium density), and the community vector in area i, at time j, and with sea otter density k.

This model partitions variation in community structure into components that are conceptually and analytically useful. The area and time components imply no cause and effect; they simply describe all variation in community structure except that related to otter predation.

Our task is to define the relative importance of each of these three sources of variation. In concept, the design protocol necessary to accomplish this is straightforward. For example, random plots might be established in areas with and without sea otters, or at sites in which sea otter density varies from zero to its maximum possible value. These sites could then be monitored at random time intervals, and the variation contributed by time, space, and sea otters determined by a three-way ANOVA, random effects model. Variations on this theme, such as fixed time or space effects, might be preferable, although these would not greatly affect the design protocol or the sampling problems.

In practice, however, this is a broad order for two reasons. First, there are no available data from properly arrayed and sampled plots in areas with and without sea otters. Ideally, this should be done by replicating the proposed experimental protocol at various locations throughout the geographic range of the species. Past studies, in which sea urchin barrens have been reported to occur in areas lacking sea otters, provide few data on variation in space and time, from areas with and without sea otters. Without such data, it is impossible to quantify the importance of variation contributed by the effects of sea otters in our model. For the same reason, reports indicating spatial and/or temporal variation in community structure in areas with or without sea otters only demonstrate the not surprising fact that such variation occurs. However, available data here as well are inadequate to rigorously evaluate the relative importance of these effects. Second, there is a range of scales (Dayton and Tegner 1984a) on which the three treatment effects may act. We do not know the sample numbers, plot sizes, or sample intervals necessary to account for these. If large sample sizes (hundreds or thousands of plots), large plots (~ 1 m2 ), or short sample intervals (weeks or months) are appropriate, can these be managed? When incorporating the effects of sea otters into our study design, what limits to space, time, and replication are appropriate? Our present concern is with this second group of problems. Specifically, we ask:

I. How spatially variable are kelp forest communities? This question is considered on a range of scales, from within I m2 patches, to among geographic provinces.

2. How temporally variable are kelp forest communities? This question also is addres-sed on a range of sclaes, from within years to over geologic time.

In considering these questions, two points should be kept in mind. First, extensive spatial and temporal variation in community structure, in systems with or without sea otters, does not necessarily exclude significant added effects from sea otter predation. It does, however, greatly complicate the problem of analyzing those effects, and thus bears directly on the nature of appropriate study designs. Second, it is important to distinguish two kinds of variation in community structure: (1) that which occurs within

The Evidence 119

algal-dominated assemblages or sea urchin barrens, and (2) that which occurs between them (Fig. 6.2).

6.3 The Evidence

Existing evidence for the influence of sea otters on plant/herbivore interactions is of three kinds: (1) comparative observations of areas with and without sea otters; (2) experimental manipulations of small habitat patches, in which the effects of sea otters are mimicked; and (3) historical records of change that occurred following the natural range expansion of sea otters. Studies by Estes et a1. (1978) in the western Aleutian Islands, Duggins (1980) in southeast Alaska, and Breen et al. (1982) in British Columbia, were based largely on the comparative approach. In all three accounts, benthic communities from areas with sea otters were reported as being algal-dominated, whereas similar habitats lacking sea otters were characterized by extensive sea urchin barrens.

Manipulative experiments to mimic the influence of sea otters on plant-herbivore interactions were employed by Palmisano and Estes (1977) in the western Aleutian Islands, and by Duggins (1980) in southeast Alaska. Sea urchin densities were artificially reduced in both cases, presumably Simulating the effect of sea otter predation. Palmisano's and Estes' experiments were done in intertidal pools; Duggin's were done in subtidal habitats. In both instances community structure in the experimental plots shifted from urchin barrens to algal dominated assemblages, whereas unmanipulated control plots remained unchanged. Numerous other studies in the North Pacific Ocean, in which sea urchin grazing has been shown to limit algal assemblages, provide similar evidence (Harrold and Pearse, in press).

Duggins (1980) further demonstrated that competitive interactions became an important ecological force in structuring the plant association following removal of sea urchins. Studies by Dayton (1975) at Amchitka Island, and by Reed and Foster (1984) at Stillwater Cove, in central California, both areas where sea otters occur, similarly demonstrated the importance of competitive interactions to the structure of plant associations.

Historical reconstructions were used by VanBlaricom (1984) to demonstrate an areal expansion and change in species composition of the kelp canopy in central California following range expansion of sea otters (Fig. 6.3). This finding was based on detailed maps of kelp canopies assembled by the U.S. Department of Agriculture in 1911 and 1912, a time when the sea otter population in California was limited to a small number of animals in the vicinity of Point Sur. In addition to areal increases of the kelp canopy, there was a shift in the relative abundance of species from Nereocystis leutkeana (bull kelp) to Macrocystis pyrifera (giant kelp). Yellin et a1. (1977), Dayton et a1. (1984) and VanBlaricom (1984) have provided evidence that undisturbed stands of Macrocystis can limit, and often exclude, Nereocystis by way of competition for light or space.

Each approach suffers certain drawbacks. The comparative studies lack appropriate controls for spatial variation, which may be considerable (see Sect. 6.4.2). Thus, the possibility cannot be excluded that such variation confounded the findings of

a

b

120 Questions of Scale

Fig. 6.2a-d. Underwater photographs of kelp beds at a Am chitka Island, Alaska (the kelp understory was cleared from the area around the diver); b San Nicolas Island, California; and sea urchin barrens at c Shemya Island, Alaska; d San Nicolas Island, California

The Evidence 121

c

d

Fig.6.2c,d

[ 1 KM

I

122

00

Soberanes Pt.

1911 1912 1981

,,,., '. .t '. . • ",t

~ ...

1973

Questions of Scale

A .. '\\;: . . ,.." .. :.;

~

• • ., ,

1981

Fig. 6.3. Kelp canopy distribution in central California before and after the arrival of sea otters. (VanBlaricom 1984)

Estes et al. (1978), Duggins (1980) and Breen et al. (1982). VanBlaricom's (1984) historical reconstructions also lack controls, in this case for temporal variation, which is considerable in some kelp forest communities (see Sect. 6.4.3). The manipulative studies are more nearly definitive in the sense that they were properly controlled for spatial and temporal variation. However, they present other difficulties having to do with the scales on which ecologically important processes in kelp forests occur. For obvious practical reasons, the experiments were done in small patches of habitat, yet the effects of sea otter predation are imparted over much larger areas. Considering that the larvae of sea urchins may disperse widely (Strathmann 1978), while spore dispersal of many marine plants is thought to be limited by comparison (Anderson and North 1966; Dayton 1973; Schiel 1981 ; Deysher and Norton 1982), it is easy to imagine artifactual effects from small scale manipulations. Additionally, sea otters appear to be size-selective predators on the larger urchins (l.A. Estes, unpubl.). This probably is a subtle effect, difficult to portray exactly under experimental conditions.

These shortcomings notwithstanding, there are now a number of independent supporting accounts, suggesting that the otter/urchin/kelp scenario may occur widely. If

Variation in Space and Time 123

other studies of similar concept and scope are used as a standard of comparison (e.g., Paine 1974; Breen and Mann 1976; Laws et al. 1975), evidence for the otter/urchin/ kelp scenario is as good as most and better than many. However, the collective results are still far from definitive evidence for the generality of the interaction. But to put this problem in perspective, there are few, if any, examples of community-level interactions that do not suffer similar drawbacks. This is an understandable outcome of the workings of modern science. Funding agencies typically support projects that investigate new ideas or unstudied processes. Most researchers thus shift focus with each new proposal by asking a progression of new questions of a system, or old questions of different systems. This approach often leaves undocumented the extent to which results apply over larger areas, sometimes leading to controversy if the issue at question is disputed among scientists or achieves the status of wide public concern. Disagreement over the importance of competition in natural communities is a prime example (Salt 1984). A more balanced approach, in which the generality of apparently important ecological phenomena is rigorously determined, will most likely prove necessary as ecology matures as a scientific discipline.

The effects of sea otter predation are disputed (Foster and Schiel, this Vol.) and are becoming highly visible public issues. Sea otter foraging is conflicting with shellfisheries. To date, the perceived effect of otter foraging has been small compared with what it eventually will be as human and otter populations continue to expand. In dealing with this problem, resource managers ultimately must determine where sea otters will and will not be. To make that decision intelligently, however, they must be able to answer the following questions: (1) To what extent are sea otters responsible for shellfishery declines? (2) To what extent will sea otters cause a shift from sea urchin barrens to algaldominated assemblages? and (3) How do we value the alternatives? Knowing the generality of sea otters effects is a prerequisite to answering these questions.

6.4 Variation in Space and Time

We now turn to the question of how extensively, and on what scales, kelp communities in the North Pacific vary in space and time. This discussion is based mainly on results from our own studies at Attu Island, westernmost of the Aleutian archipelago, and San Nicolas Island in the southern California Bight (Fig. 6.4). These locations represent geographical near-extremes in the recent distribution of sea otters, and consequently they might be expected to include extremes in community-level effects attributable to large-scale geographical influences. Both areas are being studied to document changes in the structure and organization of coastal communities in response to the reestablishment and growth of sea otter popUlations. Sea otters re-inhabited Attu Island in the mid-1960's, apparently by way of westward dispersal from the Rat Islands (Jones 1965). From the mid-1970's to the early 1980's, when the data reported herein were obtained, most of the coast of Attu remained uninhabited by sea otters.

San Nicolas Island presently is uninhabited by sea otters, and probably has been for a century 0,[ more. The U.S. Fish and Wildlife Service has identified San Nicolas Island as the preferred alternative site for the translocation of sea otters. Anticipating this


PACIFIC OCEA • tudy itt

Fig. 6.4. The North Pacific Ocean, indicating place names referred to herein and the locations of study areas at Attu and San Nicolas islands

event, we began research in 1980 to describe the nearshore communities at San Nicolas Island.

6.4.1 Methods

Similar kinds of data have been gathered at Attu and San Nicolas Islands. Six sublittoral study sites were selected at San Nicolas Island. A 50 m long transect was run along the bottom at each site at a depth of about 10-15 m. Ten 1 m2 quadrats were placed at random intervals along each transect. These were marked by embedding stainless steel bolts using marine epoxy into holes drilled in the substratum. The quadrats were used to describe the epibenthic "turf" assemblage by 20 randomly selected points , chosen and measured in situ (see Foster et al. 1985, for a description of the sampling technique). Five 2 m X 10 m swaths were selected at random intervals along the main transect, running perpendicular to it and marked at both ends by epoxy-embedded bolts . The swaths were used to describe densities of the larger solitary invertebrates and algae. Beginning in 1980, samples were usually taken twice annually, once during spring and once during late summer/early autumn.

The same procedures have been used at Attu Island, with the following exceptions: there are four rather than six study locations; none of the swaths or quadrats are per-

Variation in Space 125

manently marked, but a starting point for the swaths is determined by lowering a weighted line from a boat to the bottom at a location determined by shore line-ups; samples are taken at 3 m (10 ft) depth intervals, from about the 0.0 tidal level to a depth of 15 m (50 ft); and the sites were sampled irregularly during summers only from 1976 to 1983.

We use selected examples from these studies to illustrate some of the kinds of spatial and temporal variation that occur in kelp forest communities. The sampling regime at Attu does not permit description of change in specific patches of habitat over time, although it does permit the description of spatial variation at any point in time. Specific analytical techniques are described where appropriate.

Four quadrats and two swaths were selected at random from each study site at both locations. Only the 1 O-m depth data from Attu Island are considered here. Our analyses of spatial variation are based on samples taken from these swaths and transects during August 1980 and April 1985 at San Nicolas Island (the first and most recent sample periods), and July- August 1983 at Attu Island (the most recent sample period). These data are presented in Appendices 6.1,6.2, and 6.3. Our analyses of temporal variation are limited to several of the common species at San Nicolas Island, but include data from each sample period. Temporal changes within particular quadrats or swaths are based on data from the subsamples described above. Island-wide averages are based on the full complement of swaths and transects.

We measured sampling error by once replicating, over an interval of several days, selected quadrats and swaths at San Nicolas Island. We assume that population changes were negligible over this brief interval. Replicate samples were taken in September, 1985. Three swath counts were resampled at each of four study sites; five quadrats were resampled at each of three study sites. Differences between the first and second sample were determined for each species in each swath and quadrat that was resampled. Means and standard errors, across swaths or quadrats, were calculated from these data for each species.

6.4.2 Variation in Space

In this section we examine patterns of spatial variation in popUlation density or percent cover at three levels: among plots (quadrats and swaths) within sites; among sites within geographical regions (i.e., San Nicolas and Attu); and between regions. Except for sea urchin densities at Attu, these analyses are based on the data in Appendices 6.1,6.2, and 6.3.

6.4.2.1 Quadrats

The assemblage of benthic dwelling species in kelp forests is diverse and patchily distributed. Such variation was evident in most of the I-m samples at both study locations. For example, at Attu Island the number of species sampled per quadrat ranged from 1 to 8; at San Nicolas Island the range was 3 to 17 (Fig. 6.5). Thus, there is extreme variation in both the number of species occupying small patches of habitat, and the species composition of those patches.


6 A

4

2

0 II) 6 - B 0 ~

"0 0 :::3 4 0

...... 0 ~ 2 Q) .0

E :::3

0 z

4 c

2

0 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 1920

Number of Species/Quadrat Fig. 6.5. Frequency distributions of the number of species sampled per quadrat in 1 m point·contact samples from a Attu Island, 1983; b San Nicolas Island, 1980; c San Nicolas Island, 1985

In order to evaluate the distribution of species within and between plots, we tested the fit of these data to one of the simplest imaginable conditions: that at a given location and time, all sample categories (= species hereinafter) are equally common and randomly distributed. Consider the situation in which there are a total of x recorded species from 20 randomly selected points in y randomly selected quadrats. The probability that a particular species will occur beneath any point in any quadrat is 1/x. Among the 20 points sampled, the probability of the species not occurring is (X_1/X)20; hence the probability that it will occur at least once is 1 - (x_l/x)20. If y quadrats are sampled, the probability of any particular species being present (Le., that it will occur one or more times) in z quadrats is distributed as a binomial with the probability density function

p(z) = yCZ (x-1/x)Z [1- (x-1/x)]Y-z .

Ofthe x species, the number expected to occur one or more times in z quadrats = xp(z). We have calculated these expected distributions based on the total number of spe

cies recorded at each of the three site/time combinations (Attu Island, 1983; San Nicolas


Island 1980, 1985), and plotted them next to the observed number of species that actually were recorded one or more times in z quadrats for the respective data sets (Fig. 6.6). Departures between the observed and expected distributions are substantial, and similar in each case. Most species were rarer (occurred in fewer quadrats) than expected, and a few were much more common (occurred in more quadrats) than expected (Fig. 6.6).

At Attu Island the observed and expected distributions were almost mutually exclusive. The most common species was encrusting coralline algae (principally Clathromorphum spp.), which was uniformly abundant within quadrats, within and among sites. All other species, except rock, sand, and the kelp, Agarum cribrosum, at the Caso Pt. site, were too rare to determine whether or not there was significant variation among quadrats within sites, or among sites. Agarum was abundant in all quadrats at the Casco Pt. site, but was rare or absent at all of the other sites. It thus provided.the single clear example of among site variation from samples of the turf community at Attu Island.

The observed and expected distributions at San Nicolas Island, although different from one another, overlapped to a greater extent than they did at Attu. The most common species, Calliarthron spp. (an articulated coralline alga), encrusting coralline algae, rock and sand, were widely distributed, but highly variable in abundances among quadrats. Certain other species (algae in particular), that were less widely distributed among quadrats, also were highly variable in their abundance among the quadrats in which they occurred. The fucoid, Cystoseira osmundacea, the kelp, Laminaria spp., and the holothurian Pachythyone rubra are such examples. Abundances of several species were more consistent within sites but varied among sites; the best examples are Pachythyone rubra, the bryozoans Cdlaria sp. and Crisia sp., and an unidentified pink encrusting bryozoan. Most species were too rare to determine whether or not they varied significantly within quadrats, among quadrats, or among sites.

6.4.2.2 Swath Counts

Algal densities at Attu Island were fairly consistent within sites, but highly variable among sites (Appendix 6.3). For this reason, among-site variation was Significant (Table 6.1) for each of the three species of large fleshy brown algae (Alaria fistuiosa, Agarum cribrosum, and Desmarestia sp.) that occurred in the swaths. Two additional species of brown algae common to the western Aleutians, Laminaria spp. and Thaiassiophyllum clathrus, did not occur in any of the swaths at Attu. Large fleshy algae were rare in the swaths at two sites on Attu (x = 0.1 individual m-2 ). The Caso Pt. site, in contrast, had an almost unbroken canopy of Agarum cribrosum (x = 12.6 individuals m-2). The Pisa Pt. (exposed) site had a low to intermediate algal cover comprised of Alaria fistuiosa, Agarum cribrosum, and Desmarestia sp. (x = 2.0 invididual m-2).

Sea urchin densities ranged from 14 individuals m-2 at Casco Pt. to 195 individuals m- l at Pisa Pt. (exposed). Among-site variation in sea urchin densities was highly significant during two sampling periods (ANOVA: 1976 - F2 •27 = 175.1, P ~ 0; 1983 -F3,36 = 12.l99,P = 0.00005).

These data show substantial variation in community structure that cannot be explained by the presence or absence of otters. To a large extent, this is because of the

128

6

4

2

0

16

14

f/l 12 Q)

u c: Q) ... ... 10 :3 U U 0

8 ..... 0 ... Q)

6 .&J E :3

Z 4

2

0

12

10

8

6

4

2

0

A

B

c

o Observed

T Expected

Number of Quadrats/Species

Questions of Scale

Fig. 6.6. Observed vs. expected frequency distributions of the number of quadrats/species in 1 m point-contact samples from a Attu Island, 1983; b San Nicolas Island, 1980; c San Nicolas Island, 1985


Table 6.1. Analyses of variance of algal and invertebrate densities at Attu and San Nicolas islands

Location/data Species MS MS F P (among (within sites) sites)

Attu/1983 Alaria [istulosa 470.3 37.1 12.66 ~o

Agarum cribrosum 36687.6 1569.0 23.38 ~O

Desmarestia sp. 17.2 0.6 29.83 ~O

San Nicolas/ Cystoseira/Halidrys 6974 266 26.22 ~O

1980 Eisenia arborea 4.1 1.6 2.61 0.05 Laminaria sp. 827 213 3.87 0.Q1 Macrocystis pyrifera < 1 m 65.8 36.7 1.79 0.15 M. pyrifera > 1 m 32.4 15.7 2.05 0.11 Pterygophora californica 111 109 1.01 0.43 Small unidentified kelps 27.8 22.0 1.26 0.3 Strongylocentrotus franciscanis 21665 1672 12.95 ~O

S. purpuratus 4842 4809 1.01 0.43

San Nicolas/ Cystoseira/Halidrys 24687 4608 5.36 0.002 1985 Eisenia arhorea 13.5 4.4 3.05 0.029

Laminaria sp. 540 57.6 9.37 ~O

Macrocystis pyrifera < 1 m 150 161 0.93 0.52 M. pyrifera > 1 m 13.3 17.3 0.77 0.58 Pterygophora californica 103 20.9 4.94 ~O

Small unidentified kelps 16.4 10.8 1.52 0.22 Strongylocentrotus franciscanis 15822 2699 5.86 ~O

S. purpuratus 29022 7730 3.75 0.Q1

high algal abundance at one site (Casco Pt.) outside the otters' range, and the persistence of extensive urchin barrens (due to high densities of small individuals) at both sites [Pisa Pt. (protected) and Pisa Pt. (exposed)] within the otters' range. However, by 1986 sea otters had dispersed into the Casco Pt. and Massacre Bay sites. This was followed by development of dense algal stands at the Massacre Bay site (J .A. Estes, unpubl.).

With the exception of the small rockweed, Cystoseira/Halidrys, brown algal densities at San Nicolas Island were much lower than at Attu (Appendix 6.2). Many of the species listed in Appendix 6.2 were locally abundant, but none were consistently abundant at all six sites. Densities of about half of the seven categories of large, fleshy brown algae were significantly different among sites (Table 6.1). The data indicate substantial among-site variation in all cases, but in some instances this variation apparently was rendered insignificant because of high within-site variation. Sea urchin densities also varied greatly both within and among sites at San Nicolas Island. Among-site variation was significant for red sea urchins in 1980 and 1985, and for purple sea urchin densities in 1985.

6.4.2.3 Regional Variation

There are a number of evident differences in kelp forest communities between San Nicolas and Attu islands. Perhaps the most obvious of these is taxonomic. San Nicolas


Island is biotically more diverse than Attu, both in terms of species richness and variation among habitats. There are differences in the size and distribution of species assemblage mosaics, or patches, between Attu and San Nicolas islands. Attu is characterized by low variation within sites and high variation among sites, whereas San Nicolas is characterized by high variation both within and among sites. High within-site variation is the probable reason that the abundances of more species are not significantly different among sites at San Nicolas Island (Table 6.1).

There are other differences between these two regions. For example, Alaria fistulosa, the surface canopy kelp at Attu, appears to be an annual species. During summer it forms dense and extensive canopies that are entirely absent from about mid-September to mid-May. The surface canopy at San Nicolas Island is formed by Macrocystis pyrifera, a perennial species, individuals of which may survive for 4 or more years (Rosenthal et al. 1974; Dayton et al. 1984). The Macrocystis canopy at San Nicolas Island fluctuates substantially, but can be extensive at any time of the year.

Benthic and epibenthic assemblages also differ between the two regions. The epibenthic plant assemblage in the western Aleutians is dominated by fleshy brown algae. This algal assemblage is typically very dense, often forming an unbroken canopy of one or several species about a meter above the sea floor (e.g., see data in Dayton 1975 and Estes et al. 1978, from Amchitka Island). Although the structures of these plant assemblages may vary among sites, they are remarkably uniform over hundreds of square meters. Epibenthic brown algae at San Nicolas Island form a less conspicuous component of the community. Although species such asPterygophora californica, Eisenia arborea, and Laminaria spp., can be locally abundant, they usually occur in small patches.

San Nicolas Island supports dense populations of red and purple sea urchins at many locations, but sea urchin barrens occur only on the extensive rocky platform along its exposed western and northern shores. Even here, the urchin barrens are interspersed among large Macrocystis beds. No obvious correlates have yet been found to this mosaic pattern; physically, hydrographically, and biologically the urchin barrens and kelpdominated areas appear similar. Shifts between urchin barrens and kelp-dominated states occur in some areas, for unknown reasons (Harrold and Reed 1985). Overall, about one-third to one-half of the rocky habitat at San Nicolas Island is sea urchin barrens (Harrold, unpublished data).

Sea urchin barrens probably occur all around Attu Island, although we have not estimated the proportion of rocky habitat they occupy. Kelp beds at Attu occur mainly in two kinds of habitat; (1) shallow, exposed locations, such as at the sublittoral fringe or on the tops of pinnacles, where water motion may limit urchin movement to such an extent that destructive grazing seldom occurs (Estes, unpublished data), or (2) areas subject to large salinity changes near river outflows.

6.4.3 Variation in Time

Our discussion of temporal change pertains only to San Nicolas Island since we did not establish permanently marked plots at Attu. We looked for patterns of temporal change at two levels: (1) within quadrats or swaths, and (2) in averages within the island-wide

Variation in Time 131

100 Pachythyone rubra - Individual Quadrats

sampl ing error- x +2 sd

~ 50 > o u .c: Q)

<.I .. Q)

a..

EDH-R25

WDH-L20 P .. · .. -----EDH-Ll5

", ..... a w.~~~_.L:.:~~b~~..:;,dI~ .. WDH- R 15

205:~....--......'. tl~J. ~~t .~

F S F SF SF SF S 80 81 82 83 84 85

Date

Fig. 6.7. Temporal variation in percent cover of Pachythyone rubra at San Nicolas Island, CA., 1980-1985. Hatched area and T·bar are average sampling error + 2 s.d.

sample. We illustrate these patterns with data from five species: the holothurian Pachythyone rubra, the articulated coralline alga Calliarthron spp. (percent cover), the kelps Laminaria spp. and Macrocystis pyrifera, and the red sea urchin Strongylocentrotus franciscanus (densities).

The island-wide average abundance for Pachythyone was greatest during fall, 1980. It declined thereafter, until by spring 1984 it was rare or absent. The species then began an increase until spring 1985 (Fig. 6.7). Pachythyone abundance in each of the four quadrats changed in a similar manner to the island-wide average, although these varied in specific detail. For example, Pachythyone cover in quadrat WDH-R15, although greatest of the four in fall 1980, was absent by spring 1983, and remained absent thereafter.

The island-wide average abundance for Calliarthron spp. underwent an initial small increase from fall 1980 to fall 1981; was fairly constant thereafter until spring 1984; then declined slightly (Fig. 6.8). In this instance, temporal changes were inconsistent among the individual quadrats. For example, Calliarthron cover in quadrat WEK-RlO remained unchanged (within the range of sampling variation) throughout the study period. Other quadrats fluctuated more erratically, and in all instances these changes were substantially greater than expected from sampling error.

The island-wide abundance of Laminaria gradually increased and then declined over the study period (Fig. 6.9). However, temporal trends in densities varied remarkably


100 Cal/iarthron spp. -Individual Quadrats

WEK-RIO

GI

~ 50 u

/ .............. .

...A, \ .. , ... \ ... \ ·ilWEU - L05 \

c: Q

GI , U

, , \ NS-R35 Q ,

\ GI a..

2:p_~ : F SF SF SF S F S

80 81 82 83 84 85 Date

Fig. 6.8. Temporal variation in percent cover of Calliarthron spp. at San Nicolas Island, CA., 1980-1985

120 Laminaria sp-Individual Swaths

100

80

60 N E o 40 C\I ... ~ 20

~ A,,,.ge 2:r-:-:-1 I F S F S F S F

80 81 82 83 Date

I S F

84

I S 85

Fig. 6.9. Temporal variation in density of Laminaria sp. at San Nicolas Island, CA., 1980-1985

Variation in Time

60

50

40

30 N

E 20 0 N ... 10 CII c. ... QI

.t:I 0

E ::J Z

20

10

0

Mocrocystis pyrifero (> 1m) Individual Swaths

DB-32.6L

~ __ ~~ __ ~ __ L-~ __ -L __ ~ __ ~~ __ _

Average

Date

133

WEK-39.1 R WDH-9.6R

Fig. 6.10. Temporal variation in density of Macrocystis pyrifera > 1 m high at San Nicolas Island, CA., 1980-1985

300 Mocrocystis pyrifera «I m) -WEU-45.1 L Individual Swaths

250

200

N 150 E 2 100 ... -i=0.5 ~ 50 ..... , DB-32.6L ... CII

.t:I E ::J

Z

':~E A="'. ~ ok;--+r--'-r -....!~~

F SF SF SF SF S 80 81 82 83 84 85

Date

Fig. 6.11. Temporal variation in density of Macrocystis pyrifera < 1 m high at San Nicolas Island, CA., 1980-1985


within individual swaths, exceeding those expected solely from sampling error in all instances. In one instance (swath WDH45.1L), Laminaria densities ranged between zero and more than 100 individuals 20 m-2 over the study period. Temporal changes in the other swaths were less remarkable, but inconsistent.

The average density of Macrocystis > 1 m in height did not change substantially over the sample period, ranging between about 5 and 12 individuals 20 m-2 (Fig. 6.10). In this instance, individual swaths were consistent with one another, and with the islandwide average. Large increases occurred in several swaths during fall 1983 and spring 1984, although densities in these plots rapidly returned to "normal". Average recruitment was low but consistent from fall 1980 through spring 1983, followed by a strong pulse by fall 1983, and then a complete absence of recruitment from fall 1984 through spring 1985 (Fig. 6.11). Similar temporal patterns occurred in all of the swaths, except that the magnitude of the fall 1983 recruitment pulse ranged among swaths between about 60 and 300 individuals 20 m-2 •

The average density of red sea urchins was rather constant over the study period, ranging between about 48 and 62 individuals 20 m-2 (Fig. 6.12). However, densities in the individual swaths fluctuated more than would be expected from our measure of sampling error. In one instance (EDH-9.6L) changes in density were large but erratic. In others (e.g., WEK-39.lR and WDH-9.6R), there appear to be longer-term trends to the temporal changes.

These patterns suggest the following:

1. Island-wide averages for most species lacked abrupt inflections. Temporal trends occurred in some species whereas others appeared constant.

300 Strongylocentrotus franciscanus-Individual Swaths

250

200

C\I 150 E 0 C\J 100 ... Q) 0. ... 50 Q)

.0 E 0 ::J

Z

':: F':"':=:::::: FS F SF SF SF S 80 81 82 83 84 85

Date

Fig. 6.12. Temporal variation in density of Strongylocentrotus franciscanus at San Nicolas Island, CA., 1980-1985


2. In contrast with island-wide averages, extreme temporal variation and abrupt inflections occurred at the level of individual habitat patches (e.g., quadrats and swaths). This was true of most species.

3. Few generalizations can be made among species concerning the patterns of variation over time. For some species, temporal trends were qualitatively different among quadrats or swaths. Swath counts of Laminaria and S. franciscanus, and quadrat samples of Calliarthron, exemplify this situation. Data on island-wide trends, in such instances, bear little or no relation to the individual samples from which they were derived. Other species varied more consistently among quadrats or swaths over time, for example Pachythyone and Macrocystis recruits.

4. Even species that underwent similar patterns of temporal variation differed considerably in abundance among quadrats or swaths. For example, although there were consistent peaks in Macrocystis recruitment that appeared in fall 1983, the magnitude of these peaks varied by more than six fold among the four swaths examined.

5. In no instance did seasonality (Le., spring vs. late summer/early fall) contribute noticeably to temporal variation in this system. Longer-term patterns of change were evident in some instances, but none appeared to be cyclic, at least over the 6-year study period.

These analyses leave us with a view of the system which could be described as a paradox of scales. At the level of the entire Island (mean of quadrats or swath counts), the system appears stable through time. At the level of specific quadrats or swaths, high and inconsistent variation is typical. This condition seems analogous to the physical properties of a lake or ocean, which from a distance gives the impression of stability. On closer examination, however, such large bodies of water are very dynamiC, with internal and surface waves, complex currents, and at the smallest scale, random molecular motion.

6.4.3.1 Temporal Variation in Kelp Canopies

During the time of our study, the surface canopy cover at San Nicolas Island has varied by nearly an order of magnitude; from about 200 ha in late winter/spring of 1983 to almost 2000 ha in late winter/spring of 1985 (Fig. 6.13). One of the most interesting

Fig. 6.13. Temporal variation in the areal extent of Macrocystis pyrifera surface canopy at San Nicolas Island, CA., 1981-1985

2000

ffi 1000 ~ (.)

~ o z '" (.)

500

1981 1982 1983 1984 1985

DATE


aspects of this record is that it bears no relation to seasonal events, despite the fact that we tried to gather the data during winter/spring minimums and late summer maximums, as estimated from patterns in central California where seasonal fluctuations are obvious (Reed and Foster 1984; G.R. VanBlaricom, unpublished data). In contrast, seasonal fluctuations in surface canopy are extreme in the western Aleutian Islands where dense and extensive canopies of Alaria fistulosa form during spring and summer, but disappear completely during fall and winter.

Some of the most extreme and best-known examples of changes in surface kelp canopies are those associated with the recent EI Nino Southern Oscillation (ENSO) event, which caused world-wide anomalies in climate and oceanographic conditions (Barber and Chavez 1983, 1986; Philander 1983). The west coast of North America experienced unusually severe winter storms during this time, and due to a weakened southward flow of the California current, there was a shift from the normal pattern of coastal upwelling during spring/summer to conditions that were anomalously warm and nutrient poor. The winter storms dislodged an inordinately large number of kelp plants. Settlement and growth of kelp during the following spring and summer was depressed in some areas by warm, nutrient-poor waters. In southern California, the end result was a spectacular reduction in kelp canopies (Dayton and Tegner 1984b). Further south, along the Pacific coast of Baja California, where presumably the coastal waters were even warmer and more nutrient-depleted, Macrocystis beds disappeared entirely from some areas (J.A. Estes, R.K. Cowen, and R. VanWagenen, unpubl.). In central California, winter storm disturbances to the Macrocystis canopy were followed by an unusually dense and wide-spread surface canopy of Nereocystis leutkeana, a known competitive subordinate to Macrocystis pyrifera (Vadas 1968; Dayton et al. 1984; VanBlaricom 1984 and unpubl.).

6.4.3.2 Variation Between A [gal Assemblages and Sea Urchin Barrens

Shifts between algal-dominated assemblages and sea urchin barrens occur in areas of the North Pacific Ocean lacking sea otters. Perhaps the most well known of these was the widespread destruction of southern California kelp beds by grazing fronts of sea urchins (Leighton et al. 1966; Dayton et al. 1984). Precise reasons for the change are unknown, although a number of factors were most likely involved, including human exploitation of predatory fishes and lobsters; human exploitation of competitors, particularly abalones; increased coastal pollution; and climatic change. In other, more localized situations, the cause of such changes are better understood. For example, Cowen et al. (1982) attributed a brief reduction in sea urchin density and grazing at a site north of Santa Cruz, California, to disturbance from storm-generated waves. Pearse and Hines (1979) demonstrated that a shift from urchin barrens to a kelp forest, also at a site near Santa Cruz, was caused by the disease-induced mortality of sea urchins. Duggins (1981a) documented similar changes at Surge Bay in southeastern Alaska, in this instance because of a shift in the urcr-Jns' diet from algae to salps and diatoms, following episodic increases in these alternate food resources. Harrold and Reed (1985) observed the shift from urchin barrens to an algal-dominated assemblage at one of their study sites on the west end of San Nicolas Island. In this case, the shift was associated


with an increase in detrital input, thus causing the urchins to become more stationary. There were no changes in sea urchin population density, and ultimate reasons for the phenomenon were unknown. A final example comes from the long-term studies of Ebeling and his associates at Naples Reef, near Santa Barbara, California (Ebeling et al. 1985; Ebeling and Laur, this Vol.). In this instance, the area was dominated by an algal assemblage at the beginning of the study. Severe winter storms in January 1980 removed most of the Macrocystis canopy. In response to the resultant decrease in algal detritus, the sea urchins began to move more extensively in search of food, destructively grazing the remaining plants, and ultimately shifting the system to a sea urchin barrens by late 1980. The urchin barrens persisted for several years. During 1983 there were again unusually severe winter storms. However, this time it was the sea urchins that were directly affected, probably because of their vulnerability to physical disturbances due to the more exposed habitats they had come to occupy in the course of their movements in search of food. Sea urchin densities were reduced, grazing intensity declined, and the system shifted again to an algal-dominated assemblage. In this remarkable example, the same kind of disturbance produced opposite effects, depending on the starting point.

These accounts demonstrate that in the absence of sea otters, rocky sublittoral habitats in the North Pacific Ocean can shift between sea urchin barrens and algaldominated assemblages, and that they do so for a variety of reasons. The discoveries, in all instances, were fortuitous. Due to the localized nature of the studies, it is not possible to determine how broadly occurring or synchronous the shifts were.

6.4.3.3 Long-Term Changes

Our discussion to this point has emphaSized temporal variation over relatively brief periods; in most cases, 5 years or less. This is the nature of all such direct evidence for obvious practical reasons. Yet it is surely true that events of important and lasting consequence occur over much longer time periods.

Even short-duration climatic trends often span decades or centuries (Hastings and Turner 1965), carrying with them physical changes beneficial to some species and harmful to others. In some instances, these changes may lead to events of profound significance. For example, Cowen (1985) has argued that the density and population structure of sheephead (Semicossyphus pulcher), an important predator on benthic invertebrates in warm-temperate kelp communities, is affected by large-scale patterns of larval transport by ocean currents, which change with climate and oceanographic conditions. This species' ecological importance at some sites may not be determined so much by local conditions as by larger-scale and longer-term phenomena.

Important events occur on even longer time scales. The late Cenozoic was a period of great climatic change. During much of the Tertiary, the northernmost reaches of the Pacific basin were subtropical (Hopkins 1967) and probably unsuitable for the extensive development of kelp. A polar cooling trend began in the late Miocene, and subsequently the North Pacific Ocean was to become an arena for the evolution of many temperate marine organisms, including sea otters (Repenning 1976a), sea cows (Domning 1978), and probably kelps (Estes and Steinberg, unpubl.). The Pleistocene was a period of

l38 Questions of Scale

particularly great change, with advances and recessions of vast ice sheets. During the advances, sea ice permanently covered extensive regions of the eastern North Pacific Ocean that are entirely ice-free at present. Through scouring and limited light penetration, many forms, including the shallow-water-inhabiting kelps, probably became locally extinct. During the recessions, there were probably major northward expansions.

These changes probably had important influences on the organisms and communities we see today, although little is known of this matter. Imagine, for example, how one brief event, the extinction of Stellar's sea cow (Hydrodamalis gigas), may have affected our present view of kelp forests in the North Pacific Ocean. These animals reportedly fed on surface canopy or shallow sublittoral kelps (Domning 1978). The dense kelp assemblages that now occur, and their influence on such processes as light penetration (Neushul 1971; Reed and Foster 1984), water movement (Jackson and Winant 1983), and detritus production (Gerard 1976; Harrold and Reed 1985; Duggins, this Vol.), may have been substantially different before the sea cows were gone. Sadly, we will never even marvel at the sight of these wondrous creatures, much less know how their extinction affected kelp forest ecosystems. Apparently sea cows disappeared because of human exploitation (Doming 1978). The last known population survived until the mid-1700's at the Commander Islands, reportedly in large numbers. This probably was the only location within the range of the species that was never inhabited by indigenous people, and these last animals were quickly exterminated by the early fur hunters.

The history of the sea otter was similar in some ways to that ofthe sea cow (Riedman and Estes, this Vol.). Its earliest known ancestors (Enhydritherium) occurred broadly over the northern hemisphere in the late Miocene/early Pliocene (Berta and Morgan 1985). By the early Pleistocene Enhydra, the modern sea otter, had appeared in the North Pacific, and probably was abundant in kelp forest communities throughout the area. There is evidence that indigenous people hunted and locally exterminated the species (Simenstad et al. 1978), although this effect was surely not so extreme as it appears to have been for sea cows. Nevertheless, fur hunters nearly finished the job by the late 1800's. The kelps probably radiated into an environment where herbivory was unimportant, due largely to the predatory influences of sea otters during the late Cenozoic (Estes and Steinberg, unpubl.). In any case, changes that occurred within kelp forest ecosystems following the near extinction of sea otters were most likely widely occurring and of substantial ecological importance. How differently we might view North Pacific kelp communities had sea otters gone the way of the sea cow.

6.5 Directions for Future Research

From the preceding examples, it is evident that kelp forest communities vary substantially in space and time, and that they do so on a wide range of scales. Insufficient information is available to determine the relative importance of the three components of variance in Eq. (1); clearly, however, variation in space and time must figure strongly both in the interpretation of past results and in the design of future studies.

We foresee several approaches that might be used to resolve the contentious question of how generally sea otters affect plant-herbivore interactions. Perhaps the most straight-

Directions for Future Research 139

forward of these would be to describe extensively systems with and without sea otters. This might be done, for example, by selecting points at random along the shore in both areas, and then running transects seaward from the sublittoral fringe to the rocksand interface, or to the maximum depth at which kelps normally are able to grow -about 25 m depth seems a reasonable limit. Distances along the transects that were kelp-dominated assemblages or sea urchin barrens could be quantified, and proportions of the two habitat types in each area calculated from these measurements. An adequate sample size might be determined in the following way: First by selecting some arbitrary confidence interval; by convention, say ± 5%. Then by picking an arbitrary but small sample size, say 10 or 15 transects. Then, with order chosen at random, by calculating the cumulative proportion of sea urchin barrens from the samples as transects are added. If the standard error around the asymptote converged to within the 5% confidence limit before the sample was exhausted, then the preliminary sample size could be considered adequate. If not, additional random samples could be gathered until this limit was achieved. Under the null hypothesis that sea otters have no measurable influence on plant-herbivore interactions, there would be no a priori reason to expect the constrasted habitats to differ in the extent to which they were sea urchin barrens. The comparison would provide a legitimate test of this hypothesis, at the same providing an unbiased quantification of the generality of the interaction. This approach could be employed at any location where there were nearby areas with and without sea otters. Results would be most convincing if the contrast were replicated in several geographical areas.

Although useful and legitimate, this approach suffers from lack of control for spatial variation. Based on earlier discussions herein, confounding influences could be substantial. Better control of spatial variation could be achieved by documenting changes in community structure over large areas (e.g., islands or large expanses of coastline) following the translocation or natural reestablishment of sea otters, or by removing sea otters from areas where they presently occur. Control for temporal variation could be achieved by gathering similar data from nearby areas at which there were no changes in sea otter populations over the same time period.

Ideally both approaches should be applied at several locations spanning the sea otter's natural range, thus providing a view of geographic variation in the extent to which sea otters affect kelp forest communities. Due to the extensive small-scale variation in space and time that typifies kelp forest ecosystems, measurements using the second approach should be made from a large number of permanently marked study plots at each study location. This would permit separation of spatial and temporal variation from variation caused by sea otter predation. Although expensive and time-consuming, substantial benefits would be realized. For one, the proposed research should provide a definitive evaluation of the sea otters' influence on plant-herbivore interactions over a range of spatial scales, from within small patches, to among patches within kelp beds, to among kelp beds within regions, to among different regions. The second approach would further demonstrate how that influence occurs over time, as otter populations grow from low initial densities to eventual eqUilibria. Also, measures of change in other components of the system would indicate the breadth of these interactions.

There are presently several locations in the North Pacific Ocean where experiments of this nature could be done (Fig. 6.4). One is the Channel Islands of southern California,


using the translocation of sea otters proposed by the U.S. Fish and Wildlife Service. Studies funded by the Fish Wildlife Service, the National Park Service and the Channel Islands Research Project have laid the necessary groundwork in that area. In several other areas, necessary information could be obtained simply by making the appropriate measures at the correct places and times. These areas include the State of Washington, Vancouver Island, and southeast Alaska, where translocation efforts during the late 1960's and early 1970's resulted in now established and growing sea otter populations (Jameson et al.1982; MacAskie 1984). There are other locations in western Alaska that provide similar opportunities. One is the Kodiak archipelago, where sea otters are well established in the northeast, but there is still unoccupied habitat in the southwest. Another is the Near Islands, westernmost of the Aleutians, where we and our colleagues have an ongoing study. Properly done, these simple studies should be of applied and theoretical Significance. It is neither practical nor necessary to study all the potential sites. We point out, however, that once sea otter populations recolonize these habitats, unique opportunities for learning will be forever lost, unless a decision is made to remove otters from large areas.

Although the essential elements for generalizing the importance of sea otter preda· tion in kelp forest communities are contained in the recommended approaches, we wonder at what point the sampling effort will be considered adequate. We challenge our critics to help define that criterion.

Despite the fact that important questions remain unanswered, we have little doubt that sea otter predation is an interaction of great ecological importance. Based on this premise, one might also expect natural selection to alter the species involved over evolutionary time. Such alteration might occur, for example, in the behavior, morpho· logy, and life history of sea otter prey, and in the nature of plant/herbivore interactions. Questions concerning these potential effects have barely been considered. Perhaps one reason is that it is difficult to put specific hypotheses to definitive tests, by neontological standards. However, we believe there are exciting discoveries to be made from this line of study. We are reminded that modern sea otters never occurred outside the North Pacific Ocean, and that the Enhydritherium/Enhydra lineage was limited to the northern hemisphere (Riedman and Estes, this Vol.). Thus, as a general approach we suggest comparative studies among temperate systems of the world. If sea otter predation had significant evolutionary consequences, we might expect predictable contrasts reflecting these effects among such regions.

6.6 Summary

In this paper we consider a scenario for the interactions between sea otters, sea urchins, and kelp beds, which holds that sea otters maintain kelp beds by limiting herbivorous sea urchins, and conversely, that sea urchin barrens develop where sea otters are absent. Specifically, we ask (but do not answer) the question: how general is this interaction? We approach the problem through a components of variance model, in which community structure is affected by three factors: space, time, and sea otter predation.

Summary 141

Evidence for the otter/urchin/kelp scenario is of three kinds: (1) comparisons of areas with and without sea otters; (2) historical accounts of areas before and after the establishment of sea otters; and (3) experimental studies in which the effects of sea otter predation are mimicked in small habitat patches. Each of these approaches suffers certain drawbacks. Despite these problems, the collective evidence suggests that the scenario occurs widely in the eastern North Pacific Ocean.

Evidence for spatial variation in benthic community structure was obtained from our ongoing sampling programs at San Nicolas Island,in the Channel Islands of southern California, and Attu Island, western-most of the Aleutian Islands. Considerable variation in the turf community was due simply to the co-occurrence of a rich species assemblage. In general, densities of the larger invertebrates and plants were more variable among swaths within sites at San Nicolas than at Attu. There was substantial variation among sites for all species that were abundant enough to measure at both San Nicolas and Attu; however, for some species at San Nicolas Island, among-site variation was not statistically significant due to high within-site variation.

Temporal variation in community structure was measured from permanently marked quadrats and swaths at San Nicolas Island over a 6-year period. Island-wide population trends occurred in some species, whereas others appeared constant. Species abundances in individual quadrats or swaths varied greatly in most cases. Temporal trends were qualitatively similar among plots for some species; for other species there was no correlation in temporal variation among plots. There did not appear to be a strong seasonal component to temporal variation in any of the species examined.

Shifts between algal and urchin dominated states have been reported from some areas lacking sea otters. A review of several published examples indicates that such shifts proceed in both directions and for a number of reasons (e .g., physical disturbances from storms, changes in food availability, and disease). Due to the fortuitous nature of all studies, it is not possible to determine how widely occurring such shifts have been. Longer-term temporal changes also occur in kelp forest ecosystems due to such factors as variation in climate, and the extinction and introduction of ecologically important species.

We conclude that kelp forest communities vary considerably in space and in time, and they do so on a broad range of scales. For this reason, care must be given in the interpretation of past studies, and in the design offuture work, especially with reference to estimating the generality of the otter/urchin/kelp scenario. We propose two means whereby the importance of this scenario might be better estimated. The first is simply by thoroughly and representatively comparing areas with and without sea otters. The second is by documenting changes over time following the reestablishment and growth of sea otter populations either through translocation or natural recolonization, or by removing them experimentally. Because of the extensive spatial and temporal variation that characterizes kelp forest communities, a large number of plots will be required in both approaches, and permanently marked plots would be desirable in the latter approach.

App

endi

x 6.

1a. A

ttu

Qua

drat

sam

ples

198

3. E

ntri

es a

re t

he n

umbe

r o

f co

ntac

ts in

20

rand

om p

oint

s. C

olum

n to

tals

exc

eed

20 b

ecau

se o

f ve

rtic

al l

ayer

ing

.... .j>

. tv

Stu

dy s

ite

Spec

ies

Pisa

Pt.

Pi

sa P

t.

Cas

co P

t.

Mur

der

Pt.

P

rote

cted

E

xpos

ed

Qua

drat

No.

1-5

1

-15

2

-25

2

-15

1

-5

1-1

5

2-2

5

2-1

5

1-5

1

-15

2

-25

2

-15

1

-5

1-1

5

2-2

5

2-1

5

Aga

rum

cri

bros

um

1 1

15

20

10

16

Ala

ria

/ist

ulos

a C

lath

rom

orph

um s

pp.

3 6

18

8 17

20

16

12

12

20

19

16

12

3

12

14

Con

stan

tine

a sp

. 1

1 C

oral

lina

rubb

le

1 2

Cor

allin

a sp

p.

Cry

pton

emia

sp.

2

1 1

Des

mar

esti

a sp

. 4

Enc

rust

ing

cora

llin

es

3 1

1 2

2 E

ncru

stin

g re

d al

ga

1 1

Lam

inar

ia l

ongi

pes

Lit

hoth

amni

on s

pp.

1 M

usse

l she

ll 8

1 1

Phy

codr

ys s

p.

2 2

Pot

amil

la s

p.

Pur

ple

Bry

ozoa

n 3

Rho

dogl

ossu

m s

p.

1 R

hodo

men

ia s

p.

2 2

2 R

ock

1 1

2 5

2 S

and

12

13

12

2 15

5

4 S

tipi

tate

Lam

inar

ia s

pp.

to

I::

Stro

ngyl

ocen

trot

us p

olya

cant

hus

1 1

3 2

2 1

1 2

1! T

hala

ssio

phyl

lum

cla

thru

s o·

Uni

dent

ifie

d br

yozo

an

1 S! 0

Uni

dent

ifie

d fi

lam

ento

us r

ed a

lga

1 1

....,

Vl

Uni

dent

ifie

d sp

onge

tl

Uni

dent

ifie

d tu

nica

te

..

App

endi

x 6.

1h.

San

Nic

olas

Isl

and

Qua

drat

sam

ples

198

0. E

ntri

es a

re t

he

num

ber

of

cont

acts

in 2

0 ra

nd

om

poi

nts.

Col

umn

tota

ls e

xcee

d 20

bec

ause

of

ver

tica

l la

yeri

ng

Day

tona

Bea

ch

Spe

cies

Eas

t D

utch

Har

bor

Wes

t D

utch

H

arbo

r

Stu

dy

sit

e

Wes

t en

d U

rchi

n W

est

end

Kel

p N

orth

sid

e

Qua

drat

L05

R

15

L25

R35

R

05

LI5

R

15

R25

R

15

L20

L

25 R

35

L05

L

IO

R15

L

25

RIO

R20

L25

L

40

R05

R25

L

25 R

35

Alg

ae

Ani

socl

adel

la p

a.

2 2

4 B

ossi

ella

sp.

5

2 1

4 4

3 B

otry

ocla

dia

pseu

do-

dic

ho

tom

a

1 C

alli

arth

ron

tube

r-cu

losu

m

8 6

6 1

7 3

7 7

15

3 2

3 8

Cal

loph

ylli

s fl

abel

lula

ta

-1

Cor

alli

na s

pp.

2 6

2 2

2 E

ncru

stin

g co

rall

ine

alga

e 2

1 4

3 3

4 3

1 3

14

19

6 11

14

20

8

13

7 6

10

11

C

rypt

onem

ia o

b.

1 C

ysto

seir

a os

mun

dace

ae

-9

3 1

Dic

tyo

ta s

pp.

4 1

Ei.

enia

arb

orea

K

ally

men

ia p

acif

ica

Kel

p ho

ldfa

st

7 6

You

ng k

elp

Lam

inar

ia s

p.

1 11

11

19

9

3 M

acro

cyst

is p

yrif

era

<1

m

2 N

ienb

urgi

a an

ders

onia

na -

Plo

cam

ion

cart

ilag

ineu

m

1 2

2 4

Pol

ysip

honi

a sp

. P

rion

itis

lanc

eola

ta

Vl ~ E

'<

..- ~

w

App

endi

.x 6

.1b.

(co

ntin

ued)

.....

. .j

>.

.j>

.

Stu

dy

sit

e

Day

tona

Bea

ch

Eas

t D

utch

Har

bor

Wes

t D

utch

W

est

end

Urc

hin

Wes

t en

d K

elp

Nor

th s

ide

Har

bor

Spec

ies

Qua

drat

LOS

R15

L

25 R

35

R05

L

IS

R15

R25

R

15

L20

L

25 R

35

LOS

LI0

R

15

L25

R

IO R

20

L25

L

40

R05

R25

L

25 R

35

Alg

ae

Pte

rygo

phor

a ca

lifo

rnic

a -

Pte

rysi

phon

ia s

p.

2 2

Enc

rust

ing

red

alga

e 3

1 3

4 R

ed a

lgal

tu

rf

2 2

1 1

2 R

hody

men

ia c

alif

orni

ca

2 3

4 1

2 4

3 R

ock

2 7

3 3

3 7

2 2

2 3

1 1

4 2

San

d 2

1 6

12

9 4

4 12

3

6 2

7 2

Spo

nges

A

carn

us e

rlth

acus

2

Hal

icho

ndri

a sp

. H

ymen

amph

iast

ra

cyan

ocry

pta

1 Sp

heci

ospo

ngia

con

-fo

eder

ate

2 O

rang

e en

crus

ting

sp

onge

P

ink

encr

usti

ng s

pong

e W

hite

enc

rust

ing

spon

ge

-2

,0

c:

Tet

hya

awra

ntia

" ~

Cni

dari

ans

o· ~ A

biet

inar

ia s

pp.

4 2

0

Agi

aphe

nia

sp.

2 2

1 8

....,

til

Bal

anop

hyll

ia e

lega

ns

1 3

2 <>

Bar

nacl

e ~

Cor

ynac

tis

cali

forn

ica

1

Epi

acti

s pr

olif

era

Stye

la m

on

tere

yen

sis

2 Te

alia

spp

. 2

Bry

ozoa

ns

C/J

c

Pin

k en

crus

ting

bry

ozo

an -

1 1

a C

ella

ria

sp.

5 2

2 2

a e; C

risi

a sp

. 1

2 '<

Did

emn

um

car

nule

ntum

-

Dip

losm

a m

acdo

nald

i H

ippo

dipl

osia

ins

culp

ta

-M

embr

anip

ora

fl/s

ca

2 5

1

Mol

lusc

s Se

rpul

orbi

s sq

uam

iger

us

2 2

2 3

2

Pol

ycha

etes

D

iopa

tra

orna

ta

4 4

5 2

Dod

ecac

eria

few

kesi

2

2 P

hrag

mat

opom

a ca

lifo

rnic

a Sp

irob

ranc

hia

Ech

ino

der

ms

Cuc

umar

ia l

ubri

ca

3 3

2 5

7 C

. sa

lma

1 E

upen

tact

a qu

in-

qu

esem

ita

1

Lis

soth

ruia

nu

trie

nts

1

1 P

ach

yth

yon

e ru

bra

7 5

3 6

4 8

10

4 2

3 S

tro

ng

ylce

ntr

otu

s fr

anci

scan

us

1

Tun

icat

es

Ap

lid

ium

spp

. A

rch

idis

tom

a m

oll

i A

rch.

ru

be

Bug

ula

spp.

C

ysto

dyt

es l

obat

us

1 M

etan

droc

arpa

dur

a 2

Syn

oci

um

- """ '"

App

endi

x 6.

1c.

San

Nic

olas

Isl

and

Qua

drat

sam

ples

198

5. E

ntri

es a

re t

he

num

ber

of

cont

acts

in 2

0 ra

ndom

poi

nts.

Col

umn

tota

ls e

xcee

d 20

bec

ause

of

ver-

I-'

.j>

.

tica

l lay

erin

g 0

\

Stu

dy s

ite

Day

tona

Bea

ch

Eas

t D

utch

Har

bor

Wes

t D

utch

W

est

end

Urc

hin

Wes

t en

d K

elp

Nor

th s

ide

Har

bor

Spec

ies

Qua

drat

L05

R15

L

25 R

35

R05

L

l5

R15

R25

R

15

L20

L25

R35

L

05

LlO

R

15

L25

R

IO R

20 L

25

L40

R

05 R

25

L25

R35

Alg

ae

Enc

rust

ing

red

alga

e 4

1 1

5 2

5 B

ossi

ella

sp.

1

1 3

4 1

1 C

alli

arth

ron

spp.

5

4 1

14

1 1

2 2

1 8

2 5

17

8 2

7 7

Cor

allin

a sp

p.

5 1

2 3

1 1

8 E

ncru

stin

g co

rall

ine

alga

e 4

2 2

1 1

2 1

1 7

1 10

3

11

10

17

12

12

3 19

9

13

Cys

tose

ira

osrn

unda

ceae

2

16

10

1 2

4 2

14

Dic

tyo

ta s

pp.

3 1

6 E

isen

ia a

rbor

ea

6 K

ally

rnen

ia p

acif

ica

Kel

p ho

ldfa

st

4 1

2 1

5 3

3 La

rnin

aria

sp.

7

1 20

9

3 Y

oung

Lar

nina

ria

Mac

rocy

stis

pyr

ifer

a <

1m

5

3 3

Nie

nbur

gia

ande

rson

iana

-1

Plo

carn

ion

cart

ilag

ineu

rn -

I:)

Pol

ysip

honi

a sp

. c (I

I

Pri

onit

is l

ance

olat

a 1

'" .-to

Pte

rygo

phor

a ca

lifo

rnic

a 4

3 o· ::s '"

Pte

rysi

phon

ia s

p.

0 .... R

hody

rnen

iII

cali

forn

ica

2 4

2 1

1 v.>

Roc

k 1

~ Sa

nd

2 2

18

6 5

2 4

8 7

1 1

1 1

12

4 3

Spo

nges

V

l

Aca

rnus

erl

thac

us

'" 3 H

ymen

amph

iast

ra

3 cy

anoc

rypt

a e; '<

:

Ora

nge

encr

usti

ng

spon

ge

3 3

1 P

ink

encr

usti

ng s

pong

e 1

1 T

eth

ya a

wra

ntia

Cni

dari

ans

Abi

etin

aria

spp

. 1

Agl

aphe

nia

sp.

5 3

4 9

Bal

anop

hyll

ia e

lega

ns

2 6

1 1

Bar

nacl

e 1

Cor

ynac

tis

cali

forn

ica

Epi

acti

s pr

olif

era

Teal

ia s

pp.

Bry

ozoa

ns

Pin

k en

crus

ting

bry

ozo

an -

2 1

4 2

2 1

2 1

Cel

lari

a sp

. 3

3 2

3 1

3 C

risi

a sp

. 2

1 H

ippo

dipl

osia

ins

culp

ta

-1

1 M

embr

anip

ora

fusc

a 1

Mol

lusc

s Se

rpul

orbi

s sq

uam

iger

us

2 2

3 2

3 2

1 2

3 2

4 3

Pol

ycha

etes

D

iopa

tra

orna

ta

8 2

7 2

2 D

odec

acer

ia f

ewke

si

1 P

hrag

mat

opom

a ca

lifo

rnic

a 2

4 6

11

2 Sp

irob

ranc

hia

4

Ech

ino

der

ms

Cuc

umar

ia l

ubri

ca

C.

pipe

rata

1

-.j:>. ...,

App

endi

x 6.

1c.

(con

tinu

ed)

Day

tona

Bea

ch

Spe

cies

Eas

t D

utch

Har

bor

Wes

t D

utch

H

arbo

r

Stu

dy

sit

e

Wes

t en

d U

rchi

n W

est

end

Kel

p N

orth

sid

e

Qua

drat

L05

R

15

L25

R35

R

05

LI5

R

15 R

25

R15

L

20

L25

R35

L

05

LIO

R

15

L25

R

IO R

20 L

25

L40

R

05 R

25

L25

R35

Ech

inod

erm

s E

upen

tact

a qu

in-

ques

emit

a L

isso

thur

ia n

utri

ents

P

achy

thyo

ne r

ubra

St

rong

ylce

ntro

tus

fran

cisc

anus

Tun

icat

es

Arc

hidi

stom

a m

olli

A

rch.

nub

e B

ugul

a sp

p.

Cys

tody

tes

loba

tus

Met

andr

ocar

pa d

ura

Syno

cium

4 2

3 3

2 6

1 3 2

1 1

1

1 3

- "" 00 ,0

~ ~ o· ~

o ..., en ~

App

endi

x 6.

2. S

an N

icol

as I

slan

d sw

ath

coun

ts 1

980

and

1985

. E

ntri

es a

re t

he

nu

mb

er o

f in

divi

dual

s 20

m-2

en

~ a

Stu

dy s

ite

a e; '-<

Day

tona

E

ast

Dut

ch

Wes

t D

utch

W

est

end

Urc

hin

Wes

t en

d K

elp

Nor

th s

ide

Sw

ath

1980

Spe

cies

1

2 2

4 5

4 5

2 3

2

Cys

tose

ira/

Hal

idry

s 3

5 2

2 3

120

77

l3

9 E

isen

ia a

rbor

ea

2 1

2 1

Lam

inar

ia s

p.

2 5

2

40

1 30

7

Mac

rocy

stis

pyr

ifer

a <

1 m

4

25

M.

pyri

fera

> 1

m

3 5

15

3 7

3 11

7

2 2

Pte

rygo

phor

a ca

lifo

rnic

a 1

2 4

3 10

9

3 3

Stro

ngyl

ocen

trot

us fr

anci

scan

us

10

2 11

41

14

1 37

17

2 2

90

24

16

13

S.

pur

pura

tus

243

l30

95

14

24

4

0

175

75

31

11

18

119

You

ng K

elps

2

1

1985

Cys

tose

ira/

Hal

idry

s 14

59

15

16

18

8 1

20

l3

0

162

66

50

E

isen

ia a

rbor

ea

6 1

2 3

3 3

Lam

inar

ia s

p.

24

27

2 2

2 2

20

7 20

l3

M

acro

cyst

is p

yrif

era

< 1

m

1 M

. py

rife

ra >

1 m

7

5 1

10

4 5

2 3

9 5

6 P

tery

goph

ora

cali

forn

ica

1 12

28

2

2 St

rong

yloc

entr

otus

fran

cisc

anus

14

29

43

1

79

29

95

22

0 9

3

38

5 S.

pur

pura

tus

226

221

112

24

81

l37

19

8 1

46

2

76

76

63

20

Y

oung

Kel

ps

1 1

15

.... oj:>

'-

D


Appendix 6.3. Attu Island swath counts (Kelps only) 1983. Entries are the number of individuals per 20 m-2

Agarum cribrosum Alaria fistulosa Desmarestia sp. Stipitate Laminaria spp. Thalassiophyllum clathrus

Pisa Pt. Protected

1

1

2

2 1

Pisa Pt. Exposed

1

18 7 5

Site

Casco Pt. Murder Pt.

Swath

2 2 1 2

10 204 300 37

3

7 Effects of Sea Otter Foraging on Subtidal Reef Communities off Central California D. R. LAUR, A. W. EBELING, and D. A. COON

7.1 Introduction

As a major predator of sea urchins, the sea otter (Enhydra lutris) plays a determining role in structuring many subtidal kelp-forest communities in the North Pacific (reviewed in Estes and Harrold, Riedman and Estes, this Vol.). In the absence of otters, patches of kelp and other erect algae survive only on shallow pinnacles, crests, and other highrelief zones of turbulence (Estes et al. 1978; VanBlaricom 1984), which are unstable platforms for sea urchins (Dayton 1985b; Laur et al. 1986). Thus, after otters return to reef habitats previously overgrazed by exposed urchins (Strongylocentrotus spp.), large stands of kelp reappear as the remaining urchins become restricted to cracks and crevices (Estes and Palmisano 1974; Dayton 1975; Estes et al. 1978; Duggins 1980; Breen et al. 1982). As otters remove the destructive grazers, furthermore, enhanced productivity creates an environment favorable for greater numbers of finfish (Simenstad et al. 1978; Estes et al. 1981,1982; VanBlaricom 1984; Estes and VanBlaricom 1985).

The indirect relationship between sea otters and kelp beds in central California may require further clarification, however (see Estes and Harrold; Foster and Schiel, this Vol.). For instance, North (1965) attributed enhancement of canopies of the giant kelp Macrocystis pyrifera to the otters' presence, but Miller and Geibel (1973) pointed out that kelp also regrew during the same period off southern California outside the otters' range. Cowen et al. (1982) suggested that winter-storm disturbances may be more important than otter-grazer interactions in structuring nearshore kelp communities. On the other hand, VanBlaricom (1984) compiled historical evidence that the reduction of urchin numbers when otters enter overgrazed habitats initiates a predictable succession of kelps: a sequence beginning with stands of the annual canopy kelp Nereocystis luetkeana and various understory kelps eventually culminates in a forest of perennial Macrocystis pyrifera, which is apparently the more favorable habitat for fish production (Bodkin 1986).

Yet the initial direct effects of sea otter foraging are indisputable. Soon after moving into new areas, otters drastically reduce exposed populations of their preferred prey - sea urchins, abalone, and rock crabs (e.g., Ostfeld 1982), leaving the cast shells as evidence of their activity (Hines and Pearse 1982).

The main purpose of the present study was to measure both the direct and indirect effects of sea otters as they exploited new reef environments in central California. Until 1976 when their populations may have begun to stabilize, peripheral groups of otters were moving southward into previously unforaged habitats (e.g., Estes et al. 1981). Thus, we took advantage of a natural manipulation by observing two different reef

152 Effects of Sea Otter Foraging on Reef Communities

communities just before and after otter foraging. We compared the effects of otters immigrating to a small high-relief reef, which had not been badly overgrazed by urchins but may have never supported a persistent surface canopy of kelp, with effects on a large, mostly low-relief reef, which had been grazed nearly bare of erect algae except for a remnant of kelp canopy surviving on reef crest (Le., was mostly an urchin-dominated barren ground, sensu Lawrence 1975). Hence, the central Californican sites provided two important constrasts: (1) between pre-otter and post-otter communities, and (2) between responses of communities inhabiting different reef types (sensu Foster and Schiel, this Vol.).

A secondary objective was to simulate effects of sea otters, on a reduced scale, by excluding urchins from small plots in a barren ground on a reef off southern California, beyond the otters' range. This was done to provide some experimental verification of the descriptive results.

Events beyond our control severely compromised our sampling schedule off central California. Due to delays in obtaining research support, adequate baseline observations of the pre-otter condition could not be completed before otters arrived at our sites. Thus, only brief surveys, including movies, still photos, and macroinvertebrate counts, could be made in October 1976 before winter's rough weather made further scuba operations impossible. Improving weather conditions finally allowed a resumption of work in July 1977, about 5 months after otters were first seen in the area during February (S. Benech, Benech Biological & Assoc., Ventura, Ca., pers. commun.). In addition, much of the post-otter monitoring suffered from design by hindsight: We had concentrated most of our time and effort at the high-relief reef before we realized that the major response was occurring at the larger, deeper, and flatter site. To meet deadlines and budgetary limits, therefore, we often sampled reduced quadrat sizes as quickly as possible with minimum replication.

Nonetheless, our study was successful to the extent that most post-otter changes were large enough to be detectable by these methods. Results were consistent with the previous predictions of direct and indirect effects: (1) foraging otters virtually eliminated all exposed sea urchins and abalone; (2) a surface canopy of annual Nereocystis kelp reappeared at the deeper, low-relief reef; (3) subsurface canopies of algal turf and understory kelp returned; (4) the high-relief reef, subject to stronger water motion, was inherently less vulnerable to urchin grazing and did not sustain a surface canopy of kelp; and (5) young fish recruited to kelp canopies, where greater numbers of adult fish were observed as well.

7.2 Study Sites

In 1976, at the outset, the southern periphery of the sea otter's range was between Pt. Buchon and Avila, San Luis Obispo County in central California (Estes et al. 1981; S. Benech, Benech Biological & Assoc., Ventura, Ca., pers. commun.). Hence, we explored several possible areas downcoast. Two remote sites, Lone Black Reef and Santa Rosa Reef, were chosen because they were located directly in the path of the otters' southward advance and appeared to be suitable for supporting diverse communities

Study Sites

Fig. 7.1. Location of study reefs (upper arrow) where the effects of sea otters on reef communities were observed in central California north of Pt. Conception and where (lower arrow) the effects were simulated by macroinvertebrates exclusion cages in an urchin-dominated barren ground in southern California near Santa Barbara

Pt. Arguello

'========I 20 km

153

(Fig. 7.1). The experiment to simulate otter effects in southern California was conducted at Naples Reef, Santa Barbara County, a site that had been previously monitored for several years (see Ebeling and Laur, this Vol.).

Lone Black Reef (LBR) was selected first because it is circumscribed and easily surveyable. Located 475 m offshore and 2.0 km west of the Avila Breakwater, LBR is a rectangular monolith of rock measuring about 20 by 60 m, with its long axis extending east and west. Its reef crest, which is 4.25 - 6.0 m deep and shoals to 3 m at one end, is influenced by heavy wave surge. It bore stands of algal turf and patches of understory kelp, but no surface canopy. The reef drops off steeply to its 10-12 m deep base, which is surrounded by patches of coarse sand, cobble, boulders, and high-relief rock.

Santa Rosa Reef (SRR) is located 1500 m farther offshore from LBR. Measuring about 100 by 500 m,it is an irregular expanse of flat rock, sand channels, and boulders, with its long axis extending east and west. Most of its high-relief crest and slope is 7.5 to 9.0 m deep, shoaling to less than 5 m at one end. The crest retained patches of the understory kelp Laminaria dentigera, together with bushy stands of Gelidium robustum and other leafy red algae. The reef apparently had had no surface canopy for at least 10 years before the sea otters came (R. Burge, Dep. of Fish and Game, pers. commun.). Its rocky flats are crossed by sand channels, average 14.5 m in depth, and are about


20 times greater in area than the combined crest and slope. Initially, sea urchins dominated the flats, which, having been grazed nearly bare of erect algae, were paved with crustose coralline species.

Naples Reef (NR) is located 1.6 km offshore in the Santa Barbara Channel west of Santa Barbara (Fig. 7.1). It is an irregular offshore mound of shale outcrops and ledges measuring 75 by 300 m and descending from 5 m depth at reef crest to 15 m at its base (Ebeling et a1. 1980b). In 1979 a much smaller area of urchin-dominated barren ground was limited to the reef's west end. The extensive surface canopy of Macrocystis was destroyed by a severe storm in 1980, after which exposed urchins consumed the remaining large stands of understory kelp, Pterygophora california (Ebeling et a1. 1985; Ebeling and Laur, this VoL).

7.3 Methods

7.3.1 Sampling Schedule

The central Californian sites (LBR, SRR) were sampled during a total of 10 dive-days by four scuba observers in yearly sets from 1976 through 1980. The pre-otter baseline was limited to an initial reconnaissance, as winter's rough weather prevented further diving before otters arrived at the sites sooner than expected. Hence, pre-otter sampling in 1976 was limited to a total of only 2 days spent either counting macroinvertebrates along permanent transects installed previously at LBR, or completing a photographic survey of SRR. In addition, almost all effort budgeted for site surveying and transect construction was expended at LBR, which was to have been the only site monitored. Consequently, post-otter sampling (macroinvertebrates, algae, benthic cover) was performed, as originally designed, at LBR only, covering 4 days in July- August 1977 and 2 days in July 1978. By then it was obvious that LBR could show few indirect effects of otter foraging, and that our remaining limited resources could be better spent elsewhere. Therefore, post-otter monitoring of SRR was begun much later on an ad hoc basis, as effort was necessarily confined to rapid photographic surveys and selective counts during single days in December 1979 and October 1980.

At the southern Californian site (NR), urchin exclusion experiments and macroinvertebrate densities were monitored quarterly for 3 years. Observations were begun in 1979, when treatments were installed in a barrens at the reefs west end, and were continued through 1981, after all remaining kelp was lost following a severe storm in 1980 (see Ebeling et al.1985).

7.3.2 Macroinvertebrate Sampling

For all three sites, yeady samples of sea urchin, sea star, and abalone densities were compiled from counts usually made firsthand in band transects. At LBR, exposed individuals were counted in six 1 X 37 m permanent transects (sampling units), each positioned along the crest, slope, or base on either side of the reef. Animals in holes

Algae and Sessile Invertebrates 155

and crevices were counted separately. At SRR, however, band-transect or photoquadrat sampling units were positioned about midreef at points determined from a randomnumber table as compass direction and number of swimming kicks. In 1979 and 1980, counts were made directly from 1 X 13 m band transects extended from each of four points. In 1976, however, they were made later from 36 photoquadrats, each taken by a Nikonos camera with a 28-mm lens held at the length of a slender rod above a different point, such that about 1 m2 of surface was covered. To measure the temporal consistency of macroinvertebrate densisties in the persistent urchin barrens at NR, 27-31 counts per year were made in a single 1 X 13 m band located near the experimental plots.

7.3.3 Algae and Sessile Invertebrates

7.3.3.1 Central Californian Sites

Benthic cover was sampled yearly by photoquadrat. At LBR, 36-54 0.25 m2 quadrats were positioned at random intervals along the transect line on reef slope. Coverage of each quadrat was from six 35 -mm color slides taken by the Nikonos camera and 0 .042 m 2

close-up framer. Sampling by photoquadrat was reliable to the extent that the species dominating the percent cover as estimated from photographs also dominated the algal biomass as measured from four destructive air-lift collections made at reef crest and base during 1977. Photographic sampling of SRR was carried out in a similar way, except that the sampling units were 24-63 individual slides (0.042 m2 ) taken at random pOSitions along the transect lines.

To measure percent cover of different taxa, slides were analyzed collectively in groups of six (LBR) or singly (SRR) by projecting them, one at a time, onto a screen under a grid of 24 crossed lines and counting the number of point intercepts falling on each taxon. Taxa were pooled into four functional categories: fleshy red algae (leafy and filamentous species), crustose coralline algae (pavement-like species), sessile invertebrates (hydroids, tunicates, sponges, worms, and bryozoans), and tubicolous amphipods (species used as an index of food available for fishes). Percent cover of each category was estimated as the category's proportion of the total intercepts (144 per group of six, or 24 per Single-slide sampling unit).

At SRR, we measured understory kelp density on reef flat and reef crest in 1976 and 1980 by counting all stipes, mostly of Laminaria dentigera, in 18-35 1.0 m2 bar quadrats positioned randomly along transects through the respective habitats. Percent cover of kelp blades was estimated roughly by eye before stipes were counted in each quadrat.

7.3.3.2 Urchin Exclusion Experiment

At NR off southern California, treatments either excluding or not excluding sea urchins and other macroinvertebrates were randomly aSSigned to 0.25 m2 plots of flat overgrazed substrate. In addition, macroinvertebrate densities were monitored yearly out-

156

A. Macroinvertebrate exclusion

/ X

Effects of Sea Otter Foraging on Reef Communities

Fig. 7.2. Structure of caged treatments for the urchin exclusion experiment at Naples Reef in southern California. A macro invertebrate exclusion; 8 fish exclusion; C total exclusion

side the plots as explained above. Over the plots, pyramidal cages of 1.2S-cm Vexar plastic mesh supported by 2.5-cm diameter PVC plastic pipe were secured to the bottom by PVC strips bolted to the rock through lips of mesh extending out from the 1.0 m2 base of the pipe frame. The three caged treatments consisted of topless cages for macroinvertebrate exclusion, bottomless cages for fish exclusion, and complete cages for total exclusion (Fig. 7.2). The uncaged controls included two free plots, and two plots with mesh fitted only on the south and east sides as shade controls. Since urchin grazing overwhelmed the other factors and partial caging made little or no difference, all results were classified simply as either exclusion (macroinvertebrate plus total exclusion treatments) or nonexclusion (all other treatments and controls) for the present analysis. Percent cover of the four categories of organisms was compared between exclusion and non~xclusion groups, represented by samples of 23 and 32 photoquadrats taken within the plots and pooled among years (see above).

7.3.4 Fish Sampling

Reef fish assemblages were sampled at LBR and SRR in cinetransects: 2.5-min, super-8, high-speed color movie films taken by divers swimming unidirectionally (Ebeling et al.

Data Analysis 157

1980b). The course for each film (sampling unit) was begun in a direction selected at random, but was occasionally altered to stay within the same deep or shallow habitat. It was covered at a constant rate at a depth about 1-2 m over the bottom or in midwater under the kelp canopy (7 of 19 films at SRR 1980) and was never reversed so as not to record fish that tend to follow. The camera housing was directed forward or slightly downward, and panned as steadily as possible in a 10-degree arc, although occasionally stopped to film all fish Sighted in a school. Coverage averaged some 211 m2

of projected area (see Ebeling et a1. 1980b). Transects per sample ranged from 4-19, as time permitted.

7.3.5 Data Analysis

This opportunistic sampling of very heterogeneous reef systems presented unavoidable problems for proper statistical analysis. In the restricted space of LBR, large sampling units could not be randomly placed. Macroinvertebrate transects, for example, had to be stratified to cover different microhabitats (reef base, crest, etc.) but confined to fields large enough to accommodate transect tracks; any randomization of both position and direction of tracks would have underrepresented microhabitats or extended surveys into open water and across sand. Hence, positions along tracks could be randomized for small-scale sampling of sessile benthos, but not the tracks themselves for large-scale sampling of sea urchins, etc. For repeated sampling, as of fish, that most likely covers the same limited number of tracks, however, the mobile animals will eventually randomize themselves over the sampling units, such that sampling conforms to a "restricted systematic design" (Venrick 1978). Fish may accomplish this to a certain extent between cinetransects.

The varying sizes of sampling units presented further problems in statistical comparison, especially between sites. For example, sampling units for macroinvertebrates were either 37 m2 (LBR) and 13 m2 (SRR, NR) transect bands or 1 m2 photoquadrats (SRR). For comparison, therefore, all individual observations had to be standardized to number m-2 , such that those from transects are rates and those from quadrats are discrete counts. Cross comparisons of percent cover by sessile benthos is also subject to the bias of different unit sizes: 0.25 m2 (LBR) vs. 0.042 m2 (SRR). It is doubtful that the two units sample the same scale of patchiness. Nonetheless, the more important comparisons - between years within sites - are more reliable because sampling units were usually of the same size within sites.

To facilitate comparisons despite these problems of bias and scale, samples were simply contrasted in a gross way by graphing all statistics in the same format: means and their 95% confidence intervals based on transformed values converted back to their geometric or percentage equivalents. [Because their statistical distributions were strongly skewed, most counts (x) were transformed to log(x + 1) for calculation of parametric statistics; analogously, values of percent cover of sessile organisms were arcsine transformed (Sokal and Rohlf 1981)]. These are more realistic measures of central tendency than arithmetic means because they approximate medians (with asymmetric intervals), the better descriptors of skewed distributions (Sokal and Rohlf 1981). Furthermore, they may be contrasted informally by rule of thumb. If the inter-


val of one mean covers the value of another, the two means probably do not differ significantly by t-test; if the two intervals themselves do not overlap or barely overlap, the difference is highly significant (Simpson et al. 1960).

7.4 Results and Discussion

7.4.1 Macroinvertebrates

In April 1977, evidence of sea otter foraging - scattered broken red abalone (Haliotis rufescens) shells and red sea urchin (Strongylocentrotusfranciscanus) tests - was found about the base of LBR (pers. observ.). In addition, a raft of 58-97 otters had been sighted near LBR and SRR in February 1977, and six animals were seen foraging over LBR in June (S. Benech, Benech Biological & Assoc., Ventura, Ca., pers. commun.).

Subsequent results indicated that the otters continued harvesting abalones and urchins from LBR during the entire 15-month study period. Total exposed red abalones counted in all six transects declined from 26 (0.12 m-2) in 1976 before otters arrived, to 4 (0.02) in 1977 3 months after, and zero a year later in 1978, when the only abalone seen were 11 individuals located deep in cracks and crevices. This compares with a drop of from 0.75 to 0.01 m-2 after otters first arrived at Pt. Estero, about 40 km to the north (Estes and VanBlaricom 1985). Post-otter densities of exposed red sea urchins also decreased to nil (0.55 to 0 m-2), and purple sea urchins (Strongylocentrotus purpuratus) were always rare (Fig. 7.3). Like abalones, the surviving red urchins found refuges, with numbers in crevices increasing from zero in 1976 to 27 in 1977 and 37 in 1978 (as compared with the total of 123 exposed individuals counted in 1976). Similarly in the 1960's after otters returned to reefs near Monterey, broken shells and tests littered the bottom, as abalones and urchins were eventually restricted to crevices (Lowry and Pearse 1973; Cooper et al. 1977; Hines and Pearse 1982).

Apparently, however, the otters had not depleted available supplies of abalones and urchins to the point of diversifying their diet to include sea stars (see Riedman and Estes, this Vol.). Annual fluctuations in densities of Patiria miniata and Pisaster spp. (mostly Pisaster ochraceus and P. giganteus) were unrelated to otter presence at LBR or SRR (Fig. 7.3). Presumably, otters take sea stars only after the preferred urchins, abalones, and crabs are no longer available (Ebert 1968; Estes et al. 1981 ;Ostfeld 1982).

Even before otter foraging, LBR supported a much lower density of urchins than the larger sites (Fig. 7.3). This may be typical of small high-relief reefs where strong wave action (Lissner 1980) and steep vertical walls with patches of the stinging anemone Corynactis califomica restrict urchin positioning and movement (Foster and Schiel 1985; Laur et al. 1986). In contrast, pre-otter SRR had high densities of exposed red urchins (5.2 m-2), but within a year after otters arrived at SRR, densities had dropped to zero as urchins virtually disappeared along with the abalones.

At the southern Californian site NR, densities of all exposed macroinvertebrates remained high throughout the study in an environment without otters, kelp, or drift algae. Even large abalones remained scattered about. Many individuals appeared to be wasting from starvation, in that their muscular foot had regressed noticeably in size and

Algae and Sessile Invertebrates

Fig. 7.3. Yearly densities of macroinvertebrates at the three study sites. Species are, from left to right: S.p. purple sea urchin S. purpuratus; Sf red sea urchin Strongylocentro· tus franciscanus; P.m. bat star Patiria miniata; Pisaster Pisaster sea stars pooled. Bars measure geometric mean numbers m-2 with 95% confidence intervals: Lone Black Reef, clear for 1976 before sea otters arrived (N = 6 transects), stippled for 1977 (N = 6), and hatched for 1978 (N = 6); Santa Rosa Reef, clear for 1976 before otters (N = 361.0 m2 photoquadrats), stippled for 1979 (N = 4 transects), and hatched for 1980 (N= 4); Naples Reef, clear for 1979 (N = 31 transects), stippled for 1980 (N = 27), and hatched for 1981 (N = 29)

4

3

2

1

0

N 12 E '-

C/)

(jj .0 E ::l Z

8

4

0

16

12

8

4

159

Invertebrates

Lone Black Reef

Santa Rosa Reef

Naples Reef

D Before otters

II] 5 months after

~ 17 months after

D Before otters

CJ 21 months after

~ 33 months after

D 1979

01980

~ 1981

OL-~~4-~~~~~~~~~--_

s. p. S. f. P. m. Pis aster

they could be easily dislodged from rocks by hand. Although both urchins and abalone prefer to feed on drift kelp when available, abalones are both morphologically and physiologically less able to exploit alternative limited food sources when drift is absent (Tegner and Levin 1982).

7.4.2 Algae and Sessile Invertebrates

7.4.2.1 Central Californian Sites

The flora of LBR was typical of a small exposed rugged reef without a permanent kelp canopy (see Foster and Schiel, this Vol.). Although scattered individuals of Nereocystis luetkeana occurred in the area throughout the study, no obvious surface canopy formed on or about the reef during the post-otter period. Thus, the erect algal cover consisted mostly of low bushy forms and turf, as pavements of crustose coralline algae remained inconspicuous under the cover of erect forms (Fig. 7.4). Dominate species of fleshy red algae included Gelidium robustum, Gigartina exasperata, Callophyllis pin nata, Rhodymenia pacifica, and Botryoglossum farlowianum, all occurring together with articulated coralline algae such as Calliarthron cheilosporoides. Species common at reef crest - Gelidium, Calliarthron, Gigartina, Callophyllis, and the understory kelp Lam inaria dentigera - flourish in zones of strong water motion and high light levels (Burge and Schultz 1973; Foster and Schiel 1985).

160

50

40

30

20

10

Benthic cover

Lone Black Reef

rn 5 months after otters

~ 17 months after

O~~~~----~~~L-~~~

80

~ 60 o u C 40 Q) u Q; Il.

50

40

30

20

10

Santa Rosa Reef

Naples Reef

D Caged

fm Uncaged

O~~~~~~L---~L-~~~

Fleshy Coralline Sessile Tubicolus red- crust inverts. amphipods algae


Fig. 7.4. Yearly percent cover of four categories of benthic organisms at the three study sites. Bars measure back-converted mean percentages with 95% confidence intervals: Lone Black Reef, clear for 1977 (N = 54 six-slide photoquadrats) and stippled for 1978 N = 36); Santa Rosa Reef, clear for 1976 before sea otters (N = 24 single-slide photoquadrats), stippled for 1979 (N = 34), and hatched for 1980 (N = 63); Naples Reef, clear for the nonexclusion that allowed urchins to enter (N = 32 six-slide photoquadrats, pooled among years) and stippled for the urchin-exclusion plots (N = 192)

At LBR, benthic cover showed little response because invertebrates had not overgrazed the pre-otter reef. Most components remained essentially unchanged (Fig. 7.4). A slight increase in fleshy red algae and other cover may have occurred independently of any effect of sea otters, as species filled space vacated by barnacles. An initial dense cover of barnacles (Balanus crenatus) on parts of the reef in 1977 had declined to nil by 1978; M. Foster (pers. commun.) observed similar recruitment pulses in other areas both with and without otters along the central Californian coast. In addition, J. Estes and G. VanBlaricom (pers. commun.) " ... observed an extraordinary settlement of B. crenatus at Pt. Piedras Blancas (about 80 Ian N of our sites) in fall 1977 ."

In contrast, SRR showed a dramatic post-otter increase in erect-algal cover (Fig. 7.5). A carpet of fleshy red algae and sessile invertebrates, harboring greater abundances of tubicolous amphipods, overgrew the nearly continuous pavement of exposed crustose corallines (Fig. 7.4). A thick canopy of Nereocystis measuring some 80 X 100 m had reached the surface within 2 years. Stands of understory kelp (Laminaria dentigera), previously restricted to shallow areas of high relief and strong water motion, had spread over all deeper areas of reef flat within three years (Fig. 7.6).

Fig. 7.sA-C. Photographs showing expansion of erect benthic cover on Santa Rosa Reef flats after sea otters arrived in winter, 1977 . A exposed pavement of crustose coralline algae before otters; B fleshy red algae and amphipod tube mats overgrowing the pavement after otters; C stipes of the understory kelp Laminaria dentigera after otters. A and B include 0.042 m-2 of area

Algae and Sessile Invertebrates 161

A

B

c

Fig. 7.5A~

162

40

'" 30 I

E I/) 20 'E <'CI

0:: 10

0

100

~ 80 o o 'E 60 Q)

~ 40 0..

20

Understory kelp

D Before otters m 33 months after

Deep Shallow

OL..----Deep Shallow

7.4.2.2 Urchin Exclusion Experiment


Fig. 7.6. Density of understory kelp (Laminaria dentigera) on Santa Rosa Reef before (1976, N= 351.0 m2 quadrats) and after (1980, N = 18) sea otters arrived in 1977. Percent cover is the back-converted mean with 95% confidence interval

This post-otter effect was simulated experimentally on a smaller scale at NR. All three categories of erect cover increased markedly in urchin-exclusion plots as the exposed crust was overgrown (Fig. 7.3). Even small kelp plant (Macrocystis pyriJera and Pterygophora caliJornica) appeared in the exclusion cages and were eventually grazed by fish in the topless treatments. Continually grazed by urchins, the remaining plots retained a much larger exposed cover of coralline pavement.

7.4.3 Fish

The large drop in total fish abundance at LBR reflected the virtual disappearance of young rockfish (Sebastes mystinus and S. serrano ides) from the post-otter reef between 1977 and 1978 (Fig. 7.7). Such crashes in rockfish recruitment are not unusual along California and may result from climatic shifts and variations in larval abundance, as well as from loss of plant cover in which the young hide from predators (reviewed in Ebeling and Laur, this Vol.). The young fish had aggregated about the low foliage on the pre-otter reef. A noticeable increase in numbers of young surfperch coincided with a vigorous growth of fleshy red algae, especially Gelidium robustum, on reef crest. Striped surf perch (Embiotoca lateralis) foraged about the crest, where they can pick amphipods and other smaIl prey from the algal surfaces (Laur and Ebeling 1983; Schmitt and Holbrook 1984; Ebeling and Laur 1986).

At SRR, a large maximum in fish abundance was recorded between 1977-1979 on the post-otter reef (Fig. 7.7). Although this may have been a reproductive response to a favorable climatic shift of which we were unaware, it is tempting to attribute the initial rise to an otter effect for two reasons. First, a canopy assemblage of fishes (sensu Ebeling et al. 1980a) had begun to form: increasing numbers of senorita (Oxyjulis cali-

General Discussion and Conclusions

r

,..-

b 150 t-O) en

Fish

Lone Black Reef

m 5 months after otters

~ 17 months after

c rn .... +"

100 t- ,.,I:!:

£ 5:Lt-_~~~Z"~~:-~~L-~li~l~a:~ __ ~~T~~~~~ __ ~T~~~ __ ~~ __

Santa Rosa Reef

,..- D Before otters I 300 m u 21 months after 0)

~ en 200 33 months after c rn .... +"

.J:: 100 en .. u..

0 .:. ::: :!:

O.j. S.m.- S.m. S.s.- Fl. O.V. y.o.y. y.o.y.

300 250 200 150 100

163

50 0L......JE.;i;.:;,;~L-

500

400

300

200

100 0

Total fish

Fig. 7.7. Yearly densities of fishes at the three study sites. Species are, from left to right: OJ seflorita Oxyjulis californica; S.m. yoy blue rockfish Sebastes mystinus young; S.m blue rockfish adults; S.s. yoy yellow rockfish S. serrano ides young; E.I. striped surfperch Embiotoca lateralis; D.v. pile perch Damalichthys vacca. Bars measure geometric mean numbers per cinetransect with 95% confidence intervals: Lone Black Reef, clear for 1977 (N = 20 cinetransects) and stippled for 1978 (N = 12); Santa Rosa Reef, clear for 1976 before sea otters (N = 4), stippled for 1979 (N = 4), and hatched for 1980 (N = 19). Totalfish includes other species as well

[ornica) and blue rockfish (Sebastes mystinus) had moved higher in the water column; kelp surfperch (Brachyistius [renatus), which are obligate canopy dwellers (Bray and Ebeling 1975), were seen for the first time. Second, the increases included young rockfish (Sebastes) and surfperch (Embiotoca, Damalichthys), which use kelp for refuge, as well as adult surfperch, which respond to greater prey availability as kelp litcer refuels the detrital food chain (Ebeling and Laur, this Vol.). The 1979 abundance of Embiotoca lateralis included many young and subadult recruits probably born on or near the reef. Nonetheless, we have no good explanation, other than sampling error, for the observed decrease in 1980. Perhaps in 1980, fish were just generally less visible in algal thickets and patchy aggregations; D. Laur (pers. observ.) noted many fish in the kelp understory and in large schools beyond the range of our cinetransects.

7.5 General Discussion and Conclusions

Sea otter foraging brought about changes in the Santa Rosa Reef (SRR) community similar to changes observed elsewhere off central California. After otters returned to


the Monterey area during the 1940's and 1950's, for example, large gaps in kelp beds were filled in byboth canopy and understory species (Mclean 1962). Surviving urchins were driven to deep crevices and other refuges, where they fed on accumulating drift kelp instead of living plants (Lowry and Pearse 1973). Likewise at SRR, exposed abalones and sea urchins - the otters' preferred food - disappeared, as turf, kelp, and sessile animals filled the space once dominated by urchins and paved with crustose coralline algae. Abalones and urchins became restricted to crevices; algal turf overgrew the corallines; a canopy of Nereocystis luetkeana formed at the water surface; and an understory of Laminaria dentigera spread over the reef bottom. Similar changes occurred on a much smaller scale in our urchin-exclusion plots in a persistent urchindominated barren ground at Naples Reef off southern California, where there are no otters: rich algal turfs and small kelp plants soon covered the coralline pavements wherever urchins were excluded.

The indirect effects of sea otter foraging that we observed at SRR resembled previously described early stages in a lengthy successional process of kelp-forest development. North (1965) suggested that since perennial Macrocystis is competitively dominant over annual Nereocystis, a return of sea otters should herald the ultimate reestablishment of Macrocystis canopies on barren grounds off central California. Using historical data collected from the area, VanBlaricom (1984) composed a qualitative model of the process. The surface-canopy kelp Nereocystis is among the first colonizers because it persists on marginal substrates; during the initial few years, an understory of perennial kelps, including Laminaria, develops; finally, the maturing system is invaded by Macrocystis, which forms another surface canopy, has the potential of rapid vegetative regrowth, and may persist indefinitely. Hence at SRR, the Nereocystis canopy, which formed within 2 years after exposed urchins were eliminated, may be eventually replaced by Macrocystis invading rifts in the Laminaria understory.

Thus, the establishment of a persistent Macrocystis canopy may not occur for a decade or more after the exposed sea urchins are removed (VanBlaricom 1984). Similarly, Dayton and Tegner (1984a) concluded that seral replacement in understory patches requires a minimum of 10 years in the Pt. Loma kelp off southern California, and major changes in Macrocystis cover occur on even larger times scales.

PhYSical disturbance may disrupt the long-term process, however. Although the taller "climax" canopy of perennials dominates the competition for light, it is more vulnerable to wave stress, so that the dominance hierarchy may be reversed in areas exposed to storm action (Cowen et al. 1982; Dayton and Tegner 1984b; Dayton et al. 1984; VanBlaricom 1984). The canopies surviving in areas of greater wave stress often release their gametes into stronger currents and thereby may have the greater dispersal and colonizing potentials (Neushul 1972). Thus, annual Nereocystis and perennial understory canopies are typical of shallower cobbly or more exposed central Californian localities subject to frequent wave disturbance, while Macrocystis dominates deeper, more protected sites with hard, stable substrata (Dayton et al. 1984; Reed and Foster 1984).

In addition to demonstrating marked post-otter changes within the reef community, our study supports Foster and Schiel's (this Vol.) contention that reef "type" should be considered in predicting the indirect effects of sea otters on the community. Whereas the larger, deeper, and more low-relief type at SRR showed a dramatic post-otter

General Discussion and Conclusions 165

changeover, the smaller, rugged type at LBR was less affected. Perhaps the LBR system is controlled mostly by physical disturbance. Thus, LBR may have never supported a permanent kelp canopy or large urchin population.

Perhaps small protruding reefs like LBR present exposed and unstable platforms for conspicuous and sedentary organisms with large hydrodynamic drag because such reefs are relatively isolated and create strong water motion. Carter et al. (1985a) concluded that small modular piles of boulders making up Pendleton Artificial Reef (PAR) off southern California harbored few urchins because the structures generated turbulence and were surrounded by barriers of sand; also, transplanted kelp was either torn off by the surge or eroded away by grazing fishes attracted to the conspicuous foliage. Previously, North (1971) and Dayton and Tegner (1984a) had suggested that small isolated patches of kelp may attract high densities of encrusting organisms and browsing fishes, thereby setting a lower size threshold to the establishment of a stable kelp forest.

A likely reef type for Macrocystis establishment may be indicated by the aspect ratio (depth of surrounding water/model reef height). The optimal aspect ratio for producing a favorable lee wave in a moderate current is 10 (Nakamura 1985), comparable to that of the SRR flats. However, the aspect ratios of the LBR and abovementioned PAR prominences are all less than 3. With steep sides buffeted by a heavy swell, therefore, the latter reefs will tend to maximize hydrodynamic pressures and turbulence to the point where plants and animals with large drag, such as surface kelp and exposed sea urchins, are likely to the torn loose or pulled off the bottom (see Foster and Schiel 1985).

The post-otter regrowth of the SRR kelp forest may have enhanced some fish populations. By the end of the second year (1979), a canopy group of fishes (sensu Ebeling et al. 1980a), including an obligate canopy species (Brachyistius frenatus), had assembled, together with large schools of young rockfish and surfperch, in and about the kelp stands (although, perhaps due mostly to sampling error, the resulting increase in fish abundance was not documented by transects made during the single day's sampling for 1980). As also observed in other central Californian localities (Miller and Geibel 1973; Bodkin et al. 1986; Bodkin 1986), blue rockfish (Sebastes mystinus) were by far the most abundant species in the new SRR kelp forest (Fig. 7.6).

Future standing stocks of fish may be larger if Macrocystis eventually replaces the Nereocystis canopy. Bodkin (1986) estimated that Macrocystis beds near Pt. Piedras Blancas (about 80 km N of our sites) supported more than twice the fish biomass as occurred in equivalent Nereocystis beds, due largely to greater abundances of blue rockfish. This difference was attributed to the perennial Macrocystis forest's greater persistence, foliage biomass, and structurally diverse habitat (VanBlaricom 1984 ; Bodkin 1986).

Nonetheless, changes in fish abundance due to revegetation may be confounded with changes due to spawning or recruitment success associated with large-scale shifts in ocean-current and temperature patterns (Stephens et al. 1984; Cowen 1985; Bodkin et al. 1986; Ebeling and Laur, this Vol.). Adult fishes that tap the detritus-based food chain or young fishes that recruit to the shelter of the algal canopies are more likely to respond to revegetation (Quast 1968c; Miller and Geibel 1973; Burge and Schultz 1973; Estes et al. 1978; Simenstad et al. 1978; Ebeling and Laur, this Vol.). Some are species like surfperches that seek cover in the canopies as young and eat tiny


prey living on bushy algae and in the carpeting turf as adults (Stouder 1983; Ebeling and Laur 1986). For instance, overall abundance of Embiotoca lateralis increased at SRR as young found refuge in the understory kelp (see Ebeling and Laur 1985) and adults may have encountered greater food supplies related to the increasing algal abundance (see Hixon 1980; Laur and Ebeling 1983; Schmitt and Holbrook 1984). The large fluctuations in abundances of young rockfish may have reflected widespread recruitment failures associated with northerly ocean currents or increasing water temperature, as well as changes in cover (reviewed in Ebeling and Laur, this Vol.). Yet to our knowledge, no large shifts in weather pattern, such as EI Nino episodes, occurred during the period of study (see Seymour et al. 1984; Ebeling and Laur, this Vol., Fig. 7.2).

The potential for kelp-related enhancement of reef-fish assemblages may be greater in central than southern California, so long as a complex rocky substratum is provided (review in Ebeling and Laur, this Vol.). During a 3-year cycle ofloss and regeneration of a southern Californian kelp forest at Naples Reef, for example, total fish density varied by only about 20% because losses in some species were offset by gains in others (Ebeling and Laur, this Vol.). In fact, abundances of some species that do not always rely on the kelp forest for food and cover actually increased (see also Quast 1968b). These planktivores (Chromis punctipinnisj and "switch-feeders" (Girella nigricans, Medialuna californiensis, Paralabrax clathratusj were southern species, rare in central Californian waters north of Point Conception. Although rank orders of fish species abundance for LBR and SRR correlated significantly (tau = 0.40, P = 0.05), those for either northern site and Naples did not (0.03-0.15, NS). Hence, this kind of species replacement may not occur in waters to the north (see also Burge and Schultz 1973; Bodkin 1986).

In perspective, Californian kelp forest vary in response to a number of factors -storm disturbance, water temperature, light levels, kelp harvesting, and local pollution - besides overgrazing by herbivores (recent reviews in VanBlaricom 1984; Dayton 1985a; Foster and Schiel 1985, this Vol.; Estes and Harrold, this Vol.; Schiel and Foster 1986). For example, Cowen et al. (1982, p. 200) concluded of a central Californian kelp forest: "Once the urchins are removed, the maintenance of community structures will be a function of physical disturbance and variations in algal recruitment." The obverse is also true: in the midst of grazing urchins, new growth may not survive even in a favorable disturbance regime with sufficient supplies of potential recruits (Ebeling et al. 1985). Thus, "(b)y removing sea urchins as important grazers, ... sea otters eliminate a major source of kelp forest variability and may strengthen the resilience of the kelp community to extrinsic disturbance" (VanBlaricom 1984, p. 23). Doubtless the extent of the otter effect will vary among areas with different disturbance regimes and susceptibilities to overgrazing (Estes and Harrold, this Vol.; Foster and Schiel, this Vol.).

7.5.1 Future Research

As pointed out by other authors in this Volume, a wealth of descriptive and experimental evidence has proven that kelp and other algae regrow where the high densities

Summary 167

of exposed sea urchins are removed from a barren ground. Hence, further experiments in which urchins are excluded from small barren plots are no longer needed to demonstrate this outcome (Foster and Schiel, this Vol.). There is also adequate evidence that otters are most capable of eliminating exposed urchins, and so can bring about the regeneration or preservation of kelp forests on a local scale (Estes and Harrold, this Vol.). Nonetheless, it should be emphasized that the successional development of a kelp forest into its most productive form may be a lengthy process, requiring a decade or more of observation to confirm (Dayton and Tegner 1984a; VanBlaricom 1984).

What remains to be done from our perspective is to predict the vulnerability of a reef to urchin domination and to clarify indirect effects of overgrazing on higher trophic groups such as fishes. Thus, more research is needed on how best to classify reefs by environmental "type," as some types will be more vulnerable to overgrazing than others (Foster and Schiel, this Vol.). What physical settings present barriers to urchin settlement, recruitment, and immigration? (see Laur et al. 1986). Perhaps a reefs aspect ratio (depth of surrounding water/modal reef height), average vertical slope (inclination from the horizontal), and percent unconsolidated substrate (cobble and sand) could be measured for a beginning. Once different sites are identified by these characteristics, urchins could be transplanted at high density to plots and their survival compared with natural controls. But this can be done only after major problems in stabilizing the transplanted animals are solved (see Laur et al. 1986). Ultimately, the comprehensive effect of otters moving into a new area may be predictable from the distribution of reef types.

As recommended by Duggins (this Vol.), the contribution of the kelp detritusbased food chain to fish production should be determined by tracing energy sources and transplanting secondary producers. Experiments manipulating kelp canopies and measuring fish-recruitment responses should be supported as well (M. Carr, Univ. California Santa Barbara, pers. commun.). Determining the role of the canopy's "edge effect" in enhancing fish populations is equally important (Dayton and Tegner 1984a; Dayton 1985). Experiments could be designed to test the idea that kelp forests with open spaces (barren patches) support larger and more diverse fish assemblages because their canopies have more "edge" (M. Carr, Univ. California Santa Barbara, pers. commun.). If this is so, otters eliminating urchins that maintain such patches (see Harrold and Reed 1985) might occasionally cause a modest decline in fish abundance by bringing about the ultimate restoration of a continuous thick canopy.

7.6 Summary

Annual transect and photographic surveys of two subtidal reef communities in San Luis Obispo County, central California were made, often opportunistically, from 1976-1980, before and after sea otters arrived in winter, 1977. Results generally confirmed previous models of the effect of otter foraging on kelp-forest communities overgrazed by sea urchins. As indicated by the litter of broken shells and tests as well as the survey counts, otters directly eliminated most exposed abalones and sea urchins, which were eventually confined to crevices. Indirectly, the resulting relaxation from


urchin grazing pressure brought about a marked change in community structure at the larger, deeper, and flatter reef. Within 2 years, canopies of algal turf, understory kelp (Laminaria dentigeraj, and surface kelp (Nereocystis luetkeanaj overgrew pavements of crustose coralline algae in an urchin-dominated "barren ground." This "otter effect" was simulated on a smaller scale in urchin-exclusion plots on an equivalent barren ground at a reef near Santa Barbara, southern California, where there are no otters. Yet other investigators have pointed out that an effect such as we observed may only include the early stages in a lengthy successional process lasting a decade or more.

The magnitude of the otter effect depended on the physical setting. At the shallower, high-relief reef, vertical walls and strong water motion probably provided refuges from urchin grazing, but created unsuitable substrata for development of a surface kelp canopy. Thus, the indirect consequences of otter foraging were minor because subsurface algal stands had not been overgrazed before otters arrived and the reef may have never supported a persistent surface canopy of kelp.

Observed numerical responses of fish to the new canopy refuge and to a resurgence of the detrital food chain may have been confounded by widespread recruitment fluctuations. Nonetheless, the reef-fish assemblage may be more vulnerable to defoliation in central than in southern California, where a large indigenous array of southern species includes abundant planktivores and switch-feeders not requiring products of an intact kelp forest.

Thus, as emphasized by other authors in this Volume, local conditions must be considered in predicting otter effects. In particular, we recommend that physical properties of reef should be measured as indicators of supportable plant cover and vulnerability to grazing. For example, reef depth and shape affect water motion and, therefore, the stability of platforms for sessile organisms with large hydrodynamic drag like kelp plants and sea urchins. Future work should also include experiments to test the importance of kelp canopies to reef-fish recruitment and survival.

Acknowledgments. We thank J. Estes and G. VanBlaricom for their good services, and, along with M. Hixon and an anonymous referee, helpful manuscript reviews. M. Carr, D. Reed, and M. Foster provided stimulating disucssions. R. Bray, A. DuBois, M. Hixon, R. Larson, M. Love, R. Rowley, and G. Tribble helped with the field work. M. Hixon and D. Reed assisted with data analyses. N. Lammer and S. Anderson gave technical assistance with equipment and boating operations. The Marine Science Institute provided administrative services. This material is based on support by the California Department of Fish and Game under Agreement No. 1193 for the central California study and by the National Science Foundation under Grant No. OCE79·25008 for the urchin exclusion experiment.

8 Fish Populations in Kelp Forests Without Sea Otters: Effects of Severe Storm Damage and Destructive Sea Urchin Grazing A. W. EBELING and D. R. LAUR

8.1 Introduction

As the top predator in a three-level system, the sea otter (En hydra lutris) can prevent local extinction of kelps and other algae at the bottom level by reducing numbers of sea urchlns and other grazers at the second level (Estes and Palmisano 1974; Estes et al. 1978; Duggins 1980; Estes and Harrold, thls Vol.; Laur et aI., thls Vol.). A high algal biomass, in turn, may create an environment favorable for greater numbers offinfish at the thlrd level by sustaining a characteristic detritus-based food web (Simenstad et al. 1978; VanBlaricom 1984; Ebeling et al. 1985; Duggins, thls Vol.) and providing cover for newly born or settled young (Coyer 1979; Miller and Geibel 1973; Ebeling and Laur 1985; M. Carr, Univ. California Santa Barbara, pers. commun.).

Even though sea otters have proven capable of controlling urchln numbers and enhancing kelp-forest communities, these systems are, however, so naturally variable in space and time over different scales that present data are inadequate to demonstrate the generality of thls effect (Estes and Harrold, thls Vol.; Foster and Schiel, this VoL). Along central California, for instance, this otter effect may be eclipsed by regular seasonal events. Both withln and outside the otter's range, the thinning of mature kelps by winter storms together with spring upwelling of nutrients and recruitment of new plants causes a marked seasonal regeneration of kelp forests (Dayton 1975; Foster 1982; Reed and Foster 1984; Foster and Schiel 1985; VanBlaricom 1984, this Vol.). The predictably strong wave action can reduce urchin populations that otherwise might destroy the new plant growth (Cowen et al. 1982).

Off southern California, the growth and development of large kelp communities are less noticeably keyed to seasonal events, whlch are not so prominent there as farther north (Clarke and Neushul 1967; Dayton et al. 1984). Perhaps this, together with the lack of sea otters and enough other sea urchln predators, make regenerating forests much more vulnerable to grazing (Leighton et al. 1966; Leighton 1971; Dean et al. 1984; Ebeling et al. 1985; Harrold and Reed 1985), especially near urban areas in polluted environments during warming trends (North and Pearse 1970; Pearse et al. 1970; Wilson et al. 1980). In addition, abundant species of subtropical fishes (kyphosids and labrids), whlch are rare or absent in more northern forests, browse the algae and occasionally might inhlbit forest regeneration (Rosenthal et al. 1974; North 1976; Bernstein and Jung 1979; Harris et al. 1984; Carter et al. 1985b). Major El Nino climatic episodes occur unpredictably and may be accompanied by severe storms and warm, nutrient-poor water, whlch weaken or destroy the kelp (Dayton and Tegner 1984b;

170 Fish Populations in Kelp Forests Without Sea Otters

Gerard 1984; North 1985; Foster and Schiel, this Vol.). All these kinds of erratic disturbances contribute to the formation of urchin-dominated barren grounds (barrens) lasting for indeterminate periods (Leighton 1971 ; North 1971).

Nonetheless, lush kelp forests persist in many areas off southern California without universal regulation of urchin numbers by sea otters or other agents (North 1963; Tegner 1980; Foster and Schiel, this Vol.). A complex of alternative predators and other factors may control urchin populations in some areas or times but not in others (Tegner 1980; Cowen 1983; Ebeling et al. 1985). Under these circumstances, urchin numbers may increase locally and create a mosaic of forest and barrens (Dean et al. 1984; Harrold and Reed 1985). This and the physical setting of kelp forests produce different "types" of community structure in a "dynamic range" from extensive barrens to continuous cover (Schiel and Foster 1986; Foster and Schiel, this Vol.).

Perhaps partly because of this, there has been considerable debate over the importance of kelp in structuring southern Californian reef-fish assemblages (Quast 1968a,b; Ebeling et al. 1980a,b; Stephens and Zerba 1981; Larson and DeMartini 1984; Stephens et al. 1984; Patton 1985; Patton et al. 1985). Faced with a mosaic of forested and barrens patches, reef fishes may "hedge their bets" with an alternative behavioral repertoire that allows a successful life without kelp.

This paper examines fish abundance and assemblage structure in fluctuating kelp forests without sea otters near Santa Barbara. We monitored fish densities during a period of kelp-bed loss and regeneration: In winter 1980, an unusually severe storm removed kelp, beginning the process of transformation of kelp forest to barrens on an offshore reef (Harris et al. 1984; Ebeling et al. 1985), while sparing an inshore site. A similar storm in 1983 reversed the process at both sites. Thus, we took advantage of a controlled natural experiment (sensu Connell 1980), and sampled fish densities during periods of kelp forest and barrens at both sites while each was in a condition opposite to that of the other. Since the period included an El Niiio warm water episode, we interpreted the results in light of a possible response of the fish assemblage to climatic change (see Stephens et al. 1984; Patton 1985). More generally, we discuss the conceptual difficulty of detecting a general otter effect on reef-fish communities along the continuous mainland coast, but predict some likely local effects in the patchy and fluctuating environment, and suggest topics for futUre research.

8.2 The System

Both sites are located 23 km west of Santa Barbara, southern California (34°25'N, 119°57'W) (Ebeling et al. 1980a,b). The offshore site (Naples Reef) is a semi-isolated mound of shale outcrops and ledges about 1.6 km offshore. Covering about 2.2 ha, the reef surface averages 8-12 m deep, although crests rise to within 5 m of the surface. The reef drops off gradually or sharply on all sides to a flat bottom of sand, cobble, or pavement rock with occasional low rocky ridges at 14-16 m depth. About 6 km to the northeast, the inshore site (Devereux Point) extends from the lower intertidal zone to about 10 m depth. Less monolithic than Naples, the site was arbitrarily limited to about 2 ha of a continuous oil shale bottom with pronounced ridges and crevices. Thus,

Methods 171

the offshore site is semi-isolated rock, while the inshore site is part of a continuous longshore area of reef and kelp.

During the entire 1970's decade, both sites were heavily forested by kelps and supported similar abundant fish assemblages (Ebeling et al. 1980a,b). Beds of giant kelp (Macrocystis pyrifera) extended the substrate vertically, allowing many fish use of the entire water column. For example, many young such as those of various rockfishes (Sebastes) lived in the canopy, while those of other species such as blacksmith (Chromis punctipinnis) and surfperch (Embiotocidae) remained near shelter just above the bottom. Unlike blacksmith, which sheltered in reef crevices (Bray 1978), baby surfperch sought refuge, usually from large predatory kelp bass (Paralabrax clathratus) in dense understory foliage (Ebeling and Laur 1985). Offshore, the understory refuge was mostly among broad blades of Pterygophora califomica, a low kelp; inshore, the understory was more diverse, consisting of various fucoids, kelps, and large leafy red algae. Adult blacksmith often assembled up current near the margin of the kelp canopy to eat oceanic plankton (Bray 1981). Adult surfperch swam about in the open, gleaning small prey from a mat of algal turf and colonial animals (Laur and Ebeling 1983; Ebeling and Laur 1986). Kelp bass and grazing opaleye (Girella nigricans) or halfmoon (Medialuna californiensis) often foraged among the kelp stipes in midwater (Love and Ebeling 1978; Ebeling et al. 1980a).

Such a long episode of continuous forest cover may be unusual for Naples Reef, however. Kelp was reportedly absent in 1963, but returned luxuriantly for unknown reasons between 1967-1970 (Miller and Geibel 1973). Then, in the early 1980's, two unusually severe winter storms catalyzed opposite effects, transforming the offshore forest (Fig. 8.1 A) into a barrens and the barrens back again into a forest (Ebeling et al. 1985). In winter 1980, unusually large waves (storm I) tore away most giant kelp (Macrocystis pyrifera) and much of the understory foliage. This deprived grazing sea urchins (Strongylocentrotus franciscanus and S. purpuratus) of their preferred food, detached drift kelp. In the absence of effective predators such as sea otters, these urchins safely moved out over the reef and consumed the remaining plant growth of understory kelp and all new kelp recruits (Fig. 8.1 B). Due to continued high urchin grazing pressure, the reef remained an urchin-dominated barrens for three years (Fig. 8.1 C). In winter 1983, the second storm (storm II) initiated reforestation by eliminating exposed urchins and clearing rock surfaces for widespread kelp settlement and growth. This kelp survived a period of warm, nutrient-poor water during the 1982/4 El Nino episode through benign winters in 1983 and 1984 to produce a dense forest with a thick surface canopy. Hence, the pre-storm I and post-storm II habitats typified kelp forests where sea otters and other predators prevent destructive grazing by urchins. The interim habitat was grazed nearly bare by unregulated populations of urchins. The inshore site, on the other hand, was initially spared by storm I, but was mostly stripped of its kelp canopy and laden with sand and silt by storm II.

8.3 Methods

We counted fish along SCUBA transects in different areas of Naples Reef (the offshore site) at weekly or longer intervals from 1977-1985, and in similar areas of Devereux


Fig.8.1A-C

Methods 173

Point (the inshore site) from 1979-1984 as time permitted. Thus, offshore counts made before 1980 (storm I) were in a maturing kelp forest; those from spring 1980 to winter 1983 (storm II) were in a progressive urchin barrens; and those made thereafter were in a regenerating (1983-1984) or mature (1984-1985) forest. In contrast, inshore counts taken before winter 1983 were in a mature forested habitat, which then deteriorated. Along with each series of counts, we measured an array of physical variables, including water temperatures at surface and bottom, bottom concentration of nitrates (by colorometric determination from water samples taken in situ during 1980 and 1981 only), and wave height (estimated from trough to crest by eye). Physical variables were plotted as monthly averages.

At Naples Reef we counted fish in transects at four locations identified in Ebeling and Laur (1985): at two deeper ones (11-14 m) near the flat base of the reef and gently sloping lower sides at east (location E) and west (W) ends, and at two shallower locations (6-8 m) about steeper slopes and crest near midreef (M, M'). At E, W, and M, two ropes, 60 m long and secured 10- 20 m apart, marked parallel paths over the bottom; the lines ran along either side of a prominent crest at M. At M', large nails marked an equally long path in and out of three rills paralleling the main axis of the reef. Thus, transects representing all principal microhabitats totalled 480 (4 X 120) m in length. We counted all fish in an imaginary 3-m-wide band centered on the lines upward to about 3 m into the water column (the limit to which we felt we could usually estimate fish numbers accurately).

Counts per species or growth stage were summed over sites to compute bimonthly or seasonal averages. Since frequency distributions of these sums were significantly skewed, they were transformed as log (x + 1) to compute back-converted geometric means and their 95% (asymmetric) confidence intervals as recommended by Sokal and Rohlf (1981). For comparative purposes, therefore, the measure of central tendency approximated the median.

Counts were adjusted for varying effort during the 7.s years of observations. Whenever possible, observations were made weekly from 1977 through 1981; samples were analyzed by bimonthly interval, with sample size usually six to nine transect sets. Since midreef sites were not monitored until 1979, however, earlier pooled counts per species were of counts from E+W sites only. Pre-1979 densities were extrapolated reefwide as products of E+W sums times the 1979 annual ratio for the species: (M+M')/ (E+W). For example, 4.06 times as many Embiotoca jacksoni adults were counted at midreef than at reef end locations during 1979; hence, bimonthly averages of E + W counts made during 1977 and 1978 were increased by a factor of 4.06 for reefwide estimates (with no confidence intervals). Fish counts were resumed in fall 1982 for another project, but continued sporadically through 1984. After 1982, therefore, means were computed for seasonal (3 month) samples of three to nine transect sets.

We began inshore fish counts in November 1978. Before 1982, we counted fish one to six times per bimonthly period in undirected surveys from about 9-m depth shore-

.. Fig. B.IA-C. Naples Reef in different community states. A Kelp forest with giant kelp (Macrocystis pyrifera) in left foreground, understory kelp (Pterygophora californica) in right background, and adult surfperches. B Partly grazed with moving front of larger red sea urchins (Strongylocentrotus franciscanus) and smaller purple sea urchins (S. purpuratus) destroying a stand of understory kelp. C Urchin-dominated barrens


ward to 3 m. The duration was fixed at 60 min to match the effort offshore, where ten timed sets of all four transects averaged 60±6 (SD) min to complete. Inshore counts were also resumed in fall, 1982, but along six fixed transects totalling about 500 m in length and lasting about 60 min. Included were segments in deeper water (7-9 m), shallower water (3-4 m), and between. Because they were based on samples of variable sizes taken in different ways, the inshore averages are presented without confidence intervals.

For analysis by species, we emphasized common fishes making up most of the total fish biomass and representing functionally distinct foraging guilds. The guild of "demersal microcarnivores" included five large surfperches (Embiotocidae: Embiotoca jacksoni, E. latera lis, Hypsurus caryi, Damalichthys vacca, Rhacochilus toxotesj, which harvest food from benthic turf employing relatively inflexible foraging tactics (Laur and Ebeling 1983; Ebeling and Laur 1986). These were treated as adults separately and as juveniles or subadults collectively. We also distinguished migratory surfperches as likely to be most sensitive to habitat change. The abundant blacksmith Chromis punctipinnis was chosen as the only "midwater planktivore" with an extrinsic food source unaffected by kelp beds (see Hobson and Chess 1976; Bray 1981). Among more generalized foragers, the kelp bass Paralabrax clathratus represented a guild of "switchfeeding predators" (Love and Ebeling 1978). The opaleye Girella nigricans and halfmoon Medialuna californiensis exemplified "plant-cropping omnivores", species exploiting kelp where it is present but also occurring abundantly on high-relief reefs where it is absent (Stephens et al. 1984; Patton et al. 1985). [The very few obligatory kelp associates (Quast 1968a), such as the kelp perch Brachyistius frenatus, were not analyzed because they were rare or absent at Naples Reef (Ebeling et al. 1980b). This is a local phenomenon peculiar to offshore reefs, like Naples, that are subject to (1) periodic loss of their surface (Macrocystisj kelp canopy and (2) swifter currents that pull the kelp stipes beneath the surface, making it difficult for the small fish to maintain station amongst the blades (Ebeling et al. 1980a). Kelp perch were commonly seen in the thick surface canopies inshore at Devereux PoinL]

Modifying a method used by Patton (1985), we tested to see if the fish assemblages were responding to a climatic shift in water temperatures during the period of study. Temperature data from Naples Reef corroborated Patton's distinction between a pre-1983 cold water episode and a 1983-1984 episode of warm water associated with a major El Nino intrusion of southern water masses. Patton distinguished "southern species" of reef fishes with significantly greater abundances south of the Los Angeles area from "northern species" more abundant to the north and "central species" equally abundant north and south. We adapted his method of comparing biogeographic categories by sums of standardized species densities. Our standardization procedure differed, however, in that all counts were transformed to logarithms and divided by the species' mean. For each observation, these standardized densities were summed over the species making up a biogeographic category: of four northern species (Damalichthys vacca, Rhacochilus toxotes, Embiotoca lateralis, Sebastes mystinusj, four central species (Embiotoca jacksoni, Paralabrax clathratus, Hypsurus caryi, Phanerodon furcatus) , or five southern species (Oxyjulis californica, Chromis punctipinnis, Girella nigricans, Medialuna californiensis, Semicossyphus pulcher). We compared categories among periods as summed standardized densities averaged over three sum-

Physical Variables 175

mer-fall (July through October) periods: 1979 (before storm 1),1981 (before storm II and in the midst of the offshore barrens period), and 1983/4 (after storm II during the deterioration of the inshore site and recovery offshore). Data for 1983 and 1984 were pooled to increase sample size to equivalency with the previous periods.

To compare structures of the fish assemblage among these six site-period samples (two sites X three periods), we ranked the individual species abundances and calculated between-sample rank correlations (rs).

8.4 Results and Discussion

8.4.1 Physical Variables

Episodic events characterized seasonal weather patterns at Naples Reef (Fig. 8.2). Mean monthly wave height (dashed line) showed varying periods of calm and disturbance, with peak values identifying periods shortly after storm I (February 1980) and storm II (March 1983) when we could again go to sea safely. Peak averages during the storms were much greater, exceeding 6 m, the highest recorded in the Santa Barbara Channel since regular records began a decade earlier (Ebeling et al. 1985). Similar peaks between calmer periods were recorded from central California (Bodkin et al. 1987; VanBlaricom, this Vol.) and throughout the Southern California Bight (Seymour et al. 1984). At Naples Reef, short-term drops in water temperature (Fig. 8.2: solid lines) marked seasonal influxes of upwelled nutrient-rich water (arrows) during all years. [The assumed negative relationship between temperature and nutrient levels was verified by the distribution of nitrate concentrations measured during 1980-1981 (dashed line).] The general coincidence of upwelling and periods of maximum sunlight (spaces between

A Storm I I Storm II

" " /1

E • I' / I we: 'I I' II /I /1

1 5~ II I I I I

,'I. I I \ \ ;;: I 1.0 I I I I I 1/1 / I I I I " > ~ I \ I \\/ .... / .. ,/'./~ \ .•. ,\ I ..... , 1\ I \ I :,.", 1\ I > W I \ ,"1 v'\,/ v \ \./ I \ I ... ./

I 0.5 I \ I V v

I V I I " V"-\," ..

\,.... I \ \ \../ ,I

U 18 0

u.i a: a: 16 w:J f-f-«« 14 sffi

Q. 12 :2

Mf t t

w f- 10

Fig. 8.2. Seasonal and long-term pattern of weather variables and bottom nitrate concentrations recorded at Naples Reef. Mean monthly wave height (dashed line) shows periods of cahn and disturbance (storms I and II). The relations between water temperatures (solid lines) and upwelling (arrows), periods of maximum light for photosynthesis (spaces between hatched blocks), and nitrate levels (dotted line) predict seasonal pulses of productivity


hatched blocks) signalled seasonal pulses of productivity occurring with variable magnitude throughout the study period. With the 1982/4 El Nino episode, temperatures increased and, discounting seasonal cooling and spurts of upwelling, remained high through much of 1984. Temperatures then dropped abruptly during intense upwelling in spring 1985 after two benign winters without severe storm disturbance. Inshore conditions at the shallower Devereux site closely paralleled the offshore weather: surface temperature and wave height were about the same, although bottom temperature was occasionally higher.

8.4.2 Total Fish Density

By the middle of the barrens period at Naples Reef offshore (July-Oct., 1981), the collective density of all 32 species counted had decreased significantly (although not drastically) from that estimated during the forested period (July-Oct., 1979) before defoliation by storm I. Average density fell by 20% from 527± 129 (SD, N = 18) individuals per transect to 422±232 (N = 16, t between means = 2.3, log-transformed variates, P < 0.05). This was more likely due to the local habitat change offshore rather than any global shift in fish abundance because at the inshore forested control site at Devereux Point, average density was unchanged between the two sampling intervals: 335±116 (N = 9) and 380±149 (N = 14), respectively (t = 0.6, log-transformed variates, P> 0.5). Then Devereux itself was ravaged by storm II, so subsequent contrasts were not made.

However, the net decline offshore oversimplifies the effect of habitat change from forest to barrens because decreases did not occur across-the-board. Much of the losses by some species was offset by gains in others either unaffected or actually facilitated by the change. Later, in fact, an enormous rise in blacksmith during 1982 probably erased the overall Naples deficit. A species-by-species analysis is more revealing.

8.4.3 Species Densities

8.4.3.1 Surfperch Adults

Relative to Devereux counts, offshore counts of surfperch adults fell gradually between storms during the barrens period at Naples Reef (Fig. 8.3). Initially, the abundant blackperch Embiotoca jacksoni contributed most to the collective decrease in adult numbers. Of the four species represented, it may be most dependent on prey winnowed from the superficial carpet of algal turf (Laur and Ebeling 1983; Stouder 1983) that was sporadically eroded from the reef bottom by sea urchins (Ebeling et al. 1985). Yet, E. jacksoni and the less common striped perch Embiotoca lateralis were even more abundant during 1977 and 1978, well before the barrens period. Adept at extracting less superficial prey (Laur and Ebeling 1983; Ebeling and Laur 1986), adults of the rubberlip perch Rhacochilus toxotes and the pile perch Damalichthys vacca varied more like species not dependent on kelp forests; Patton et a1. (1985) noted than the pile perch was the only surfperch more abundant on breakwaters than on the natural reefs

Species Densities 177

(f) ...J -0: ::J o >

80

SURFPERCH ADULTS (ex Hypsurus caryi)

--- Inshore - Naples Reef

!2:F~te1i; o

~;;:' I;.!.;,+-~ I I L

TOTAL I n 1'1 .~2,236633265333216862

U L......I ~ore

4 2 6 10 4 2 1 1

80 N, Naples Reef

I -9- - -646997276 -99884810876898866 633

1 ~1~1~lil~I~I~I~I~lil~I~I~I~I~lil~I~I~I~I~lil~I~I~I~1~lil~I~1 I:I~I~I~I~I~I~I~I~I 1977 1978 1979 1980 1981 82 1198311984 185

40

o

Fig. 8.3. Seasonal offshore (solid bars) and inshore (open dashed bars) densities of resident surfperches, i.e., adults of the four species that do not migrate inshore. For offshore samples at Naples Reef, bars measure geometric means with 95% confidence interval (N usually 6-9) of counts per set of four transects or about 1 h of underwater survey; hatched bars indicate densities pooled for only two of the transects before all four were established; open bars above hatched bars measure extrapolated reefwide densities. For inshore samples at Devereux Point, open dashed bars measure geometric means without confidence intervals (N usually 1-6) of counts per hour survey. Arrows (storms I and II) indicate the periods of severe storm disturbance (see Fig. 8.2)

they surveyed. Nonetheless, numbers of even these two declined near the end of the barrens period. Densities of all species then increased abruptly in spring and summer 1985 (pers. observ.) after the mature forest had regrown and as the oceanographic conditions returned to pre-El Nino conditions.

Inshore at Devereux Point, however, a sharp drop in total surfperch abundance in winter 1983 followed storm II, as the benthic environment deteriorated due to sanding and silting over.

8.4.3.2 Sur/perch Young, Subadults, and Summer Transients

These groups responded most to barrens formation (Fig. 8.4). All showed marked seasonal cycles in densities offshore and weak indications of complementary trends

178

120 if) ;;! 100 => 0

80 :; 0 ~ 60 u. 0 40 0: W

20 CD :2 => Z z <{ 80 w :2

" 60 0: f- 40 w :2 0 20 w (:J

40 0:

~ 20

Fish Populations in Kelp Forests Without Sea Otters

SURFPERCH YOUNG-OF-YEAR (all 5 spp.)

Inshore ---Naples Reef -

SURFPERCH SUBADUL TS (ex Hypsurus caryi)

j140)

I t I '-l 1_.:--1 ! r-! : : I I I I I I

r-I I I I I I I I I I 1-' I I I I I I 1 I I P.._

: : : I ~ L_ r-' r : : : : I ~J-~

I I I I .. _, !.._I

i...J-J J

-, J

J J J J J J J J J J J J : ~-L_ .. r-..-- I

t Storm I t Storm"

SUMMER TRANSIENT SURFPERCHES (Subadults and adults) Inshore --

Naples Reef -

:2 =>if) OL-~=W.w ______ ~~~~~~~~ __ ~.-~~_-_-==-~-~--~-~~~~~

~ ~ 100 <{ 5 Hypsurus caryi ~ 2: 80

,,~ 60 0:f-u. ~ 0 40 o ~ 20

Fig. 8.4. Seasonal densities of young-of-year surfperches of all five species, subadults of the four species whose adults are residential, and of subadults + adults of the two migratory species. See Fig. 8.3 for sample sizes and explanation of format

inshore. Perhaps this reflected inshore-offshore migrations of parts of populations: offshore during spring and summer periods of accelerated offshore productivity, and inshore during the fall and winter when offshore cover and food supplies decrease. The same individuals may return to Naples Reef yearly; of 31 adult rainbow surfperch (Hypsurus caryi) tagged in 1975, one was seen again in 1977 (M. Hixon, Oregon State Univ., pers. commun.). With barrens formation, the offshore habitat remained unsuitable all year and summer migrant popUlations dwindled. This was demonstrated explicitly for young-of-year surfperch (Ebeling and Laur 1985). Before storm I, young were born on Naples Reef during spring and early summer and remained until fall; during the barrens period they disappeared along with their shelter of understory kelp, but then reappeared after storm II as their refuge regenerated. likewise, their abundance varied predictably with the loss and revegetation of macro algae after storm II. The less vulnerable subadults continued their seasonal appearances but at reduced densities as the barrens progressed. Yet, numbers of adult transients (Hypsurus caryi, Plumerodon

Species Densities

(fJ

;;i 240 :::J

e 200 >

~ 160 LL 0120 a: w 80 CD :2; :::J 40 z z « 0 w :2; 100 ()

a: f-

80 W :2; 40 0 w CJ 0

MIDWATER PLANKTIVORES (Chromis punctipinnis)

iO co '£ Young-of-year

179

t + ,Storm I {Storm II

---Inshore - Naples Reef

;;:- 00;:- iO '" "'''' 0 ~ ",co

f t r::·

Fig. 8.5. Seasonal densities of young-Qf-year and subadults of the blacksmith Chromis punctipinnis, an obligate planktivore. See Fig. 8.3 for sample sizes and explanation of format

furcatus) fell abruptly as well, indicating that migrants as a group were most sensitive to habitat change. All groups reappeared on schedule as the forested habitat returned.

8.4.3.3 Midwater Plilnktivores

Loss of kelp offshore did not destroy habitat for blacksmith (Chromis punctipinnis), which are among the most abundant reef fish, eat only oceanic plankton, and shelter in rocky holes and crevices (Bray 1978, 1981). In fact, recruitment to the reef of these residential fish increased remarkably during the barrens period (Fig. 8.5 , top) before a warming trend (Fig. 8.2) signalled the E1 Nino intrusion of water from the south. Yet there was no concomitant increase at the forested inshore site, where density remained low. The young do not seek refuge in the plant understory, but hover near the bottom and quickly hide in small cracks and crevices when danger threatens (Bray 1981). Adults are mostly invulnerable during the day and pass the night in rocky holes and crevices. Thus, neither young nor adults need change their basic anti-predator and sheltering behaviors between barrens and forested periods.

Admittedly, some counts may be biased upward, as observed in the barrens habitat. For blacksmith and other mid water species, counts made along bottom transects may be unrealistically low when a well-developed kelp canopy is present at the surface. The canopy extends their vertical scope of orientation, and so much of the population may be overlooked during bottom transects (Bray 1978; Ebeling et al. 1980b; Larson and DeMartini 1984). When a kelp bed is present, furthermore, adult blacksmith assemble at the upcurrent edge where plankton is most abundant (Bray 1981). Yet, our estimates

180

SWITCH-FEEDING PREDATORS (Paralabrax clathratus)

(j) Subadults and small adults ;:;!1 ::J e 120 > Inshore ---~ 100 Naples Reef -


(Storm I Both

Storm II ~

j"i I I I I

200 (j) -' .0:

180 6 >

160 0 z

140 u. o

120 ffi CD

100 ~ z

80 ~ w

60 ::;; o

40 ~ w

20 ~

r-'--+--L...J.....L...I--'--.L...t-' 0

Flw slslflslslfl s

w Cl

alpuWpuwp

82 198311984185

Fig. 8.6. Seasonal densities of large and small sizes of the kelp bass Paralabrax clathratus, a switchfeeding predator. See Fig. 8.3 for sample sizes and explanation of format

of blacksmith young were accurate because the small individuals seldom stray from the immediate vicinity of their rocky shelters. Also, we included only the counts of subadults, which range less widely from shelter than adults (Hobson and Chess 1976; Bray 1981), and counts were made well up into the water column. Counts made at Devereux Point were probably accurate because the observer was in shallower water, usually in visual range of the canopy; inshore fish were at much lower density and scattered in smaller groups.

8.4.3.4 Switch-Feeding Predators and Plant-Cropping Omnivores

Although subject to upward bias as explained above, the observed densities of these species corroborated previous evidence (Stephens et al. 1984) that they can or not take advantage of kelp as the occasion arises. Among the switch-feeding predators, subadult and small adult kelp bass (Paralabrax clathratus), which can eat plankton, nekton, or substrate-oriented prey from plants (Love and Ebeling 1978), increased in density both inshore in the kelp forest and offshore in the barrens after storm I (Fig. 8.6). Numbers of large adults rose as well, showing similar seasonal peaks, perhaps as aggregations moved offshore to breed (Feder et al. 1974). Undiminished numbers of these larger predators assured the demise of any shelterless surfperch young born on Naples Reef. These density trends of all post-juvenile kelp bass counted together after 1982 indicated that inshore abundance first dropped and then recovered after the storm II disturbance in 1983.

Temporal trends of the two plant-cropping omnivores Girella nigricans andMedialuna californiensis resembled those of kelp bass (Fig. 8.7). During the barrens period offshore, fish of both species browsed about bushes of the red alga Gelidium robustum

Biogeographic Species Groups

PLANT-CROPPING OMNIVORES 60

GirefJa nigricans

181

--- Inshore - Naples Reef

Fig. 8.7. Seasonal densities of the opaleye Girella nigricans and halfmoon Medialuna californiensis, both plant browsers who may switch to other prey. See Fig. 8.3 for sample sizes and explanation of format

that had escaped sea urchin grazing at reef crest. In the offshore kelp forest before storm I, both species ate mostly algae, including kelp encrusted with bryozoans and other attached animals (pers. observ.). During the subsequent barrens period, Medialuna picked plankton and even cropped dense clones of the sea anemone Corynactis cali/ornica (R. Rowley, Univ. California Santa Barbara, pers. commun.). Girella's gut contents were not examined.

Perhaps these two fishes (as well as kelp bass) are actually attracted to disturbed and newly altered habitats. They often occur at higher densities on artificial reefs where they browse transplanted kelps (Turner et al. 1969; Carter et al. 1985b).

8.4.4 Biogeographic Species Groups

Standardized sums of species densities provided inconsistent evidence of a faunal shift between cool (I 979-1981) and warm episodes (I983-1984). Perhaps the best evidence is that northern species declined as southern species increased offshore at Naples Reef during the warm period (Fig. 8.8). Otherwise, any such shift was confounded by the fishes' responses to the storm-induced changes in structural habitat. Inshore at Devereux Point, across-the-board declines in southern as well as other species indicated a response to habitat deterioration after storm II. The decline in abundance of central species at Naples Reef within the cool episode simply reflected the predominance of surfperch members in the group: of the total of four central species, the three surfperches (Embiotoca jacksoni, H. caryi, Phanerodon /urcatus) decreased in abundance as sea urchins destroyed patches ofturf containing their prey, while only the kelp bass (Paralabrax clathratus) increased.

Recruitment patterns of juveniles provided clearer evidence of a response to the El Nino-induced climatic shift. Among northern species at Naples Reef, densities of


8 4 Northern species

6 ------- ,

4 ' (J) , W , ~ , iii 2 0 Devereux Point (inshore) "'1 ill • Naples Reef (offshore) o o o~~~--------~--~--------~~--w N rr: ~ 6 o z ~

t) 4 o w ~ 2 ~ ::l (J)

u. o W 8 C) ~ rr: ~ 6 ~

4

2

t- ---+ ......... "

, 1

5 Southern species

" , " , , , " ,

1979 1981 1983/84

Fig. 8.8. Average summed standardized densities with 95% confidence intervals for biogeoggraphical groups of reef-fish species observed at the offshore and inshore study sites. See text

young blue rockfish (Sebastes mystinus) fell from 169±77 (SD) per transect set during July-Oct. 1979 to 6.5±1.6 in 1981, and only 1.3±1.1 in 1983-1984. Alternatively, their disappearance in 1981 may have coincided with the loss of kelp canopy, an apparent refuge where the young may initially settle (Miller and Geibel 1973), but they did not reappear in noticeable numbers as the canopy returned, and Bodkin (1986) observed a similar decline under a continuous canopy during the later part of the same period. Among southern species, the remarkable rise in recruitment of blacksmith (Chromis punctipinnis) was described earlier (Fig. 8.5). Young of the tropical family Labridae also recruited to Naples Reef in greater numbers during the warm episode. Recruits of the senorita Oxyjulis californica and California sheephead Semicossyphus pulcher increased from negligible numbers in 1979 to 116±80 and 11 ± 16, respectively, in 1983-1984. We also saw many rock wrasse (Halichores semicinctus) young for the first time since our Naples studies began in 1970. These sightings coincided with unusually heavy recruitment of southern species along northern coasts (see also Bodkin 1986),


an event more likely due to enhanced transport of planktonic larvae northward during the 1982/4 El Nino episode than to rising water temperatures per se (Cowen 1985).

8.4.5 Fish Assemblage Structure

Comparison of ranked species abundances also indicated that effects of storm-initiated habitat changes were confounded with those of a climatic shift (Table 8.1, Fig. 8.9). Correlation analysis resolved two clusters of similar species arrays: one from the inshore and offshore sites during the cool episode (1979,1981) before the sites were defoliated, the other from offshore in the midst of the barrens period (1981) and in the warm episode during reforestation (1983/4). This reflected primarily the increase in rank abundances of southern species (e.g., Chromis punctipinnis, Girella nigricans, Medialuna californiensis) offshore at the peak of the barrens period before the warm episode.

8.5 General Discussion and Conclusions

Current and historical evidence indicates that where sea otters prevent sea urchins from dominating food and space in the Aleutian Islands, fish are much more abundant

Table 8.1. Rank order of fish (subadults + adults) abundances and correlations among summer-fall periods of 1979,1981, and 1983/4 at offshore Naples Reef (N) and inshore Devereux Point (D)

N1979 D1981 D1979 D1983/4 N1983/4 N1981 (18)a (14) (10) (15) (19) (16)

Embiotoca jacksoni 1 3 4 6 Oxyjulis californica 2 3 2 2 2 5 Chromis punctipinnis 3.5 6 8 11.5 1 Paralabrax clathratus 3.5 2 3 3 2 Hypsurus caryi 5 4 4 7 10 12 Damalichthys vacca 6 7 5 6 8 7 Gire/la nigricans 7 5 10 11.5 5 3 Medialuna californiensis 8 10 12 4 6 4 Rhacochilus toxotes 9 9 7 8 9 8 Embiotoca lateralis 10 11 10 9 12 10 Phanerodon furcatus 11.5 8 6 5 11 13 Sebastes mystinus 11.5 13 13 11.5 13 9 Semicossyphus pulcher 13 12 10 11.5 7 11

Correlations (r s) N1979 0.90 0.74 0.58 0.74 0.62 D1981 0.86 0.61 0.66 0.44 D1979 0.71 0.43 0.09 D1983/4 0.33 0.18 N1983/4 0.81

a Sample size.

184

FISH ARRAYS

0.4


N-1979

0-1981

Forest L..---0-1979

L--------0-1983

0.6

..----N-1981

Barren '---- N-1983

0.8

Fig. 8.9. Relations among the spatio· temporal arrays of reef fishes identified in Table 8.1, as measured by averaged correlations between arrays of ranked species densities

RANK CORRELATION

as products of greater habitat complexity and an enhanced detritus-based food chain (Estes and Palmisano 1974; Estes et al. 1978; Simenstad et al. 1978). Fewer fish occur in the urchin-dominated barrens about islands without otters, and those that do are of mostly open-water species instead of the typical reef species that abound in kelp forests fringing islands with otters. Through 2500 years of strata in aboriginal refuse heaps, furthermore, the abundance of fish bones correlates positively with that of otter bones, but negatively with urchin tests (Simenstad et al. 1978). After otters returned to some islands a few decades ago, they ate progressively more fish, mostly of sluggish bottomsitting species, after exposed urchins and other primary prey organisms were consumed and the kelp forest regrew (Estes et al. 1978, 1981; Riedman and Estes, this Vol.).

Off central California, where the otter population has also redeveloped, this otterkelp-fish relationship is less clear. Otters rarely, if ever, eat fish (e.g., Estes et al. 1981; Ostfeld 1982), and opinions vary as to the generality of the otter effect there: Mclean (1962) reported that previous large gaps in kelp beds were filled in by dense growth after otters returned to the Monterey area, and Lowry and Pearse (1973) observed that urchins were then relegated to deep crevices and other refuges, where they ate algal drift. Laur et al. (this Vol.) documented an increase in fish abundance on an offshore reef farther south, as newly arrived otters eliminated urchin grazers and kelp regrew. Miller and Geibel (1973) showed that kelp canopy removal destroys suitable habitat for many young fishes and for small species adapted to a specialized life among the fronds. On the other hand, they pointed out that the return of giant kelp (Macrocystis pyrifera) canopies to the Monterey area in the late 1960's was not limited to areas foraged by sea otters (see also Foster 1982; Foster and Schiel 1985). They concluded that even though otter control of kelp herbivores may create an environment where kelp responds quickly to favorable conditions for recruitment and growth, there is no obvious correlation between presence of otters, kelp, and larger numbers of adult reef fishes.

VanBlaricom (1984) concluded, however, that sea otters chronically reduce urchin densities in barrens, thereby permitting a predictable succession of kelps that culminates in dominance by perennial Macrocystis after several years. In addition, Bodkin (1986) showed that the Macrocystis forests may contain significantly greater numbers of rock-


fishes (Sebastes), the most productive kelp-bed species off central California, than do forests of Nereocystis luetkeana, an annual species often preceding Macrocystis. Thus, the perceived beneficial effects of sea otters may not be fully evident from short-term studies (VanBlaricom 1984). Moreover, win ter-storm disturbances may sustain a diverse mosaic of kelp assemblages in different successional stages (Cowen et al. 1982; Dayton et al. 1984; Reed and Foster 1984), which may obscure a general otter effect on the fish populations.

These differences in perceived otter effects between the Aleutian and central Californian systems may be partly due to a basic difference between island and mainland environments. The island systems tend to be more isolated and homogeneous (Estes et al. 1981); thus all effects on islands, whether from domination by urchins or otters, may be confined to, and concentrated in, a limited amount of uniformly rocky space. Hence, sublittoral macroalgae are universally rare about islands without otters, but continuously abundant about islands where otters have reached equilibrium density (Estes et al. 1978). Along coastal California, however, the otter effect may be preempted or dwarfed by a variety of other processes in the heterogeneous, though contiguous, mainland space, where energy and materials can flow between urchin -dominated barrens and kelp forests.

Perhaps, therefore, the mainland otters of central California have eschewed finfish because they could move on to unforaged habitats after supplies of their favored shellfish had been depleted (Ostfeld 1982). In contrast, the otters at equilibrium densities in the Aleutian Islands feed more opportunistically on alternative prey, requiring greater effort per calorie, because they have exhausted all unexploited supplies of preferred prey resources (Estes et al. 1981, 1982). Yet the Californian otter population has not expanded during the last decade (Riedman and Estes, this Vol.).

Estes and VanBlaricom (1985) concluded that the addition of fish to otter diets enhances otter abundance in the Aleutians to higher levels not sustainable by invertebrate prey alone. Off California, however, otters may have more difficulty exploiting this additional food resource because the kelp-bed fish fauna has lower densities of the kinds of sluggish, bottom-sitting species (bottom rockfish, greenlings, cottids) that otters can profitably catch (see Riedman and Estes, this Vol.). Densities of these bottomsitters in temperate kelp forests of central California were estimated to total only 0.04m-2 (from data in Miller and Geibel 1973,plus Bodkin 1986, pooled over 257 transects), compared to 0.13 m-2 in boreal kelp beds of, for example, the Gulf of Alaska (from data in Rosenthal 1980, pooled over 81 transects). There is, in fact, some evidence that otters do respond functionally to decreasing densities of available fish prey: the boreal otters reportedly eat fewer fish during the winter (cited in Riedman and Estes, this Vol.) when fishes move from the kelp beds into deeper water (Rosenthal 1980).

Considering all this, it is difficult to say what the general effect of sea otters on fish assemblages would be in southern California, where kelp forests (1) have persisted without otters for more than 100 years (Tegner 1980), (2) show less predictable seasonal turnover (see Introduction), and (3) endure alongside barrens patches of variable ages and sizes (Dean et al. 1984; Ebeling et al. 1985; Harrold and Reed 1985). In these warm-temperate environments, the density of potential bottom-sitting fish prey may be lower yet (for example, estimated as 0.025 m-2 from data in Ebeling et aI. 1980b, pooled over 168 transects at Naples Reef).


Opinions vary, furthermore, on the importance of kelp to survival and growth of most fishes in southern Californian kelp beds. Although kelp and other macro algae serve as refuges for a variety of young fishes and constitute essential resources for a few specialized species such as the kelp perch (Brachyistius frenatus) and kelp goby (Lethrops connectens) (Miller and Lea 1972; Coyer 1979), deforested natural reefs and artificial structures, such as breakwaters that have never borne kelp, support abundant and diverse fish assemblages (Stephens and Zerba 1981). Stephens et al. (1984) found that the return of kelp to a high-relief rocky reef had little effect on fish abundance. In comparing fish populations on natural reefs, Quast (1968a,b) showed that among high-relief sites, fish abundance was about the same whether kelp was pre· sent or not (though fish made fuller use of the water column in kelp), but among low· relief reefs, abundance was more than three times higher where kelp was present. Patton et al. (1985), furthermore, concluded that beyond a low average relief height (0.4-0.7 m), fish abundance and diversity no longer increase with greater turf cover and kelp density (during the summer productive season, at least); on high -relief reefs, therefore, the algal forest and attendant resources seem to be superfluous. From an analysis of transects taken at three depths, Larson and DeMartini (1984) showed that abundances of typical reef fishes over flat cobble were much greater where kelp diversified the water column. Hence, any control of grazing urchin populations may enhance fish populations in such low-relief areas.

However, even at Naples Reef - certainly a high relief site by Patton's criterion -the fish assemblage sustained a modest decline in abundance and significant change in structure when urchin domination after a storm disturbance turned the kelp forest to barrens. Comparison with a control site inshore indicated that most of the change was due to habitat transformation, not a climatic shift. Some species declined ("losers") while others were unaffected ("break-evens") or even prospered ("winners"). Among the losers were micro carnivorous species with fixed requirements indirectly or directly linked to kelp-forest products; winners or "break-evens" were either never dependent on the forest products or able to switch to other resources.

Adult surfperches were "losers" via indirect effects. Their populations declined as sea urchins prevented forest regeneration after wave action destroyed the giant kelp. Without forest litter to eat, urchins eliminated all remaining understory kelp and substantial portions of algal turf harboring the invertebrate prey so important to fishes; thus, fish and turf densities were strongly correlated within transect segments (Ebeling et al. 1985). The surfperch species that exploited more superficial prey experienced the greatest immediate loss of food, and decreased in abundance sooner than species eating buried prey. Yet, none altered its foraging behavior or diet as all converged on the remaining food patches (Stouder 1983).

Sub adults and adults of summer transientsurfperches, which undergo seasonal migrations, were particularly sensitive to the change at Naples reef. Perhaps they suffered from increased competition with dominant resident species for depleted food supplies (Stouder 1983); for instance, Hypsurus caryi is not only aggressively submissive to the residents (M. Hixon, Oregon State Univ., pers. commun.), but has greatest dietary overlap with them as well (Ebeling and Laur 1986).

The effect of kelp loss on surfperch young born on the reef was direct. The small fishes' fixed refuging behavior prevented them from surviving the destruction of the


understory kelp cover (Ebeling and Laur 1985). Likewise, tiny rockfish and kelp bass recruits, which also disappeared during the barrens period, need kelp or other large algae in which to settle from the plankton (M. Carr, Univ. California Santa Barbara, pers. commun.).

The loss of a nursery area at Naples Reef may not affect fish recruitment and population structure globally, however. Even before forest loss, for example, all young surfperch left the reef in the fall as cover and food declined somewhat; subadult immigran ts replenished adult stocks during the following spring and early summer (Ebeling and Laur 1985). Then during the barrens period, young disappeared entirely as pregnant females may have left the reef to give birth elsewhere in suitable cover, and subadult recruitment continued but with decreasing amplitude. A nursery area was provided inshore where widespread foliage persisted all year, fewer of their large kelp bass predators (Ebeling and Laur 1985) occurred (see Fig. 8.6), and young fishes increased in abundance. Young surfperch may also find shelter in heaps of bottom drift in deeper water (pers. observ.; G. Cailliet, Moss Landing Marine Laboratories, pers. commun.). Likewise, Miller and Geibel (1973) concluded that fish recruit to offshore reefs near Monterey via both annual settlement of young and immigration from nearby habitats. They found dense concentrations of young surfperch and rockfish in heavily vegetated rocky zones inshore, as well as in the offshore kelp canopies.

We conclude that any decreased capacity of a barrens patch to support fishes like surfperches may be compensated by increased use of kelp-forest patches because seasonal migrations link the populations. Populations of resident adults on reefs without nursery areas are maintained by immigrant subadults, while transient adults can sample patches until they find favorable forage. Consequently, the system of nursery and foraging grounds persists as an open mosaic, consisting of assemblages in habitat patches, all connected by migrations (see Caswell 1978). Because these fish minimize the costs of environmental variability by shifting across a spatio-temporal mosaic (Levin, this Vol.), therefore, it is difficult to assess effects of the loss of a few nursery or feeding areas on the total fish population.

Other common kelp-bed fishes were apparent "winners" or at least "break-evens" during the barrens period. As midwater planktivores, blacksmith flourished and young recruited to the reef in record numbers. These fish were apparently unaffected by diminished stocks of locally generated food because they ate plankton swept along by prevailing currents unimpeded by kelp. In contrast, when Naples Reef supported a dense kelp bed, fewer blacksmith foraged less successfully at the down current edge, as plankton density decreased significantly between upcurrent and down current edges (Bray 1981). Adult populations of kelp bass, opaleye, and halfmoon also remained strong. These switch-feeding predators and plant-cropping omnivores may even be attracted to new opportunities in disturbed areas. They soon appear on newly set artificial reefs, for example (Turner et al. 1969; Carter et al. 1985b). Some eat plankton; others forage in surviving algal patches or exploit burgeoning populations of alternative prey such as sea anemones. Choat (1982) pointed out that, unlike their tropical counterparts, these "herbivorous fishes" of temperate waters are not highly specialized browsers with advanced jaw structures.

Populations of large benthic fishes were probably not enhanced by the increasing availability of sea urchins as potential prey. As the best example, the California sheep-

188 Fish Populations in Kelp Forest Without Sea Otters

head (Semicossyphus pulcher) reportedly feeds on urchins and may contribute to controlling their numbers elsewhere (Tegner 1980; Cowen 1983). On Naples Reef, the density of adult sheephead more than quadrupled between 1981 and 1986 to exceed 200 ha-1 , as results of record recruitment during the 1982-1984 El Nino episode (see also Cowen 1985) and subsequent growth (work in progress). Currently, however, these fish show little, if any, evidence of eating urchins of any size, and are not controlling the large number of urchins also recruited during El Nino (work in progress).

The preceding discussion implies that reef fishes cannot fully utilize the total available production from kelp-forest communities because the output is distributed over locally variable patches and subject to rapid and unpredictable change (see Choat 1982). Response-time lags are such that consumers may track but seldom overtake their food supply during favorable periods (Boyce 1979). Recruitment success of the reef fishes fluctuates as well (Cowen 1985), so the effects of favorable periods may be stored in declining populations of long-lived adults (see Warner and Chesson 1985).

The differential effect of climate on biogeographic species groups complicates the issue. Since the Southern California Bight is an ecotone where northern, central, and southern species mingle, a climatic shift is likely to favor one group over another (Stephens and Zerba 1981; Stephens et al. 1984; Patton 1985). The barrens period at Naples Reef coincided with a shift from cool to warm (El Nino) episodes. Therefore, effects of changes in structural habitat on abundances of southern species - all planktivores or multivores - were confounded by change in climate. Recruitment of young blacksmith and senorita (Oxyjulis cali/ornica) began to increase just before the warm episode (see also Bodkin 1986); the first appearance of noticeable numbers of young sheephead coincided with it. Since recruitment of young blue rockfish (Sebastes mystinus) crashed both at Naples Reef and the forested site inshore, the decline was not entirely due to loss of refuge at Naples; it may have followed a general decline in northern species during a long-term warming trend (Stephens et al. 1984).

Alternatively, Cowen (1985) presented convincing evidence that recruitment of southern species to reefs in the Santa Barbara Channel depends on the sporadic transport of larvae northward, perhaps independent of temperature change. Hence, the persistence of reef populations may have less to do with physiological acclimation than with episodes of larval settlement associated with El Nino events. Such episodes would explain the occurrence near Monterey of aging adult populations of southern species that are infrequently replaced (Miller and Geibel 1973 ; Hubbs 1974).

8.5.l Predicted Behavior of Our System in the Presence of Sea Otters

Foster and Schiel (this Vol.) pointed out that an otter ("keystone species") effect in southern California would most likely be local, not general. They reviewed evidence that kelp communities are naturally quite variable due to actions of several factors, and are of different physical types, some more vulnerable to urchin grazing than others. Hence, they reasoned, there is no compelling evidence that sea otter predation is necessarily the most pervasive factor determining kelp abundance and community structure. We therefore direct our predictions to offshore reefs like Naples, where urchin outbreaks have been common during the 1980's (see Laur et al. 1986).

Predicted Behavior of Our System in the Presence of Sea Otters 189

If sea otters occupied such areas (assuming they could survive the chronic oil slicks off Santa Barbara - see VanBlaricom 1984), they would tend to stabilize the kelpforest community by preventing the outbreaks of overgrazing by sea urchins that may follow (1) a ioss of drift-kelp production after a severe storm or (2) an unusually strong episode of urchin recruitment (see Carter 1985; Dayton 1985; Ebeling et al. 1985; Davis 1986; Estes and Harrold, this Vol.). Thus, the effect of otters would strengthen the resilience of this community, which is otherwise vulnerable to major perturbation (VanBlaricom 1984). The resulting persistence of large kelp stands would tend to enhance fish stocks by expanding suitable substrates for cover, foraging, and detrital production (VanBlaricom 1984; Duggins, this Vol.); however, this effect would perhaps not be crucial to the entire fish assemblage because many members have flexible behaviors allowing them to exploit alternative resources (see also Stephens et al. 1984). Only the few specialized canopy species, certain juveniles, and benthic micro carnivores that require outputs from the detrital food chain would be critically affected.

Otters may initially compete with meso carnivorous fishes for limited supplies of macroinvertebrate prey, then eventually include fish in their diets as supplies of favored macroinvertebrates are further depleted (see Estes et al. 1978, 1981, 1982; Simenstad et al. 1978; Estes and VanBlaricom 1985). As discussed previously, however, the density of sluggish benthic fishes that otters could profitably catch in southern Californian kelp beds may be too low to constitute a major alternative source of food.

As otters eliminated exposed urchins, the resulting sharp reduction in grazing pressure could eventually depress production of other kelps during a prolonged climatically benign period by letting Macrocystis kelp shade out understory canopies (see Dayton et al. 1984). This, in turn, would impact the forage base of several abundant epibenthic fishes by reducing algal turf and other refuges of their microinvertebrate prey. In addition, the expanding surface canopy would slacken water flow to central regions (Jackson and Winant 1983), where settlement of pelagic larvae and secondary production may be depressed (Dayton 1985), even though the kelp structure may help retain food particles within the bed (Duggins, this Vol.). The accumulated kelp litter from a dense canopy in stagnant water may tend to foul rills or other rocky depressions and render them unsuitable for foraging (pers. observ.). Nevertheless, any such adverse effects would probably be offset by gains in surface canopy production (see VanBlaricom 1984) and by regulation of macroinvertebrate densities (Estes et al. 1981).

In sum, total fish production and diversity in a kelp bed with sea otters would, over the long term, exceed those in one without otters and subject to long periods of urchin overgrazing. The former habitat would sustain a nursery for juveniles of all species and a robust detrital base for production of fish prey for epibenthic microcarnivores; the latter habitat would suffer periodic declines in populations of the dependent species, probably without compensatory gains in populations of others having less specific requirements. The net effect would be to decrease spatio-temporal variability of fish assemblages within the mosaic of inshore and offshore areas of reef and kelp.

This otter effect would be unique. Alternative predators are less effective urchin disposers and are not abundant enough in many areas to control outbreaks (Cowen 1985). The increasing harvest by humans of red sea urchins leaves destructive populations of purple urchins and white urchins (Lytechinus anamesus) unchecked (reviewed in Foster and Schiel 1985); and because only populations of large well-fed


urchins with full ripe gonads are exploited (Wilson et al. 1980), the populations of "starved," poor-quality urchins will remain to dominate all established barrens (Laur et al. 1986). Although urchins are occasionally eliminated by disease, such control may be sporadic and limited to unusual situations of high urchin densities (Pearse and Hines 1979; Miller and Colodey 1983).

8.5.2 Future Research

Other symposium contributors have already suggested several island and mainland localities where observations of different scale can be made to quantify general otter effects, measure contributions of kelp forests to secondary productivity, and determine mechanisms that drive the succession of kelp-forest communities (Duggins; Estes and Harrold; VanBlaricom, this VoL). As for the fishes, their assemblages should be compared between more "open" mainland systems, where a general otter effect may be confounded by interactions between patches, and "closed" island systems where the effect may be more pervasive.

We also agree with Foster and Schiel (this Vol.) that studies of the setting, composition, and variability of different types of kelp-forest communities are needed to find out how each persists in a fluctuating environment. Observations of the age structure of their fish populations could be included, for example. Yearly counts could be made to measure the amount to which the variability in birth (recruitment, settlement) rate exceeds that in adult death rate. This difference in variability can be used to assess the contribution of long-lived adults to the survival of populations through unfavorable environmental periods (Warner and Chesson 1985). If the contribution is high, sea otters may help, for example, by assuring the availability of plant cover for recruitment during a favorable period. If the contribution is low, patch dynamics and interactions on a broader scale may be more important. In addition, tagging sub adult fishes to determine the contribution of inshore-nurtured fishes to offshore recruitment should give a measure of assemblage resilience.

There is also need to consider any possible direct interactions between sea otters and fish. For instance, many macroinvertebrate-eating fishes live in kelp forests about San Nicolas Island (Cowen 1983), to where otters may be transplanted (Estes and Harrold, this Vol.). We should compare diets and abundances of these fishes before and after otters have exploited their mutual prey resources. It would be important to know, for example, if the benefits of a potentially enhanced habitat structure and detritus-based food chain outweigh any costs of otter-fish competition or of size-selective predation by otters on benthic species (C. Simenstad, Univ. Washington, unpubl.).

We should also explore experimentally the effects of the burgeoning sea urchin fishery on forest production and reef-fish abundance. Kelp, urchin, and fish abundances should be monitored at harvested and similar unworked sites, perhaps before and after new operations begin. Of particular interest is the effect of the remaining urchins that are unfit for harvest.

Summary 191

8.6 Summary

Southern Californian kelp forests without sea otters disappear and reappear with episodic events such as severe storms, outbreaks of sea urchin grazing, and widespread plant recruitment. We observed the response of reef-fish assemblages to this variability by monitoring species abundances at two sites off Santa Barbara, one of which was transformed from forest to urchin-dominated barrens for 3 years. Compared to a "control" group at the kelp-forested site inshore, the semi-isolated offshore assemblage sustained a net loss in total fish abundance during the barrens period and underwent a change in structure. While most of the change was due to kelp-forest loss, some was attributable to differential recruitment success at the onset of an El Nino episode of southern water intrusion.

We conclude that all but a few southern Californian reef fishes are adapted to local, unpredictable losses of kelp and other stands of macro algae if rocky relief is high. Any transformation of a reef from forest to urchin-dominated barrens alters the spectrum of resources available to fishes, however. As food production from plant and detrital sources falls, the availability of planktonic food may expand because currents flow more freely over the reef. Shelter in plants is lost, but refuge in rocks remains. Therefore, planktivorous fishes may never use the vegetation as a principal source of food and shelter as they experience a mosaic of forests and clearings; multivores may use the kelp habitat when present but switch to alternative resources when it disappears, but if fish have a fixed suite of requirements that includes the kelp-forest products, they may move to richer patches in an unsaturated environment. Kelp forests without sea otters or other effective predators of sea urchins will vary in structure much more than their fish populations as long as rock relief can replace plants as a source of refuge and substrates for food. Nonetheless, episodic recruitment may alter the complexion of fish assemblages despite the presence or absence of kelp.

We predict that invading sea otters would tend to stabilize the offshore kelp-forest community by checking any destructive sea urchin grazing brought about by loss of drift kelp supplies or episodic urchin recruitment. Thus, the long-term effect of otters would be to decrease spatio-temporal variability of fish assemblages within the mosaic of inshore and offshore areas of reef and kelp. Locally, at least, this otter effect would bring about a net increase in fish abundance and diversity, although some macroinvertebrate eaters might experience increased competition for available food. The otter effect would be unique, in that other natural predators of sea urchins are much less effective in controlling urchin numbers, and urchin harvesting by humans is selective.

Future research is needed to see if the otter effect is more pervasive about islands than along a contiguous mainland. Monitoring the age structure of fish populations would measure their resilience as the extent to which reproductive potential accrued during favorable environmental periods is stored in long-lived adults. Baseline studies should be made of fish abundance and feeding behavior before otters are introduced to island habitats, and in areas outside the otters' range subject to urchin fisheries.

Acknowledgments. We thank J. Estes and G. VanBlaricom for their direction, encouragement, and manuscript reviews; G. Cailliet, M. Foster, and C. Sirnenstad for further critiques; and D. Reed for thoughtful discussions. R. Bray, A. DuBois, C. Ebeling, M. Hixon, R. Rowley, and G. Tribble helped with the field work. S. Anderson provided technical assistance with equipment and boating operations. The Marine Science Institute provided administrative services. This material is based on support by the National Science Foundation under Grants Nos. OCE76-23301, OCE79-25008 and OCE82'{}8183.

9 The Effects of Kelp Forests on Nearshore Environments: Biomass, Detritus, and Altered Flow D. O. DUGGINS

9.1 Introduction

The structure and the dynamics of kelp-<iominated communities have been areas of intense ecological interest since the advent of modem SCUBA equipment made the habitat accessible. In general, research has focused on (1) the biological and physical forces which act either to disturb existing kelp beds or to prevent their development (e.g., herbivory, storms); (2) factors which modify these disturbance events (e.g., sea otter predation on herbivores); (3) competition between species of benthic macrophytes; (4) the physiological ecology and demography of the kelps themselves; and (5) descriptions of species associated with kelp beds. Research in these areas is frequently predicated upon the assumption that kelp beds play an "important role" in nearshore ecosystems and that consequently the presence or absence of kelps will greatly affect organisms including, but not limited to, those utilizing kelps directly as food or substrate. While this "important role" paradigm has been cited frequently by those of us studying kelp-dominated communities, its foundation is, in many instances, based on assumption rather than solid empirical evidence. This is not to suggest that kelps are of little consequence to the majority of nearshore organisms, but rather that the mechanisms linking kelps to the diversity, standing stock, and productivity of most nearshore communities are not well understood. For example, we frequently are told that kelps are important because of their high potential productivity, the inference being that this primary productivity fuels the nearshore food web; but the number of organisms known to feed directly upon the kelps is small, and the indirect utilization of kelps (hypothesized to occur through detritus based food chains) has not been demonstrated.

In this paper I will examine what is hypothesized and what is known about the role that kelps play in the nearshore ecosystem. While I do not directly address questions relating to the role of sea otters in structuring kelp beds, it is my hope that this discussion will help place into perspective the changes that may result from the reestablishment of sea otters along the west coast of North America. Where the reappearance of sea otters leads to a large increase in kelp biomass (as it does in the Aleutian Islands of Alaska, southeast Alaska, British Columbia, and central California), we might expect that otters playa uniquely important role in the nearshore ecosystem. I use the term ecosystem to specifically include organisms not strictly associated with kelp-<iorninated communities (such as pelagic zooplankton and marine vertebrates), but which may be influenced by the presence or absence of kelps. I will organize my thoughts around three general "models", each of which addresses a particular attribute of kelps.

Habitat Model 193

9.2 Habitat Model

Kelp beds are spatially complex communities. The three-dimensionality they impart to an otherwise two-dimensional habitat, and the primary substrate they provide for benthic and epibenthic organisms are thought to lead to an increase in the abundance, biomass, and diversity of nearshore organisms. The potential for microhabitat specialization is great within a kelp bed in that kelps provide distinctly different habitats both within a single plant (holdfast, stipe, etc.) and at different horizontal and vertical positions within a bed. Thus, even pelagic organisms may be segregated within a canopy level or at the edge or center of a bed. Such regional, intra-bed spatial partitioning may be facilitated by creation of distinct canopy layers (overstory versus understory), or a change in the dominant kelp from the edge to the center of a large bed. However, while microhabitat specialization may be a strong mechanism leading to high diversity within kelp beds, it appears that most organisms observed within beds are facultative rather than obligate associates.

Evidence for the utilization of kelps as primary substrate is clear and abundant. Kelp holdfasts (Andrews 1945 ; Jones 1971; Ghelardi 1971; Moore 1971, 1974; Edwards 1980; Sheppard et al. 1980; Cancino and Santelices 1981) and stipes and blades (Wing and Clendenning 1971; Jones 1971; Bernstein and Jung 1979; Harrold and Pearse, in press) may contain literally tens of thousands of individuals belonging to hundreds of species. The giant kelp Macrocystis pyrifera can produce over 15 m2 of surface area for every square meter of rock substrate (Wing and Clendenning 1971); much of the older portions of this surface may be covered by epiphytic plants and animals. Holdfasts are thought to also serve as nursery areas for organisms which are not necessarily associated with kelps as adults (McKenzie and Moore 1981; Hines 1982). Many of these infaunal and epifaunal organisms utilizing kelp as substrate are probably important links in the nearshore food web.

Pelagic organisms utilize kelp habitats in a somewhat different way, and may be so strongly associated with kelp that some species are rarely observed outside the habitat. Certain shrimp (Lowry et al. 1974), mysids (Clarke 1971) and fish (Quast 1968a; Miller and Geibe11973; Bray and Ebeling 1975, Hobson and Chess 1976;Carr 1983; Larson and DeMartini 1984; Jones 1984a,b; Robertson and Lenanton 1984; Holbrook and Schmitt 1984; Schmitt and Holbrook 1985) are highly correlated with the presence of kelp, but perhaps the strongest evidence comes from experimental (Hixon 1980; Carr 1983; Ebeling and Laur 1985) or natural (Ebeling et al. 1980b, Ebeling and Laur 1985; Ebeling et al. 1985) manipulation of kelp densities, leading to changes in species richness or abundance of associated organisms. The relationship between kelp and fish seems to be particularly strong for juvenile and young-of-the-year individuals (Coyer 1979; Carr 1983; Ebeling and Laur 1985), with some species losing their association as adults. The mechanism most frequently hypothesized to drive the association between fish and structural elements increasing habitat heterogeneity (e.g., kelps) is avoidance of predation (Hubbs et al. 1970; Mitchell and Hunter 1970; Helfman 1981; Carr 1983; Ebeling and Laur 1985). Corroborating evidence that predation (by other fish, birds, and marine mammals) is lower in spatially complex environments such as kelp beds is suggested by work on fish assemblages on artificial platforms (Gooding and Magnuson

194 Kelp Forests and Nearshore Environments

1967; Wickham and Russell 1974), and seagrass beds (VanDolah 1978; Nelson 1979; Coen et al. 1981; Heck and Thoman 1981; Stoner 1982; Shulman 1985). Again, the predation-avoidance mechanism seems particularly important in juvenile fish (see above citations). Some fish associate with kelps in order to feed on organisms that use kelps as primary substrate (Bray and Ebeling 1975; Hixon 1980), or that concentrate along the edge of the bed (Bray 1981).

Several authors discuss resource partitioning when addressing the distribution patterns of organisms within kelp beds. Within guilds of ecologically similar species there is strong evidence for segregation of guild members along depth gradients correlated with kelp canopy height (Hallacher and Roberts 1985), but few studies have empirically demonstrated that such segregation is caused by competition. Hixon (1980) was able to experimentally remove one species member of a fish guild from its microhabitat (a canopy layer) and found that a second species from an adjacent canopy layer expanded in distribution to take its place. Bernstein and Jung (1979) demonstrated that epiphytic species reduce their spatial overlap through mechanisms such as selective larval settlement, and that these species have a clear competitive hierarchy. Spider crabs segregate in different portions of a Macrocystis canopy (Hines 1982), and trochid gastropods at different depths in a kelp bed (Watanabe 1984b), but there was no evidence that competition led to these distribution patterns. Branch (1975) points out that intraspecific spatial partitioning occurs in the limpet Patella compressa, with different aged limpets utilizing different parts of the same plant (Ecklonia maxima).

The habitat model (Fig. 9.1), which predicts that kelps enhance the abundance and diversity of nearshore organisms by producing a distinct and complex physical environment, is certainly the most strongly documented of the three models discussed here. However, the fact that most kelp bed organisms can also be found in comparable habitats without kelps suggests that further work is warranted in order to establish quantitative relationships between these plants and their associates.

9.3 Trophic Model

The trophic model (Fig. 9.2) predicts that very high rates of primary production in kelps lead to enhanced secondary production through four pathways: (1) direct consumption by grazers such as sea urchins; (2) consumption of drift material by benthic herbivores such as abalone and urchins; (3) consumption of particulate detritus (particulate organic carbon, POC) by suspension feeders; and (4) utilization of dissolved organic carbon (DOC) by a wide range of organisms. While intuitively appealing, this model is supported by direct evidence for only the first two of the above pathways; it is the last two pathways that may, in my opinion, turn out to be the most important in understanding the flow of energy through the kelp bed food web.

Kelps are plants capable of high rates of primary production (Duggins 1980; Mann 1982 and papers cited within). It is the fate of this production which is not well known. In some kelp-dominated communities, where an equilibrium exists between plants and herbivores (e.g., in many California Macrocystis beds), a relatively small fraction of macrophyte biomass (less than 10%) may be consumed directly by kelp bed grazers

Trophic Model

KElP 810MASS < Fig. 9.1. Habitat model

KELP PRODUCTION

Fig. 9.2. Trophic model

HABITAT MODEL

SURFACE AREA FOR BENTHIC AND EPIBENTHIC ORGANISMS

195

ABUNDANCE AND DIVERSITY Of NEARSHORE OR6AHISMS

SPATIAL (HABITAT) COMPLEXITY

I. REfUGES 2. MICROHABITIITS 3. RESOURCE PARTITIONING

TROPH I C MODEL

DIRECT UTILIZATION

/

(

I. BENTHIC HERBIVORES "'z. DRIfT fEEDERS ~

SECONDARV PRODUCTION

INDIRECT UTILIZATION /

I. PIIRTlCULiln MllnRIIIL Z. DISSOLVED ORGIINICS

(Miller et al. 1971; Gerard 1976; Newell et al. 1982). Some portion of this consumed plant biomass will find its way into detrital food chains because some grazers such as urchins pass significant amounts of algal material through their guts undigested (Vadas 1977; Santelices et al. 1983). Estimates of the fraction of kelp production consumed by grazers are probably high (and estimates of kelp productivity low) because estimates of total kelp production usually ignore their release of dissolved organic carbon. There is a wide range of estimates for this important parameter (Sieburth 1969; Khailov and Burlakova 1969; Fankboner and de Burgh 1977; Hatcher et al. 1977; Newell et al. 1980); but even if we could agree on a reasonable value for DOC production, the degree to which it is utilized by nearshore organisms is completely unknown. The same is true for particulates (POC). Some estimates of kelp production have taken particulates in to consideration, but as is true for DOC, many questions remain unanswered: (1) Are these fractions retained in or near kelp beds long enough to be consumed there? (2) To what degree do they contribute to growth and reproduction in suspension feeders, particularly relative to phytoplankton, the other autotrophic sources of organic carbon in the system? (3) To what degree are microbial consumers (bacteria, protozoa) necessary


before kelp DOC and POC are of nutritive value to suspension feeders? It seems reasonable that DOC and POC characterize major pathways for kelp production in that both are constantly released by kelps as they grow, and that both are probably end products of the degradation of kelp biomass. When entire plants (or portions thereof) are removed from their point of attachment by storms or grazers, they face three possible fates, two of which involve the recycling of their biomass as POC and DOC. If these plants are exported into deep, offshore waters, they are probably of no further consequence to nearshore organisms; but many of these plants sink to the bottom in the nearshore zone where they are degraded, with some fraction reentering the trophic web as POC and DOC. The third fate for unconsumed drift kelp is to be washed ashore, where very large quantities are frequently observed. Some ofthis beach drift will be consumed by intertidal and supratidal organisms (Robertson and Lucas 1983) which are in turn potentially important components of the nearshore food web. Some unknown and potentially large portion will be degraded on the beaches and return to the sea (as POC and DOC) in runoff or during high spring tides.

Thus it would appear that the hypothesis that kelps are a Significant contributor to nearshore food webs through detritus is an eminently reasonable one. When one considers that in some kelp-dominated communities more than 70% of the consumer standing stock is composed of suspension feeders (Newell et al. 1982), the potential trophic importance of kelp is undeniable. Furthermore, POC and DOC may provide a trophic linkage between kelps and higher level pelagic consumers (such as fish) through suspension feeding zooplankton. The problem is that we have no direct evidence as to how much kelp derived carbon enters higher trophic levels. To date the only approach to this problem has been a modeling one, based on estimation or extrapolation of all the relevant parameters necessary to create an energy budget for the entire nearshore ecosystem (see Newell et al. 1982; Wulff and Field 1983 for the best example of this approach). The difficulties inherent in this approach should be obvious from the above discussion. Most of the numerous critical parameters in such energy flow models are complex, difficult to evaluate, and are undoubtedly highly variable over time and space. A realistic model must consider not only all the pertinent producer and consumer productivities, fluxes, and conversion efficiencies, but also the oceanographic parameters governing the import and export of water from kelp beds. The Wulff and Field (1983) model predicts that kelp POC will contribute (assuming 50% of all particles are consumed and assimilated at 50% efficiency) from about 5% (during periods of downwelling and high water column turnover) to about 70% (upwelling, high turnover) to the diet of nearshore suspension feeders. Even the best of these models are very constrained by their assumptions, estimations, and simplifications.

A very different approach to this problem is being used by C. Simenstad, J. Estes and myself, taking advantage of a unique "natural experiment" and natural biochemical tracers for kelp-derived carbon. In the Aleutian archipelago of Alaska, islands which are separated by only a few hundred miles (and which we hypothesize to be oceanographically and climatologically Similar) differ greatly in the amounts of kelp present in the shallow subtidal (Estes and Palmisano 1974; Estes et al. 1978). Islands with sea otters (Amchitka, Adak) are virtually free of the grazing influence of sea urchins and consequently have extensive and luxuriant kelp beds. Islands without sea otters (Shemya) or with recently established otter populations (Attu) exhibit very large sea urchin pop-

Hydrodynamic Model 197

ulations and significantly reduced subtidal kelp biomass. Above the upper limit of sea urchin grazing, intertidal kelps achieve high biomass in both urchin-dominated and sea otter-kelp dominated islands.

We are directly measuring growth and reproduction in a number of secondary consumer species at all four islands (which presumably differ only in the biomass of kelp available for direct or indirect consumption). We are also transplanting suspension feeders (mussels, barnacles) from a common population (San Juan Islands, Washington) to all four islands, and subsequently monitoring growth and reproduction. If kelp is important, these indicators of secondary production should be considerably higher at Amchitka and Adak. Conversely, if nearshore phytoplankton is not a limiting resource or if kelp biomass (for whatever reason) is not available to nearshore suspension feeders, we would not expect to see differences in growth and reproduction of suspension feeders (or their predators) at the four islands.

In addition, we are utilizing stable carbon isotope techniques (see Fry and Sherr 1984 for a review of these techniques) to determine if the carbon in consumers matches that of kelp or of phytoplankton (in the Aleutians, there is very little contribution of carbon from terrestrial sources). This is possible because these two sources of organic carbon fix the stable isotopes of carbon C 2C and 13C) in different ratios, and because there is little alteration of the ratio as carbon is passed up the food chain (the fractionation that does occur in consumers is predictable and allows extrapolation back to the source). Consequently, if kelp is an important source of carbon, nearshore consumers should reflect kelp stable carbon isotope values where kelp is available (islands with otters), and phytoplankton values where kelp is not available (islands without otters). If isotope signatures in secondary consumers reflect kelp values at all four islands, we will conclude that kelp production (primarily from low intertidal plants), even in the presence of subtidal grazers, is sufficient to dominate nearshore secondary production.

By coupling stable carbon isotope analyses with direct measures of consumer growth and reproduction, we hope ultimately to address the roles of kelp, sea urchins, and sea otters in governing nearshore secondary production. We believe these measurements will, for the first time, provide direct qualitative and quantitative evidence for the trophic role played by kelps in nearshore ecosystems.

9.4 Hydrodynamic Model

Because of their high productivities, large size, and frequently high densities, kelps can generate very large standing stocks. The projection of kelp biomass into the water column will extract momentum from moving water via drag, and consequently alter turbulent flow patterns in the nearshore region. The biological ramifications of this hydrodynamic influence are potentially very important to a wide range of nearshore organisms. While a number of studies have examined flow regime as a factor influencing algal morphology and physiology (Charters et al. 1969, 1972; Neushul 1972; Steneck and Adey 1976; littler 1979; Wheeler 1980; Norton et al.1982), very little is known about the reciprocal effect that large benthic algae exert on their fluid medium. Jackson


(1984) and Jackson and Winant (1983) reported on the effects of Macrocystis forests on both long-shore and cross-shore components of currents off Point Lorna, California. They further examined kelp effects on these components within discrete frequency bands: a low frequency « 0.6 cycles/day) band occupied primarily by wind-driven, long-shore, barotropic currents; a medium frequency (0.6- 6 cycles/day) band occupied primarily by diurnal and semi-diurnal rhythms; and a high frequency (> 6 cycles/day) band occupied primarily by internal waves. They found that higher frequency currents in both the long-shore and cross-shore directions are strongly attenuated within kelp stands, whereas semi-diurnal and diurnal currents are still detectable within the forest's interior. The low frequency, long-shore currents are deflected in bulk about the kelp forest. Thus kelp forests on the open coast of southern California preferentially pass diurnal and semi-diurnal signals. Although they did not study the effects of winddriven gravity waves, Jackson (pers. commun.) believes that higher frequency "chop" will be damped through Macrocystis forests far more effectively than will relatively lower frequency swell (periods of approximately 10- 20 s).

Indirect support for the potentially important hydrodynamic role of kelps comes from several sources. Seagrass beds and arrays of animal tubes projecting above the substrate are known to reduce near-bottom water velocities (Fonseca et al. 1982; Eckman 1983; Peterson et al. 1984), with significant biological ramifications. Also, a considerable literature exists from terrestrial studies on the effects of crop plants and forests on wind velocities and particle transport (Raynor et al. 1974; Petit et al. 1976). The hydrodynamic effects of prostrate and stipitate (sensu Dayton et al. 1984) understory kelps may be particularly important in that these plants are capable of creating very dense canopies just above the substrate which may act to retain sub canopy water, and greatly reduce sub canopy flushing rates.

The hydrodynamic model (Fig. 9.3) predicts that the disruption of flow by structural arrays, such as a kelp bed, will have Significant effects on biological parameters that are influenced by water movement such as feeding and growth (in suspension and deposit feeders), and dispersal and recruitment. This prediction arises from the assump-

HYDRODYNAM I C MODEL

REDUCED LONGSHORE ~ <CURRENT FLOW PARTICULATE

KELP BIOMASS flUXES (SURFACE AREA) / +

SEDIMENTATION

REDUCED WATER TURBULENCE

BIOLOGICAL RAMIFICATIONS

I. SUSPENSION FEEDING 2. DISPERSAL 3. RECRUITMENT

Fig. 9.3. Hydrodynamic model

Hydrodynamic Model 199

tion that, for the most part, food and dispersal stages of many kelp bed organisms (including dispersal stages of the kelps themselves) are passively dispersed, and their transport and settling characteristics will be determined in large part by the movement of the fluid medium in which they are suspended. There are, however, few empirical data to allow us to predict precisely what these effects might be.

The retention of water within a kelp bed and the reduction of turbulence by kelps might be expected to bring more particles into contact with benthic suspension feeders as these particles fallout of suspension. On the other hand, Sebens (1984) has demonstrated that the suspension-feeding octo coral Alcyonium siderium grows more rapidly in high current-turbulence regimes, probably due to higher encounter rates with suspended particles in a high flow environment. For hard-substrate organisms living on horizontal surfaces, reduction in flow could ultimately lead to sedimentation rates so high as to actually interfere with feeding, and accumulation of sufficient sediment to interfere with recruitment. Eckman (in prep.) and Peterson et al. (1984) have both demonstrated that eelgrass (Zostera marina) reduces current velocities relative to adjacent control areas. However, in Eckman's study, growth of juvenile Bay scallops (Argopecten irradians) attached by byssal threads to eelgrass blades was reduced, while in Peterson's study growth of the infaunal bivalve Mercenaria mercenaria was increased in areas of reduced current.

The hydrodynamic effects of kelps on recruitment and dispersal are also difficult to predict. Kelp beds may act to retain larvae released within the bed, and the strong deceleration of flow at the margins of the bed could facilitate settlement of larvae imported from outside the bed. Thus Bernstein and Jung (1979) hypothesize that the restriction of Membranipora membranacea to the outer edge of large Macrocystis beds results from poor penetration of currents carrying bryozoan larvae into the bed. The concentration of zooplankton at the upcurrent edge of a Macrocystis bed, and the consequent higher densities and feeding rates of fish in this area (Bray 1981) are probably the result of alterations of current flow by kelps.

The complex relationship between kelp standing stock (or perhaps more appropriately surface area), water flow, and particulate fluxes is the subject of a research effort by J. Eckman and myself at the University of Washington's Friday Harbor Laboratories. We are restricting our efforts to kelp beds composed entirely of understory (stipitate) plants, and are performing field measurements and experiments in three current and two exposure regimes. Our approach is to measure fluid and particulate fluxes at replicated sites where we are simultaneously manipulating kelp density. Concurrently we are measuring growth and reproductive output of organisms transplanted into the manipulated beds, and such physical parameters as sediment accumulation as a function of kelp density. Recruitment and dispersal phenomena are being examined at these sites by (1) following development on settling plates, and (2) measuring the flux of neutrally buoyant particles (propagule mimics) into beds (release of particles upstream of the bed) and out of beds (release within the bed). These experiments should elucidate the physical and biological ramifications of kelp-mediated hydrodynamics.

The hydrodynamic model is the most speculative of the three models discussed here, but could prove to be as significant to the nearshore ecosystem as the habitat and trophic models. With the vast majority of organisms living in kelp beds being suspension feeders, and with many kelp bed organisms dispersing passively (Hannan 1984), any


physical obstruction that reduces fluid momentum is likely to have wide-ranging biological effects.

9.5 Discussion

The reintroduction and expansion of sea otter populations along the west coast of North America will undoubtedly lead to an increase in kelp biomass in the nearshore region (Estes et al. 1978; Duggins 1980; Breen et al. 1982; Foster and Schiel 1985; Harrold and Pearse, in press). By removing herbivorous sea urchins, sea otters remove a major (although not exclusive) source of kelp mortality. Even in areas where urchins are not observed to be feeding extensively on large plants, their scraping of the rock surface and the consequent removal of kelp gametophytes and small sporophytes are likely to inhibit the development of kelp stands. Given the previous discussion, it should be apparent that the expansion of existing kelp beds and the creation of new ones may have ecological ramifications that affect a wide range of organisms through direct and indirect interactions.

The concept of a "keystone predator", a term frequently applied to sea otters, usually relates to species diversity at prey trophic levels. If the ultimate consequence of sea otter predation is a dramatic increase in macrophyte productivity and biomass, and if the models discussed above prove to be correct, the consequences of sea otter predation will go far beyond production and diversity at the prey (herbivore) trophic level. Actually, diversity in the grazer guild may be little affected in that in some kelp-dominated communities there is no evidence that grazer species are competing (Duggins 1981b). Interestingly, diversity at the primary producer level may actually be decreased, as reduction in grazing pressure may lead to competitive dominance of a single kelp species (Duggins 1980). Alternatively, sources of disturbance other than grazing can create patches in kelp beds which are rapidly invaded by opportunistic algal species, leading to an overall increase in diversity (on a large scale) relative to areas dominated and overgrazed by herbivores.

A broader consequence of otter predation is likely to be manifested outside the kelp-grazer-otter food chain. These manifestations are achieved through the mechanisms discussed above and ultimately may affect the entire nearshore ecosystem. These secondary and tertiary effects must be evaluated cautiously in that for two of the three models discussed above, direct evidence bearing on the extent of the effect is lacking. Only the habitat model has been conclusively demonstrated, but its predictions alone give credence to the important consequences of sea otter predation. The trophic model, while not presently supported by direct evidence, implies that the basis for the entire nearshore food web may be altered by the activity of sea otters. In the absence of otters (reduced kelp biomass), nearshore primary productivity will be dominated by planktonic autotrophs; in the presence of otters, it will be dominated by benthic macrophytes through detritus-based food chains. The consequence of kelp production to production at higher trophic levels will depend on many variables, and ultimately on whether or not phytoplankton (in the absence of kelps) is a limiting resource. The hydrodynamic model is so poorly understood that prediction of the nature, much less

Summary 201

the extent, of kelp-mediated hydrodynamic influence is difficult. Existing data on the physical influence (fluid and particulate fluxes) are intriguing, but too preliminary to even formulate solid predictions of the biological ramifications.

The picture is complicated by the possibility that the mechanisms operating in the three models may act synergistically or antagonistically with one another in determining ecological consequences. For example, the trophic model predicts that kelp-derived detritus is an important food source for kelp bed suspension feeders, and the hydrodynamic model predicts that this particulate material may be held within the bed and within "reach" of these organisms. In this case, the mechanisms may act synergistically to enhance production in these suspension feeders. On the other hand, the habitat model predicts that certain planktivorous fish may be more common in kelp beds, where they have a refuge from their predators. In this case, the ultimate consequence of kelp abundance to suspension-feeding zooplankton would have to be weighed in light of opposing processes (increased food and increased predation). The observation by Gaines and Roughgarden (1985) that many larvae are "filtered" by fringing kelp beds may be the result of predation on them by kelp bed fishes (as in the habitat model) or of the disruption of flow that transports larvae across the bed (as in the hydrodynamic model). The implication of these models is that the reintroduction of sea otters and the consequent expansion of kelp beds is likely to have considerable and wideranging impact - but we are just beginning to understand just what or how large that impact is likely to be.

9.6 Summary

With the expansion of sea otter populations all along the west coast of North America, a Significant increase in the standing stock of subtidal kelps is likely to occur. The importance of kelp beds to nearshore secondary production is discussed in light of three, largely hypothetical models. The habitat model predicts that kelp beds provide unique, structurally complex habitats, leading directly to increases in the density and standing stock of benthic, epibenthic and pelagic consumers. The trophic model predicts that kelp production fuels nearshore food webs through direct and indirect (detrital) pathways. The hydrodynamic model predicts that kelp will significantly affect the flow and turbulence of water within and adjacent to kelp beds, with wide-ranging biological ramifications.

The degree to which these models are supported by data varies, with the habitat model being the best supported and the hydrodynamic model the most speculative.

10 Sea Otters and Nearshore Benthic Communities: A Theoretical Perspective S. A. LEVIN

10.1 Introduction: Regulatory Issues

The central problem in environmental management is the balancing of competing societal values. Virtually any proposed management plan necessarily involves actions that will be viewed positively by some, negatively by others, and with mixed feelings perhaps by the majority. The management of the California sea otter population represents a classic example of such conflict, one in which multiple uses of the ecosystem are at odds and in which the diverse interests of different segments of society must be accommodated equitably.

Those concerned with protection of the otter view it as endangered by oil exploration and other factors, and vulnerable to potential spills. Thus, as is often the case in environmental decision-making, a conflict exists between the protection of a natural population and the development of a valuable economic resource. In response to the perceived threat of local catastrophe owing to an oil spill, the U.S. Fish and Wildlife Service has proposed to introduce a second population of the otter at San Nicolas Island (U.S. Fish and Wildlife Service 1986). This proposal, in tum, raises concerns for other segments of society. The otter is viewed by shellfishermen as a competitor for a common resource; moreover, the otter's role as a top predator could lead to fundamental alterations in the community structure and dynamics of the nearshore ecosystem, engendering changes in offshore kelp beds and associated shellfisheries and finfisheries (Estes and VanBlaricom 1985; Laur et al., this Vol.; Duggins, this Vol.).

Were the scientific facts clear, evaluation of the appropriateness of the proposed action would be difficult enough; but the dilemma could be resolved on the basis of legal arguments and the perceived wishes of society. Unfortunately, as is generally the case in environmental management, there is considerable uncertainty associated with prediction. In such a situation, the responsibility of the scientist is to make explicit the range of possible outcomes and the level of uncertainty involved, and thus to provide the scientific foundation for decision-making (Limburg et al. 1986a). Ultimately, the conflict cannot be resolved on purely scientific grounds, but must involve decisions by regulators regarding what weights to assign to different societal interests and to possible inconsistent scientific claims. The determination of those weights must involve interpretation of what the laws and statutes mandate, and the degree to which science permits the quantification of likely effects.

The sea otter also provides a textbook example for illustrating basic ecological principles. In some of the ecosystems in which it occurs, the otter has been shown to

Ecological and Regulatory Parallels 203

be a keystone species, (sensu Paine 1966), playing a pivotal role in the maintenance of the structure of the community (Estes and Palmisano 1974; Duggins 1980, this Vol.; Laur et al., this Vol.; Kvitek and Oliver, this Vol.). The otter, as a major predator of sea urchins and other herbivores, modifies competition and coexistence among them; this can lead to qualitative changes in the nature of the kelp bed, which potentially is a dynamic patchwork of successional stages (Ebeling and Laur, this Vol.). The variability of the kelp bed in space and time is of major importance to associated invertebrates and to fish populations, which minimize the costs of variability by foraging widely over the spatio-temporal mosaic (Ebeling and Laur, this Vol.; Laur et al., this Vol.). The introduction of the otter, by modifying grazing pressure on the kelp bed, could alter the structure of this mosaic, shifting it toward larger kelps and changing the composition of the associated faunal community. How should one evaluate the importance of such potential changes, which could shift the harvestable resource base from shellfish to kelp and finfish?

The above discussion makes clear that the major issues to be confronted in evaluating alternative management strategies for the California sea otter relate to the tradeoffs among preservation of an endangered species in the face of significant environmental risk; the maintenance of the health of the ecosystem, however this is to be measured; and economic issues involving existing and potential shell- and fin fisheries , kelp harvesting, and the exploration and exploitation of oil resources. One must balance a view of the system as a natural environment, whose indigenous popUlations must be maintained, against one in which exploitation of resources assumes primacy, and in which the major consequences of perturbations are evaluated in terms of shifts in the relative availability of various economic resources.

The particular combination of issues bearing on this management decision may be unique, but the general considerations are not. Environmental legislation often is phrased in vague terms, in order to assure its broad applicability (Limburg et al. 1986a). The Marine Mammal Protection Act of 1972, as amended, mandates the maintenance of the optimum sustainable population consistent with the carrying capacity of the habitat and maintenance of the health and stability of the ecosystem. However, terms such as ecosystem health and stability, carrying capacity, and optimum sustainable population do not translate easily into quantitative endpoints for regulation. Similar difficulties exist in other applications of environmental legislation, in which, for example, phrases like balanced indigenous population and significant environmental degradation must be made operational. The quantification of such terms is the basic challenge facing the applied ecologist, and thus the dilemma presented by management ofthe otter population is neither trivial nor isolated.

10.2 Ecological and RegUlatory Parallels

One of the most important developments in the study of ecological communities over the past 1 ° to 15 years has been the recognition of the importance of spatial and temporal variability, and the relationship of variability to the scale of observation. Long ago, Gleason (1917, 1926) emphasized the importance of stochastic phenomena in

204 A Theoretical Perspective

community dynamics; and Watt (1947), in his prescient presidential address to the British Ecological Society, developed the essential elements of the theory of gap-phase phenomena in forest communities. Regrettably, the insights that were contained in that paper were lost for decades on most community theorists, who remained obsessed with the monotonous homogeneous asymptotics of Lotka-Volterra theory and its generalizations (see for example Levin 1970). The revitalization of the study of "patch dynamics" came from a variety of sources; experimental and theoretical studies in the rocky intertidal (Paine 1966; Levin and Paine 1974; Paine and Levin 1981) coral reefs (Connell 1978), prairies (Platt 1975), and tropical and temperate forests (Connell 1978 ; Bormann and Likens 1979; Whittaker and Levin 1977; Runkle 1979). The study of patch dynamics, and associated developments in landscape ecology, now represent the cutting edge of research in community theory, and have been discussed in depth in a variety of recent books and reports (pickett and White 1985; Forman and Godron 1986; Risser et al. 1984).

This general theory has great relevance to the present debate, especially considering that rocky shores provided the major stimulus to the recognition of the importance of localized disturbance in structuring communities and in maintaining species diversity. Paine (1966) demonstrated that predation by the carnivorous sea starPisaster ochraceus could reduce levels of the dominant competitor Mytilus californiflnus sufficiently that a wide variety of other species could coexist; otters can perform the same function in other locales (VanBlaricom, this Vol.; also see discussion in Duggins, this Vol.). Other studies (Levin and Paine 1974; Paine and Levin 1981) documented the similar role played by physical disturbance, which by removing the dominant competitor locally reinitiates successional sequences, leading to the maintenance of a spectrum of successional stages. In the same way, sea urchins can foster the maintenance of particular successional mosaics of algal species by modifying competition, and otters can influence the structure of the entire community through their preferential feeding on urchins. A fundamental objective must be the quantification of patchiness in these systems, and the elucidation of its major determinants (Estes and Harrold, this Vol.).

In addition to these ecological parallels, we can learn also from a variety of management experiences, especially those concerned with estuaries and the coastal zone. In the case of environmental management applied to the Hudson River, for example, similar issues had to be confronted, although the relevant legislation was different (Limburg et al. 1986a,b). To a large extent, in many of the controversies on the Hudson, the health of the ecosystem came to be measured in terms of the health of the major fisheries, especially the striped bass. More or less by mutual agreement among the various litigants, the fisheries became the battleground, and arguments focused upon effects on fish populations. From an ecosystem perspective, this was somewhat disconcerting, because ecosystem effects were not considered directly and the relations between ecosystem changes and the health of the fisheries were not well understood; but from a pragmatic management perspective, some such agreed-upon surrogate for ecosystem health is essential in transforming a socioeconomic battle into a quasi-scientific debate, and in the determination of workable regulatory endpoints.

In estuarine and near-shore systems in general, similar issues must be confronted. How is the health of the community to be assessed? What are appropriate regulatory endpoints? Which species are of most value? What critical roles are performed by such

Measures of Ecosystem Health 205

system components as aquatic vegetation and the benthic community? The increasing sophistication of ecosystem science has made clear the inadequacy of studying populations without attention to their ecosystem context (Levin and Kimball 1984). Environmental managers recognize the importance of submerged aquatic vegetation, the benthic community, and adjacent wetlands as structural elements, as basic habitat and food source for species of direct interest, and as processors of materials, including wastes (Limburg et al. 1986a). All of these elements are present in the debate at hand (see Duggins, this Vol.; VanBlaricom, this Vol.). For example, kelp communities serve as structural elements, buffering the ecosystem by attenuating high frequency waves (Duggins, this Vol.); as substrate and cover for organisms, especially in the absence of rocky relief (Ebeling and Laur, this Vol.); directly and indirectly as sources of food and nutrients; and as uptake systems for toxicants. Thus, it is clear that major challenges exist for environmental scientists in determining direct or indirect measures of ecosystem health, in finding ways to measure and anticipate effects, and in developing understanding of the critical factors and elements maintaining ecosystem integrity. Whereas the general issues can be discussed in general terms, environmental management cannot proceed until these considerations are made specific to the systems and stresses of in terest.

10.3 Measures of Ecosystem Health

As already discussed, the assessment of ecosystem effects of proposed actions must be predicated upon accepted measures of system health and value. Primary decisions, therefore, involve the definition of the system: What measures are to be used? What is the reference behavior of the system, against which the effects of perturbations are to be compared? Just as in any geopolitical border dispute, one must decide to what extent the past is prologue to the future. Is the system to be maintained as it is now; as it was before the otters began to reoccupy habitats from which they had been extirpated; as it was before the otters were exploited (if that could be determined); or for maximum socio-economic benefit? What baseline is to be accepted?

In deciding upon appropriate measures of ecosystem health, one must make decisions a priori concerning the importance to be ascribed to the otter population per se, to the kelp beds that might be affected substantially by otter densities, to the shellfisheries, and to present and potential finfisheries. It is important to recognize that such definitional decisions are not primarily scientific, but involve a balancing of the rights and interests of different sectors of society.

More subtle is the issue of the determination of the status of anyone of the elements described above, and detection of important changes. One needs to determine the temporal context, and to resolve associated questions of spatial scale. The kelp and associated invertebrate communities are patchy, diverse, and successional (Dayton 1985a); both otter predation on grazers and urchin grazing on algae can contribute directly to this diversity and ultimately to the stability of the system (Estes and Harrold, this Vol.).

Diversity, expressed in spatiotemporal heterogeneity, most generally has stabilizing effects, tending to reduce fluctuations. There is no mandate to maintain the system in


a state of maximal diversity; indeed, that objective well might require active intervention to displace the system from its "natural" state. It might be foolhardy, however, to tamper substantially with system heterogeneity. When the importance of heterogeneity and diversity has been ignored in other systems, as in the case of forest fire suppression, the price has been high in terms of the loss of system stability and productivity. Thus, in characterizing the system's reference behavior, we must find ways to measure patchiness, and management strategies to maintain it (Estes and Harrold, this VoL).

10.4 Risk Assessment: Predicting Fate, Transport, and Effects of Otters

Once measures of system health have been determined, and reference behavior established, one can turn to the problem of prediction, the starting point for risk assessment and management. As in the case of the release into the environment of chemicals (Levin and Kimball 1984) or of genetically engineered organisms (Gillett et al. 1986; Levin and Harwell 1986), the first step in assessment is evaluation of the fate and transport of the released material; specifically, where does it go? To what extent can we predict popUlation growth and spread?

Although prediction of otter growth and spread must remain somewhat problematical, some general patterns do emerge (Lubina and Levin, in press). Conventional theory suggests that populations spreading within homogeneous habitat should establish fronts that advance at a constant rate within any homogeneous habitat; that rate is twice the square root of the product of the population's intrinsic rate of natural increase (r) and its diffusion coefficient (D). Thus, within any particular habitat type, one should expect constant patterns of advance; and at the borders between habitats, one should expect discontinuities in rates of spread. That is exactly what was observed for the otter until recent years, and the observed changes in rates of advance seem explicable in terms of what is known about the various habitats. More recently, patterns have broken down somewhat, perhaps due to increased incidental take in fishing nets (Estes and Jameson 1983a; U.S. Fish and Wildlife Service, 1986; Jameson and Estes, unpubl. data; Cal. Dept. of Fish and Game, unpubl. data). What is needed for effective prediction is more accurate information concerning habitat dependent birth rates, mortality (both natural and unnatural), and movement (Lubina and Levin, in press).

Solving the problem of predicting rates of spread would not guarantee a capability to predict population growth; it is possible to give estimates of the likely range expansion of the otter without knowing what the densities will be within that range. The problem is that range expansion is governed largely by linear phenomena, in particular by growth at low densities. Thus, estimation of (r) is central, and in theory can be done by using the standard methods of life table analysis. For example, Kenyon (1969) estimated that sea otter populations increased at about 15% per year in the years immediately following the cessation of hunting, when the population began to recover; and Estes (1981) estimated that, both in California and in Alaska, there are 0.2 to 03 pups produced per adult per year, providing an upper bound consistent with Kenyon's estimate. For the California population, Lubina and Levin (in press) estimate that (r) is about 0.056 (5.6% per year).

Risk Assessment: Predicting Fate, Transport, and Effects of Otters 207

Prediction of population densities, on the other hand, raises harder and nonlinear problems: understanding of the influence of densities upon recruitment, and estimation of the carrying capacity of the environment. It is well known (May 1973) that even the simplest nonlinear growth models can exhibit erratic ( chaotic) behavior, and that this phenomenon can play havoc with parameter estimation and with population growth prediction. Thus, even in fisheries, in which the problem of prediction has been the focus of research for decades, long-term prediction remains an elusive target. Typically, short-term prediction is more reliable, and effective management couples such shortterm prediction with monitoring and adaptive management.

The problem of predicting indirect effects on other species and on the ecosystem is, of course, the most difficult of all. Ecological community theory rarely is mistaken for being a predictive science, because the number and complexity of possible interactions makes impossible the task of predicting possible effects on all system components. The agricultural literature is full of examples in which effects of pesticides upon nontarget species have led to consequences opposite in direction from what was intended; horror stories associated with the indirect effects of the application of chemicals such as DDT to control pest populations can be traced to complex chains of influence that cause effects to appear at points far removed in the trophic network from the point of application (Levin and Kimball 1984). In practice, these complex causal pathways are difficult to predict.

The nearshore ecosystems of the California coast are not simple ecosystems, so complicated pathways of influence are both possible and likely. Otters feed on turban snails, octopus, urchins, crabs, abalone, and fish, among other species, and thus can modify the competitive balance among them (Lowry and Pearse 1973; Riedman and Estes, this Vol.; Kvitek and Oliver, this Vol.). Urchins are major consumers of the kelp beds, so that otter reduction of urchin densities indirectly benefits abalone and fish populations, despite the fact that both also are consumed by otters. When one popUlation affects another in opposite directions along different pathways of influence, the problem of determining the net influence becomes very difficult, and cannot be resolved without good quantitative estimates of the various influences. Few would claim that otters will benefit abalone populations; but the net effect on finfish populations likely will be positive, especially given that finfish are a minor component of the diets of California otters (Ebert 1968; Ostfeld 1982; Laur et al., this Vol.). Similarly, the effects of urchins upon the kelp beds are not described simply. Moderate grazing by urchins favors the growth of understory plants, but heavy grazing leads to an overall reduction in algal cover (Foster and Schiel 1985 ; Dayton 1985a).

In general, to the extent that community theory supports defensible predictions, it does so by simplifying the description of communities; that is, by identifying the critical factors controlling community structure. In practice, this often involves the identification of keystone or trigger species or groups of species (Paine 1966); the compartmentalization of the community into tightly interacting modules, groups of species or populations that interact strongly with one another, but weakly with other modules (Paine 1980); the elucidation of patch structure (Levin and Paine 1974; Paine and Levin 1981); and the implementation of other devices for discovering the hierarchical control of community and ecosystem dynamics. Thus, it is essential to discover the role of the otter in controlling community structure and ecosystem dynamics, its role in influencing patchiness, and its most tightly linked competitors and prey.


As the other chapters in this Volume make clear, however, precise prediction of the potential ecosystem level effects of an introduction of sea otters would be very difficult. The best tool available in determining the otter's role is extrapolation from other systems, but this is not as easy as it may sound. Without doubt, otters are keystone species in some systems; but it is not clear to what extent those patterns are general, and thus the basis for extrapolation is weakened (Foster and Schiel, this Vol.). The situation is complicated considerably by the fact that otters have labile, opportunistic foraging patterns; moreover, their effects are influenced by the presence of other disturbing agents, such as starfish and storms, and by other system characteristics (Riedman and Estes, this Vol.; Kvitek and Oliver, this Vol.).

Even the effects of disturbance on diversity are equivocal. As already mentioned, moderate grazing favors understory development and increases diversity, whereas heavy grazing leads to reduction in diversity. This can have important consequences for the fish populations that exploit the kelp mosaic, but it is not possible to state in general what those consequences will be.

The general problem of uncertainty in prediction is a frustrating one. Certainly, steps can be taken to reduce that uncertainty, and the program proposed by Estes and Harrold (this Vol.) represents a hopeful step in that direction; but it must be recognized that, as for almost any environmental management strategy, there will be uncertainty. This uncertainty by itself should not be a barrier to action. One must proceed on the basis of the best available information, but expect surprises (Holling, in press). Thus, any management action should have associated with it a plan for monitoring, and the flexibility to modify the management activities (Holling, in press; Walters 1986).

10.5 Summary

The proposal to release sea otters at San Nicolas Island is a classic example of the kinds of decisions typically faced in environmental management of multiple-use ecosystems, in which multiple interests of different sectors of society create conflicts and introduce multiple objectives. The problems of risk assessment and risk management involve decision-making under uncertainty. In such situations, we know that there are no easy scientific answers, because societal values are in competition. The scientist is entitled to an opinion on such a proposed action; but in presenting his conclusions to regulators or to the public, he must endeavor to separate issues of science from issues of personal taste and preference.

The role of the scientist is to reduce uncertainty to the greatest extent possible by experiments such as are described by Estes and Harrold (this Vol.), by drawing on other scientific experiences as appropriate, and by developing and applying general theory. In making predictions, one must not fall prey to the temptation to sweep uncertainty under the rug in a misguided attempt to give the regulator and the public the impression of certitude that they may seek. Rather, in any prediction, one must state the level of uncertainty clearly, through the presentation of ranges and probability distributions of possible outcomes. Ultimately, the decisions in such situations cannot be made on scientific grounds alone, and the job of the scientist is to provide the decision-maker and the public with the best available scientific information.

Summary 209

Acknowledgments. This is publication ERC-126 of the Ecosystems Research Center, Cornell University. This research was supported by U.S. Environmental Protection Agency Cooperative Agreement CR-811060 with Cornell University, and by NSF Grant DMS-8406472. The ideas and conclusions expressed in this paper are those of the author, and do not necessarily represent the views of the Environmental Protection Agency or the National Science Foundation.

The author acknowledges many helpful comments by Daniel Goodman, and by the editors.


J. A. ESTES and G. R. VANBLARICOM

11.1 Introduction

In organizing this symposium on the community ecology of sea otters, we hoped to accomplish two results. One was that the important scientific questions concerning direct and indirect influences of sea otter predation on nearshore communities would surface and be discussed. The other was that needs and directions for future research would be identified. Here we seek to integrate and discuss the major points raised for us in the symposium. We divide these concluding remarks into four sections. The first and longest (Sect. 11.2) will emphasize generalities. What are the similarities and differences in the otter's influence among rocky subtidal, rocky intertidal, and soft-bottom communities? Second (Sect. 11.3), we discuss variation in the structure and organization of nearshore communities within the historical range of the sea otter. Rocky subtidal communities will be emphasized in this section because (1) there have been enough studies in these systems to allow such treatment, and (2) these results are the focus of considerable controversy. The third section (Sect. 11.4) discusses directions for future research that have been recommended by the contributors to this Volume. Finally (Sect. 11.5), we venture briefly into the realm of philosophy to examine how some of the issues raised in this volume apply more broadly to community ecology, and to science in general.

11.2 Patterns, Processes and Paradigms in Communities Occupied by Sea Otters - A View Among Systems, and Through Space and Time

The paradigms for community-level interactions involving the sea otter bear little similarity among rocky subtidal, rocky intertidal and soft-bottom systems. As stated by Levin (this Vol.), the sea otter is a "textbook example" of a keystone species, "playing a pivotal role in the maintenance of structure in the community ... ". This view is based largely on rocky subtidal communities, in which sea otters are thought to maintain kelp beds in some areas by limiting sea urchins (Chapts. 6, 7 and 8, this Vol.). The so-called "keystone" role of sea otters in rocky subtidal systems emanates from the observation that food webs in such systems appear to be "strongly linked" (sensu Paine 1980). Thus, sea otters effectively limit sea urchins; uncontrolled populations of sea urchins limit kelp and other fleshy macro algae ; and kelp beds have a variety of important effects on other species in the system (Duggins, this Vol.).

Patterns, Processes and Paradigms in Communities Occupied by Sea Otters 211

If, for the moment, we accept this paradigm, then in what ways and for what reasons does it hold for rocky intertidal and soft-bottom systems? In each system the direct effects of otter predation are similar (although as will be discussed shortly, they vary in certain details) in that sea otters limit prey densities (and in most instances, sizes) to levels well below those that would occur in the absence of sea otters. However, the similarities end here, as far as we know. Plant/herbivore interactions are mediated by sea otter predation in rocky subtidal systems, and when the herbivores are uncontrolled, they may be limited by food (plant biomass). In contrast, space is the most important limiting resource to prey populations (including mussels) in rocky intertidal systems, so indirect community-level effects of sea otter predation in these systems probably are mediated by space competition. Free space allows the establishment of competitively subordinate species such as turfy algae and molluscan herbivores (Paine 1984), although the various consequences of space provision are most likely quite different from those that are related to the presence or absence of kelp. There is evidence from one study (Sousa 1984a) that gap size influences the extent to which herbivorous limpets can limit the recruitment and growth of algae. However, the breadth of effects resulting from the provision of space by sea otters in a mussel bed probably is less than that caused by the presence of kelp on a rocky subtidal reef. Indeed, VanBlaricom (this Vol.) argues that aside from perhaps adding temporal consistency to the frequency of gap creation in Mytilus californianus beds, the consequences of sea otter predation are not fundamentally different from those of wave shear.

Species interactions that emanate from prey limitation by sea otters in soft-bottom systems are the most poorly known, although Kvitek and Oliver (this Vol.) list some intriguing possibilities. For example, the sea otter's various prey in soft-bottom systems probably have several community-level functions, although in contrast to both rocky subtidal and rocky intertidal systems, the limiting resources to undisturbed populations of these prey are unknown. Decapods may be important consumers themselves, so that sea otter predation on crabs could be functionally similar to predation on urchins, except that the crab species in question are predators or scavengers rather than herbivores. Thus, the effect of sea otter predation on crabs in soft-sediment systems may have less severe influences on the base of the food web.

Resources or processes that limit undisturbed populations of infaunal bivalves are not presently known because it is difficult to work with these species in situ without disrupting their habitat. However, sea otters may uniquely influence these systems by excavating sediment. Sediment disruption that occurs while digging for clams disturbs the habitat of the smaller infauna. Pits created by digging in the soft sediment may trap detritus, thus leading to the attraction of detritivores and their predators. The deposition of shells on the bottom may create the solid substrate required for settlement by some organisms. Although none of these potential effects has yet been demonstrated for communities occuppied by sea otters, all are known in related instances. If they also result from sea otter predation, then the otter's role in soft-sediment systems is likely to be different from those proposed for rocky subtidal and intertidal communities.

A unifying theme among these three systems in the community ecology of sea otters is that otters can limit their prey populations in each. Beyond the reduction of prey numbers and biomass, a recurrent pattern seems to be the selection of larger in-


dividuals. This has been reported for sea urchins and abalones in rocky subtidal systems (Lowry and Pearse 1973; Estes et al. 1978; Hines and Pearse 1982), mussels in rocky intertidal systems (VanBlaricom, this Vol.), and crabs and certain clam species in softsediment systems (Kvitek and Oliver, this Vol.). We note that for urchins and mussels at least, size selection appears to be absolute rather than species-specific. The maximum shell length of mussels, and the maximum test diameter of urchins is about 35 - 40 mm in populations that are heavily exploited by sea otters. This is true for bay mussels in Prince William Sound (VanBlaricom, this Vol.), green urchins in the Aleutian Islands (Estes et al. 1978; Simenstad et al. 1978), and Prince William Sound (Irons et al. 1984; VanBlaricom, unpubl.), and purple urchins in California (M. Kenner, unpubl.). Size selection may optimize energy gain because digestible tissue weight varies immensely over the size range of most prey, due to exponential relationships between exoskeletal lengths and internal volumes. However, the demographic and community consequences of size-selective predation may vary among systems and areas. Based on the length of sea urchin demipyramids in otter spraints, otters seldom consume individuals < about 25 mm test diameter (J .A. Estes, unpubl.), thus perhaps providing a refuge from predation in smaller-sized individuals. Such a demographic effect may be further enhanced for green sea urchins at many sites in the western Aleutian Islands by what appears to be a predictable annual recruitment (J .A. Estes, unpubl.) and the fact that small individuals are left largely undisturbed in situ. The co-occurrence at Amchitka Island of a high density sea otter population and high densities of small sea urchins is probably explainable in large part by these two processes. In contrast, red and purple sea urchins recruit episodically along much of the west coast of North America (Ebert 1983), apparently because of larval transport offshore in upwelled areas (Ebert and Russell, unpubl.) , and between recruitment events, sea urchins are comparably rare in areas inhabited by sea otters, except within substratum crevices and other cryptic habitats (Lowry and Pearse 1973).

The demographic consequences of size-selective predation on mussels may be different still. Whereas recruitment of bay mussels in many areas is more frequen t than it is for red and purple sea urchins (Suchanek 1981, 1985), sea otters remove clumps of mussels, discarding the smaller individuals in areas where the probability of their survival and return to the intertidal zone is almost nil (VanBlaricom, this Vol.). Thus, sea otters may selectively consume large mussels, while effectively removing all size classes. Where conditions are suitable, this combination of processes can cause substantial reductions in mussel populations.

From the perspective of a foraging sea otter, the excavation of deep-burrowing prey in soft-sediment systems probably adds time and energy costs not incurred in rocky subtidal or intertidal communities. Kvitek and Oliver (this Vol.) suggest that such costs provide a partial refuge from predation for deep-burrowing species, with the consequences that: (1) population reductions from sea otter predation may require decades, if they occur at all, in contrast to the months or years required for similar reductions to occur in shallow burrowing or surface dwelling species; and (2) size selection may not be employed on such prey because the otters cannot judge prey size prior to excavation, and the cost of excavation makes it uneconomical not to return to the surface and consume even small prey.

Variation in Community Structure 213

Prey species that have a refuge in small size by virtue of size-selective predation, and that recruit predictably, possess a demographic mechanism for persisting at high population levels in the face of intensive sea otter predation. There is evidence that urchin populations, recently exploited by sea otters at Attu Island, occur at higher densities than do unexploited populations (Simenstad et al. 1978; Estes 1981, unpubl.). However, the absence of either size selection or predictable recruitment may lead to population declines when subjected to sea otter predation, unless spatial refuges are available, such as those that occur for deep-burrowing species in soft-sediment systems. For example, abalones, pismo clams, and red and purple sea urchins recruit eposodically in many areas; populations of these species typically are dominated by large individuals and usually are reduced to low levels within months following the dispersal of sea otters into a previously unexploited area. Similarly, populations of bay mussels in Prince William Sound, which recruit more predictably, but have no size refuge from sea otter predation, can be quickly reduced to low densities by sea otter predation.

These demographic processes also may have important community-level consequences. In areas where red or purple sea urchins are the principal herbivores, a shift in community structure from urchin barrens to kelp beds appears to occur within a year or less following the arrival of sea otters (Duggins 1980; Breen et al. 1982; Laur et al., this VoL). In contrast, sea urchin barrens have persisted in the sublittoral zone for 20 or more years following the invasion of sea otters along the northeast coast of Attu Island where green sea urchins are the principal herbivores (Simenstad et al. 1978; Estes 1981, unpubl.).

11.3 Variation in Community Structure

Throughout this volume we have seen emphasis on variation in population and community-level processes, in time, space, and on a range of scales of both. Foster and Schiel (this Vol.) point out that the extent to which sea urchins destructively graze macro algae varies substantially outside the present range of the sea otter in California. Thus, they conclude that a keystone predator's role (for sea otters) has been overemphasized and overgeneralized. Evidence for the importance of spatial and temporal variation is provided in the report by Estes and Harrold (this Vol.) of changes in algal and invertebrate abundance, frequently with no correlation among species or within species among habitat patches. The contributions by Laur et al. (this Vol.) and Ebeling and Laur (this Vol.) both emphasize variation. Laur et al. demonstrated that whereas sea urchin densities declined remarkably at two of their study sites shortly after the arrival of otters, one site shifted from a sea urchin barrens to a kelp bed, with a concordant increase in total fish abundance, just as the paradigm for rocky subtidal systems predicts, while the other site, which may have never borne an extensive kelp canopy, changed relatively little because urchin densities were low to begin with. Substrate topography was flat at the former site, whereas the latter was smaller in area, of high relief, and subject to strong water motion. From these findings the authors concluded that substrate configuration figures prominently in the extent to which the paradigm for the role of sea otters in rocky subtidal communities is true in California. Bodkin


(in press) reported that removal of the kelp canopy (Macrocystis pyri!era) from a patch reef with high substrate relief in central California resulted in significant population declines of several fish species compared with a nearby unmanipulated site. Bodkin's results again emphasize the importance of variation in community-level processes - in this case regarding the interaction effect between substrate relief and kelp abundance on fish abundance. Ebeling and Laur (this Vol.) found that although the development of a kelp canopy on a high-relief reef off southern California correlated with modest overall increases in fish, this response varied by species, with some shOwing large increases, others little change, and still others declines. Variation in response among fish species was related mainly to the extent to which kelp canopy afforded the fish a refuge from predators or affected food supply.

Thus, there is evidence for variation in space and time for several of the fundamental interactions that go together to form the paradigm for how sea otters influence rocky subtidal communities. The extent to which subtidal habitats are denuded by grazing sea urchins is highly variable among areas outside the range of sea otters in California (Foster and Schiel, this VoL). The extent to which kelp beds control fish abundance is variable among both habitats and species offish (Ebeling and Laur, this Vol.; Laur et aI., this VoL). Even the foundation of the paradigm - the extent to which sea otters limit sea urchins - may be variable, at least between areas or species where recruitment is predictable and those where it is not.

Levin (this Vol.) pointed out that whereas variation in some processes, such as range increases of growing sea otter populations, may be predictable, variation in others, such as ecosystem-level phenomena, are more problematic. This raises questions about how to describe and understand variation in the community ecology of sea otters. Recognizing that there are physical differences among rocky sublittoral habitats, and that these probably account for much of the variation seen in kelp forest community structure, Foster and Schiel (this Vol.) suggest a "habitat-type" approach. There have been few studies relating variation in kelp forest communities to such factors as the relief or rock type of the substratum, or wave exposure, and further efforts to understand such relationships should prove enlightening, especially with reference to large-scale patterns. Foster and Schiel (this Vol.) point out that substantial differences occur among several of the areas they have studied in central California, which appear to recur among specific sites and thus may well be regarded as community types. Although there is no doubt that kelp forest community structure varies substantially in response to variation in physical factors, the nature of this variation will most likely prove as difficult to document over large expanses of coastline as the effects of sea otter predation, if not more so. The major sampling problems in both instances are identical. We are reminded of the long-standing debate among plant ecologists regarding whether terrestrial vegetation varies over physical gradients as a continuum (Curtis 1959; Maycock and Curtis 1960; Whitaker 1956), or as identifiable habitat types (Daubenmire 1966), that are separated by punctuated changes. The debate is still unresolved, which is surprising because spatial variation in terrestrial systems should be easy to observe and measure over large areas, and in fact many workers have done this. Large-scale observations of benthic systems, even under the best of conditions, are difficult to make by comparison. Thus, without the development of new technologies, the extent and quality of data on community variation over large areas (Le., hectares or » in benthic marine systems

Future Research Needs 215

will never be what it is in terrestrial systems, and the empirical basis for judging whether marine communities vary as discrete types or continua may always be wanting.

11.4 Future Research Needs

Although quite a lot is known about the community ecology of sea otters, the chapters in this Volume raise several important questions. Knowledge of sea otter predation in rocky intertidal and soft-bottom communities is at the beginning stage. Further documentation, by area and by species, of the extent and rate of change resulting from sea otter predation, would be useful for both these systems. Other than the report by VanBlaricom (this VoL), there are no well-documented accounts of the effects of sea otter foraging on intertidal mussel beds. Since VanBlaricom's data indicate rather different community-level effects between his exposed outer coast site in central California and the protected site in Prince William Sound, Alaska, we would like to know how the effect of sea otter foraging varies among other sites, and why. If variation in effects among sites is substantial, three general causes are possible, based on what is now known: (1) variation in the physical and biological environment, across which sea otters have a range of abilities or needs to forage on mussels; (2) variation in foraging patterns by age, sex, or reproductive condition such that areas inhabited by different social or reproductive groupings are affected differently; and (3) variation in foraging patterns by individuals (as shown by Lyons and Estes, unpubl.), possibly transmitted matrilineally, or by example in unrelated individuals, such that one or a few animals in particular areas develop the habitat of feeding on mussels. Further detailed study of foraging in recognizable individuals would help unravel these mysteries.

For the reasons discussed above, additional documentation of sea otter foraging in soft-bottom systems is needed. We would like to know, by prey species and physical habitat, both the extent and time course of change caused by sea otter predation. This could best be done simply by following the size and abundance of prey species in specific areas through time with the arrival and establishment of sea otters. Kvitek and Oliver (this Vol.) raise several specific questions regarding the influences of sea otters in soft-bottom communities, such as: (1) How does the excavation of sediment that occurs when sea otters dig for deep-burrowing bivalves affect the assemblage of infaunal organisms? (2) What indirect effects on infaunal assemblages may result from the reduction of crab populations by sea otter predation? (3) How does the deposition of bivalve shell fragments, discarded by foraging sea otters, affect organisms in the upper sediment layers? These questions could be addressed through field experiments, especially if coupled with comparative studies.

The cost of excavating sediments may be substantial to foraging sea otters, as suggested by Kvitek's and Oliver's (this Vol.) finding that populations of deep-burrowing bivalves are reduced by sea otter predation at a slower rate than are shallow-burrowing species. Further understanding of this system might come from modeling energy and time budgets of sea otters, whose behavior when foraging follows a predictable sequence of events: dive, search, pursuit/capture, return to surface, consume prey. Each of these activities costs the animal time and energy, which must be repaid by the value


of the prey obtained. Foraging for deep-burrowing prey in soft sediments adds the additional cost of sediment excavation to pursuit/capture activities. These various costs might be measured using recent technical advances like the time-depth recorder (Gentry and Kooyman 1986). Once cost estimates were available, optimization models could be developed to predict, for example, optimum prey size. Predictions from such models might be used to contrast foraging behavior between soft-sediment and rocky-bottom habitats.

Important questions concerning sea otter/community interactions in rocky subtidal communities relate to spatial and temporal variation and although this is a major theme developed by Foster and Schiel (this Vol.) and Estes and Harrold (this Vol.), these authors suggest different approaches to the problems it presents. Foster and Schiel recommend that future studies be directed toward understanding factors that contribute to spatial and temporal variation in kelp forest community structure. This, they suggest, can be accomplished best by performing small-scale studies designed to test specific hypotheses about factors regulating community structure. It is their opinion that a better understanding of the influences of sea otters will only be obtained through recognition of spatial and temporal variation, and the factors that contribute to it. Estes and Harrold recommend a different approach. They argue that further understanding the influence of sea otters in kelp forest communities would be best accomplished through large-scale experiments, in which changes in community structure are related to changes in the distribution and numbers of sea otters. They suggest that this could be achieved by study before and after the introduction of otters to areas from which they once occurred but are presently absent, or by removal of sea otters from areas where they presently occur. Estes and Harrold argue that if the data were gathered from an adequate number of fixed-location plots that were representatively distributed over the sea floor, then it would be possible to determine both the generality of community-level influences related to sea otter predation, and the variation in those influences among different types or kinds of habitats. Foster's and Schiel's view is broadly directed at understanding the range of factors that affect variation in kelp forest community structure. Estes and Harrold's view is specifically directed at documenting the generality, breadth, and time course of effects from sea otter predation. The former view may be likened to a multi-factoral design, in which as much variation in community structure as possible is accounted for by simultaneously evaluating the effects of all factors thought to be important, and their interactions. The latter view may be likened to a one-way analysis of variance, in which factors other than sea otter predation, of known or suspected importance to kelp forest community structure, are accounted for by blocking, or subsumed in the error term.

Other suggestions for further research relate mainly to possible indirect effects of sea otter predation, especially the importance of kelp beds. Duggins (this Vol.) points out that changes in the abundance of kelp may affect coastal communities by creating habitat, increaSing food production, and altering water flow regimes. Because of the possible multiplicity of general avenues by which kelp beds affect ecological processes in coastal systems, it has proven difficult to document their effects, to say nothing of the mechanisms by which they occur. To the extent that sea otters predation effects a shift from sea urchin barrens to kelp beds, the various influences of kelp beds are indirectly attributable to sea otters. Further understanding of these processes is clearly

The Approach to Variation - A Philosophical Perspective 217

needed to understand the breadth of influences that sea otters may have on rocky subtidal communities. The paradigm for the role of sea otters in rocky subtidal communities holds that fish population increases are linked to increases in the abundance of kelp. Ebeling and Laur (this Vol.) have shown that in southern California this probably is true only for some species and in certain habitats. They suggest that studies focusing on the demography and feeding ecology of kelp bed fishes be done before and after the arrival of sea otters in presently unoccupied habitats. Demographic studies should focus on longevity and recruitment. Ebeling and Laur hypothesize that species with a large "storage effect" (Warner and Chesson 1985), that is long-lived adults that recruit infrequently, may be enhanced by sea otters via increased recruitment success. Feeding studies might clarify mechanisms of change in fish populations. For example, species whose food resources were increased via the detritus-based food web might be expected to increase with the arrival of sea otters, whereas species whose food resources overlapped those of sea otters might be expected to decline, because of exploitation competition. Laur et al. (this Vol.) recommend experiments in which the algal canopy is manipulated, noting differences in recruitment and survival of fish between treatments and controls. Bodkin (in prep.) recently completed just such a study.

11.5 The Approach to Variation - A Philosophical Perspective

Description of the physical universe can be done from two very different perspectives reflecting fundamentally different views on the organization of natural systems. The focus from one perspective is on recurrent pattern, while from the other it is on variation. For disciplines that involve "well-behaved systems" (i.e., those for which functional relationships are characterized by coefficients of determination that approach unity), the former view has almost become a working principle. The so-called "hard sciences", such as mathematics and physics, have developed accordingly.

Modern science has been so influenced by the quest for predictability that in some circles "good science" and "bad science" have become nearly synonymous with the extent to which predictable outcomes derive from a new discovery (Platt 1964). The notion of "strong inference", based upon the hypothetico-deductive approach to scientific inquiry, is founded largely on the assumption that important processes in our physical universe are not only recurrent, but predictable. It would be convenient for scientists if all of nature worked this way. But it may not, and in concept at least, one can imagine systems that lack predictable outcomes, or appear at least to lack predictable outcomes because numerous important factors are involved, most of which are subsumed in the error term. The qualities of being exactly predictable or completely unpredictable are, of course, extremes: most natural phenomena fall somewhere in between.

Ecologists have endeavored to make their science predictable, even though ecological events are inherently complex with many unexplained components of variability. After all, if the results of a study done in some small patch of habitat cannot be extended to the rest of the world, or at least to that community or habitat type, what good is it? This is the ethic ecology has inherited from the "hard" sciences, and for the most part ecologists have accepted it.


Levin (this Vol.) emphasizes that natural communities are most often "patchy" places. All that is required is a brief look at the world to see the truth in this statement. It follows that ecological processes may not yield predictable outcomes, especially over a range of varying conditions because there are simply too many unexplained potential sources of variability in complex ecosystems. The outcomes of competition or consumer-prey interactions, even among the same or similar species, may not be the same everywhere these species co-occur. This phenomenon is broadly relevant to community ecology. However, we are not aware of a single case in which variation in space or time in the outcome of species interactions has been the main theme of a field study. Ecologists replicate their measurements or manipulations because statisticians have told them to, but they hope that the outcomes are similar because similar outcomes indicate that the processes under investigation produce predictable outcomes over larger areas and longer times; and this is "good science". Ecological processes that produce repeatable outcomes are somehow more interesting than those that do not.

Few people would take issue with the conclusion that sea otters are important predators. Similarly, no one would argue that sea otter predation explains all variation in benthic community structure within the range of the species. What we would like to know is this: within the complex milieu of processes that influence community structure, just how important is sea otter predation; what are its indirect effects; and how do these vary in space and time? In attempting to face these challenging questions, we must not lose track of standards set by the discipline of ecology, for the problems faced are similar to those of documenting the importance of any species interaction when the interacting species occur widely over a heterogeneous environment.

References

Abbott lA, Hollenberg GJ (1976) Marine algae of California. Stanford University Press, Stanford, California

Aller RC (1982) The effects of macrobenthos on chemical properties of marine sediment and overlying water. In: McCall PL, Tevesz MJS (eds) Animal-sediment relations. The biogenic alteration of sediments. Plenum, New York, p 53

Aller RC (1983) The importance of the diffusive permeability of animal burrow-linings in determining marine sediment chemistry. J Mar Res 41 :299-322

Aller RC, Yingst JY, Ullman WJ (1983) Comparative biogeochemistry of water in intertidal Onuphis (Polychaeta) and Upogebia (Crustacea) burrows: temporal patterns and causes. J Mar Res 41: 571-604

Ambrose WG Jr (1984) Increased emigration of the amphipod Rhepoxynius abronius (Barnard) and the polychaete Nephtys caeca (Fabricius) in the presence of invertebrate predators. J Exp Mar BioI EcoI80:67-75

Anderson EK, North WJ (1966) In situ studies of spore production and dispersal in the giant kelp Macrocystis. Proc Int Seaweed Symp 5 :73-86

Andrew NL, Choat JH (1982) The influence of predation and conspecific adults on the abundance of juvenile Evechinus chloroticus (Echinoidea:Echinometridae). Oecologia (Berl) 54 :80-87

Andrews HL (1945) The kelp beds of the Monterey region. Ecology 26:24-37 Antonelis GA Jr, Leatherwood S, Cornell LH, Antrim JG (1981) Activity cycle and food selection

of captive sea otters. Murrelet 62:6-9 Armstrong JJ (1979) The California sea otter: emerging conflicts in resource management. San

Diego Law Rev 16:249-285 Barabash-Nikiforov II, Reshetkin VV, Shidlovskaya NK (1947) The Sea Otter (Kalan). Trans from

Russian by A Birron and ZS Cole, 1962. National Science Foundation & US Department of the Interior, Washington DC (Israel Program for Science Translation), 227 pp

Barber RT, Chavez FP (1983) Biological consequences of El NiUo. Science 222:1203-1210 Barber RT, Chavez FP (1986) Ocean variability in relation to living resources during the 1982-83

El Nino. Nature 319 :279-285 Baxter JL (1962) The Pismo clam in 1960. Calif Fish Game 48 :35-37 Baxter RE (1971) Earthquake effects on clams of Prince William Sound. In: National Research

Council. The great Alaska earthquake of 1964. Biology. National Academy of Sciences, Washington DC, p 238

Bell SS, Coull BC (1978) Field evidence that shrimp predation regulates meiofauna. Oecologia (Berl) 35:141-148

Bell SS, Watzin MC, Coull BC (1978) Biogenic structure and its effects on spatial heterogeneity of meiofauna in a salt marsh. J Exp Mar BioI Ecol 35 :99 -1 07

Benech SV (1981) Observations of the sea otter Enhydra lutris population between Point Buchon and Rattlesnake Creek, San Luis Obispo, CA, January through December 1980. Unpublished report. ECOMAR, Ventura, California

Bernstein BB, Jung N (1979) Selective pressures and coevolution in a kelp canopy community in southern California. Ecol Monogr 49:335-355

Berta A, MorganGS (1985) A new sea otter (Carnivora: Mustelidae) from the late Miocene and early Pliocene (Hemphillian) of North America. J Paleontol59 :809-819

220 References

Bickel PMcW (1978) Changing sea levels along the California coast: Anthropological implications. J Calif Anthro 5 :6-20

Bigg MA, MacAskie IB (1978) Sea otters reestablished in British Columbia. J Mammal 59:874-876 Blake JA (1979) California marine waters areas of special biological significance reconnaissance

survey report: Duxbury Reef Reserve and Extension, Marin County. California State Water Resources Control Board, Water Quality Monitoring Report 79-14, Sacramento

Bodkin JL (in press) Fish assemblages in Macrocystis and Nereocystis kelp forests off central California. US Natl Oceanic Atmos Adm Fish Bull 84:799-808

Bodkin JL, VanBlaricom GR, Jameson RJ (1987) Mortalities of kelp forest fishes associated with large oceanic waves off Central California, 1982-1983. Environ BioI Fish 18:73-76

Boolootian RA (1961) The distribution of the California sea otter. Calif Fish Game 47 :287-292 Bormann FH, Likens GE (1979) Pattern and process in a forested ecosystem. Springer, Berlin

Heidelberg New York Boudreau BP (1984) On the equivalence of nonlocal and radial-diffusion models for porewater ir

rigation. J Mar Res 42:731-735 Boyce MS (1979) Seasonality and patterns of natural selection for life histories. Am Nat 114:

569-583 Boyd MJ (1980) California marine waters areas of special biological significance reconnaissance

survey report: kelp beds at Trinidad Head, Humboldt County. California State Water Resources Control Board, Water Quality Monitoring Report 79-19, Sacramento

Boyd MJ, DeMartini 10, Pic'l G (1981) California marine waters areas of special biological significance reconnaissance survey report: Redwoods National Park, Del Norte and Humbolt Counties. California State Water Resources Control Board, Water Quality Monitoring Report 81-5, Sacramento

Branch GM (1975) Mechanisms reducing intraspecific competition in Patella spp: migration, differentiation and territorial behavior. J Anim EcoI44:575-600

Bray RN (1978) Daily foraging migrations of the blacksmith, Chromis punctipinnis, a planktivorous kelp-bed fish. Thesis, University of California, Santa Barbara

Bray RN (1981) Influence of water currents and zooplankton densities on daily foraging movements of blacksmith, Chromis punctipinnis, a planktivorous reef fish. US Nat! Oceanic Atmos Adm Fish Bull 78:829-841

Bray RN, Ebeling AW (1975) Food, activity and habitat of three "picker-type" microcarnivorous fishes in the kelp forests off Santa Barbara, California. US Natl Oceanic Atmos Adm Fish Bull 73:815-829

Breda VA, Foster MS (1985) Composition, abundance, and phenology of foliose red algae associated with two central California kelp forests. J Exp Mar Bioi EcoI94:115-130

Breen PA, Mann KH (1976) Destructive grazing of kelp by sea urchins in eastern Canada. J Fish Res Board Can 33:1278-1283

Breen PA, Carson TA, Foster JB, Stewart EA (1982) Changes in subtidal community structure associated with British Columbia sea otter transplants. Mar Ecol Prog Ser 7: 13-20

Brooks J, Dodson SI (1965) Predation, body size, and composition of plankton. Science 150:28-35 Brusca RC, Wicksten MK (1979) California marine waters areas of special biological significance re

connaissance survey report: Heisler Park Ecological Reserve, Orange County. California State Water Resources Control Board, Water Quality Monitoring Report 79-2, Sacramento

Brusca RC, Winn RN (1979) California marine waters areas of special biological significance reconnaissance survey report: Newport Beach Marine Life Refuge, Orange County. California State Water Resources Control Board, Water Quality Monitoring Report 79-4, Sacramento

Brusca RC, Zimmerman R (1979) California marine waters areas of special biological significance reconnaissance survey report: Irvine Coast Marine Life Refuge, Orange County, California State Water Resources Control Board, Water Quality Monitoring Report 79-3, Sacramento

Burge RT, Schultz SA (1973) The marine environment in the vicinity of Diablo Cove with special reference to abalones and bony fishes. Calif Dep Fish Game Mar Resour Tech Rep 19:1-433

California Department of Fish and Game (1976) A proposal for sea otter protection and research, and request for the return of management to the state of California. Unpublished report, California Department of Fish & Game, Sacramento

References 221

Calkins DG (1978) Feeding behavior and major prey species of the sea otter, Enhydra lutris, in Montague Strait, Prince William Sound, Alaska. US Natl Oceanic Atmos Adm Fish Bull 76: 125-131

Cancino J, Santelices B (1981) The ecological importance of kelp-like holdfasts as a habitat of in-vertebrates in central Chile. Proc Int Seaweed Symp 10:241-246

Carefoot T (1977) Pacific seashores. University of Washington Press, Seattle Carlisle JG Jr (1973) Results of 1971 Pismo clam census. Calif Fish Game 59:138-139 Carr MH (1983) Spatial and temporal patterns of recruitment of young-of-the-year rockfishes

(genus Sebastes) into a central California kelp forest. Thesis, San Francisco State University, San Francisco

Carroll JC (1982) Seasonal abundance, size composition, and growth of rock crab, Cancer antennarius Stimpson, off central California. J Crustacean Bioi 2:549-561

Carter JW (1985) Kelp and subtidal habitats. In: Yuge JE, Elliott JL (eds) Marine environmental analysis and interpretation, San Onofre Nuclear Generating Station, Report on 1985 data 85-RD-37. Southern California Edison Co, Pasadena, pp 5-11

Carter JW, Carpenter AL, Foster MS, Jessee WN (1985a) Benthic succession on an artificial reef designed to support a kelp-reef community. Bull Mar Sci 37 :86-113

Carter JW, Jessee WN, Foster MS, Carpenter AL (1985b) Management of artificial reefs designed to support natural communities. Bull Mar Sci 37 :114-128

Caswell H (1978) Predator-mediated coexistence: a nonequilibrium model. Am Nat 112:127-154 Cayot RF, North WJ (1971) Investigations related to the effect of a thermal power plant on the

marine environment at Davenport, Santa Cruz County, California. Pacific Gas & Electric Co, Emeryville, California

Chan G (1980a) California marine waters areas of special biological significance reconnaissance survey report: Point Reyes, Headland Reserve and Extension, Marin County. California State Water Resources Control Board, Water Quality Monitoring Report 80-1, Sacramento

Chan G (1980b) California marine waters areas of special biological significance reconnaissance survey report: Double Point, Marin County. California State Water Resources Control Board, Water Quality Monitoring Report 79-15, Sacramento

Chanin P (1985) The natural history of otters. Facts on File, New York Charters AC, Neushul M, Barilotti C (1969) The functional morphology of Eisenia arborea. Proc

Int Seaweed Symp 6:89-105 Charters AC, Neushul M, Coon DA (1972) Effects of water motion on algal spore attachment. Proc

Int Seaweed Symp 7 :243-247 Choat JH (1982) Fish feeding and the structure of benthic communities in temperate waters. Annu

Rev Ecol Syst 13 :423 -449 Cicin-Sain B (1982) Sea otters and shellfish fisheries in California: the management framework. In:

Cicin-Sain B, Grifman PM, Richards JB (eds) Social science perspective on managing conflicts between marine mammals and fisheries. Univ California Cooperative Extension, San Luis Obispo, P 195

Cimberg RL, Costa DP (1985) North Aleutian shelf sea otters and their susceptibility to oil. In: Proceedings of the 1985 Oil Spill Conference (Prevention, Behavior, Control, Cleanup). American Petroleum Institute, Washington, DC, p 211

Cirnberg RL, Costa DP, Fishman PA (1984) Ecological characterization of shallow subtidal habitats in the North Aleutian Shelf. Final Report, Research Unit 623, VTN Oregon, Portland, Oregon

Clarke TA (1971) Territory boundaries, courtship, and social behavior in the garibaldi, Hypsypops rubicunda (Pomacentridae). Copeia 2:295-299

Clarke WD, Neushul M (1967) Subtidal ecology of the southern California coast. In: Olson TA, Burgess FJ (eds) Pollution and marine ecology. Interscience, New York, p 29

Coe WR (1932) Season of attachment and rate of growth of sedentary marine organisms at the pier of the Scripps Institution of Oceanography, La Jolla, California. Bull Scripps Inst Oceanogr Tech Ser 3:37-86

Coe WR, Fitch JE (1950) Population studies, local growth rates and reproduction of the Pismo clam (Tivela stultorum). J Mar Res 9 :188-210

Coen LD, Heck KL, Abele LG (1981) Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62:1484--1493

222 References

Connell JH (1961) Effects of competition, predation by Thais lapillus and other factors on natural populations of the barnacle Balanus balanoides. Ecol Monogr 31 :61-104

Connell JH (1978) Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310 Connell JH (1980) Diversity and coevolution of competitors, or the ghost of competition past.

Oikos 35:131-138 Connell JH (1983) On the prevalence and relative importance of interspecific competition: evidence

from field experiments. Am Nat 122:661-696 Connell JH, Sousa WP (1983) On the evidence needed to judge ecological stability or persistence.

Am Nat 121 :789-824 Conner EF, Simberloff D (1986) Competition, scientific method, and null models in ecology. Am

Sci 74:155-162 Cook J (1784) A voyage to the Pacific Ocean. Undertaken by the command of His Majesty, for

making discoveries in the northern hemisphere. To determine the position and extent of the west side of North America; Its distance from Asia; and the practicability of a northern passage to Europe. Lords Commissioners of the Admiralty, London

Cooper J, Wieland M, Hines A (1977) Subtidal abalone populations in an area inhabited by sea otters. Veliger 20:163-167

Corbet GB, Hill 1£ (1980) A world list of mammalian species. British Museum and Cornell University Press, Ithaca, New York

Costa DP (1978a) The ecological energetics, water, and electrolyte balance of the California sea otter, Enhydra lutris. Thesis, University of California, Santa Cruz

Costa DP (1978b) The sea otter: its interaction with man. Oceanus 21 :24-30 Costa DP (1985) The ecological energetics of sea otters. Paper presented at the 4th International

Otter Symposium, Santa Cruz, California Costa DP, Kooyman GL (1982) Oxygen consumption, thermoregulation, and the effect of fur oil

ing and washing on the sea otter, Enhydra lutris. Can J ZooI60:2761-2767 Cowen RK (1983) The effect of sheephead (Semicossyphus pulcher) predation on red sea urchin

(Strongylocentrotus franciscanus) popUlations: an experimental analysis. Oecologia (Berl) 58: 249-255

Cowen RK (1985) Large scale pattern of recruitment by the labrid, Semicossyphus pulcher: causes and implications. J Mar Res 43:719-742

Cowen RK, Agegian CR, Foster MS (1982) The maintenance of community structure in a central California giant kelp forest. J Exp Mar Bioi EcoI64:189-201

Coyer JA (1979) The invertebrate assemblage associated with Macrocystis pyrifera and its utilization as a food source by kelp forest fishes. Thesis, University of Southern California, Los Angeles

Crusafont-Pairo R, Golpe JM (1962) Nuevos hallazgos de Lutridos aonicoides(Sivaonyx, Enhydri· odon) en el Pikermiense espanol (Venca de Tervel). Inst Gee Minero Espanas Nofas Commons 67:5-15

Curtis JT (1959) The vegetation of Wisconsin. University of Wisconsin Press, Madison, Wisconsin Daubenmire R (1966) Vegetation: identification of typal communities. Science 151 :291-298 Davis BS (1977) The southern sea otter revisited. Pac Discovery 30:1-13 Davis GE (1986) Kelp forest dynamics in Channel Islands National Park, California 1982-85.

Channel Islands National Park and National Marine Sanctuary Natural Science Study Reports CHIS-86~01. National Park Service, US Department of the Interior, Ventura, California

Dayton PK (1971) Competition, disturbance, and community organization: The provision and subsequent utilization of space in a rocky intertidal community. Ecol Monogr 41 :351-389

Dayton PK (1973) Dispersion, dispersal and persistence of the annual intertidal alga, Postelsia pal· maeformis Ruprecht. Ecology 52:433-438

Dayton PK (1975) Experimental studies of algal canopy interactions in a sea otter-dominated kelp community at Amchitka Island, Alaska. US Nat! Oceanic Atmos Adm Fish Bull 73:230-237

Dayton PK (1985a) Ecology of kelp communities. Annu Rev Ecol Syst 16:215-245 Dayton PK (l985b) The structure and regulation of some South American kelp communities. Ecol

Monogr 55 :447-468 Dayton PK, Tegner MJ (1984a) The importance of scale in community ecology: A kelp forest ex

ample with terrestrial analogs. In: Price PW, Siobodchikoff CN, Gaud WS (eds) A new ecology: novel approaches to interactive systems. Wiley, New York, p 457

References 223

Dayton PK, Tegner MJ (1984b) Catastrophic storms, EI Nino, and patch stability in a southern California kelp community. Science 224 :283-285

Dayton PK, Currie V, Gerrodette T, Keller BD, Rosenthal R, Ven Tresca D (1984) Patch dynamics and stability of some California kelp communities. Ecol Monogr 54 :25 3-289

Dean TA, Schroeter SC, Dixon JD (1984) Effects of grazing by two species of sea urchins (Strongy· locentrotus franciscanus and Lytechinus anamesus) on recruitment and survival of two species of kelp (Macrocystis pyrifera and Pterygophora californica). Mar Bioi 78:301-313

DeMartini JD (1979) California marine waters areas of special biological significance reconnaissance survey report: Gerst!e Cove. California State Water Resources Control Board, Water Quality Monitoring Report 79-17, Sacramento

Deysher LE, Norton TA (1982) Dispersal and colonization in Sargassum muticum (Yendo) Fensholt. J Exp Mar Bioi EcoI56:179-195

Domning DP (1978) Sirenian evolution in the North Pacific Ocean. Univ Calif Publ Geol Sci 118: 1-176

Duggins DO (1980) Kelp beds and sea otters: an experimental approach. Ecology 61 :447 -453 Duggins DO (1981a) Sea urchins and kelp: the effects of short-term changes in urchin diet. Limnol

Oceanogr 26:391-394 Duggins DO (1981b) Interspecific facilitation in a guild of benthic marine herbivores. Oecologia

(Berl) 48:157-163 Durham JW (1950) Cenozoic marine climates of the Pacific coast. Geol Soc Am Bull 61 :1243-1264 Ebeling AW, Laur DR (1985) The influence of plant cover on surfperch abundance at an offshore

temperate reef. Environ Bioi Fishes 12:169-179 Ebeling AW, Laur DR (1986) Foraging in surfperches: resource partitioning or individualistic re

sponses? Environ Bioi Fishes 16:123-133 Ebeling AW, Larson RJ, Alevizon WS (1980a) Habitat groups and island-mainland distribution of

kelp-bed fishes off Santa Barbara, California. In: Power DM (ed) The California Islands: Proceedings of a multidisciplinary symposium. Santa Barbara Museum of Natural History, Santa Barbara, California, p 403

Ebeling AW, Larson RJ, Alevizon WS, Bray RN (1980b) Annual variability of reef-fish assemblages in kelp forests off Santa Barbara, California. US Nat! Oceanic Atmos Adm Fish Bull 78:361-377

Ebeling AW, Laur DR, Rowley RJ (1985) Severe storm disturbances and reversal of community structure in a southern California kelp forest. Mar Bioi 84:287 -294

Ebert EE (1967) Foraging activity of sea otters in the San Simeon Cambria region, November 13 through 16, 1967. California Department of Fish & Game, MRO Reference 67 -31, Sacramento

Ebert EE (1968) A food-habits study of the southern sea otter, Enhydra lutris nereis. Calif Fish Game 54:33-42

Ebert TA (1983) Recruitment in echinoderms. In: Jangoux M, Lawrence JM (eds) Echinoderm studies, vol I. Baikema, Rotterdam, p 169

Eckman JE (1983) Hydrodynamic processes affecting benthic recruitment. Limnol Oceanogr 28: 241-257

Edwards A (1980) Ecological studies of the kelp Laminaria hyperborea and its associated fauna in southwest Ireland. Ophelia 19:47 -60

Engle JM, Coyer JA (1981) California marine waters areas of special biological significance reconnaissance survey report: Santa Catalina Island - subarea three Farnsworth Bank Ecological Reserve Los Angeles County. California State Water Resources Control Board, Water Quality Monitoring Report 81-4, Sacramento

Estes JA (1977) Population estimates and feeding behavior of sea otters. In: Merritt MC, Fuller RG (eds) The environment of Amchitka Island, Alaska. US Energy Research and Development Administration, Springfield, Virginia, p 526

Estes JA (1980) Enhydra lutris. Mamm Species 133:1-8 Estes JA (1981) The case of the sea otter. In: Jewell PA, Holt S (eds) Problems in management of

locally abundant wild animals. Academic Press, London, p 167 Estes lA, Jameson RJ (1983a) Size and status of the sea otter population in California. Unpublished

report, US Fish and Wildlife Service, Santa Cruz, California Estes JA, Jameson RJ (1983b) Summary of available population information on California sea ot

ters. US Minerals Management Service, POCS Technical Paper 83-11, Los Angeles

224 References

Estes JA, Palmisano JF (1974) Sea otters: their role in structuring nearshore communities. Science 185 :1058-1060

Estes JA, VanBlaricom GR (1985) Sea-otters and shellfisheries. In: Beddington JR, Beverton RJH, Lavigne DM (eds) Marine mammals and fisheries. George Allen & Unwin, London, p 187

Estes JA, Smith NS, Palmisano JF (1978) Sea otter predation and community organization in the western Aleutian Islands, Alaska. Ecology 59 :822-833

Estes JA, Jameson RJ, Johnson AM (1981) Food selection and some foraging tactics of sea otters. In: Chapman JA, Pursley D (eds) The Worldwide Furbearer Conference Proceedings. Worldwide Furbearer Conference, Frostburg, Maryland, p 606

Estes JA, Jameson RJ, Rhode EB (1982) Activity and prey selection in the sea otter: influence of population status on community structure. Am Nat 120:242-258

Estes JA, Underwood KE, Karmann MJ (1986) Activity-time budgets of sea otters in California. J Wildlife Manage 50:626-636

Fager EW (1964) Marine sediments: effects of a tube-building polychaete. Science 143:356-359 Fankboner PV, de Burgh ME (1977) Diurnal exudation of 14C-labeled compounds by the large

kelp Macrocystis integrifolia Bory. J Exp Mar Bioi EcoI28:151-162 Feder HM, Turner CH, Limbaugh C (1974) Observations on fishes associated with kelp beds in

southern California. Calif Dept Fish Game Fish Bull 160:1-144 Ferson S, Downey P, Klerks P, Weissburg M, Kroot I, Steward S, Jacquez G, Ssemakula J, Malenky

R, Anderson K (1986) Competing reviews, or why do Connell and Schoener disagree? Am Nat 127 :571-576

Fisher EM (1939) Habits of the southern sea otter. J Mammal 20 :21-36 Fitch JE (1950) The Pismo clam. Calif Fish Game 36 :285-312 Fitch JE (1965) A relatively unexploited population of Pismo clams, Tivela stultorum (Mawe,

1823) (Veneridae). Proc Malacol Soc Lond 36:309-312 Fleminger A, Klein HT (1963) Atlas of 10-meter temperatures and salinities, 1949 through 1959.

Calif Coop Oceanic Fish Invest Atlas 1 Fonseca MS, Fisher JS, Zieman JC, Thayer GW (1982) Influence of the seagrass, Zostera marina L.,

on current flow. Estuarine Coastal Shelf Sci 15 :351-364 Forman RTT, Godron M (1986) Landscape ecology. Wiley, New York Foster MS (1975) Algal succession in a Macrocystis pyrifera forest. Mar Bioi 32:313-329 Foster MS (1982) The regulation of macroalgal associations in kelp forests. In: Srivastava L (ed)

Synthetic and degradative processes in marine macrophytes. de Gruyter, Berlin, p 185 Foster MS, Reed D (1980) Studies of the effects of sea otter foraging on kelp forest communities

in central California: Year three. Report to Department of Fish and Game, State of California. California Department of Fish & Game, Sacramento

Foster MS, Schiel DR (1985) The ecology of giant kelp forests in California: a community profile. US Fish & Wildlife Service Biological Report 85 (7.2), Slidell, Louisiana

Foster MS, Mullins HT, Reilly M, Reed D (1980) California marine waters areas of special biological significance reconnaissance survey report: kelp beds at Saunders Reef. California State Water Resources Control Board, Water Quality Monitoring Report 80-3, Sacramento

Foster MS, Carter lW, Schiel DR (1983) The ecology of kelp communities. In: Bascom W (ed) The effects of waste disposal on kelp communities. Southern California Coastal Water Research Project, Long Beach, p 53

Foster MS, Dean TA, Deysher LE (1985) Subtidal techniques. In: Littler MM, Littler DS (eds) Handbook of phycological methods: ecological methods for macroalgae. Cambridge University Press, Cambridge, England, p 199

Foster MS, DeVogelaere AP, Harrold C, Pearse IS, Thurn AB (in press) Causes of spatial and temporal patterns in rocky intertidal communities of central and northern California. Memoirs of the California Academy of Sciences

Fry B, Sherr EB (1984) .suC measurements as indicators of carbon flow in marine and freshwater ecosystems. Contrib Mar Sci 27 :13-47

Gaines S, Roughgarden J (1985) Variable settlement in intertidal invertebrates: the perilous life behind kelp forests. Paper presented at the 66th annual meeting. Western Society of Naturalists, Monterey, California

References 225

Garshelis DL (1983) Ecology of sea otters in Prince William Sound, Alaska. Thesis, University of Minnesota, Minneapolis

Garshelis DL, Garshelis JA (1984) Movement and management of sea otters in Alaska. J Wildl Manage 48:665-678

Garshelis DL, Johnson AM, Garshelis JA (1984) Social organization of sea otters in Prince William Sound, Alaska. Can J Zool 62 :2648-2658

Geibel JJ, Miller DJ (1984) Estimation of sea otter, Enhydra lu tris, population, with population bounds, from air and ground counts. Calif Fish Game 70:225-233

Gentry RL, Kooyman GL (1986) Fur seals: maternal strategies on land and at sea. Princeton University Press, Princeton, New Jersey

Gerard VA (1976) Some aspects of material dynamics and energy flow in a kelp forest in Monterey Bay, California. Thesis, University of California, Santa Cruz

Gerard VA (1984) Physiological effects of El Nino on giant kelp in southern California. Mar BioI Lett 5 :317 -322

Ghelardi RJ (1971) Species structure of the animal community that lives in Macrocystis pyrifera holdfasts. In: North WJ (ed) The biology of giant kelp beds (Macrocystisj in California. Nova Hedwigia 32:381-420

Gillett JW, Stern AM, Levin SA, Harwell MA, Andow DA, Alexander M, Staff of the Ecosystems Research Center (1986) Potential impacts of environmental release of biotechnology products: Assessment, regulation, and research needs. Environ Manage 10:437 -563

Gleason HA (1917) The structure and development of the plant association. Bull Torrey Bot Club 43:463-481

Gleason HA (1926) The individualistic concept of the plant association. Bull Torrey Bot Club 53: 7-26

Gooding RM, Magnuson JJ (1967) Ecological significance of a drifting object to pelagic fishes. Pac Sci 21 :486-497

Goodwin CL (1976) Observations on spawning and growth of subtidal geoducks (Panope generosa Gould). Proc Nat! Shellfish Assoc 65 :49-58

Gotshall DW (1977) Stomach contents of northern California Dungeness crabs, Cancer magister. Calif Fish Game 64:24-37

Gotshall DW, Lea RN, Laurent LL, Hoban TL, Farrens GD (1974) Mendocino power plant site ecological study final report. California Department of Fish & Game, Administrative Report 74-7, Sacramento

Grigg RW, Kiwala RS (1970) Some ecological effects of discharged waste on marine life. Calif Fish Game 56:145-155

Hall CA (1976) Geologic map of the San Simeon - Piedras Blancas region, San Luis Obispo County, California. Miscellaneous Field Studies Map MF-784, US Geological Survey, Reston, Virginia

Hall ER, Schaller GB (1964) Tool-using behavior of the California sea otter. J Mammal 45 :287 -298 Hallacher LE, Roberts DA (1985) Differential utilization of space and food by the inshore rock

fishes (Scorpaenidae: Sebastes) of Carmel Bay, California. Environ Bioi Fishes 12:91-110 Hannan CA (1984) Planktonic larvae may act like passive particles in turbulent near-bottom flows.

Limnol Oceanogr 29:1108-1116 Harger JRE (1972) Competitive coexistence. Maintenance of interacting associations of the sea

mussels My tilu s edulis and Mytilus californianus. Veliger 14:387-410 Harper JL (1969) The role of predation in vegetational diversity. Brookhaven Symp Bioi 22:48-62 Harris LG, Ebeling AW, Laur DR, Rowley RJ (1984) Community recovery after storm damage: a

case of facilitation in primary succession. Science 224:1336-1338 Harris RF (1977) Feeding and other activities of the sea otter, Enhydra lutris, along Cannery Row,

Monterey, California. Thesis, University of California, Santa Cruz Harrold C, Hardin D (1986) Foraging on land by the California sea otter, Enhydra lutris. Mar Mamm

Sci 2:309-313 Harrold C, Pearse JS (in press) The ecological role of echinoderms in kelp forests. In: Lawrence

JM (ed) Echinoderm studies, vol II. Balkema, Rotterdam Harrold C, Reed DC (1985) Food availability, sea urchin grazing, and kelp forest community struc

ture. Ecology 66:1160-1169

226 References

Hastings JR, Turner RM (1965) The changing mile. University of Arizona Press, Tucson, Arizona Hatcher BG, Chapman ARO, Mann KH (1977) An annual carbon budget for the kelp Laminaria

/ongicruris. Mar Bioi 44:85-96 Heck KL, Thoman TA (1981) Experiments on predator-prey interactions in vegetated aquatic

habitats. J Exp Mar Bioi Eco153:125 -134 Helfman GS (1981) The advantages to fishes of hovering in shade. Copeia 1981 (2):392-400 Hennessey SL (1972) The intestinal parasites and diet analysis of the southern sea otter. Thesis,

California State University, Hayward Highsmith RC (1982) Induced settlement and metamorphosis of sand dollar (Dendraster excentri

cus) larvae in predator-free sites: adult sand dollar beds. Ecology 63:329-337 Hines AH (1982) Coexistence in a kelp forest: Size, population dynamics and resource partitioning

in a guild of spider crabs (Brachyura, Majidae). Ecol Monogr 52:179-198 Hines AH, Loughlin TR (1980) Observations of sea otters digging for clams at Monterey harbor,

California. US Natl Oceanic Atmos Adm Fish Bull 78:159-163 Hines AH, Pearse JS (1982) Abalones, shells, and sea otters: Dynamics of prey populations in central

California. Ecology 63:1547-1560 Hixon MA (1980) Competitive interactions between California reef fishes of the genus Embiotoca.

Ecology 61:918-931 Hobson ES, Chess JR (1976) Trophic interactions among fishes and zooplankters nearshore at Santa

Catalina Island, California. US Nat! Oceanic Atmos Adm Fish Bull 74 :567 -598 Holbrook SJ, Schmitt RJ (1984) Experimental analysis of patch selection by foraging black surf

perch, Embiotoca jacksoni Agazziz. J Exp Mar Bioi Ecol 79 :39-64 Holling CS (in press) Resilience of ecosystems; local surprise and global change. In: Clark WC,

Munn RE (eds) Sustainable development of the biosphere. Cambridge University Press, New Rochelle, New York

Hone EB (1984) Talesofthe sea otter. Pac Discovery 37:10-15 Hopkins DM (1967) The Cenozoic history of Beringia - a synthesis. In: Hopkins DM (ed) The

Bering Land Bridge. Stanford University Press, Stanford, California, p 451 Houk JL, Geibel JJ (1974) Observations of underwater tool use by the sea otter, Enhydra lutris

Linnaeus. Calif Fish Game 60 :207 -208 Hubbard JD (1971) Distribution and abundance of intertidal invertebrates at Olsen Bay in Prince

William Sound, Alaska, one year after the 1964 earthquake. In: National Research Council. The great Alaska earthquake of 1964. Biology. National Academy of Science, Washington DC, p 13 7

Hubbs CL (1974) "Marine Zoogeography" by John C Briggs. Copeia 1974:1002-1005 Hubbs CL, Kelly AL, Limbaugh C (1970) Diversity in feeding by Brandt's cormorant near San Diego.

Calif Fish Game 56:156-165 Hulberg LW, Oliver JS (1980) Caging manipulations in marine soft-bottom communities: The im

portance of animal interactions and sedimentary habitat modifications. Can J Fish Aquat Sci 37:1130-1139

Irons DB (1984) Sea otter distribution and abundance in western Prince William Sound, Alaska. Progress report 1984. Unpublished report, US Fish & Wildlife Service, Anchorage, Alaska

Irons DB, VanBlaricom GR, Bodkin JL (1984) Sea urchin populations in northwestern Prince William Sound, Alaska. Unpublished report, US Fish & Wildlife Service, Anchorage, Alaska

Islieb ME, Kessel B (1973) Birds of the north Gulf coast - Prince William Sound Region, Alaska. Bioi Pap Univ Alaska 14:1-149

Iverson JA, Krog J (1973) Heat production and body surface area in seals and sea otters. Norw J ZooI21:51-54

Jackson GA (1984) Internal wave attenuation by coastal kelp stands. J Phys Oceanogr 14:1300-1306

JacksonGA, Winant CD (1983) Effect of a kelp forest on coastal currents. Cont Shelf Res 2:75-80 Jameson RJ (1975) An evaluation of attempts to reestablish the sea otter in Oregon. Thesis, Oregon

State University, Corvallis Jameson RJ, Kenyon KW, Johnson AM, Wight HM (1982) History and status of translocated sea

otter popUlations in North America. Wildl Soc Bull 10:100-107 Jameson RJ, Kenyon KW, Jeffries S, VanBlaricom GR (in press) Status of a translocated sea otter

population in the state of Washington, with notes on habitat. Murrelet

References 227

Janzen DH (1970) Herbivores and the number of species in tropical forests. Am Nat 104:501-528 Johansen HW (1971) Effects of elevation changes on benthic algae in Prince William Sound. In:

National Research Council. The great Alaska earthquake of 1964. Biology. National Academy of Sciences. Washington DC, p 35

Johnson AM (1982) Status of Alaska sea otter populations and developing conflicts with fisheries. Trans N Am Wildl Nat Resour Conf 47 :293-299

Johnson AM, Jameson R, Schmidt T, Calkins D (1983) Sea otter survey, southeast Alaska, 1983. Unpublished report, US Fish & Wildlife Service, Anchorage, Alaska

Jones GP (1984a) Population ecology of the temperate reef fish Pseudolabrus celiodotus (Pisces, Labridae). 1. Factors influencing recruitment. J Exp Mar Bioi Ecol 75 :257-276

Jones GP (1984b) Population ecology of the temperate reef fish Pseudolabrus celiodotus (Pisces, Labridae). 2. Factors influencing adult density. J Exp Mar Bioi Ecol 75 :277 -303

Jones LC (1971) Studies on selected small herbivorous invertebrates inhabiting Macrocystis canopies and holdfasts in southern California kelp beds. In: North WJ (ed) The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia 32:343-367

Jones RD Jr (1965) Sea otters in the Near Islands, Alaska. J Mammal 46:702 Kain JM (1979) A view of the genus Laminaria. Oceanogr Mar Bioi Annu Rev 17 :101-161 Kastendiek IE (1982) Factors determining the distribution of the sea pansy, Renilla kollikeri, in a

subtidal sand-bottom habitat. Oecologia (Bed) 52:340-347 Kenyon KW (1969) The sea otter in the eastern Pacific Ocean. N Am Fauna 68:1-352 Khailov KM, Burlakova ZP (1969) Release of dissolved organic matter by marine seaweeds and

distribution of their total organic production to nearshore communities. Limnol Oceanogr 14: 521-527

Kilmer FH (1972) A new species of sea otter from the Pleistocene of northwestern California. Bull South Calif Acad Sci 71 :150-157

Kimker A (1982) Prince William Sound management area data report 17, 1982 Orca Inlet dungeness crab survey. Unpublished report, Alaska Department of Fish & Game, Cordova

Kimura RS, Foster MS (1984) The effects of harvesting Macrocystis pyrifera on the algal assemblage in a giant kelp forest. Hydrobiologia 116/117 :425-428

King JD (1976) Sea otters and their habitat north of Santa Cruz, California. Thesis, University of California, Santa Cruz

Kirkpatrick CM, Stulken DE, Jones RD Jr (1955) Notes on captive sea otters. Arctic 8:46-59 Kleiman DG, Malcolm JR (1981) The evolution of male parental investment in mammals. In:

Gubernik DJ, Klopfer PH (eds) Parental care in mammals. Plenum, New York, p 347 Kobayashi BN (1979) California marine waters areas of special biological significance reconnais

sance survey report: San Diego-La Jolla Ecological Reserve. California State Water Resources Control Board, Water Quality Monitoring Report 79-1, Sacramento

Kovnat GD (1982) Intertidal foraging by the southern sea otter, Enhydra lutris. Thesis, Stanford Univ, Stanford, California

Kuznetsov AP (1963) The bottom invertebrate fauna of the Kamchatka waters of the Pacific Ocean and the northern Kurile Islands. Publ AN SSR, Moscow, USSR (In Russian, original paper not seen, citation based on description in Shitikov AM, Lukin VI, Chaplygina SF (1973) The character of the upper sublittoral of Paramushir Island as a foraging habitat of the sea otter. Proc Pac Res Inst Fish Ocean [TINRO) 87 :130-134 [Transifrom Russian by Pierson MOD

Kvitek RG, Oliver JS (1986) Side-scan sonar estimates of the utilization of gray whale feeding grounds along Vancouver Island, Canada. Cont Shelf Res 6:639-654

Kvitek RG, Fukuyama AK, Anderson BS, Grimm BK (in press) Sea otter foraging on deep-burrowing bivalves in a California coastal lagoon. Mar Bioi

Larson RJ, DeMartini EE (1984) Abundance and vertical distribution of fishes in a cobble-bottom kelp forest off San Onofre, California. US Nat! Oceanic Atmos Adm Fish Bull 82:37-53

Laur DR, Ebeling AW (1983) Predator-prey relationships in surfperches. Environ Bioi Fish 8 :217-229

Laur DR, Ebeling AW, ~eed DC (1986) Experimental evaluations of substrate types as barriers to sea urchin (Strongylocentrotus spp.) movement. Mar Bio 93:209-215

Lawrence JM (1975) On the relationships between marine plants and sea urchins. Oceanogr Mar Bioi Annu Rev 13 :213-286

228 References

Laws RM, Parker ISC, Johnstone RCB (1975) Elephants and their habitats. Clarendon, Oxford Leighton DL (1971) Grazing activities of benthic invertebrates in southern California kelp beds. In:

North WJ (ed) The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia 32: 421-453

Leighton DL, Jones LG, North WJ (1966) Ecological relationships between giant kelp and sea urchins in southern California. Proc Int Seaweed Symp 5 :141-153

Lensink CJ (1960) Status and distribution of sea otters in Alaska. J Mammal 41 :172-182 Lensink CJ (1962) The history and status of sea otters in Alaska. Thesis, Purdue University, Lafayet

te, Indiana Levin SA (1970) Community equilibria and stability, and an extension of the competitive exclu

sion principle. Am Nat 104:413-423 Levin SA (1976) Population dynamic models in heterogeneous environments. Annu Rev Ecol Syst

7:287-310 Levin SA (1978) Pattern formation in ecological communities. In: Steele JH (ed) Spatial pattern in

plankton communities. Plenum, New York, p 433 Levin SA (1981) The role of theoretical ecology in the description and understanding of popula

tions in heterogeneous environments. Am Zool21 :865 -875 Levin SA, Harwell MA (1986) Environmental risks and genetically engineered organisms. In:

Panem S (ed) Biotechnology - Implications for Public Policy. Brookings Inst, Washington DC, p 56

Levin SA, Kimball K (1984) New perspectives in ecotoxicology. Environ Manage 8:375-442 Levin SA, Paine RT (1974) Disturbance, patch formation, and community structure. Proc Nat!

Acad Sci USA 71:2744-2747 Levin SA, Paine RT (1975) The role of disturbance in models of community structure. In: Levin

SA (ed) Ecosystem analysis and prediction. Society for Industrial & Applied Mathematics, Philadelphia, p 56

Lewis JR (1964) The ecology of rocky shores. English Universities Press, London Lewontin RC (1969) The meaning of stability. Brookhaven Symp Bioi 22:13-24 Limbaugh C (1961) Observations of the California sea otter. J Mammal 42:271-273 Limburg KE, Levin SA, Harwell CC (1986a) Ecology and estuarine impact assessment: Lessons

learned from the Hudson River USA and other estuarine experiences. J Environ Manage (England) 22:255-280

Limburg KE, Morgan MA, McDowell WH (1986b) The Hudson River ecosystem. Springer, Berlin Heidelberg New York

Lissner AL (1980) Some effects of turbulence on the activity of the sea urchin Centrostephanus coronatus Verrill. J Exp Mar BioI EcoI48:185-193

Littler MM (1979) The effects of bottle volume, thallus weight, oxygen saturation levels, and water movement on apparent photosynthetic rates in marine algae. Aquat Bot 7 :21-34

Loughlin TR (1977) Activity patterns, habitat partitioning, and grooming behavior of the sea otter, Enhydra lutris, in California. Thesis, University of California, Los Angeles

Loughlin TR (1979) Radio telemetric determination of the 24-hour feeding activities of sea otters, Enhydra lutris. In: Amlaner CJ, MacDonald DW (eds) A handbook on biotelemetry and radiotracking. Pergamon, New York, p 717

Love MS, Ebeling AW (1978) Food and habitat of three switch-feeding fishes in the kelp forest off Santa Barbara, California. US Natl Oceanic Atmos Adm Fish Bull 76 :257 -271

Lowry LF, Pearse JS (1973) Abalones and sea urchins in an area inhabited by sea otters. Mar Bioi 23:213-219

Lowry LF, McElroy AJ, Pearse JS (1974) The distribution of six species of gastropod molluscs in a California kelp forest. BioI Bull (Woods Hole) 147:386-396

Lubchenco J, Menge BA (1978) Community development and persistence in a low rocky intertidal zone. Ecol Monogr 48 :67 -94

Lubina JA, Levin SA (1987) The spread of a reinvading species: range expansion in the California sea otter. The American Naturalist (in press)

Lyons KJ, Estes JA (1985) Individual variation in diet and foraging strategy in the female California sea otter, Enhydra /utris. Paper presented at the 66th Annual Meeting. Western Society of Naturalists, Monterey, California

References 229

MacArthur RH (1972) Strong, or weak, interactions? Trans Conn Acad Arts Sci 44:177-188 MacAskie IB (1984) Sea otter census, Vancouver Island, British Columbia. Unpublished Report,

West Coast Whale Research Foundation, Vancouver, British Columbia MacGinitie GE, MacGinitie N (1968) Natural history of marine animals. 2nd ed. McGraw-Hill, New

York Maminov MK, Shitikov AM (1970) On the food of the sea otter on Paramushir Island. Proc Pac Res

Inst Fish Ocean (TINRO) 70:176-179 (Translfrom Russian by Pierson MO) Mann KH (1973) Seaweeds: their productivity and strategy for growth. Science 182:975-981 Mann KH (1982) Ecology of coastal waters. University of California Press, Berkeley Martin PS (1973) The discovery of America. Science 179:969-974 Mattison JE, Trent JD, Shanks AL, Akin TB, Pearse JS (1977) Movement and feeding activity of

red sea urchins (Strongylocentrotus franciscanus) adjacent to a kelp forest. Mar Bioi 39 :25 -30 May RM (1973) Stability and complexity in model ecosystems. Princeton University Press, Prince

ton, New Jersey Maycock PF, Curtis IT (1960) The phytosociology of boreal conifer-hardwood forests in the Great

Lakes region. Ecol Monogr 30:1-35 McCleneghan K, Ames JA (1976) A unique method of prey capture by the sea otter, Enhydra lutris

Linnaeus. J Mammal 57 :410-412 McKenzie JD, Moore PG (1981) The microdistribution of animals associated with the bulbous hold

fasts of Saccorhiza polyschides (Phaeophyta). Ophelia 20:201-213 McLean JH (1962) Sublittoral ecology of kelp beds of the open coast area near Carmel, California.

Bioi Bull (Woods Hole) 122 :95 -114 Meistrell JC, Montagne DE (1983) Waste disposal in southern California and its effects on the rocky

subtidal habitat. In: Bascom W (ed) The effects of waste disposal on kelp communities. Southern California Coastal Water Resources Project, Long Beach, p 84

Menge BA (1976) Organization of the New England rocky intertidal community: Role of predation, competition, and environmental heterogeneity. Ecol Monogr 46:355-393

Merrill RJ, Hobson ES (1970) Field observations of Dendraster excentricus, a sand dollar of western North America. Am MidI Nat 83 :595 -·624

Miller DJ (1974) The sea otter, Enhydra lutris: Its life history, taxonomic status, and some ecological relationships. Marine Resources Leaflet 7, California Department of Fish & Game, Sacramento

Miller DJ (1980) The sea otter in California. Calif Coop Oceanic Fish Invest Rep 21 :79-81 Miller DJ, Geibel 11 (1973) Summary of blue rockfish and lingcod life histories; a reef ecology

study; and giant kelp, Macrocystis pyrifera, experiments in Monterey Bay, California. Calif Dep Fish Game Fish Bull 158:1-137

Miller DJ, Lea RN (1972) Guide to the coastal marine fishes of California. Calif Dep Fish Game Fish Bull 157:1-235

Miller DJ, Hardwick JE, Dahlstrom WA (1975) Pismo clams and sea otters. Calif Dep Fish Game Mar Resour Tech Rep 31:1-49

Miller RJ, Colodey AG (1983) Widespread mass mortalities of the green sea urchin in Nova Scotia, Canada. Mar Bioi 73:263-267

Miller RJ, Mann KH, Scarratt DJ (1971) Production potential of a seaweed-lobster community in eastern Canada. J Fish Res Board Can 28:1733-1738

Mitchell CT, Hunter JR (1970) Fishes associated with drifting kelp, Macrocystis pyrifera, off the coast of southern California and northern Baja California. Calif Fish Game 56 :288-297

Moffit FH (1954) Geology of the Prince William Sound region, Alaska. US Geol Surv Bull 989-E: 225-310

Moore PG (1971) The nematode fauna associated with holdfasts of kelp (Laminaria hyperborea) in northeast Britain. J Mar Bioi Assoc UK 51 :589-604

Moore PG (1974) The kelp fauna of northeast Britain. 3. Qualitative and quantitative ordinations and the utility of a multivariate approach. J Exp Mar Bioi Eco116:257 -300

Morin JG, Harrington A (1979) California marine waters areas of special biological significance reconnaissance survey report: Mugu Lagoon to Latigo Point. California State Water Resources Control Board, Water Quality Monitoring Report 79-5, Sacramento

Morin JG, Kastendiek JE, Harrington A, Davis N (1985) Organization and patterns of interactions in a subtidal sand community on an exposed coast. Mar Ecol Prog Ser 27 :163-185

230 References

Morris RH, Abbott DP, Haderlie EC (1980) Intertidalinvertebrates of California. Stanford University Press, Stanford, California

Morrison P, Rosenmann M, Estes JA (1974) Metabolism and thermoregulation in the sea otter. Physiol Zool4 7 :218-229

Moss IE, Lawrence JM (1972) Changes in carbohydrate, lipid, and protein levels with age and season in the sand dollar Mellita quinquiesperforata (Leske). J Exp Mar Bioi EcoI8:225-239

Murie OJ (1940) Notes on the sea otter. J Mammal 21:119-131 Muus K (1973) Settling, growth and mortality of young bivalves in the Oresund. Ophelia 12:79-116 Nakamura M (1985) Evolution of artificial fishing reef concepts in Japan. Bull Mar Sci 37:271-278 National Ocean Service (1985) Tide tables 1986. High and low water predictions. West coast of

North and South America, including the Hawaiian Islands. US Department of Commerce, Rockville, Maryland

Nelson WG (1979) Experimental studies of selective predation on amphipods: consequences for amphipod distribution and abundance. J Exp Mar Bioi Ecol 38 :225 -245

Nerini MK, Oliver JS (1983) Gray whales and the structure of the Bering Sea benthos. Oecologia (Ber!) 59:224-225

Neushul M (1965) SCUBA diving studies of the vertical distribution of benthic marine plants. In: Proceedings of the fifth marine biology symposium. Acta Universitatis, Goteburg, p 161

Neushul M (1971) Submarine illumination in Macrocystis beds. In: North WJ (ed) The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia 32:241-254

Neushul M (1972) Functional interpretation of benthic marine algal morphology. In: Abbott lA, Kurogi M (eds) Contributions to the Systematics of Benthic Marine Algae of the North Pacific. Japanese Society of Phycology, Kobe, p 47

Newby TC (1975) A sea otter (Enhydra lutris) food dive record. Murrelet 56 :19 Newell RC, Lucas MI, Velimirov B, Seiderer LJ (1980) Quantitative significance of dissolved organic

losses following fragmentation of kelp (Ecklonia maxima and Laminaria pallida). Mar Ecol Prog Ser 2:45-59

Newell RC, Field JG, Griffiths CL (1982) Energy balance and significance of micro-Qrganisms in a kelp bed community. Mar Ecol Prog Ser 8:103-113

North WJ (1963) Ecology of the rocky nearshore environment in southern California and possible influences of discharged wastes. Int J Air Water Pollut 7 :721-736

North WJ (1965) Annual report of the kelp habitat improvement project, 1964-1965. Keck Lab, Calif Inst Tech, Pasadena

North WJ (1971) Introduction and background. In: North WJ (ed) The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia 31 :1-97

North WJ (1974) Annual report of the kelp habitat improvement project, 1973-1974. Keck Lab, Calif Inst Tech, Pasadena

North WJ (1976) Aquaculture techniques for creating and restoring beds of giant kelp, Macrocystis spp. J Fish Res Board Can 33 :1015-1023

North WJ (1983) The sea urchin problem. In: Bascom W (ed) The effects of waste disposal on kelp communities. Southern California Coastal Water Resources Project, Long Beach, p 147

North WJ (1985) Health of kelp beds. In: Yuge IE, Elliott JL (eds) Marine environmental analysis and interpretation, San Onofre Nucleur Generating Station, Report on 1985 data 85-RD-37. Southern California Edison Co, Pasadena, pp 6-1

North WJ, Pearse JS (1970) Sea urchin population explosion in southern California coastal waters. Science 167:209

North WJ, Anderson EK, Chapman FA (1979) Marine ecological transect studies. In: Environmental investigations at Diablo Canyon, 1975 -1977 , vol I. Pacific Gas & Electric Company, San Ramon, California (unpaginated)

Norton TA, Mathieson AC, Neushul M (1982) A review of some aspects of form and function in seaweeds. Bot Mar 25 :501-510

Nybakken JW (1969) Pre~arthquake intertidal ecology of Three Saints Bay, Kodiak Island, Alaska. Bioi Pap Univ Alaska 9:1-117

Nybakken JW (1982) Marine biology: an ecological approach. Harper & Row, New York Ogden A (1941) The California sea otter trade, 1784-1848. University of California Press, Berkeley Oliver JS, Slattery PN (1985) Destruction and opportunity on the sea floor: effects of gray whale

feeding. Ecology 66 :1965-1975

References 231

Oliver IS, Slattery PN, Hulberg LW, Nybakken IW (1980) Relationships between wave disturbance and zonation of benthic invertebrate communities along a subtidal high-energy beach in Monterey Bay, California. US Nat! Oceanic Atmos Adm Fish Bull 78:437-454

Oliver IS, Slattery PN, O'Connor EF, Lowry LF (1983) Walrus, Odobenus rosmarus, feeding in the Bering Sea: a benthic perspective. US Nat! Oceanic Atmos Adm Fish Bull 81 :501-512

Oliver IS, Kvitek RG, Slattery PN (1985) Walrus disturbance: Scavenging habits and recolonization of the Bering Sea benthos. I Exp Mar Bioi Ecol91 :233-246

Ostfeld RS (1982) Foraging strategies and prey switching in the California sea otter. Oecologia (Berl) 53:170-178

Paine RT (1966) Food web complexity and species diversity. Am Nat 100:65-75 Paine RT (1969) A note on trophic complexity and community stability. Am Nat 103:91-93 Paine RT (1971) A short-term experimental investigation of resource partitioning in a New Zea-

land rocky intertidal habitat. Ecology 52:1096-1106 Paine RT (1974) Intertidal community structure: Experimental studies on the relationship between

a dominant competitor and its principal predator. Oecologia (Berl) 15 :93-120 Paine RT (1976) Size-limited predation: An observational and experimental approach with the

Mytilus-Pisaster interaction. Ecology 57 :858-873 Paine RT (1977) Controlled manipulations in the marine intertidal zone and their contributions to

ecological theory. Spec Publ, Acad Nat Sci Philadelphia 12:245-270 Paine RT (1980) Food webs: linkage interaction strength and community infrastructure. I Anim

Ecol49 :667 -685 Paine RT (1984) Ecological determinism in the competition for space. Ecology 65 :1339-1348 Paine RT, Levin SA (1981) Intertidal landscapes: Disturbance and the dynamics of pattern. Ecol

Monogr 51:145-178 Paine RT, Castillo IC, Cancino J (1985) Perturbation and recovery patterns of starfish-dominated

intertidal assemblages in Chile, New Zealand, and Washington state. Am Nat 125 :679-691 Palmisano JF (1975) Sea otter predation: its role in rocky intertidal community structure at Am

chitka and other Aleutian Islands. Thesis, University of Washington, Seattle Palmisano JF (1983) Sea otter predation: its role in structuring rocky intertidal communities in the

Aleutian Islands, Alaska USA. Acta Zool Fenn 174:209-211 Palmisano JF, Estes JA (1977) Ecological interactions involving the sea otter. In: Merritt MS,

Fuller RG (eds) The environment of Amchitka Island, Alaska. US Energy Research & Development Administration, Springfield, Virginia, p 527

Patton ML (1985) Changes in fish abundance in the southern California bight during a warm-water episode. In: Yuge JE, Elliott JL (eds) Marine environmental analysis and interpretation San Onofre Nuclear Generating Sation, Report on 1985 data 85-RD-37. Southern California Edison Co, Pasadena, p 8-1

Patton ML, Grove RS, Harman RF (1985) What do natural reefs tell us about designing artificial reefs in southern California? Bull Mar Sci 37 :279-298

Paul AJ, Feder HM (1976) Clam, mussel, and oyster resources of Alaska. University of Alaska Institute of Marine Science Report 76-4, Fairbanks, Alaska

Payne SF, Jameson RJ (1984) Early behavioral development of the sea otter, Enhydra lutris. J Mammal 65 :527 -531

Pearse JS, Hines AH (1979) Expansion of a central California kelp forest following the mass mortality of sea urchins. Mar Biol51 :83-91

Pearse JS, Clark ME, Leighton DL, Mitchell CT, North WJ (1970) Marine waste disposal and sea urchin ecology. In: North WJ (ed) Kelp habitat improvement project annual report 1969-1970. California Institute of Technology, Pasadena

Pearse JS. Costa DP, Yellin MB. Agegian CR (1977) Localized mass mortality of red sea urchins, Strongylocentrotus franciscanus, near Santa Cruz, California. US Nat! Oceanic Atmos Adm Fish Bull 75:645-648

Pennington JT, DeMartini JD, Pic'l G, Thoney D (1981) California marine waters areas of special biological significance reconnaissance survey report: Pygmy Forest Ecological Staircase. California State Water Resources Control Board, Water Quality Monitoring Report 81-3, Sacramento

Pequegnant WE (1964) The epifauna of a California siltstone reef. Ecology 45 :272-283

232 References

Peterson CH (1979a) The importance of predation and competition in organizing the intertidal epifaunal communities of Barnegat Inlet, New Jersey. Oecologia (Berl) 39:1-24

Peterson CH (1979b) Predation, competitive exclusion, and diversity in the soft-sediment benthic communities of estuaries and lagoons. In: Livingston RJ (ed) Ecological processes in coastal and marine systems. Plenum, New York, p 233

Peterson CH (1982) The importance of predation and intra- and interspecific competition in the population biology of two infaunal suspension-feeding bivalves, Protothaca staminea and Chione undatella. Ecol Monogr 52:437-475

Peterson CH (1983) Interactions between two infaunal bivalves, Chione undatella (Sowerby) and Protothaca staminea (Conrad), and two potential enemies, Crepidula onyx (Sowerby) and Cancer anthonyi (Rathbun). J Exp Mar Bioi Ecol 68:145-158

Peterson CH, Andre SV (1980) An experimental analysis of interspecific competition among marine filter-feeders in a soft sediment environment. Ecology 61: 129-139

Peterson CH, Summerson HC, Duncan PB (1984) The influence of seagrass cover on population structure and individual growth rate of a suspension-feeding bivalve, Mercenaria mercenaria. J Mar Res 42:123-138

Petit C, Trinite M, Valentin P (1976) Study of turbulent diffusion above and within a forest - application in the case of S02. Atmos Environ 10:1057 -1063

Philander SGH (1983) El Nino southern oscillation phenomena. Nature 302 :937 -948 Pickett STA, White PS (eds) (1985) The ecology of natural disturbance and patch dynamics.

Academic Press, Orlando, Florida Pitcher KW (1975) Distribution and abundance of sea otters, Steller sea lions, and harbor seals in

Prince William Sound, Alaska. In: Calkins DG, Pitcher KW, Schneider K (eds) Distribution and abundance of marine mammals in the Gulf of Alaska. Alaska Department of Fish & Game, Anchorage. Appendix 1, P 1

Platt JR (1964) Strong inference. Science 146:347 -353 Platt WJ (1975) The colonization and formation of equilibrium plant species associations on badger

disturbances in a tall1P"ass prairie. Ecol Monogr 45 :285 -305 Pringle JD (1986) A review of urchin/macro-algal associations with a new synthesis for nearshore,

eastern Canadian waters. Monogr Bioi 4:191-218 Quast JC (1968a) Fish fauna of the rocky inshore zone. In: North WJ, Hubbs CL (eds) Utilization

of kelp-bed resources in southern California. Calif Dep Fish Game Fish Bull 139 : 35 -55 Quast Je (1968b) Effects of kelp harvesting on the fishes of the kelp beds. In: North WJ, Hubbs

CL (eds) Utilization of kelp-bed resources in southern California. Calif Dep Fish Game Fish Bull 139:143-149

Quast JC (1968c) Estimates of the populations and the standing crop of fishes. In: North WJ, Hubbs CL (eds) Utilization of kelp-bed resources in southern California. Calif Dep Fish Game Fish Bull 139 :57 -79

Radinsky LB (1968) Evolution of somatic sensory specialization in otter brains. J Comp Neurol 134 :495 -506

Ralls K, Siniff D, Monett C, Eagle T, Ferm L (1985) Summary of information pertaining to California permit to capture sea otters for scientific research. Report, California Fish & Game Commission, San Rafael

Raynor GS, Hayes JV, Ogden EC (1974) Particulate dispersion into and within a forest. BoundaryLayer Meteorol 7 :429-456

Reed DC, Foster MS (1984) The effects of canopy shading on algal recruitment and growth in a giant kelp forest. Ecology 65 :937 -948

Reise K (1981) Gnathostomulida abundant alongside polychaete burrows. Mar Ecol Prog Ser 6: 329-333

Reise K (1985) Tidal flat ecology. An experimental approach to species interactions. Springer, Berlin Heidelberg New York (Ecological studies, vol 54)

Repenning CA (1976a) Enhydra and Enhydriodon from the Pacific coast of North America. J Res US Geol Surv 4:305-315

Repenning CA (1976b) Adaptive evolution of the sea lions and walruses. Syst Zool 25 :375 -390 Ribic CA (1982) Autumn activity of sea otters in California. J Mammal 63 :702-706

References 233

Ricketts EF, Calvin J, Hedgpeth JW (1985) Between Pacific tides, 5th edn. Stanford University Press, Stanford, California

Riedman ML, Hines AH, Pearse JS (1981) Spatial segregation of four species of turban snails (Gastropoda: Tegula) in central California. Veliger 24:97 -102

Risser PG, Karr JR, Forman RTT (1984) Landscape ecology. Directions and approaches. Illinois Natural History Survey Special Publication 2, Champaign, Illinois

Ristau DA, Tarp C, Hand C (1979) California marine waters areas of special biological significance reconnaissance survey report: Bodega Marine Life Refuge. California State Water Resources Control Board, Water Quality Monitoring Report 79--16, Sacramento

Robertson AI, Lenanton RCJ (1984) Fish community structure and food chain dynamics in the surf-zone of sandy beaches: the role of detached macrophyte detritus. J Exp Mar Bioi Eco184: 265-283

Robertson AI, Lucas JS (1983) Food choice, feeding rates, and the turnover of macrophyte biomass by a surf-zone inhabiting amphipod. J Exp Mar Bioi Ecol 72:99-124

Rosenthal RJ (1980) Shallow water fish assemblages in the northeastern Gulf of Alaska: habitat evaluation, species composition, abundance, spatial distribution and trophic interaction. Unpublished Report, NOAA Outer Continental Shelf Environmental Assessment Project Office, Juneau, Alaska

Rosenthal RJ, Barilotti DC (1973) Feeding behavior of transplanted sea otters and community interactions off Chichagof Island, southeast Alaska. In: North WJ (ed) California Institute of Technology, Kelp Habitat Improvement Project. Annual Report, 1 July 1972-30 June 1973, p 74

Rosenthal RJ, Clarke WD, Dayton PK (1974) Ecology and natural history of a stand of giant kelp, Macrocystis pyr(fera, off Del Mar, California. US Natl OCEANIC Atmos Adm Fish Bull 72 :670-684

Runkle JR (1979) Gap phase dynamics in climax mesicforests. Thesis, Cornell University, Ithaca, New York

Salt GW (ed) (1984) A round table on research in ecology and evolutionary biology. Am Nat 122: 583-705

Sandegren FE, Chu EW, Vandevere IE (1973) Maternal behavior in the California sea otter. J Mammal 54 :668-679

Santelices B, Correa J, Avila M (1983) Benthic algal spores surviving digestion by sea urchins. J Exp Mar Bioi Ecol 70 :263-269

Scheffer VB (1981) The amazing sea otter. Scribner, New York Scheibling R (1986) Increased macroalgal abundance following mass mortalities of sea urchins

(Strongylocentrotus droebachiensis) along the Atlantic coast of Nova Scotia. Oecologia (Berl) 68:186-198

Schembri PJ, Jaccarini V (1978) Some aspects of the ecology of the Echiuran worm Bonellia viridis and associated infauna. Mar Bioi 47:55-61

Schiel DR (1981) A demographic and experimental evaluation of plant and herbivore interaction in subtidal algal stands. Thesis, University of Auckland, New Zealand

Schiel DR, Foster MS (1986) The structure of subtidal algal stands in temperate waters. Oceanogr Mar Bioi Annu Rev 24:265-307

Schmitt RJ, Holbrook SJ (1984) Gape-limitation, foraging tactics and prey size selectivity of two microcarnivorous species of fish. Oecologia (Berl) 63 :6-12

Schmitt RJ, Holbrook SJ (1985) Patch selection by juvenile black surfperch (Embiotocidae) under variable risk: interactive influence of food quality and structural complexity. J Exp Mar Bioi Ecol85 :269-285

Schneider KB, Faro JB (1975) Effects of sea ice on sea otters (En hydra lutris). J Mammal 56:91-101 Schoener TW (1983) Field experiments on interspecific competition. Am Nat 122:240-285 Sebens KP (1984) Water flow and coral colony size: interhabitat comparisons of the octocoral

Alcyonium siderium. Proc NatJ Acad Sci USA 81 :5473-5477 Seymour RJ, Strange RR III, Cayan DR, Nathan RA (1984) Influence of EI Nifios on California's

wave climate. In: Edge BL (ed) Nineteenth coastal engineering conference, proceedings of the international conference, 1984, Houston, Texas. American Society of Civil Engineers, New York, p577

234 References

Sheppard CRC, Bellamy DJ, Sheppard ALS (1980) Fauna inhabiting the holdfasts of Laminaria hyperborea along some environmental and geographical gradients. Mar Environ Res 4:25 -51

Shimek SJ (1977a) The underwater foraging habits of the sea otter, Enhydra lutris. Calif Fish Game 63:120-122

Shimek SJ (1977b) Sea otter predation on Pismo clams in Monterey Bay. Calif Fish Game 63: 117-120

Shimek SJ, Monk A (1977) Daily activity of sea otters off the Monterey Peninsula, California. J Wildl Manage 41 :277 -283

Shitikov AM (1971) The influence of the trophic factor on the abundance and distribution of the sea otter in the northern and middle Kurile Islands. Proc Pac Res Inst Fish Ocean (TINRO) 80: 239-277 (Trans I from Russian by Pierson MO)

Shitikov AM (1973) The feeding of sea otters at Urup and Simushir Islands. Proc Pac Res Inst Fish Ocean (TINRO) 87: 135 -138 (Transl from Russian by Pierson MO)

Shitikov AM, Lukin VI, Chaplygina SF (1973) The character of the upper sublittoral of Paramushir Island as a foraging habitat of the sea otter. Proc Pac Res Inst Fish Ocean (TINRO) 87: 130-134 (Transl from Russian by Pierson MO)

Shulman MJ (1985) Recruitment of coralreeffishes: effects of distribution of predators and shelter. Ecology 66:1056-1066

Sieburth JMcN (1969) Studies on algal substances in the sea. 3. The production of extracellular organic matter by littoral marine algae. J Exp Mar Bioi Ecol 3:290-309

Sih A, Crowley P, McPeek M, Petranka J, Strohmeier K (1985) Predation, competition, and prey communities: a review of field experiments. Annu Rev Ecol Syst 16:269-311

Simenstad CA, Estes JA, Kenyon KW (1978) Aleuts, sea otters, and alternate stable-state communities. Science 200:403-411

Simon-Jackson T, Taylor D, Schliebe S, Vivion M (1985) Sea otter survey, Kodiak Island - 1984. Unpublished Report, US Fish & Wildlife Service, Anchorage, Alaska

Simpson GG, Roe A, Lewontin RC (1960) Quantitative zoology. Harcourt Brace Jovanovich, New York

Siniff DB, Ralls K (1986) Summary of information obtained on sea otters for MMS study on population status of California sea otters. Unpublished report, University of Minnesota, Minneapolis

Sokal RR, Rohlf FJ (1981) Biometry, 2nd edn. Freeman, San Francisco Sousa WP (1984a) Intertidal mosaics, patch size, propagule availability, and spatially variable pat

terns of succession. Ecology 65 :1918-1935 Sousa WP (1984b) The role of disturbance in natural communities. Annu Rev Ecol Syst 15 :353-391 Sousa WP (1985) Disturbance and patch dynamics on rocky intertidal shores. In: Pickett ST A,

White PS (eds) The ecology of natural disturbance and patch dynamics. Academic Press, Orlando, Florida, p 101

Steneck RS, Adey WH (1976) The role of environment in control of morphology in Lithophyllum congestum, a Caribbean algal ridge builder. Bot Mar 19:197 -215

Stephens JS, Zerba KE (1981) Factors affecting fish diversity on a temperate reef. Environ Bioi Fish 6:111-121

Stephens JS, Morris PA, Zerba KE. Love MS (1984) Factors affecting fish diversity on a temperate reef: the fish assemblage of Palos Verdes Point, 1974-1981. Environ Bioi Fish 11 :259-275

Stephenson MD (1977) Sea otter predation on Pismo clams in Monterey Bay. Calif Fish Game 63: 117-120

Stephenson TA, Stephenson A (1972) Life between tidemarks on rocky shores. Freeman, San Francisco

Stoner AW (1982) The influence of benthic macrophytes on the foraging behavior of pinfish, Lagodon rhomboides (Linnaeus). J Exp Mar Bioi Ecol58 :271-284

Stouder DJ (1983) Effects of a severe weather disturbance on foraging patterns of a guild of surfperches. Thesis, University of California, Santa Barbara

Stout WE (1970) Some associates of Tresus nuttallii (Conrad, 1837). Veliger 13:67-70 Strathmann R (1978) Length of pelagic period in echinoderm larvae from the northeast Pacific.

J Exp Mar Bioi EcoI34:23-27 Suchanek TH (1978) The ecology of Mytilus edulis L. in exposed rocky intertidal communities. J

Exp Mar Bioi Ecol 31: 105 -120

References 235

Suchanek TH (1979) The Mytilus californianus community: Studies on the composition, structure, organization, and dynamics of a mussel bed. Thesis, University of Washington, Seattle

Suchanek TH (1981) The role of disturbance in the evolution of life history strategies in the intertidal mussels Mytilus edulis and Mytilus californianus. Oecologia (Bed) 50 :143-152

Suchanek TH (1985) Mussels and their role in structuring rocky shore communities. In: Moore PG, Seed R (eds) The ecology of rocky coasts. Hodder & Stoughton, London, p 70

Sutherland JP (1974) Multiple stable points in natural communities. Am Nat 108:859-873 Tegner MJ (1980) Multispecies considerations of resource management in southern California kelp

beds. In: Pringle JD, Sharp GJ, Caddy JF (eds) Proceedings of the workshop on the relationship between sea urchin grazing and commercial plant/animal harvesting. Can Tech Rep Fish Aquat Sci 954:125-143

Tegner MJ, Dayton PK (1981) Population structure, recruitment and mortality of two sea urchins (Strongylocentrotus franciscanus and S. purpuratus) in a kelp forest. Mar Ecol Prog Ser 5 :255-268

Tegner MJ, Levin LA (1982) Do sea urchins and abalones compete in California kelp forest communities? In: Lawrence JM (ed) Echinoderms: proceedings of the international conference, Tampa Bay. Balkema, Rotterdam, p 265

Tomlinson PK (1968) Mortality, growth, and yield per recruit for Pismo clams. Calif Fish Game 54:100-107

Trivers R (1985) Social evolution. Benjamin/Cummings, Menlo Park, California Turner CH, Ebert EE, Given RR (1968) The marine environment offshore from Point Loma, San

Diego County. Calif Dep Fish Game Fish Bu11140:1-85 Turner CH, Ebert EE, Given RR (1969) Man-made reef ecology. Calif Dep Fish Game Fish Bull

146:1-221 Underwood AJ, Denley EJ (1984) Paradigms, explanations, and generalizations in models for the

structure of intertidal communities on rocky shores. In: Strong DR, Simberloff D, Abele LG, Thistle AB (eds) Ecological communities: conceptual issues and the evidence. Princeton University Press, Princeton, New Jersey, p 151

US Fish and Wildlife Service (1982) Southern sea otter recovery plan. US Fish & Wildlife Service, Washington DC

US Fish and Wildlife Service (1986) Draft environmental impact statement. Proposed translocation of southern sea otters. US Fish & Wildlife Service, Sacramento, California

Vadas RL (1968) The ecology of Agarum and the kelp bed community. Thesis, University of Washington, Seattle

Vadas RL (1977) Preferential feeding: an optimization strategy in sea urchins. Ecol Monogr 47: 337-371

Vadas RL, Elner RW, Garwood PE, Babb IG (1986) Experimental evaluation of aggregation behavior in the sea urchin Strongylocentrotus droebachiensis: a reinterpretation. Mar Bioi 90: 433-448

Valiela I (1984) Marine ecological processes. Springer, Berlin Heidelberg New York VanBlaricom GR (1982) Experimental analyses of structural regulation in a marine sand community

exposed to oceanic swell. Ecol Monogr 52:283-305 VanBlaricom GR (1984) Relationships of sea otters to living marine resources in California: a new

perspective. In: Lyle V (ed) Collection of papers presented at the Ocean Studies Symposium, Nov 7 -11 1982, Asilomar, California. California Coastal Commission & California Department of Fish & Game, Sacramento, p 361

VanBlaricom GR (in press) Regulation of mussel population structure in Prince William Sound, Alaska. Natl Geogr Res

VanBlaricom GR, Estes JA (1986) Demography of intertidal mussel populations in Prince William Sound, Alaska: Effects of sea otters as mussel predators. Final Report, Grant 2883-84, National Geographic Society, Washington DC

Vanderwilt DJ, Cazanjian GV, Hurley AC, Foster MS (1979) California marine waters areas of special biological significance reconnaissance survey report: James V. Fitzgerald Marine Reserve. California State Water Resources Control Board, Water Quality Monitoring Report 79-12, Sacramento

236 References

Vandevere JE (1969) Feeding behavior of the southern sea otter. In: Proceedings, Sixth Annual Conference on Biological Sonar and Diving Mammals. Stanford Research Institute, Menlo Park, California, p 87

Van Dolah RF (1978) Factors regulating the distribution and popUlation dynamics of the amphipod Gammarus palustris in an intertidal salt marsh community. Ecol Monogr 48:191-217

Van Wagenen RF ,Foster MS,Burns F (1981) Sea otter predation on birds near Monterey, California. J Mammal 62:433-434

Van Zyll de Jong CG (1972) A systematic review of the nearctic and neotropical river otters (genus Lutra, Mustelidae, Carnivora). R Ont Mus Life Sci Contrib 80:1-104

Venrick EL (1978) Systematic sampling in a planktonic ecosystem. US Nat! Oceanic Atmos Adm Fish Bull 76:617 -627

Villalta Comella JF, Crusafont-Pairo M (1945) Enhydriodon lluecai nova sp el primer Lutrido del Pontiense espanol. Real Soc Esp Hist Nat Bolet Sec GeoI43:383-396

Virnstein RW (1977) The importance of predation by crabs and fishes on benthic infauna in Chesapeake Bay. Ecology 58:1199-1217

Virnstein RW (1978) Predator caging experiments in soft-sediments: caution advised. In: Wiley ML (ed) Estuarine Interactions. Academic Press, New York, p 261

Wade LS (1975) A sea otter possibly feeding on Pismo clams. J Mammal 56:720-721 Walters C (1986) Adaptive management of renewable resources. Macmillan, New York Warner RR, Chesson PL (1985) Coexistence mediated by recruitment fluctuations: a field guide to

the storage effect. Am Nat 125 :769-787 Watanabe JM (1984a) Food preference, food quality and diets of three herbivorous gastropods

(Trochidae: Tegula) in a temperate kelp forest habitat. Oecologia (Berl) 62:47-52 Watanabe JM (1984b) The influence of recruitment, competition, and benthic predation on spatial

distributions of three species of kelp forest gastropods (Trochidae: Tegula). Ecology 65 :920-936

Watt AS (1947) Pattern and process in the plant community. J Ecol35 :1-22 Wendell FE, Ames JA, Hardy RA (1985) Assessment of the accidental take of sea otters, Enhydra

lutris, in gill and trammel nets. Unpublished report, Marine Resources Branch, California Department of Fish & Game, Sacramento

Wendell FE, Hardy RA, Ames JA, Burge RT (1986) Temporal and spatial patterns in sea otter (En hydra lutris) range expansion and in the loss of the Pismo clam fisheries. Calif Fish Game 72:197-212

West L (1986) Interindividual variation in prey selection by the snail Nucella (=Thais) emarginata. Ecology 67 :798 -809

Whedon WF (1936) Spawning habits of the mussel Mytilus californianus with notes on the possible relation to mussel poison. Univ Calif Berkeley Publ Zool41 :35-44

Wheeler WN (1980) Effect of boundary layer transport on the fixation of carbon by the giant kelp Macrocystis pyrifera. Mar Bioi 56:103-110

Whittaker RH (1956) Vegetation of the Great Smoky Mountains. Ecol Monogr 26:1-80 Whittaker RH, Levin SA (1977) The role of mosaic phenomena in natural communities. Theor

PopuIBioI12:117-139 Wickham DA, Russell GR (1974) Evaluation of midwater artificial structures for attracting coastal

pelagic fishes. US Nat! Oceanic Atmos Adm Fish Bull 72:181-191 Wild PW, Ames JA (1974) A report on the sea otter Enhydra lutris, L. in California. Calif Dep

Fish Game Mar Resour Tech Rep 20:1-94 Wild PW, Tasto RN (1983) Life history, environment, and mariculture studies of the Dungeness

crab, Cancer magister, with emphasis on the central California fishery. Calif Dep Fish Game Fish Bull 172:1-352

Wing BL, Clendenning KA (1971) Kelp surfaces and associated invertebrates. In: North WJ (ed) The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia 32 :319-342

Williams CS (1938) Notes on food of the sea otter. J Mammal 19:105-107 Williams JG (1980) The influence of adults on the settlement of spat of the clam, Tapes japonica.

J Mar Res 38:729-741

References 237

Wilson KC, Mearns AJ, Grant JJ (1980) Changes in kelp forests at Palos Verdes. In: Bascom W (ed) Coastal water research project biennial report 1979-1980. Southern California Coastal Water Research Project, Los Angeles, p 77

Witman JD, Suchanek TH (1984) Mussels in flow: Drag and dislodgement by epizoans. Mar Ecol Prog Ser 16:259-268

Woodhouse CD, Cowen RK, Wilcoxen LR (1977) A summary of knowledge of the sea otter Enhydra lutris, 1., in California, and an appraisal of the completeness of biological understanding of the species. Report MMC-76/02. National Technical Information Service, Springfield, Virginia

Woodin SA (1976) Adult-larval interactions in dense infaunal assemblages: Patterns of abundance. J Mar Res 34:25-41

Woodin SA (1981) Disturbance and community structure in a shallow water sand flat. Ecology 62: 1052-1066

Woolfenden J (1979) The California sea otter: saved or doomed? Boxwood, Pacific Grove, California Wulff FV, Field JG (1983) Importance of different trophic pathways in a nearshore benthic com

munity under upwelling and downwelling conditions. Mar Ecol Prog Ser 12:217 -228 Yellin MB, Agegian CR, Pearse JS (1977) Ecological benchmarks in the Santa Cruz County kelp

forests before the re~stablishment of sea otters. Report MMC-76/04. National Technical Information Service, Springfield, Virginia

Young RT (1942) Spawning season of the California mussel, Mytilus californianus. Ecology 23: 490-492

Young RT (1946) Spawning and settling season of the mussel, Mytilus californianus. Ecology 27: 354-363

Subject Index

Abruone 13,20,22,39,43,57,65,92,96, 136,151-152,154,158-159,164,167, 194,207,212-213

Aboriginru man 7, 88, 90, 102, 138 Adak Island (Alaska) 196 Air-lift samples 155 Alaria fistulosa 127, 130, 136 Alaria marginata 54 Alaska Penn insula (Alaska) 9,11,26,29,37 Alcyonarium siderium 199 Aleutian Islands (Alaska) 1,9,13-18,21,

87-88,119,123,127,130,136,140-141, 183,185,192,196-197,212

Algru-dominated assemblages 119,123,130, 136-137,139,141

Alternate stable states 92-93, 106-107 Amak Island (Alaska) 42 AmchitkaIsland (Alaska) 7-9,14-15,119,

196,212 Amphipods 155,160,162 Anemones 158,181 Ano Nuevo Island (Cruifornia) 105 Aonyx 4 Aptocyclus 15 Areas of Speciru Biologicru Significance

(California) 10 1 Argopecten irradions 199 Articulated corruline algae 127,131,159 Attu Island (Alaska) 14,123-127,129-130,

141,196,213 Avila (Cilifornia) 152-153

Bacteria 195 "Bad science" 217 Baja Cruifornia (Mexico) 11,21 Balanus crenatus 160 Balanus spp. 13 Barnacles 13,85, 160, 197 Barotropic currents 198 Barren grounds 98,152, 154, 164, 167 -168,

170-171,173,175-181,183-188, 190-191

Bay mussels 29,40,212-213

Bay scrulops 199 Bering Expedition 7 Bering Island (USSR) 9 Bering land bridge 6 Bering Sea 9,26 Big Sur (Cruifornia) 105 Birds 193 Bivruves 15,21,70 Black perch 176 Blacksmith 171,174,176,179-180,182,

187-188 Blue crabs 29-30 Blue rockfish 163, 165, 182, 188 Boca de Quadra Fjord (Alaska) 29 Botryoglossum farlowianum 159 Brachyistiusfrenatus 163,165,174,186 Bristol Bay (Alaska) 2 Bryozoa 127,155,181,199 Bull kelp 119 Butter clams 36-37,40,44

Calliarthron cheilosporioides 159 Calliostoma spp. 58 Callophyllis pinnata 159 Callorhinus ursin us 8 Cancer anthonyi 29-30 Cancer magister 19, 28,46 Cancer producta 28,46 Cancer spp. 13,29-30,38-39,43,46 Canopy-dwelling fish 163 Cape Alava (Washington) 85 Capitola (California) 29 Carbon 12 197 Carbon 13 197 Carmel Bay (Cruifornia) 105 Carnassiru teeth in otters 4, 6 Carnivora 4 Casco Point (Alaska) 127, 129 Cellaria sp. 127 Channel Island (Alaska) 73 Channel Islands (Cruifornia) 139, 141 Channel Islands Research Project 140 Chione undatella 29

240

Chitons 13 Chromispunctipinnis 166,171,174,179,

182-183 Cinetransects 156-157,162 Clams 13-14,86,211-212 Clathromorphum spp. 127,131,135 Cockles 30 Cold Bay (Alaska) 9 Commander Islands (USSR) 8-9,87,138 Communities - Biomass 48 - Competition 204,218 - Consumer-prey interactions 218 - Continua 215 - Disturbance 204,208 - Diversity 204-206,208 - Ecology 210 - Gap-phase phenomona 204 - Generality 217 - Heterogeneity 206 - Important species 1 - Keystone species 207 - Organization 97

Patch dynamics 204,206-207 - Prediction 207-208,217-218

Productivity 206 - Scale 203, 213

Stability 205 -206 - Stochastic phenomona 203 - Succession 204 - Trigger species 207 - Types 170,188,190,215 - Variation in space 203,206-207,

213-214,218 - Variation in time 203,206,213-214,218 Conservation of Ecosystems 1 Coral Reefs 92, 204 Cordova (Alaska) 69,88 Corynactis attenuata 181 Corynactis califomica 158 Crabs 14,70,86-87,90,158,207,211-212 Crevices as refugia from predators 43,73.85,

158, 164, 171 Crisia sp. 127 Crustaceans 42 Crustose coralline algae 154-155,159-160,

162, 164 Cryptochiton stelleri 13 Cyclopterichthys 15 Cystoseira osmundacea 105, 127, 129

Damalichthysvacca 163,174,176 DDT 207 Decapods 211 "Demersal microcamivores" 174

Subject Index

Dendraster excentricus 22,29,31 Deposit feeders 198 Desmerestia sp. 127 Destruction Island (Washington) 85, 88 Detritivores 211 Detritus 116-117, 136-137 Detritus-based food webs 200-201,217 Devereux Point (California) 170-171,

173-174,176-177,180-181 Diatoms 136 Dissolved organic carbon 194-196 Drift algae 158-159,164,167-169,171,

184,186-187,189,191,196,200-201 Dungeness crabs 18-19,28,46 Dusignathinae 7

Echinoids 93,96-97,99, 102, 106 Echiurids 13,86 Echiurus echiurus 37, 42 Ecklonia maxima 194 Ecosystem health 203-205 Ecosystem stability 203 Eelgrass 199 Eisenia arborea 130 Elkhorn Slough (California) 14,32,37,40,

42,44 El Niii.o/Southern oscillation 136, 166,

169-171,174,176-177,179,181,183, 188, 191

Embiotoca 163 Embiotocajacksoni 173-174,176,181 Embiotoca lateralis 162-163,166,174,176 Embiotocidae 171,174 Encrusting coralline algae 117, 127 Endangered species 203 Enhydra 4, 6-7, 20, 138 Enhydra [utris (see also Sea otter) 1, 4, 20,

92-93,106,151,169 Enhydra macrodonta 7 Enhydriodon 6, 20 Enhydritherium 6, 20, 138, 140 Enhydritherium lluecai 6 Enhydritherium te"anovae 6 Environmentallegislation 203 Environmental management 202-208 Evasterias 81

Fat innkeepers 42 Finfish fisheries 202-205 Fire suppression 206 Fish 13-15,21-22,44,151-152,156-157,

162-163,165-171,173-191,193-194, 213-214,217

- Biogeography 174,181,188 - Biomass 174

Subject Index

- Competition with sea otters 189,217 - Consumption by sea otters 13-15,21 - Consumption of plants 165,174 - Demography 217 - Edge effects in kelp forests 167 - Effects of - - climate 162,166,170,174,176-179,

181-183,186,188-191 - - drift kelp 163,167-168,187 - - kelp abundance 151-152,162-163,

165-171,174,176-191 sea otters 22,44,162,167,169-170, 183-185,189-191

- - sea urchin fishery 190 - - sea urchins 169-170,176-181,

183-191 - - storms 170,175-176,178,180-181,

183,185-186,191 - - topographic relief 186, 191 - Eggs 15 - Experimental manipulation 156, 162, 168 - Foraging 174,186,191,217 - Larvae 162, 188 - Productivity 151 - Recruitment 152,162,165-169,

178-179,181-182,187-188,190-191, 217

- Spawning 165 - "Storage effect" 217 - Variationinspace 173-174,176-187,

189,191 Fisheries management 207 Flatfish 2 Fleshy macro algae 116,130,210 Fleshy red algae 153,155,159-160,162 Fort Ross (California) 8 Fucoid algae 85,171 Fucus distichus 67 Fur hunters 138 Fur-hunting era 54,66,87 Fur seals 8

Gaper clams 32,35-36,38-40,42-44 Gap-phase phenomona 204 Gelidium robustum 153, 159, 162, 180 Genetically-engineered organisms 206 Geoducks 36,43 Giant kelp (see also Kelp forests) 92,96, 119,

151,193 Gibbon Anchorage (Alaska) 67, 69-70, 73 Gigartina exasperata 159 Girella nigricans 166,171,174,180-181,183 Glazenap Island (Alaska) 29 "Good science" 217 Gray whales 39,42

241

Great earthquake of 1964 (Alaska) 37 Green Island (Alaska) 18-19,26,36-37,40,

42,67,72-73,82,85-86, 89-90 Green sea urchins 212-213 Grimes Point (California) 105 Gulf of Alaska 11, 66, 185

Halfmoon 174,187 Halichoeres semicinctus 182 Ha/idrys 129 Haliotis rufescens 96, 158 Haliotis spp. 13 "Hard science" 217 Hartney Bay (Alaska) 69,79 Hawaiian Islands 8 "Herbivorous fishes" 187 Herbivory 116, 138 Hexagrammos 15 Hexagrammos decagrammus 14 Hinnites spp. 13 Holothurians 127, 131 Hudson River (New York) 204 Hydrodamalis gigas 138 Hydroids 155 Hypothetico-deductive approach 217 Hypsurus careyi 174, 178, 181,186

Ice scouring 138 Incidental take 11, 206 Infaunal bivalves 211,215 Izembeck Lagoon (Alaska) 42

Kamchatka Penninsula (USSR) 9 Kelpbass 171,174,180-181,187 Kelp crabs 13,58 Kelp forests 1-3, 12,20,22,44,51,58-59,

88-89,92-107,116-119,122-123, 125-127,129-131,134-141,151-155, 157-160,162-171,173-208,210-211, 213-214,216-217

- Algal densities 129,131, 134 - Alteration of water flow 3,197-201 - Alternate stable states 92-93, 102,

106-107,117 - Bathymetric distributions 94, 139 - Canopies 12,58-59,89,94-96, 101, 105,

119,130,135-136,151-154,157, 159-160,164-165,167-168,193, 213-214,217

- Carbon consumption 195-197,201 - Community types 94,103-107,214,216 - Competition 119,123,136,164,189,

192,194,200 - Currents 198-199,201 - Damping of waves 198-201,205

242

Kelp forests (cont.) - Demography 106-107, 192 - Detritus 116-117,137-138,194-196,

200-201 - Detritus-based food webs 169, 184, 186,

189-191,217 - Dispersal 106, 122, 164, 199, 201 - Dissolved organic carbon (DOC) 194-196 - Disturbance 99, 103, 192, 196,200, 208 - Diversity 92,130,194,205-206 - Economic value 92, 205 - Edge effects 167 - Effects of - - aboriginal man 102 - - fish grazing 162,165, 169, 180-181,

187-188 - - ice cover 138 - - oceanographic conditions 96,99,

102,105 sea cows 138 sea otters 1-3,20,22,44,51,92-93, 96-99,116-117,119,122-123,127, 137-141,151-153,162-165,167-171,183-185,189-192,196-197, 200-205,207-208,210,213-214,216 sea urchins 2,92-93,95-99,101-102,116-117,119,122-123,129-130,136-141,151-152,154-155, 162,164-171,173,176-181,183-192,194-197,200-201,210-211, 213-214,216

- EI Nilio/Southern oscillation 136 - Evolution 138 - Fish 151,162,165-167,169-171,174,

176-191,193-194,196,201,203,208, 214,217

- Generality of community structure 92-97, 99,102-103,106-107,117,123,138-140,216-218

- Growth rates 107, 136, 197 - Habitats 3, 193, 200, 205, 216 - Hierarchical model 93,100,105-106 - Historical reconstructions 96, 119, 122 - Internal waves 198 - Larval retention 199 - Larval settlement 194,199,201 - Light 103, 106, 159, 164 - Linear Model 117 - Local extinctions 169 - Long-term physical changes 13 7, 141 - Management 92-93,103,107,203-206,

208 - Microhabitat specialization 193 - Multifactorial interactive model 93, 100,

105

Subject Index

- Nutrients 103, 105, 136 - Particle flux 199,201 - Particulate organic carbon (POC) 194, 196 - Patchiness 95,97,99,102-103,125-

130,139,141,203-207,213-214,217-218

- Pelagic organisms 193 - Physiological ecology 192 - Pollution 96-97, 136, 166, 169 - Primary production 194 - Recruitment 102-103,106-107,134,

136 - Refugia from predation 193-194,201 - Reproduction 107 - Resource partitioning 194 - Seasonal variation 130, 135-136, 141,

169, 185, 187 - Sedimentation 105 - Species richness 130 - Stability 117, 135 - Storms 2,99, 103,136-138,141,151,

154,164,166,169-171,173,175-176, 178,180-181,183,185-186,189,191-192,196

- Substrata 103,106,164,193,214 - Succession 151, 164, 167 -168 - Survival rates 107 - Temperature 173,175-176,181-183,

185 - Topographic relief 105,152,158-160,

164-165,167-168,213-214 - Turbidity 2 - Understory 105,151-154,159-160,

164,166,168,171,178-179,186-187, 189, 193, 198-199

- Variation 92-94,96-98, 102-103, 105-107,117-119,122-123,125-127,130-131,134-141,169-171,173,175-191

- - Spatial 93-94,97-99,101, 103, 105, 107, 117-1i~ 123, 125-127,13~ 135,138-141,169-170,173-174, 176-187,189,191,203,205-206, 213-214,216 Temporal 93-94,96,99, 103, 105-107,117-119,122-123,125,130-131,134-141,169,171,173,176-178,180-184,186-187,189-191, 203,205-206,216

- Water depth 105 - Water movement 151-152,154,159-

160,164-166,168,196-201,216 - Waves 105-106,214 Kelp gametophytes 200 Kelp goby 186 Kelp harvesting 166,203

Subject Index

Kelp holdfasts 193 Kelpperch 163,174,186 Kelp sporophytes 200 Kelp stipes 193 Kelps 92-94,99,101-102,192-201,203 Keystone predator 45,84,200,213 Keystone species 92-93,95-99,106-107,

203,207 -208,210 Kodiak (Alaska) 8 Kodiak Archipelago (Alaska) 11, 38, 140 Kodiak Island (Alaska) 43 Komandorskiye Islands (USSR) 8-9,87,138 Kuril Islands (USSR) 9,11,14-15,21,60,87 Kyphosidae 169

Labridae 169,182 Laminariadentigera 153,155,159-160,

164,168 Laminariaspp. 127,130-131,134-135 Laminariales 38 Landscape ecology 204 Lethrops connectens 186 Linear population models 206 Little Green Island (Alaska) 73 Littleneck clams 29-30,39,41,43-44 Lobsters 87,90,136 Lone Black Reef (LBR; California) 152-160,

162,165-166 Los Angeles (California) 97 Lotka-Volterra theory 204-Lutra 4 Lutra felina 4 Lutrinae 4 Lytechinus anamesus (see also sea urchins)

95,189

Macoma 30, 41 Macrocystis (see also kelp forests) 105, 194,

198-199 Macrocystis pyrifera (see also kelp forests)

58,92,119,130-131,134-137,151,154, 162,164-165,171,174,184-185,189, 193,214

Marine Mammal Protection Act (US) 203 Marine otter (South America) 4 Massacre Bay (Alaska) 129 Medialuna californiensis 166,174,180-181,

183 Medny Island (USSR) 8-9 Membranipora membranacea 199 Mercenaria mercenaria 199 Metridium 38 "Midwater planktivores" 174, 179 Modiolus 15 Montague Strait (Alaska) 14

Monterey (California) 13,17,81,96,102, 164, 184, 188

Monterey Bay (California) 14, 18, 29-30, 40, 105

Monterey Harbor (California) 28, 34-35, 37,39,42,44

Monterey Penninsula (California) 60,81 Morro Bay (California) 13,35-36 Moss Landing Beach (California) 30, 44 Mukkaw Bay (Washington) 61, 85 Mussel fisheries 40,44,46 Mussel-dominated intertidal communities

48-90,211-213 - Competition 48, 87 -88, 90 - Composition 54, 67 - Crevice refugia 73,85

243

- Disturbance 49,54-55,57,60-62,65, 87-89

- Effects of - - crabs 87,90 - - lobsters 87, 90

sea otters 48-90, 212-2l3 seastars 49,81

- - waves 49,54-55,57-58,60-63, 81,84,86-87

- Gaps in mussel cover 49,54-55,57-58, 60-63,84-85,89,211

- Geographic distribution 87 - Growth rates of mussels 82 - Ice scour 81 - Immigration 49 - Interstitial habitat 48,67, 88-89 - Larval recruitment 49, 84 - Life histories 49,87,90 - Mobile carnivorous invertebrates 87-

88,90 - Mussel fisheries 40, 44, 46 - Population structure 48,51,54,64-66,

70,73,79,81-83,90 - Reproductive potential 85 - Size refuge from predation 49,64-65,

82-84, 89-90 - Space utilization 48, 67, 85 - Spatial distribution 48, 51, 67, 70, 73,

81-82,90 - Spawning 85 - Substrata 67,69,73,85 Mussels 13,39,48-51,54-55,57,60-67,

69-70,72-73,79,81-90,197,211-213, 215

Mustelidae 4 Mysids 193 Mytilidae 48 Mytilus californianus (see also mussel-dominated

communities) 51,54,60, 82, 84-85, 88-90,204,211

244

Mytilus edulis (see also mussel-dominated communities) 29,40,60,65,67,70,82, 85,88-90

Mytilus spp. 13, 15,48

Naples Reef (NR; California) 99,137,153-155,158,162,164,166

National Park Service (US) 140 Near Islands (Alaska) 9, 140 Nelson Bay (Alaska) 18-19 Nereocystis luetkeana 58, 105,119,136,

151-152,159-160,164-165,168,185 Nonlinear popUlation models 207 Nucella 81

Octocorals 199 Octopus 13,207 Odiak Channel (Alaska) 69 Odiak Point (Alaska) 69 Odobenid pinnipeds 7 Oil exploration 202-203 Oil shale 170 Oil spills 189,202 Olympic Penninsula (Washington) 54,61-62,

85,87-88 Opaleye 171, 174, 187 Optimum Sustainable Population 203 Orca Inlet (Alaska) 37,67,69-70,73,79,

81-82,86,89-90 Oxyjulis californica 162-163.174, 182, 188

Pachythyone rubra 127, 131, 135 Panope generosa 36,43 Paralabrax clathratus 166,171,174,180-181 Paramushir Island (USSR) 14 Particulate organic carbon (PaC) 194-196 Patch dynamics 204 Patella compressa 194 Patiria miniata 158 Pelagic zooplankton 192 Pendleton Artificial Reef (California) 165 Pesticides 207 Phanerodonfurcatus 174,178-179,181 Phyllospadix spp. 54 Phytoplankton 195, 197,200 Pile perch 176 Pisa Point (Alaska) 127, 129 Pisaster giganteus 158 Pisaster ochraceus 54,57,84, 158,204 Pisaster spp. 13, 81, 84, 158 Pismo Beach (California) 30-31 Pismo clams 13,30-32,39-41,43-44,46,

213 Planktivorous fish 166,168,201

Subject Index

"Plant-cropping herbivores" 174,180,187 Plant/herbivore interactions 97, 119, 138-

140,211 Point Buchon (California) 152 Point Cabrillo (California) 105 Point Conception (California) 166 Point Estero (California) 158 Point Lorna (California) 198 Point Piedras Blancas (California) 17,40,

51-52,54,57-58,60,65,72,84,87,89, 160,165

Point Santa Cruz (California) 105 Point Sur (California) 119 Pollicipes polymerus 54 Polychaetes 30, 39,43 Port Moller (Alaska) 9 Postelsia palmaeformis 54 Prince William Sound (Alaska) 9, 11, 1.4,

16-18,21,26,28,36-37,40,42,44, 65-67,70,72-73,82-83,85-90, 212-213,215

Protothaca staminea 29-30,39,41,43-44 Protozoa 195 Pteronura 4 Pterygophora californica 130, 154, 162, 171 Pugettia producta 58 Pugettia spp. 13 Purple sea urchins 129-130,158, 189,

212-213

r (intrinsic rate of natural increase) 206 Rainbow surfperch 178 Rat Islands (Alaska) 123 Rays 42-43 Razor clams 37,42 Red abalone 96 Red crabs 28, 46 Red sea urchins 129-131,134,158,189,

212-213 Reefs 152,158,160,164-165,167-168 - Aspect ratio 165, 168 - High topographic relief 152, 158, 160, 168 - Low topographic relief 152, 164 - Types 164-165, 167 Rhacochilus toxotes 174, 176 Rhodymenia pacifica 159 Risk assessment 206 Risk management 206 Rock crabs 13, 39 Rock greenling 14-15 Rock wrasse 182 Rockfish 162-163,165-166,171,184-185,

187 Rockweed 129

Subject Index

Rocky intertidal communities (see also Musseldominated communities) 2-3,45, 103, 204,210-213,215

- Competition for space 211 - Effects of sea otters 2-3,45,210-212,

215 - Gaps 211 - Wave shear 211 Rocky subtidal communities 210-214,

216-217 Rough piddocks 37 Rubberlip surfperch 176

Salps 136 San Juan Islands (Washington) 197 San Luis Obispo County (California) 2,51,

152 San Nicolas Island (California) 123-127,

129-130,135,141,190,202,208 San Simeon (California) 17 Sand dollars 2,22,29,31 Sandhill Bluff (California) 104-105 Santa Barbara (California) 137, 154, 158, 170,

189,191 Santa Barbara Channel (California) 154, 188 Santa Barbara County (California) 153 Santa Cruz (California) 17,29, 136 Santa Rosa Reef (SRR; California) 152-158,

160,162-166 Saxidomusgigantea 14,36-37,40,44 Saxidomus nuttalli 14,32,35-36,39-40,

42-44 Scale of sampling 118,122-123,135,157 Scallops 13, 199 Sea cows 137 Sea ice 138 Sea otters 1-4,6-90,92-93,95-97,99,

102-103,106-107,116-119, 122-123, 127,129,137-141,151-152,158,162, 164-165,167-171,173,183-185, 188-192,196-197,200-208,210-217

- Activity patterns 16-19,54 - As keystone species/predator 45, 92,

95-97,99,107,200,202-203,207,210, 213

- Competition with fish 189-191,217 - Diet 2,12-15,18-24,26-32,34-44,

46-51,54-55,57 -58,60-65,69-70, 72-73,79,81-90,158,184-185, 189-190,207

- Effects on - abalone 2, 13,20,92,96, 151-152,

158, 164,212-213 - - clams 2,30-33,35-37,39,44,46,

211-213

245

- - community diversity 200 - - crabs 2,18-19,23,26-29,38-39,

46,151,212 - - ecosystem health 203 - - finfish fisheries 202-203 - - fish 22,44, 151, 162, 169, 183-185,

189-191,202-203,207,217 - - kelp forests 1-3,20,22,44,51,

92-93,96-97,99, 103, 106-107, 116-119,122-123,127,137-141, 169-171,173,183-185,188-192, 196-197,200-205,207,210, 213-214,216

- - kelp harvest 202-203 - - mussel-dominated communities 2-3,

23,26-29,39-40,44,46,48-90, 212-213

- - oil production 203 - - productivity of communities 200

reef communities 151-152, 164 - - rocky intertidal communities 22,

43-45,47-90,211-213 - - rocky subtidal communities 22,43-44,

47,211-212 - - seaurchins 13-14,92,95-96,99,

102,106-107,116-117,122-123, 127,129,139-141,151-152,158, 162,164-165,167-171,183-185, 188-191,196-197,200-201, 203-205,207,210,212-214,216

- - shellfish fisheries 1-2,11-12,28-31, 36-37,39-46,123,202-203

- - soft bottom communities 2-3,22-47, 211-212

- Endangered status 202-203 - Evolution 4, 6-7, 20 - Excavation of sediments 16,32,34-36,

38-39,41-46,211-212,215-216 - Exploitation by aboriginal man 7, 88, 90,

102,138 - Exploitation by fur hunters 7-8,12,21 - Food stealing 16 - Foragingbehavior 4,12-13,15-17,

19-22,30-32,34-37,39-44,46-48,51, 54-55,57-58,60-64,69,72,79,81-82, 84-90,158,185,215

- Hauling-out behavior 70 - Incidental take in fiShing nets 11, 206 - Inclusive fitness 86,90 - Individual variation in diet 20-21,37,42,

60,87,89-90,215 - Life table analysis 206 - Maternal care patterns 86-87,90 - Metabolism 16,21 - Nutritional value of prey 82, 86, 90

246

Sea otters (cont.) - Parental care patterns 86-87,90

Population dynamics 4, 7, 8-12, 20-22, 28-29,43,45,54,66-67,79,81,85,106, 123,140,151-152,206,214 Population structure 4,7-9,11,13-14, 20-21,54,66-67,69,79,85,138,206 Radio telemetric studies 18 Rafting 158 Refugia for prey 43,46-47,73,85, 151, 158,164,167,212 Risk of oil spills 202 Social interactions 66, 86, 88, 90 Spraint composition 70,73,212 Starvation 86 Tool use 16,20-21,55,65,72,82, 89-90

- Translocation 123,139-141,202,208 - Value to society 205 Sea stars 13,22,49,84, 154,204 Sea urchins 2,13-15,39,92-93,95-99,

101-103,106-107,116-117,119, 122-123,125,127,129-131,134, 136-141,151-156,158-159,162, 164-171,173,176-181,183-191, 194-197,200,203-205,207-208,21~ 212-214,216

- Association with sewage outfalls 96-97 - Behavior 95, 102, 106 - Competition with abalone 136 - Consumption of detritus 116, 137 -138,

159, 164 - Demipyramids 212 - Disease 2,99,103,136,141,190 - Distribution on boulders 97 - Effects of - - anemones 158 - - fish predation 136, 187 -189

lobster predation 136 salinity 13 0

- - sea otters 13-15,92,95-96,99, 102, 106-107,116,119,122-123,129, 137,139-141,151-152,162,165, 167-171,183-185, 189-191, 196-197,200,203-205,207,210, 212-214,216

- - storms 137 - - topographic relief 165, 168,213-214 - - water motion 151, 165, 168 - Effects on - - drift algal supply 171,184,191 - - fish 169,176-181,183-184,

186-191 - - kelp forests 2,92-93,95-99,

101-103,106-107,116-117,119,

Subject Index

123,129-130,136-141,151-152, 154-155,162,164-171,173, 176-181,183-191,194-197,200, 203-205,207-208,210,213-214,216

- Experimental manipulation 106, 119, 122,152-156.162,164,167-168

- Fishery 39,97, 189-191 Foragingbehavior 116-117,136-138, 159,164

- Larval dispersal 122 - Population structure 95 -96, 107, 125,

129-130, 134, 136 - Recruitment 107,167,189,212-213 - Value to society 205 - Variation in space 95,97,99, 101-102,

127,129-130 - Variation in time 95-96,99,131,

134-136 Seabirds 13 Seagrass beds 194, 198-199 Sebastes mystinus 162-163, 174, 182, 188 Sebastes serranoides 162 Sebastes spp. 15,163, 171, 185 Sebastichus 15 Sedimentation 199 Semicossyphus pulcher 137, 174, 182, 188 Senorita 162-163, 182, 188 Sheep Bay (Alaska) 40, 72 Sheephead 137, 182, 187 -188 Shemya Island (Alaska) 196 Shrimp 22, 29, 193 Siliqua patula 37,42 Simpson Bay (Alaska) 67,69-70,73,86,

88-90 Simushir Island (USSR) 15 Sitka (Alaska) 8 Skates 39 Soft bottom communities 2-3,16,22-47,

210-213, 215 -216 - Clam fisheries 30-31,36-37,39,44,46 - Commensal species in clam burrows 38,46 - Community structure 28-30, 32-34,

36-39,41,44,46 - Competition for space 37 -38 - Crab fisheries 28-29, 39,46 - Disturbance by sea otters 16,32, 34-36,

38-39,41-46,211-212,215-216 - Effects of clams 37 -38, 46 - Effects of crabs 29-30,211,215 - Effects of sea otters 2 - 3, 22 -47,

210-213,215-216 - Mussel fisheries 40,44,46 - Recruitment of clams 30, 32, 37,41,46 Spider crabs 194 Spike Island (Alaska) 69, 79

247

Sponges 155 Stable carbon isotopes 197 Starfish 208 Steller's sea cow 138 Stillwater Cove (California) 105, 119 Storms 136-137,141,166,169-171,173,

175-178, 180-181, 183, 185-186, 189, 191,208

Striped bass 204 Striped surfperch 162, 176 "Strong inference" 217 Strongylocentrotus franciscan us 95, 131, 135,

158,171 (see also sea urchins)

Strongylocentrotus polyacanthus 14, 60 (see also sea urchins)

S trongyloce ntro tus purpuratus 158, 17 1 (see also sea urchins)

Strongylocentrotus spp. 151 (see also sea urchins)

Sublittoral fringe 139 Surfperch 162-163,165,171,174,176-177,

181, 186-187 Surge Bay (Alaska) 136 Suspension feeders 195-199,201 "Switch-feeders" 166, 168 "Switch-feeding predators" 174, 180, 187

Tatoosh Island (Washington) 61,85,88 Tegula spp. 13,58 Tellina 15 Tellinid clams 15,30 Telmessus cheiragonus 14 Terrestrial plant communities 198,204,

214-215

Thalassiophyllum clathrus 127 Time-depth recorder 215

Subject Index

Tivela stultorum (see also Pismo clams) 13 Tool use 16,55,65,72,82,89-90 Translocation of sea otters 123,139-141,

202,208 Tresus nuttallii 14,32,35-36,38-40,42-44 Tresus spp. 38 Trigger species 207 Trochidae 194 Tubicolous amphipods 155, 160 Tunicates 155 Turban snails 13,58,207

Unimak Pass (Alaska) 9 Upwelling 7,169,175-176 Urchin barrens 117-119,123,129-130,

136-137,139,141,213,216 Urechis caupo 13,42 Urup Island (USSR) 15 US Department of Agriculture 119 US Fish and Wildlife Service 123, 140, 202

Vancouver Island (British Columbia) 29,140

Walrus 38, 42, 43 Washington clams 32,35-36,39-40,42-44 "Well-behaved ecosystems" 217 Wetlands 205 White sea urchins 95, 189 Worms 155

Zirphaea pi/sbryi 37 Zooplankton 196,201 Zostera marina 199

Documents

The Community Ecology of Sea Otters