Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of harvest and climate change on polar marine ecosystems : case studies from the Antarctic Peninsula.. 2012

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2012_spring_hoover_carie.pdf [ 15.35MB ]
Metadata
JSON: 1.0105170.json
JSON-LD: 1.0105170+ld.json
RDF/XML (Pretty): 1.0105170.xml
RDF/JSON: 1.0105170+rdf.json
Turtle: 1.0105170+rdf-turtle.txt
N-Triples: 1.0105170+rdf-ntriples.txt
Citation
1.0105170.ris

Full Text

Eects of Harvest and Climate Change on Polar Marine Ecosystems Case Studies from the Antarctic Peninsula and Hudson Bay by Carie Hoover B.Sc., The University of California Santa Barbara, 2002 M.Res., The University of St Andrews, 2005 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Resource Management and Environmental Studies) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2012 c
 Carie Hoover 2012 Abstract This thesis applies food web modelling to increase our understanding of how the interaction of climate change and exploitation have historically altered, and continue to alter, marine polar ecosystems. Understanding stressors responsible for ecosystem level changes is important not only to the people and industries reliant on the resources, but for managers to make future decisions on resource uses. The rst two chapters develop models of Hud- son Bay (Arctic) and Antarctic Peninsula (Antarctic) marine ecosystems, focused on re-creating changes in the past 30 years. Both ecosystems have undergone changes due to environmental factors, which are incorporated into the models. While the Hudson Bay model exhibits a shift from benthic to pelagic species, the Antarctic Peninsula model is identied to have more uniform declines across all species, as the main trophic link in the ecosystem, Antarctic krill declines. Model simulations are continued in the next two chapters, whereby future environmental changes are tested in conjunction with multiple exploitation levels. For Hudson Bay, continued harvest of ma- rine mammals at current conditions results in large-scale declines for some species (narwhal, eastern Hudson Bay beluga, polar bears, and walrus), in- dicating current harvest levels are too high to sustain long term. Further shifts from benthic to pelagic species in the lower trophic levels favor sh species such as capelin and sandlance. Future simulations of the Antarctic Peninsula identify large reductions in ecosystem biomass of all species due changes in environmental conditions and an overall reduction in krill, with minimal ecosystem impacts from harvest. In the last chapter, an economic model is constructed to assess the use value of hunting narwhal and bel- uga in the Hudson Bay region. The economic impact to northern residents is considered as future model simulations of Hudson Bay reveal that these species may be susceptible to population declines, and issues of food security are becoming increasingly important. Economic analysis reveals the moti- vation to hunt in Hudson Bay may not be economically-driven, there are substantial benets derived by northern communities through narwhal and beluga hunts. Results for each ecosystem are discussed as they pertain to future research and management of each ecosystem. ii Preface A version of chapter 2, co-authored with Tony Pitcher and Villy Christensen, has been re-submitted with revisions to Ecological Modelling. I constructed the model and wrote the manuscript. Villy Christensen was key in the de- velopment of the model structure, tting of the model, and other technical model aspects. Tony Pitcher provided guidance on model construction and discussions on the direction of the manuscript. A version of chapter 4, co-authored with Tony Pitcher and Villy Christensen, has been re-submitted with revisions to Ecological Modelling. I created the model simulations and wrote the manuscript. In addition to their assistance with chapter 2 which utilizes the same model, both Villy Christensen and Tony Pitcher provided guidance with the future simulations. Villy Chris- tensen also provided technical model assistance. A version of chapter 6, co-authored with Megan Bailey, Je Higdon, Steve Ferguson, and Rashid Sumaila, has been re-submitted with revisions to the journal Arctic. I conceptualized and constructed the model in addition to writing the manuscript. Megan Bailey assisted in model construction. Rashid Sumaila and Megan Bailey provided the framework for the model and assistance on economic analyses. Je Higdon and Steve Ferguson pro- vided expertise on model parameters, in addition to Je Higdon collecting input parameter values during eldwork in the north. All authors provided feedback on the submitted manuscript. Chapters 3 and 5 are co-authored with Tony Pitcher and Evgeny Pakhomov, and will be submitted to a peer-review journal. I constructed the model for chapter 3, created the model simulations for chapter 5, and wrote both manuscripts. Tony Pitcher provided the idea for the model and guidance throughout the model construction, and provided guidance with model sim- ulations. Evgeny Pakhomov was key in providing expertise to the ecology of the model which was important to the tting process, and contributed to the ecological relevance of the model. iii Table of Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . xii Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Ecosystem-Based Management . . . . . . . . . . . . . . . . . 1 1.2 Study Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Ecopath with Ecosim . . . . . . . . . . . . . . . . . . . . . . 10 1.4 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 Impacts of Hunting, Fishing, and Climate Change to the Hudson Bay Marine Ecosystem 1970-2009 . . . . . . . . . . 15 2.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3 Eects of Harvest and Climate Change on the Antarctic Peninsula Marine Ecosystem (FAO area 48.1) . . . . . . . . 47 3.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 iv Table of Contents 3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4 Future Impacts of Hunting, Fishing, and Climate Change on the Hudson Bay Marine Ecosystem . . . . . . . . . . . . 80 4.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 4.6 Hudson Bay Biomass and Morality Figures . . . . . . . . . . 116 5 Future Impacts of Fishing and Climate Change on the Antarc- tic Peninsula Marine Ecosystem . . . . . . . . . . . . . . . . 121 5.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 5.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 5.6 Antarctic Peninsula Biomass and Mortality Figures . . . . . 165 6 Estimating the Economic Value of Narwhal and Beluga Hunts in Hudson Bay, Nunavut . . . . . . . . . . . . . . . . . . . . . 172 6.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 6.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 6.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 6.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 7.1 Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 7.2 Hudson Bay . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 7.3 Antarctic Peninsula . . . . . . . . . . . . . . . . . . . . . . . 213 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 Appendices A Hudson Bay Ecosystem Model Parameters and Details . . 287 A.1 Model Parameters by Functional Group . . . . . . . . . . . . 287 v Table of Contents A.2 Fisheries Input . . . . . . . . . . . . . . . . . . . . . . . . . . 325 A.3 Model Fitting Parameters and Data Sets . . . . . . . . . . . 333 A.4 Model Parameterization and Output . . . . . . . . . . . . . . 336 B Marine Mammal Mortality Equations . . . . . . . . . . . . . 346 C Hudson Bay Bird Species . . . . . . . . . . . . . . . . . . . . . 348 D Hudson Bay Fish Species . . . . . . . . . . . . . . . . . . . . . 353 E Hudson Bay Model Vulnerabilities . . . . . . . . . . . . . . . 355 F Hudson Bay Mixed Trophic Impacts . . . . . . . . . . . . . . 358 G Hudson Bay Monte Carlo CV Values . . . . . . . . . . . . . 367 H Hudson Bay Monte Carlo Results . . . . . . . . . . . . . . . 369 I Hudson Bay Ecosim Biomass Trends by Species . . . . . . 373 J Antarctic Peninsula Ecosystem Model Parameters and De- tails . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376 J.1 Model Parameters by Functional Group . . . . . . . . . . . . 376 J.2 Ecosim Input Parameters . . . . . . . . . . . . . . . . . . . . 426 J.3 Model Parameterization and Output . . . . . . . . . . . . . . 437 K Antarctic Peninsula Model Vulnerabilities . . . . . . . . . . 453 L Antarctic Peninsula Model Mixed Trophic Impact Values 458 M Antarctic Peninsula Monte Carlo CV Values . . . . . . . . 466 N Antarctic Peninsula Monte Carlo Results . . . . . . . . . . 468 O Antarctic Peninsula Monte Carlo Graphs . . . . . . . . . . 471 P Antarctic Peninsula Model Biomass Trends By Species . 473 vi List of Tables 2.1 Harvest trends used in the Hudson Bay Ecosim model . . . . 26 2.2 Balanced Ecopath model parameters . . . . . . . . . . . . . . 29 2.3 Trophic level of the ecosystem (TLE) and catches (TLC) . . . 37 3.1 Time series data used for Antarctic Peninsula model tting . 55 3.2 Balanced Ecopath model parameters for the Antarctic Penin- sula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.3 Antarctic Peninsula trophic level of the ecosystem TLE and catches TLC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.1 Hudson Bay future climate and hunting scenarios . . . . . . . 86 4.2 Summary of harvest values and hunting/shing mortalities used for the initial Hudson Bay Ecopath model and future hunting scenarios . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.3 Trophic level of ecosystem (TLE) and catches(TLC) for each future Hudson Bay simulation . . . . . . . . . . . . . . . . . . 94 5.1 Catches and shing mortalities for each Antarctic Peninsula future scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . 129 5.2 Antarctic Peninsula future harvest and climate scenario names.130 5.3 Trophic level of ecosystem (TLE) and catches(TLC) for the Antarctic Peninsula future simulations . . . . . . . . . . . . . 133 6.1 Parameter inputs for Hudson Bay economic model equations 188 6.2 Statistics for hunting communities in Hudson Bay . . . . . . 195 6.3 Economic value including cost sharing and opportunity cost . 200 6.4 Contribution of revenue to each community . . . . . . . . . . 201 A.1 Hudson Bay Ecopath marine mammal input parameters . . . 289 A.2 Hudson Bay shing mortality based on per capita consump- tion rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 A.3 Calculated input parameters for Hudson Bay sh groups . . . 309 vii List of Tables A.4 Comparison of parameters for benthic functional groups from high latitude Ecopath models. . . . . . . . . . . . . . . . . . . 320 A.5 Arctic killer whale harvests . . . . . . . . . . . . . . . . . . . 326 A.6 Time series data for Hudson Bay Ecosim tting . . . . . . . . 333 A.7 Balanced Ecopath model parameters for Hudson Bay . . . . . 338 A.8 Coecient of variation values used for Monte Carlo estimates 339 B.1 Marine mammal survivorship curve parameters . . . . . . . . 347 C.1 Hudson Bay bird species . . . . . . . . . . . . . . . . . . . . . 348 D.1 Hudson Bay sh species by functional group . . . . . . . . . . 353 E.1 Hudson Bay model vulnerabilities . . . . . . . . . . . . . . . . 355 F.1 Hudson Bay mixed trophic impact results . . . . . . . . . . . 359 G.1 Hudson Bay Monte Carlo coecient of variation values . . . . 368 J.1 Southern Ocean cetacean estimates . . . . . . . . . . . . . . . 376 J.2 Antarctic Peninsula marine mammals parameter values . . . 378 J.3 Antarctic Peninsula Ecopath penguin parameters . . . . . . . 390 J.4 Antarctic Peninsula Ecopath sh parameters . . . . . . . . . 399 J.5 Benthic habitat by depth range for the Antarctic Peninsula . 404 J.6 Antarctic Peninsula Ecopath parameters for invertebrate groups411 J.7 Natural mortality rates of Antarctic krill . . . . . . . . . . . . 420 J.8 Multistanza parameters for krill functional groups. . . . . . . 421 J.9 Antarctic Peninsula time-series data . . . . . . . . . . . . . . 430 J.10 Antarctic Peninsula balanced Ecopath model parameters . . . 439 J.11 Antarctic Peninsula Monte Carlo estimates . . . . . . . . . . 444 K.1 Antarctic Peninsula vulnerabilities used for the tted model . 454 L.1 Antarctic Peninsula mixed trophic impact results . . . . . . . 459 M.1 Antarctic Peninsula Monte Carlo CV values . . . . . . . . . . 467 N.1 Antarctic Peninsula Monte Carlo results . . . . . . . . . . . . 469 viii List of Figures 2.1 Hudson Bay map . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2 Mean sea surface temperature and ice cover for Hudson Bay . 24 2.3 Changes in sh abundance as measured by the diets of thick- billed murres . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.4 Food web linkages in the Hudson Bay ecosystem model . . . 31 2.5 Biomass trends for functional groups tted to time-series data 34 2.6 Contribution of sh groups to total sh biomass . . . . . . . 35 2.7 Hudson Bay Monte Carlo simulation results . . . . . . . . . . 36 2.8 Changes in biomass by Hudson Bay model scenario . . . . . . 39 3.1 Map of Antarctic Peninsula (FAO area 48.1) . . . . . . . . . 51 3.2 Krill and sh catches presented by year . . . . . . . . . . . . 57 3.3 Environmental drivers used in the model tting process . . . 59 3.4 Antarctic Peninsula tted model . . . . . . . . . . . . . . . . 66 3.5 Antarctic Peninsula Monte Carlo biomass estimates . . . . . 69 3.6 Antarctic Peninsula Ecosim simulation results . . . . . . . . . 71 4.1 Hudson Bay environmental data used in future model simu- lations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4.2 Hudson Bay future changes in biomass by model scenarios . . 92 4.3 Future scenario changes in biomass for sandlance . . . . . . . 99 4.4 Future scenario changes in biomass for capelin . . . . . . . . 100 4.5 Future scenario changes in biomass for gadiformes . . . . . . 101 4.6 Future scenario changes in biomass for northern walrus . . . . 105 4.7 Future scenario changes in biomass for ringed seals . . . . . . 106 4.8 Future scenario changes in biomass for narwhal . . . . . . . . 107 4.9 Marine mammals and sh biomass results by future scenario 117 4.10 Invertebrates, zooplankton and producer biomass results by future scenario . . . . . . . . . . . . . . . . . . . . . . . . . . 118 4.11 Marine mammals and sh mortality results by future scenario 119 4.12 Invertebrates, zooplankton and producers mortality results by future scenario . . . . . . . . . . . . . . . . . . . . . . . . 120 ix List of Figures 5.1 Antarctic Peninsula sea ice and SST trends . . . . . . . . . . 127 5.2 Antarctic Peninsula future biomass changes by species and scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 5.3 Future scenario changes in biomass for copepods . . . . . . . 139 5.4 Future scenario changes in biomass for juvenile krill . . . . . 140 5.5 Future scenario changes in biomass for adult krill . . . . . . . 141 5.6 Future scenario changes in biomass for salps . . . . . . . . . . 142 5.7 Future scenario changes in biomass for myctophids . . . . . . 146 5.8 Future scenario changes in biomass for large deep demersals . 147 5.9 Future scenario changes in biomass for toothsh . . . . . . . 148 5.10 Future scenario changes in biomass for Adelie penguins . . . 152 5.11 Future scenario changes in biomass for minke whales . . . . . 155 5.12 Future scenario changes in biomass for Antarctic fur seals . . 156 5.13 Antarctic Peninsula marine mammal ending biomass by sce- nario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 5.14 Antarctic Peninsula sh and invertebrate ending biomass by scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 5.15 Antarctic Peninsula invertebrate and plankton ending biomass by scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5.16 Antarctic Peninsula marine mammal ending mortality by sce- nario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 5.17 Antarctic Peninsula sh and invertebrate ending mortality by scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 5.18 Antarctic Peninsula invertebrate and plankton ending mor- tality by scenario . . . . . . . . . . . . . . . . . . . . . . . . . 171 6.1 Map of communities in Nunavut portion of Hudson Bay hunt- ing narwhal or beluga . . . . . . . . . . . . . . . . . . . . . . 176 6.2 Distributions and 95% CI for total revenue, total cost, total use value, and total use value including opportunity cost . . . 197 6.3 Average per capita use value for beluga and narwhal hunts with cost charing and opportunity cost . . . . . . . . . . . . . 198 A.1 Reported catches of narwhal from 1977-2007 for Hudson Bay communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 A.2 Catches of beluga whales form 1970-2007 as aggregated by stock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330 A.3 Regression of community population size in Nunavut . . . . . 332 A.4 Polar bear mediation function . . . . . . . . . . . . . . . . . . 334 x List of Figures A.5 Antarctic Peninsula ending biomass results for producers and detritus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342 A.6 Antarctic Peninsula ending biomass results for zooplankton and benthic groups . . . . . . . . . . . . . . . . . . . . . . . . 343 A.7 Antarctic Peninsula ending biomass results for sh and seabirds344 A.8 Antarctic Peninsula ending biomass results for marine mammals345 H.1 Hudson Bay marine mammal Monte Carlo biomass and P/B results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370 H.2 Hudson Bay sh Monte Carlo biomass and P/B results . . . . 371 H.3 Hudson Bay plankton Monte Carlo biomass and P/B results . 372 I.1 Hudson Bay tted model biomass trends for marine mammal and sh groups . . . . . . . . . . . . . . . . . . . . . . . . . . 374 I.2 Hudson Bay tted model biomass trends for benthic and plankton groups . . . . . . . . . . . . . . . . . . . . . . . . . 375 J.1 Numbers of penguin breeding pairs at Anvers Island, Antarc- tic Peninsula . . . . . . . . . . . . . . . . . . . . . . . . . . . 388 J.2 Krill shing eort used in model tting . . . . . . . . . . . . 427 J.3 Krill catches used in model tting . . . . . . . . . . . . . . . 428 J.4 Antarctic Peninsula krill abundance and biomass trends . . . 429 J.5 Antarctic Peninsula salp abundance trends . . . . . . . . . . 429 J.6 Mediation function used for larval and juvenile krill. . . . . . 433 J.7 Mediation function used for salps. . . . . . . . . . . . . . . . 434 J.8 Antarctic Peninsula Ecosim changes in biomass for producers and detrital groups. . . . . . . . . . . . . . . . . . . . . . . . 447 J.9 Antarctic Peninsula Ecosim changes in biomass for zooplank- ton groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449 J.10 Antarctic Peninsula Ecosim changes in biomass for benthic groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449 J.11 Antarctic Peninsula Ecosim changes in biomass for sh groups.450 J.12 Antarctic Peninsula Ecosim changes in biomass for penguin and 
ying bird groups. . . . . . . . . . . . . . . . . . . . . . . 451 J.13 Antarctic Peninsula Ecosim changes in biomass for marine mammal groups. . . . . . . . . . . . . . . . . . . . . . . . . . 452 O.1 Antarctic Peninsula Monte Carlo biomass results . . . . . . . 472 P.1 Antarctic Peninsula biomass trends for SST and SOI tted models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474 xi Acknowledgements There are may people who have contributed time, support, wisdom, and encouragement. Without them this thesis would not exist. First to my advisor Tony Pitcher for allowing me the opportunity to pursue this degree, and to my committee members Evgeny Pakhomov and Andrew Trites for their advice throughout this process. To Villy Christensen for his constant advice and technical assistance throughout my degree. For all of my co- workers who provided technical assistance: Sherman Lai, and Dalai Felinto for assisting with Ecopath errors. Joreen Steenbeek for going above and beyond his job xing Ecopath bugs, and all of his help with programming. Thanks to Steve Martell and Rob Ahrens for expanding my knowledge of R and always helping to x code. To Rashid Sumaila, thank you for your endless guidance on all things economic, and for always listening to my ideas. There have been many students past and present who provided academic and personal assistance: Brooke Campbell for providing GIS maps, Megan Bailey for providing food web images for publications, and Laura Tremblay- Boyer for technical help. Thanks to Chiara Piroddi, Divya Varkey, Leigh Gurney, Cam Ainsworth, Colette Wabnitz, and Robyn Forrest for providing peer support on modeling questions. For the Hudson Bay portion of the thesis, numerous researchers pro- vided their time and resources in order to make the ecosystem model and the economics chapters possible. To them I would like to say thanks: Je Higdon, Elly Chmelnitsky, Patt Hall, Jack Orr, Blair Dunn, Tara Bortoluzzi, Lisa Loseto, Sebastian Luque, and Bruce Stewart. Travel to Hudson Bay was provided by the Cecil and Kathleen Morrow Scholarship, and provided unparalleled insight to the dynamic of the ecosystem through rst-hand ex- perience, thank you. A special thanks to Steve Ferguson for the opportunity xii Acknowledgements to work with him as part of the IPY project on Hudson Bay, for the oppor- tunity to participate in eldwork, and endless advice on Hudson Bay. In addition to academic support, many colleagues and friends have pro- vided invaluable support through my time at UBC. For that I would like to thank Megan Bailey, Rachael Louton, Shannon Obradovich, Rhona Goven- der, Brooke Campbell, Andres Cisneros, Roseti Imo, Prammod Ganapathi- raju, Sarika Cullis-Suzuki, Jennifer Jacquet, Meaghan Darcy, Erin McCul- loch, Liz Martell, Chiara Piroddi, and Maria Espinosa. xiii Dedication To William, Diane, and Travis. xiv Chapter 1 Introduction The main aim of this thesis is to address the impacts of harvest and envi- ronmental changes on two polar ecosystems; one Arctic and one Antarctic. Both regions aim to manage with an ecosystem-based approach (CCAMLR, 1980; Anonymous, 2006), implying that exploitation of target species should not cause destruction of other species. For both ecosystems, the following questions formed the chapters presented in the thesis. (1) What did the past ecosystem look like? (2) What factors caused past changes in the ecosystem? (3) How will these factors continue to impact the ecosystem in the future? And lastly, (4) For the Arctic ecosystem where people rely on harvest for subsistence, how might these changes aect these communities? This rst chapter aims to provide a background into both of the case study areas. First, I provide information on the geographic regions, and the environmental factors that shape them. Second, I address the management of each area, and the goals of the managers within the context of the ecosys- tem. Last, this chapter provides a summary of the questions and research completed in each chapter of the thesis. 1.1 Ecosystem-Based Management The focus of this thesis is on two polar ecosystems with the intent to identify prominent stressors that have, or will in the future, alter ecosystem struc- ture. The overall goal is to provide this information so that future research and management decisions can take into account the ecosystem dynamics of environmental change and exploitation. Fisheries regulation within the context of an entire ecosystem has become more prominent in recent years with the development of management strategies such as 'Ecosystem-Based 1 1.1. Ecosystem-Based Management Management' (EBM), 'Ecosystem-Based Fisheries Management' (EBFM), and 'Ecosystem Approach to Fisheries' (EAF). Despite the dierent termi- nologies, these management strategies share many of the same goals such as maintaining natural structure and function of the ecosystem, preventing declines of target and non-target species, and identifying environmentally sustainable development of resources (Ward et al., 2002; Hall and Main- prize, 2004; Pitkitch et al., 2004; Scandol et al., 2005). It has been noted that management priorities should be focused on the ecosystem as a whole, rather than just target species (Pitkitch et al., 2004). Yet as these phrases have only in recent years begun appearing in the lit- erature and management plans, the foundations of EBM are deep-rooted within many international agreements. The United Nations Convention on the Law of the Sea, an international agreement between 162 countries, notes harvest of species should be accomplished while maintaining or restoring populations of harvested and dependent species for both coastal and high seas sheries (United Nations, 1982, articles 61 and 191). The same year the Commission for the Conservation of Antarctic Marine Living Resources came into eect regarding the management of the Antarctic. Article II of the convention specically addresses harvest in that it should prevent ir- reversible changes in the ecosystem and maintain ecological relationships between harvested, dependent, and related populations (CCAMLR, 1980; Constable et al., 2000). In 1992, the Rio Declaration on the Environment and Development called for the use of the precautionary principle in order to protect the environment (United Nations, 1992, principle 15). Individual countries such as the US, Canada and Australia have also integrated aspects of EBM into their management plans for specic areas or sheries (Quinn and Theberge, 2004; Scandol et al., 2005; Pace, 2009). Evaluation of aquatic ecosystems through models such as EwE (Ecopath with Ecosim) and Atlantis allow for sheries assessments at the ecosystem level (Scandol et al., 2005), and the rst four chapters of this thesis address this. In addition to ecological aspects of EBM, social and economic goals are also considered for the success of EBM (Hilborn et al., 2004). Scandol et al. (2005) noted the importance of management to recognize that in addition to 2 1.2. Study Areas harvesting, there are additional uses and values of the ecosystem that must be considered. As human values drive the management process, and the human uses and values are an important component to EBM, these must be considered for management to be successful (Ward et al., 2002). It is for these reasons that I have included a chapter estimating the economic use value of hunts in Hudson Bay. 1.2 Study Areas Two regions were chosen as case studies for the thesis; one from the Arctic and one from the Antarctic. From the Arctic, Hudson Bay was chosen. Al- though Hudson Bay is considered sub-Arctic in location, its weather patterns are re
ective of a higher latitude region, with many high Arctic species re- siding there such as polar bears (Stirling and Parkinson, 2006). In addition, collaboration with researchers at the Department of Fisheries and Oceans Canada (DFO) Cental and Arctic Division in Winnipeg, Manitoba as part of their International Polar Year (IPY) Global Warming and Arctic Marine Mammal (GWAMM) project dictated the study area. For the Antarctic case study, the Antarctic Peninsula was selected. It is considered one of the fastest warming areas in the world (Anisimov et al., 2001; Hansen et al., 2006a), while other areas of the Antarctic have shown to be in a cooling trend (Turner et al., 2005). The central species in this ecosystem, krill (Euphausia superba), have shown declines linked to environmental changes in addition to being directly harvested (Atkinson et al., 2004; CCAMLR, 2008b). Hudson Bay Physical Environment Hudson Bay is a large, shallow, low nutrient marine area which freezes and thaws annually (Markham, 1986; Stewart and Lockhart, 2005; Stewart and Barber, 2010). Ice and temperature within this region are more re
ective of a high Arctic ecosystem, allowing species normally found at higher lati- tudes to be found within Hudson Bay (Maxwell, 1986; Stewart and Barber, 3 1.2. Study Areas 2010). The Hudson Bay watershed is the second largest in Canada, captur- ing roughly 30% of all Canadian runo (Natural Resources Canada, 1999). The timing of freshwater and nutrients inputs can have large impacts on the type and amount of annual primary production (Stewart and Barber, 2010). Seawater enters and exits via Hudson Strait, and circulates in a counter- clockwise direction (Stewart and Lockhart, 2005). The cooler deeper waters characteristic of Hudson Strait potentially act as a thermal barrier, pre- venting species from entering Hudson Bay. This divide is believed to be a choke-point for migratory species, such as killer whales, and it is believed that this divide will open as climate warms (Higdon and Ferguson, 2009). Being classied as a polar ecosystem, ice is an important component in the life cycle of many organisms, ranging from algae frozen within the sea ice (Horner et al., 1992), to top predators such as polar bears who use ice as a foraging platform (Stirling et al., 2004). Increases in temperature combined with lengthening of the ice-free season have increased concern for species residing in Hudson Bay (Parkinson et al., 1999; Gagnon and Gough, 2005; Hansen et al., 2006b). Resource Uses Hudson Bay has been used for roughly 4000 years by nomadic hunters who depended on marine mammals, sh and land animals such as caribou for subsistence (Stewart and Lockhart, 2005). European activity in the region started in the 17th century, as explorers searched for the northwest passage. Henry Hudson (Hudson Bay's namesake) was the rst recorded explorer into Hudson and James Bays (Francis and Morantz, 1983). Continued ex- peditions into the region and the abundance of available furs, primarily beaver, led to the establishment of the Hudson's Bay Company (Stewart and Lockhart, 2005). Harvest of fur-bearing animals by natives increased to meet the supply demands of the Europeans, although this con
icted with ancient spiritual beliefs (Sokolow, 2003). While other animals were also harvested for fur, beavers were specically targeted, with their pelts used as currency between Europeans and Aboriginals before populations crashed 4 1.2. Study Areas (Homren, 2004). As the fur trade declined, interest in whaling became more prominent (Francis and Morantz, 1983). Prior to the commercial whaling of bowhead whales, Hudson's Bay Company had small-scale unsuccessful at- tempts at commercial whaling operations for belugas (Reeves and Mitchell, 1987). The Northwest Company, a fur trading company formed in Montreal which later merged with the Hudson's Bay Company, and the Hudson's Bay Company operated posts in Hudson and James Bays related to whal- ing and trade (see Stewart and Lockhart, 2005, table 11-3 for a full list of settlements). American and Canadian vessels commercially harvested bow- head whales from 1860 to 1915, causing a large population decline before commercial whaling commenced in the region (Ross, 1974). Presently, subsistence harvest is allowed for many species with varying levels of regulation. Narwhal, beluga and polar bears have quotas and are harvested annually, while bowhead whales are considered endangered and are rarely harvested (DFO, 1998; Cosens and Innes, 2000; Hammill, 2001; Lunn et al., 2002). Seals, walrus, birds, sh and invertebrates are harvested by Aboriginals (Inuit and Cree) without a license and can be taken through- out the year (Berkes, 1977; Wein et al., 1996; Stewart and Lockhart, 2005). A license is required for sport hunters to harvest birds within the area (Stew- art and Lockhart, 2005). The only commercial shing operation is for Arctic char along the river mouths, but this shery yields small catches (Carder and Peet, 1983; DFO, 1997). Presently, most communities surrounding Hudson Bay are inhabited by rst nations, making up 85% of the total population in Nunavut, most of which are Inuit (Statistics Canada, 2006). Management The territory encompassing Hudson Bay is divided between the provinces of Manitoba, Ontario and Quebec, and the Nunavut territory. Within Que- bec, indigenous people (Inuit and Cree) live in Nunavik, the name for the northern third of the province. The rst major agreement between Quebec and the Inuit was the James Bay and Northern Quebec Agreement in 1978 to give environmental and social protection (Anonymous, 1975). 5 1.2. Study Areas The territory of Nunavut was established in 1999, separating it from the pre-existing Northwest Territories. Management for the Ontario and Manitoba portions of Hudson Bay is regulated by DFO, while the Nunavut portion is governed by the Nunavut Wildlife Management Board (NWMB). The Nunavut Land Claims Agreement, signed into eect in 1993, gives management authority of wildlife within Nunavut to the NWMB (Nunavut Land Claims Agreement, 1993). The NWMB consists of appointed mem- bers which are responsible for establishing, modifying, or removing levels of total allowable harvest. In 2006, the Nunavik Inuit Land Claims Agree- ment established the Nunavik Marine Region Wildlife Board (NMRWB) and granted Nunavik Inuit the right to harvest wildlife species to full their economic, social, and cultural needs (Anonymous, 2006). From 1996-2001 the NWMB conducted the Nunavut Wildlife Harvest Study to collect data for species within Hudson Bay for which it was re- sponsible (Nunavut Wildlife Managament Board, 2000). This harvest study was to help provide baseline information for all of Nunavut, for which to base total allowable harvests, primarily for marine mammal species (Priest and Usher, 2004). The total allowable harvest must be approved by the NWMB, and they retain the right to alter harvest levels in the future. The Nunavik parallel to this board, NMRWB, is responsible for the harvesting of species within the Nunavik and James Bay portions of Hudson Bay regarding Inuit harvest. The Canadian government, specically DFO, can disallow deci- sions set by the NWMB for reasons of conservation, public safety, or public health (Nunavut Land Claims Agreement, 1993). The federal government also holds the power to interfere regarding harvest in Nunavik. Antarctic Peninsula Physical Environment The Southern Ocean surrounds Antarctica, and while there are no physical barriers separating this ocean from the surrounding waters, the Antarctic Polar Front (or Antarctic convergence) is where the colder Antarctic wa- ters sink below the warmer sub-Antarctic waters forming a thermal barrier 6 1.2. Study Areas between 50S to 60S (Knox, 1994). Two main current systems occur in the Antarctic. The rst is the Antarctic Circumpolar Current (ACC) or west wind drift, which 
ows east around the continent, near the Antarctic Convergence and carries with it nutrient rich upper circumpolar deep water (Tynan, 1998; Fallon and Stratford, 2003). The second is the coastal current (east wind drift) which moves towards the west as a counter current to the ACC. It moves close to the continent, and is responsible for forming eddies close to the shelf (Knox, 1994). The Antarctic Peninsula is the only land mass to extend from the con- tinent. Along with the tip of South America, this peninsula impedes both wind and ocean currents in the Southern Ocean through Drake Passage, the area between the two peninsula tips (Fallon and Stratford, 2003). The ACC moves faster through this area and constricts to bring in warmer wa- ter originating from the Bellingshausen Sea (to the west) towards the Scotia Sea (to the east) (Hewitt et al., 2002). The constriction of the ACC in this area forces the southern boundary (southern limits of the ACC) close to the continent at the Antarctic Peninsula (Tynan, 1998). In addition to the southern boundary, wind and bathymetry also contribute to the high productivity of the area and the large biomass of Antarctic krill (Euphausia superba) (Prezelin et al., 2000). Seasonal ice conditions are also a feature of the region, with the extent of sea ice as an important factor for many ice-associated species. Observed declines in sea ice and increases in temperature are more extreme at the Antarctic Peninsula than other Antarctic locations (Doake and Vaughan, 1991; Anisimov et al., 2001; Cook et al., 2005; Hansen et al., 2006a). One of the most studied species in the Antarctic, krill, has been identied to be a key link in the Antarctic food web, in addition to having stages of its life history associated with sea ice (Marschall, 1988; Daly, 1990; Moline et al., 2000). Future changes to the environment are expected to impact krill, and subsequently, the rest of the food web. 7 1.2. Study Areas Resource Uses Resource use in the Antarctic began with the discovery of seals at South Georgia before moving on to whaling, shing, and nally krill harvest. Seal- ing in the Antarctic began after Captain Cook reported large populations of fur seals in the sub-Antarctic islands (Kriwoken and Williamson, 1993). Seals were targeted for their pelts, with Antarctic and sub-Antarctic fur seals making up the majority of catches in the late 1700s to early 1800s with over 1.2 million harvested by 1822 (Agnew et al., 2000). While fur seals were the early targets of sealing 
eets, elephant, Ross, crabeater, and Weddell seals have all been targeted, with many populations being largely reduced by harvest (Fallon and Stratford, 2003). The Convention on the Conservation of Antarctic Seals (CCAS) was established in the 1970s to set catch limits for seals, and prevents the commercial harvest of seals south of 60S (Agnew et al., 2000). During the era of seal harvest, penguins were also harvested, primar- ily for oil (Agnew et al., 2000). Whaling began as seal resources declined. Commercial whaling in the Antarctic was initiated in 1892 and continued until 1982, when the International Whaling Commission (IWC) issued a moratorium on whaling (Fallon and Stratford, 2003; International Whaling Commission, 2009). Whaling started at South Georgia before expanding to other sub-Antarctic islands and further south to the continent (Agnew et al., 2000). Humpback, minke, blue, sei, southern right and sperm whales have all been harvested in the Southern Ocean (Fallon and Stratford, 2003). Due to large declines in many whale populations, the IWC assigned humpback and blue whales protected status in 1963 and 1964, respectively (Kriwoken and Williamson, 1993). The Southern Ocean was declared a whale sanc- tuary in 1994 by the International Whaling Commission prohibiting ship or land-based whaling operations (Agnew et al., 2000). Japan objects to the moratorium and continues to harvest whales, claiming scientic whal- ing, with their primary target being minke whales in the Southern ocean (Agnew et al., 2000). A shery for nsh species; mackerel icesh (Champsocephalus gunnari), 8 1.2. Study Areas spiny icesh (Chaenodraco wilsoni), marbled rockcod (Notothenia rossi), humped rockcod (Notothenia gibberifrons), blackn icesh (Chaenocephalus aceratus), and ocellated icesh (Chionodraco rastrospinosus) was open from 1978 to 1989 in the Antarctic Peninsula area (Kock, 1998). Since the sh- ery closure there is currently some exploratory shing, but no re-opening of nsh shing. Patagonian and Antarctic toothsh (Dissostichus eleginoides and Dissostichus mawsoni) were harvested within the Southern Ocean start- ing in the mid 1980s (Agnew et al., 2000). The majority of catches from this shery are taken from South Georgia, with limited catches recorded for only a few years within the Antarctic Peninsula area (CCAMLR, 2008b). Following a decade of exploratory shing Antarctic krill, became a tar- get species when the commercial shery opened in 1972 (Nicol and Endo, 1999; Agnew et al., 2000). Japan, the Soviet Union, and Russia obtain the majority of krill catches, with large numbers harvested from the Antarctic Peninsula (Nicol and Endo, 1999; CCAMLR, 2008b). The shery operates year-round with catches closer to the continent occurring primarily in the austral summer, and catches from sub-Antarctic areas (South Georgia) in winter months (Nicol and Endo, 1999). Observed declines in krill stocks over the last 20 years are associated with changes in environmental conditions (Atkinson et al., 2004). While the quota for krill is much higher than annual catches, in 2010 catch biomass increased to nearly double the values from 1994-2009 (Nicol et al., 2012). Management The Antarctic Treaty, which entered into force in 1961, established freedom of scientic information in the Antarctic in addition to establishing its use for peaceful purposes (Anonymous, 1959). Prior to this, the International Whaling Commission was responsible for managing species in the Southern Ocean (Fallon and Stratford, 2003). In 1982 the Commission on the Con- servation of Antarctic Marine Living Resources (CCAMLR) was established (CCAMLR, 1980). It has been considered one of the rst regulating agen- cies to establish an ecosystem approach to managing resources (Constable 9 1.3. Ecopath with Ecosim et al., 2000). In 1985 CCAMLR established the EcosystemMonitoring Program (CEMP) in order to regulate harvest in accordance with the ecosystem approach. CEMP monitors both harvested and dependent species to estimate preda- tor, prey and environmental performance parameters around the Antarctic (Agnew, 1997). The monitoring program assists CCAMLR in parameteriz- ing models for use in establishing quotas. 1.3 Ecopath with Ecosim The majority of this thesis applies the Ecopath with Ecosim (EwE) ap- proach to construct ecosystem models and simulate changes over time. The Ewe approach originated with a single mass-balanced Ecopath model based in Hawaii (Polovina, 1984), and has expanded throughout development to include numerous additional features for assessing ecosystems. Temporal simulations (Ecosim) and spatial analysis abilities (Ecospace) were later added to aid in assessments of shing policies and formation of protected areas (Walters et al., 1997, 1999, 2000). Indices to explore the health of the ecosystem were developed through a series of network analyses (Christensen and Pauly, 1992; Christensen, 1995). Additional features such as automated mass-balance with incorporation of Monte Carlo for better parameter es- timation and network analysis have been added throughout development (Kavanagh et al., 2004). EwE is used in over 154 countries, with over 300 papers published, and has been named one of NOAA's top 10 breakthroughs (NOAA, 2006). An updated version is now in use to allow greater 
exibil- ity in user programming and coupling between other modelling programs (Christensen et al., 2007; Buszowski et al., 2009). Ecosystem models, specically EwE, have been developed to evaluate ecosystem eects of shing and environmental change (Christensen and Wal- ters, 2004), which are the main objectives of the thesis. Single species models may prove ecient when assessing one species, but they are unable to iden- tify potential impacts caused by linkages within the ecosystem (Fulton and Smith, 2004). Multispecies models are able to identify non-intuitive changes 10 1.4. Thesis Outline in biomass through species interactions within the model, and may assist in evaluating ecosystem impacts of management policies (Walters et al., 1997; Fulton and Smith, 2004). While other modeling tools exist (see Plaganyi, 2007, for a detailed com- parison of ecosystem modelling tools), EwE was selected over single species models due to ease of use and scope of the thesis. Ecosystem modelling tools such as Atlantis are considered to be the most complete when assess- ing entire ecosystems as it represents both biological and physical interac- tions within an ecosystem, however large amounts of data are required in addition to sub-models to address bio-geochemical interactions (Fulton and Smith, 2004; Plaganyi, 2007), which are beyond the scope of the thesis. Lim- itations to ecosystem and other multi-species approaches to modelling are rooted in the quality and availability of data (Plaganyi and Butterworth, 2004). The EwE approach allows users an existing model framework in ad- dition to the ability of the software to focus on shery and environmental issues (Christensen and Walters, 2004; Plaganyi and Butterworth, 2004). As models should aim for the 'minimum realistic' approach to avoid over- parameterization (Fulton et al., 2003), the EwE software was selected as the most capable tool for the thesis. 1.4 Thesis Outline Managers are becoming increasingly focused on policies that include an ecosystem-based management approach, meaning the context of the ecosys- tem is considered when policies are focused around a particular species. Ecosystem models can identify potential impacts that a series of single- species models cannot (Fulton and Smith, 2004). Furthermore, management policies focusing on single species have the potential to overlook important indirect trophic linkages to targeted species. Within this thesis, I investigate the impacts of harvest on all species in the ecosystem within the same time scale in conjunction with known or theorized impacts from environmental changes. The goal of this thesis is to identify important stressors to each ecosystem, and how future changes in these stressors may impact ecosystem 11 1.4. Thesis Outline structure. Chapters 2 and 3 Chapters 2 and 3 use the Ecopath with Ecosim software (Walters et al., 1997; Christensen et al., 2005) to assess past changes in ecosystem struc- ture for Hudson Bay and the Antarctic Peninsula respectively. Models were constructed based on past ecosystem structure and projected forward to the present day focusing on catch and environmental changes that have oc- curred. Methods for using Ecosim simulations to recreate past catch and environmental trends is well established, and has been explored for a mul- titude of ecosystems, including the Gulf of Alaska and Aleutian Islands, northern British Columbia, Raja Ampat Indonesia and the northern Ionian Sea (Guenette et al., 2006; Ainsworth et al., 2008b,a; Piroddi et al., 2010). In chapter 2, data from all species are combined to assess the trophic structure of the Hudson Bay food web through diet linkages. As part of the IPY project on marine mammals, this chapter explores the potential causes of changes to marine mammals and the rest of the ecosystem with respect to climate change. Declines in some stocks of marine mammals (polar bears, eastern Hudson Bay beluga and narwhal), have prompted research on the reasons for these changes, in part to determine if climate change has had an impact (Stirling et al., 1999; COSEWIC, 2004a; Stirling et al., 2004; Ham- mill et al., 2009). I rst identied the ecosystem structure through literature reviews and assistance from researchers at the Department of Fisheries and Oceans Central and Arctic Division in Winnipeg, Canada. I was able to assess gaps in data, such as the biomass of sh groups, through modeling approaches, such as the Monte Carlo routine in EwE (Christensen and Wal- ters, 2004). After the initial structure of the model was complete, re-creation of past trends in Ecosim were performed. Catch records for marine mammal species were readily available from government records, but information on other species was lacking. Changes in the diets of thick-billed murres indi- cated shifts in the sh community from benthic to pelagic species (Gaston et al., 2003). This was coupled with information on lower trophic levels 12 1.4. Thesis Outline from other Arctic ecosystems to gain an understanding of past changes to the ecosystem. In chapter 3, the past Antarctic Peninsula ecosystem was recreated in the same manner as the Hudson Bay ecosystem (chapter 2). Previous as- sessments of the Antarctic Peninsula have utilized the EwE methods (Efran and Pitcher, 2005; Cornejo-Donoso and Antezana, 2008), however, they have not included environmental factors. This chapter expands on past research to incorporate dierent environmental variables to explain declines in krill biomass, and increases in salp groups (a gelatinous tunicate and perceived competitor of krill) in conjunction with harvesting trends. This chapter tests the likelihood of dierent environmental variables as causing the changes in salp and krill abundance based on ecological studies (Marschall, 1988; Loeb et al., 1997; Brierley and Watkins, 2000; Atkinson et al., 2004; Lee et al., 2010; Flores et al., 2011). I also explore the eects of increasing the harvest of krill to quota levels. As the krill shery operates on what is considered a keystone species (Quetin and Ross, 1991; Moline et al., 2004), annual catches are only roughly 10% of the quota limits (Hewitt et al., 2002, 2004). This chapter explores the potential repercussions of harvesting krill at full quota levels. Chapters 4 and 5 Chapters 4 and 5 build on chapters 2 and 3 respectively, by extending sim- ulations into the future. Ecosim scenarios are routinely used to explore shing strategies in future scenarios (Araujo et al., 2008; Heymans et al., 2009) particularly in an economic context. However, rather than focusing on maximizing prots or other policy objectives, these chapters explore future ecosystem states and address the ecological structure rather than policy objectives. Each chapter utilizes dierent levels of harvest and environ- mental drivers previously identied to assess potential future states of each ecosystem. Data from global climate models GFDL (2010) allowed for envi- ronmental drivers to be continued into the future in conjunction with IPCC (Intergovernmental Panel on Climate Change) scenarios. Catch scenarios for 13 1.4. Thesis Outline each ecosystem are based on either current harvest levels or are increased to simulate higher quotas in the future. These chapters identify species within each ecosystem likely to be impacted by harvest or environmental changes in the future. Chapter 6 Chapter 6 focuses on the human component to the Hudson Bay ecosystem by providing an economic assessment to the harvest of narwhal and beluga. Many species are currently harvested within Hudson Bay by Inuit, however cetacean species have been a prominent focus of the Inuit diet for thousands of years (Stewart and Lockhart, 2004; Freeman, 2005). Narwhal and beluga were selected as the focus for an economic assessment of hunting. Model simulations from chapters 2 and 4 identify declines in narwhal and the east- ern Hudson Bay beluga indicating their potential lack of availability in the future. Focussing on these two hunts, the economic use value is explored primarily through the costs and revenues associated with harvesting narwhal and beluga. Past economic assessments in the north have been limited and focused on one or more aspects of individual hunts rather than an overview (Weaver and Walker, 1988; Reeves, 1992a). Previous studies have assessed the economic value of hunting in specic high Arctic communities Loring (1996), however this has not been attempted for the Hudson Bay region. This chapter provides a summary of economic components associated with the harvesting of these two species. In addition to providing an estimate on the total economic use value for each for theses hunt, costs and revenues are also assessed based on each community participating in the harvest. Chapter 7 Chapter 7 provides a summary and discusses the results of the thesis in the context of managing ecosystem. Directions for future research and applica- tions to management are presented. 14 Chapter 2 Impacts of Hunting, Fishing, and Climate Change to the Hudson Bay Marine Ecosystem 1970-2009 2.1 Synopsis An ecosystem model was created for the Hudson Bay region, Canada, for 1970-2010, aiming to identify ecosystem linkages while bringing together research from diverse research sources. The research presented was com- pleted as part of the International Polar Year Global Warming and Arctic Marine Mammal project, focusing on the impacts of climate change on ma- rine mammals. The model presented in detail here synthesizes research spanning all trophic levels for incorporation into the Ecopath with Ecosim (EwE) modeling framework. The Ecopath model, containing 40 functional groups, identies a previously unestimated sh biomass of 3.42t  km2 for the region, based on the trophic linkages and diets within the food web. Catch and abundance data for the Hudson Bay region, along with environ- mental drivers (sea surface temperature and ice cover) were used to re-create past changes to the ecosystem through the tting of individual groups. The Ecosim model captures many dynamics present in the system, while iden- tifying gaps in existing data for future research and as the basis for work simulating climate change and its impacts on the ecosystem. A general shift in lower trophic levels from a sea ice to benthos to benthic sh pathway to 15 2.2. Introduction one favoring pelagic phytoplankton to zooplankton to pelagic sh. Declines in polar bear, narwhal, and eastern Hudson Bay beluga model groups iden- ties harvest as the main stressor. Simulations testing the model sensitivity to hunting and environmental pressures indicate the biomasses of higher trophic level organisms (marine mammals) are more responsive to hunting pressures while lower trophic levels (benthos, zooplankton) are more easily in
uenced by climate drivers. 2.2 Introduction Polar regions are increasing in temperature faster than temperate areas, with Arctic temperature rising at almost twice the rate of the rest of the world (ACIA, 2004). The fourth International Polar Year (IPY) in 2007-2009 highlighted the need for research to increase our knowledge of the dynamics occurring in Polar areas. While Hudson Bay (HB) (gure 2.1) is geographically considered sub- Arctic, between 50-70N, this system re
ects high Arctic attributes such as climate, biogeography, and higher trophic level animals. For example, polar bears, are found at their lowest latitudinal range in HB, due to the cold winters and the ice available for foraging (Stirling and Parkinson, 2006). Moreover, many species present in this ecosystem have adapted to the sea- sonal ice cycle, from whales occupying the region during the ice free seasons, and seals breeding on the ice, to the ability of smaller zooplankton to survive winter months using nutrients frozen within the sea ice (Poltermann, 2001; Stewart and Lockhart, 2005). Research in HB has been limited in the past, compared to other Arctic ecosystems. Two surveys of phytoplankton and zooplankton have been com- pleted in HB assessing lower trophic levels; one in 1993 sampling from James Bay (JB) along the east coast of HB into Hudson Strait (HS) (Harvey et al., 1997, 2001), and a second in 2003 running east to west through the middle of HB (Harvey et al., 2006). The most comprehensive benthic summary from numerous locations in HB from 1953 to 1956 (Atkinsor and Wacasey, 16 2.2. Introduction 1989) recorded only the presence of benthic species. Fish are poorly under- stood, although there is the general belief that sh are not abundant in HB, a situation somewhat veried by unsuccessful commercial shery ventures in the past (Stewart and Lockhart, 2005). Marine mammals are some of the most well studied species in the region, although only a handful of surveys have been completed for each species (Ferguson et al., 2010). Surface temperatures in HB have increased by 0.5-1.5C during 1955- 2005 (Hansen et al., 2006a), and sea ice extent decreased by 2000900 km2y1 between 1978 and 1996 (Parkinson et al., 1999). These changes combined with a longer ice free season (Gough et al., 2004; Gagnon and Gough, 2005) are likely yielding large scale changes to the sympagic marine ecosystem. Ice algae, which contributes up to 57% of primary production in some Arctic regions (Gosselin et al., 1997), and roughly 25% of total production in some areas of Hudson Bay (Legendre et al., 1996), can be stored through the winter within the sea ice. Therefore, the loss of sea ice will alter the availability of algae stored within the sea ice, which will cause shifts in the ecosystem by altering energy transfer to higher trophic levels. Such shifts have already been observed in bird diets as indicated by declines in Arctic cod (Boreogadus saida) and benthic sh species such as sculpins (Family: Cottidae) and zoarcids (Family: Zoarcidae) with increases in pelagic sh such as capelin (Mallotus villosus) and sandlance (Ammodytes spp.). Polar bear populations are at their southern limit in HB, and already experience longer summers than their northern counterparts. Lengthening of the ice free summer is believed to increase nutritional stress as there is less ice to forage on, decreasing their hunting platform, and making polar bears vulnerable to sea ice declines (Stirling and Derocher, 1993; Stirling et al., 1999). Along with environmental changes, human uses of the ecosystem also have the potential to alter the abundance of species. Currently all ma- rine mammal species are hunted annually, with the exception of bowhead where harvest only occurs in specic years. Quotas are imposed on the har- vest of certain cetacean species. Seabirds and sh are also harvested, how- ever these are generally unregulated. Since the 1970, human populations 17 2.3. Methods have nearly tripled (Bell, 2002; Statistics Canada, 2006; Nunavut Bureau of Statistics, 2008; Sutherland et al., 2010) with increases in harvest levels for many species also being recorded. Understanding whether these stocks can withstand the continuous pressure of harvest is important, and even more so in conjunction with the impacts of climate change. In order to test the importance of multiple stressors on the ecosystem, we have constructed an ecosystem model to re-create the dynamics from 1970-2009. The ecosystem model was created using the Ecopath with Ecosim soft- ware (Buszowski et al., 2009; Christensen et al., 2007), to assess the Hudson Bay ecosystem with a mass-balance model. Through the construction of an Ecopath model, gaps in existing ecosystem knowledge can be identied. For example, biomass of sh populations are obtained by assessing the de- mands of predators and the amount of sh which can be supported by lower trophic levels, based on food web structure. Ecosim temporal simulations (Walters et al., 1997; Christensen and Walters, 2004) are used to re-create observed changes since 1970, helping to identify causes to changes in ecosys- tem structure. The model aims to focus on the impact of climate change and hunting on marine mammal species as part of the Global Warming and Arctic Marine Mammal International Polar Year research project, therefore giving marine mammals a greater presence in the model structure. While high and low trophic level organisms are relatively well studied in this re- gion, serious gaps regarding mid trophic level organisms (primarily benthos and sh) exist. Despite these gaps, there is an urgency to understand a sys- tem that is subjected to multiple stressors. This modeling approach allows us to infer changes likely occurring to mid-trophic level organisms through existing knowledge of predators and producers. 2.3 Methods Study Area The Hudson Bay region often includes Hudson Bay (HB), James Bay (JB), Foxe Basin (FB) and Hudson Strait (HS) (gure 2.1). This system is one of 18 2.3. Methods the largest bodies of water in the world to freeze over every winter and open up every summer. HB and JB are both categorized by shallow, less produc- tive waters, with large inputs of freshwater from rivers in the spring. Con- versely, Foxe Basin and Hudson Strait have more mixing with the Labrador Sea (Straneo and Saucier, 2008), and are thought to be an important sea ice choke-point for HB, ultimately determining which marine species have access to the region (Higdon and Ferguson, 2009). Repulse Bay Cape Dorset Sanikiluaq Arviat Rankin Inlet Baker Lake Whale Cove Peawanuck Fort Albany Attaapiskat Moosonee Waskaganish Eastmain Chisasibi Kuujjuarapik Umiujaq Inukjuaq Puvirnituq Akulivik Kimmirut Killiniq Kangiqsualujjuaq KuujjuaqTasiujaq Aupaluk Kangirsuk Quaqtaq Salluit Ivujivik Kangiqsujuaq Fort Severn 4-3 Coral Harbour Chesterfield Inlet Churchill 60°0'0"W 80°0'0"W 80°0'0"W100°0'0"W 60°0'0"N 60°0'0"N 50°0'0"N 50°0'0"N ! NU MB ON QB HB JB FB HS Figure 2.1: Greater Hudson Bay region including Hudson Bay (HB), James Bay (JB), Hudson Strait (HS), and Foxe Basin (FB). Communities in Nunavut (NU), Manitoba (MB), Ontario (ON), and Quebec (QB) are shown. Selection of the model area was based on use patterns of marine mammals as their data are more prevalent compared to sh and plankton species. JB was included in the model area due to its similarity to southern HB and the use of this area by certain stocks of polar bears, beluga, seals, and birds. HS and FB were excluded from the model area, as these deeper more productive waters are strongly in
uenced by currents (Straneo and Saucier, 2008), and are likely to host a dierent suite of species. For the remainder of 19 2.3. Methods this paper, referral to HB will include JB, an area covering roughly 900,000 km2 (Legendre et al., 1996). The Ecopath base year model describes the conditions in 1970, with the Ecosim model running from 1970-2009. The base year was chosen as there are no comprehensive estimates of marine species prior to 1970. In addition, changes in environmental conditions and harvest pressure have been documented for this period, thus making for an interesting time to examine the ecosystem dynamics. Model Equations Using the Ecopath with Ecosim (EwE) software version 6 (Christensen et al., 2007; Buszowski et al., 2009), an Ecopath or mass-balance model was con- structed for 1970. This mass-balance approach links all species or functional groups (groupings of similar species) through diets. Under this assump- tion there must be enough energy produced by each prey group to account for consumption, migration, shing mortality, and other mortalities. More specically this can be expressed as: Pi = X j Bj M2ij + Yi +Ei +BAi + Pi  (1EEi) (2.1) where Pi is the production of functional prey group i, Bj is the biomass of predator group j with predation mortality on group i of M2ij . Yi is the shery catch, Ei is the net migration rate (emigration-immigration), BAi is the biomass accumulation, and EEi is the ecotrophic eciency (proportion of production that is consumed within the system by predators or exported out of the system due to shing or migration) for prey i. Equation 2.1 can be re-written as equation 2.2: Bi  (P=Bi) = X j Bj  (Q=B)j DCji+Yi+Ei+BAi+Bi  (P=B)i  (1EEi) (2.2) Where Bi and Bj are the biomasses of prey (i) and predator (j), (P=B)i is the production to biomass ratio, generally equal to total mortality (Z) (Allen, 1971), (Q=B)j is the consumption by predator i per unit biomass, 20 2.3. Methods and DCji is the proportion of prey i in the diet of predator j. Ecopath models are balanced using an algorithm to solve a set of linear equations in the form of Equation 2.2 for each functional group. For each functional group 3 of the 4 basic parameters are imputed (B, P/B, Q/B, EE) along with shery landings and diet composition, allowing the algorithm to solve for the 4th parameter. Temporal simulations were generated for the time period of 1970-2009 in Ecosim using equation 2.3; dBi=dt = gi X j Qji  X j Qij + Ii  (MOi + Fi + ei)Bi (2.3) Where dBi=dt represents the change in biomass (B) for group i over the time interval t, with starting biomass Bi. gi represents the net growth e- ciency (production/consumption ratio), the X j Qji is the total consumption on group i, and X j Qij is the predation of all predators on group i. MOi represents the other mortality term (for mortality associated with old age), Fi is the shing mortality rate, Ii is the immigration rate, ei is the emigra- tion rate, with the combined term Bi  (ei  Ii) as the net migration rate. The consumption rate of a group, Qij is based on the foraging arena theory where the biomass Bi is further divided into vulnerable and invulnerable proportions to group i's predators (Walters et al., 1997), and the transfer rate between these two states. Ecosim is based on the foraging arena theory that describes the interactions between predators and prey attributing a vul- nerability term. Low values of vulnerability (close to 1) mean that prey pro- duction determines the predation mortality (bottom-up interaction) while high values of vulnerability (e.g., 100) mean that predator biomass deter- mines how much prey is consumed (top-down interaction)(Christensen and Walters, 2004). 21 2.3. Methods Model Inputs and Functional Groups Ecopath model parameters were set to 1970 values for the marine environ- ment only, estuary and freshwater areas were excluded from the model. A total of 40 functional groups were created; 15 marine mammal groups, 1 bird group including all birds, 9 sh groups, 7 plankton groups, 4 benthic groups, 2 producers, and 2 detritus groups (species for each functional groups are listed in appendix A with full details on input parameters). Marine mammal groups were created to represent individual species, or separate stocks within species if applicable, as changes in stocks have been identied. As there was little knowledge of sh species in the region, sh species were grouped into functional groups based on life history, feeding preferences, and taxonomic characteristics. Plankton and benthic groups were split into those important to higher predators or groups with more information available. Primary producers were split into two groups: ice associated algae and pelagic phytoplankton, with the aim to capture the dynamics of organisms, which are dependant on either one. Ice algae is an important component of the ecosystem, as plankton cells are frozen within the ice each fall and released back into the water column during the spring melt. Contribution of ice algae has been estimated at 25% of total production in parts of HB (Legendre et al., 1996) and can range from 57% in the central Arctic to 3% in surrounding sub-Arctic areas (Gosselin et al., 1997). While some species of phytoplankton and zooplankton have adapted to survive this freeze and return to the water column the following year (Horner et al., 1992), those that do not survive sink through the water column to the benthos. Within the model, exports from the ice algae group are directed to the ice detritus group, which is a major contributor to the diets of benthos. During the spring melt, algal cells are 
ushed out of the brine channels into the pelagic environment, with a minimum export of ice algae to the benthic community estimated at 20% in southeastern parts of HB (Tremblay et al., 1989). Moreover, accumulation of algal biomass within the sea ice is thought to favor an eective transfer to the benthos, as aggregated algal cells sink up to three times faster than individual algal cells, and damaged cells sink 22 2.3. Methods faster than healthy ones (Tremblay et al., 1989; Riebesell et al., 1991). It has been noted in other Arctic ecosystems that zooplankton biomass is too low during the spring melt to eciently graze the sinking ice algae, allowing it to sink to the benthos (Legendre et al., 1992). The pelagic production functional group represents all producers not as- sociated with the sea ice. This group exports to a pelagic detritus groups, which is named as it represents the detritus captured by the pelagic produc- ers, rather than its location in the water column. Pelagic production blooms generally occur after the sea ice has started to melt, and remains in the wa- ter column longer than ice algae cells (Tremblay et al., 1989). This pelagic bloom sustains pelagic sh and zooplankton into the summer months. In order to simulate changes to primary producer functional groups, data was extracted from the global Hadley Centre Sea Ice and Sea Surface Temperature model (HadISST) from the British Atmospheric Data Centre (2010) and used to force the primary production groups. Warmer temper- atures have been shown to alter the mean ice freeze-up and break-up dates by 0.8-1.6 weeks in spring and fall (Hochheim et al., 2010). The availability of ice algae within the model is contingent upon the presence of sea ice; therefore the ice algae group was driven through a forcing function (FF) in the model. The sea ice FF was applied to the ice algae group, as a multiplier of the production rate using the average % cover of sea ice of all cells in the model area. The pelagic phytoplankton functional group was also driven in the model using SST (sea surface temperature), from the same HadISST model. Figure 2.2 shows the average SST and % ice cover by month for 1970-2009 with 95% CI. See appendix A for details on model tting and selection of drivers. There are no estimates of sh biomass or community composition for HB. Changes in sh populations have been inferred from the diets of thick- billed murres, as biomass is estimated using equations 2.1 and 2.2 in order to satisfy the needs of the predators within the food web, using each group's respective production ability. There has been a shift from Arctic to sub- Arctic sh composition (gure 2.3); from Arctic and polar cod, sculpins, and zoarcids to capelin and sandlance (Gaston et al., 2003). Although the 23 2.3. Methods -1 0 1 2 3 4 5 0 20 40 60 80 100 S ea  S u rf a ce  T em p er a tu re  ( ˚C ) Ic e C o v er  ( % ) Month Ice Temp J F M A M J J A S O N D Figure 2.2: 30 year means and 95% CI for sea surface temperature (SST) and % ice cover calculated by the HadISST global model. diets were collected from the northern limits of HB, due to the gross lack of data on sh populations, diets of birds were the only indication of changes in sh community structure. For all functional groups biomass parameters were expressed in t km2, and for non-sh groups were based on surveys collected within the region. For many marine mammal species the total number of animals has been re- ported. Here, the biomass was extrapolated to the entire region area, which for HB and JB has been estimated at nearly 900,000 km2 (Legendre et al., 1996). P/B (production to biomass) and Q/B (consumption to biomass) were calculated as a yearly value (y1) from species specic empirical values if available, with P/B ratios adjusted to account for hunting and shing mortality in the Ecopath model. Expert opinion, and values from similar ecosystems were used in absence of region specic data. EE (ecotrophic eciency) was generally estimated by the model, considering the model bal- anced when the EE value was between 0 and 1 (Christensen et al., 2005). For full descriptions of data incorporated into the model see appendix A. Model tting included hunting/shing for species, which are known to be 24 2.3. Methods 0 10 20 30 40 50 60 70 80 90 100 1984 1987 1990 1993 1996 1999 2002 C o n tr ib u ti o n  t o  D ie t (% ) Year Arctic Cod Sculpins/ Zoarcids Capelin Sandlance Figure 2.3: Changes in sh abundance as measured by the diets of thick- billed murres. Graph recreated from data presented in Gaston et al. (2003). harvested table 2.1. Model Analysis and Simulations Monte Carlo simulations were run on the tted model to estimate plausible ranges of biomass using equation 2.4: Lxi = xi  2  CV  xi (2.4) where Lx represent the limits (upper and lower) of the biomass of group i. The mean biomass, xi, is taken as the value imputed Ecopath starting value. CV values were determined using a pedigree ranking, whereby input parameters are assigned a coecient of variation (CV) based on the quality of input data, using the pedigree routine in EwE version 5 (Christensen et al., 2005) (see table 2.2 for CV values used in the Monte Carlo Routine). One thousand Monte Carlo simulations were run to nd ranges of input parameters that allowed the Ecopath model to be balanced. The trophic level (TL) of each species group was calculated for the ini- 25 2.3. Methods Table 2.1: Hunting and shing trends as drivers for the Ecosim model (zindicates information also contributed by Ferguson (pers. comm.)) Fishery Functional Groups Model Drivers References SH Polar Bear Southern Hudson Bay Polar Bear Landings (Lee and Taylor, 1994; Aars et al., 2005) WH Polar Bear Western Hudson Bay Polar Bear Landings (Lee and Taylor, 1994; Aars et al., 2005) FB Polar Bear Foxe Basin Polar Bear Landings (Lee and Taylor, 1994; Aars et al., 2005) Killer whale Killer Whale Landings (Higdon, 2007)z Bowhead Bowhead Landings (Higdon, 2008)z Narwhal Narwhal Landings (DFO, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998; Stewart and Lockhart, 2005; JCNB/NAMMCO, 2009) N Walrus Northern Hudson Bay Wal- rus Landings (Strong, 1989; NAMMCO, 2005b; Stewart and Lockhart, 2005) S Walrus Southern Hudson Bay Wal- rus Landings (Strong, 1989; NAMMCO, 2005b; Stewart and Lockhart, 2005) Beluga E Eastern Hudson Bay Beluga Landings (JCNB/NAMMCO, 2009; de March and Postma, 2003) Beluga W Western Hudson Bay Beluga Landings (JCNB/NAMMCO, 2009; de March and Postma, 2003) Beluga S James Bay Beluga Landings (JCNB/NAMMCO, 2009; de March and Postma, 2003) Sealing Bearded Seal, Harbour seal, Ringed Seal, Harp Seal Eort (Stewart and Lockhart, 2005) Bird Hunting Birds (all) Eort (Stewart and Lockhart, 2005) Fishing Arctic Char, Atlantic Salmon, Gadiformes, Sculpins/Zoarcids, Capelin, Sandlance, Other Marine Fish, Brackish Fish Eort (Stewart and Lockhart, 2005; Booth and Watts, 2007) tial Ecopath model and each year of the Ecosim simulation using equation 2.3, where primary producers are assigned a TL of 1, and consumers with diets comprised of 100% primary production have a TL of 2 (Christensen et al., 2007). Consumer TL, TLi, is dependent upon the TL of prey items (TLa; TLb; TLc, for prey items a, b, c, etc.) and the percentage (X) each prey item contributes to the predator's diet (Xa; Xb; Xc). TLi = 1 + X (Xa  TLa) + (Xb  TLb) + (Xc  TLc)::::: (2.5) 26 2.4. Results Once the TL of each species group is calculated, the mean trophic level of the ecosystem (TLE equation 2.6) and the mean trophic level of the catches (TLC equation 2.7) can be calculated for each year of the simulations (1970- 2009); TLE = X Bi BE  TLi (2.6) TLC = X Ci CE  TLi (2.7) where Bi and Ci are the biomass and catch for group i, and BE and CE are the biomass and catch of the entire ecosystem, with values represented in t  km2. Using the model tted to reported trends in hunting and environmental conditions, further Ecosim simulations were run to test the sensitivity of the ecosystem to hunting and environmental conditions (SST and % ice cover). Two additional simulations were run. First, a "No Hunting" scenario was run, removing all hunting and shing mortality from the model while still using environmental drivers (SST and ice cover). Second, a "Constant Cli- mate" scenario assumed past hunting levels, but the environmental data from 1970 was repeated annually until 2009, to simulate a constant climate condition thus eliminating the declines in sea ice and increases in temper- ature. This allowed assessment of climate changes in functional groups if driven by environmental changes, hunting pressure, or both. 2.4 Results Ecopath Output Using input parameters listed in appendix A, the model was able to esti- mate the missing parameters in table 2.2. Through the balancing of the model many parameters were rened. Once Ecopath parameters (B, P/B, Q/B, EE) were calculated, P/B ratios were adjusted to account for hunting mortality. The equation used to calculate the P/B ratio for sh often un- 27 2.4. Results derestimates higher latitude species (Pauly, 1980), and the smaller P/B was causing the mass model to estimate large biomasses of sh. Consequently, these ratios were increased to the upper limits based on the species found within the functional group. Many of the zooplankton groups lacked region specic data for P/B and Q/B, therefore a P/Q ratio of 0.25 was assumed (Christensen et al., 2005), allowing the model to estimate an additional pa- rameter. The EE of birds indicated higher mortality than allowed in the model, therefore the P/B ratio was increased to allow for hunting and pre- dation mortality within the model. Food web structure is displayed in gure 2.4. 28 Table 2.2: Balanced Ecopath model parameters. Biomass (B) and catches are presented in t  km2, PB (Pro- duction/Biomass ratio), QB (Consumption/Biomass ratio), and BA (Biomass Accumulation) are presented in y1. EE (Ecotrophic Eciency) and P/Q (Production/Consumption) ratios are dimensionless. Bolded values are estimated by the Ecopath model. The CV (Coecient of Variation) values for each group are used in equation 2.4 to calculate biomass ranges. Group Name TL B PB QB EE PQ BA Catches CV WHB Polar Bear 4.857 0.0005 0.129 2.08 0.414 0.062 - 1.50E-05 0.15 SH Polar Bear 4.906 0.0004 0.154 2.08 0.506 0.074 - 2.20E-05 0.15 Polar Bear Foxe 4.927 0.0002 0.121 2.08 0.304 0.058 - 5.00E-06 0.15 Killer Whale 4.872 2.5E-05 0.151 4.998 0.265 0.03 - 1.00E-06 0.15 Narwhal 4.062 0.0019 0.084 26.182 0.271 0.003 - 3.40E-05 0.15 Bowhead 3.335 0.0109 0.021 5.475 0.384 0.004 0.007 9.00E-06 0.4 Walrus N 3.332 0.0027 0.172 47.123 0.188 0.004 - 8.00E-05 0.25 Walrus S 3.452 0.001 0.097 33.778 0.143 0.003 - 6.00E-06 0.25 Bearded Seal 3.866 0.0037 0.176 14.262 0.791 0.012 - 0.000167 0.25 Harbour Seal 3.971 0.001 0.125 18.612 0.074 0.007 - 2.00E-06 0.25 Ringed Seal 4.077 0.0469 0.158 17.272 0.413 0.009 - 0.000393 0.25 Harp seal 4.103 0.001 0.126 15.66 0.688 0.008 - 1.40E-05 0.25 Beluga E 3.694 0.0021 0.066 21.448 0.22 0.003 -0.004 3.30E-05 0.15 Beluga W 3.873 0.0247 0.064 16.713 0.133 0.004 0.01 6.05E-05 0.15 Beluga James 3.869 0.0015 0.087 16.623 0.679 0.005 - 1.40E-05 0.15 Seabirds 3.839 0.065 0.37 17.258 0.95 0.021 - 0.000325 0.4 Arctic Char 3.3 0.412 0.2 1.5 0.95 0.133 - 4.62E-07 0.1 Atlantic Salmon 3.45 0.148 0.52 7.15 0.95 0.073 - 1.32E-08 0.1 Continued on Next Page 29 Table 2.2 Continued Group Name TL B PB QB EE PQ BA Catches CV Gadiformes 3.235 0.853 0.47 1.85 0.95 0.254 - 2.64E-07 0.1 Sculpins/ Zoarcids 3.188 0.382 0.7 3.269 0.95 0.214 2.64E-07 0.1 Capelin 3.132 0.488 1.7 4.8 0.95 0.354 - 1.32E-07 0.1 Sandlance 3.128 0.705 0.85 3.45 0.95 0.246 - 3.96E-08 0.1 Sharks/Rays 4.033 3.18E-06 0.22 1.25 0.95 0.176 - - 0.1 Other Marine Fish 2.948 0.374 0.932 3.018 0.95 0.309 - 6.60E-08 0.1 Brackish Fish 3.216 0.055 3.5 5.798 0.95 0.604 - 2.64E-08 0.1 Cephalopods 3.645 0.227 1.5 5 0.95 0.3 - - 0.25 Macro-Zooplankton 2.711 7.5 1 3 0.278 0.333 - - 0.25 Euphausiids 2.787 2.148 3.3 13.2 0.8 0.25 - - 0.15 Copepods 2.05 4.015 16 64 0.472 0.25 - - 0.15 Crustaceans 2.41 1.8 3.6 14.4 0.584 0.25 - - 0.15 Other Meso-Zooplankton 2.336 1.21 10 40 0.556 0.25 - - 0.15 Micro-Zooplankton 2 2.235 15 45 0.95 0.333 - - 0.25 Marine Worms 2.275 5.93 0.6 4 0.95 0.15 - - 0.1 Echinoderms 2.575 8.708 0.3 1 0.95 0.3 - - 0.1 Bivalves 2.148 5.942 0.57 6.3 0.95 0.091 - - 0.1 Other Benthos 2.091 3.139 2.5 12.5 0.95 0.2 - - 0.1 Pelagic Production 1 8 46.865 - 0.8 - - - 0.15 Ice Algae 1 3.5 46.197 - 0.65 - - - 0.15 Ice Detritus 1 0.009 - - 0.904 - - - - Detritus 1 0.33 - - 0.224 - - - - 30 2.4. Results 5 4 3 2 1 Gadiformes Ice Detritus Pelagic Detritus Echinoderms Worms Bivalves Other Benthos Narwhal Polar Bear Bearded Seal Harbour/Harp Seal Walrus Zooplankton Ringed Seal Killer Whale Beluga Bowhead Crustaceans Copepods Capelin Ice Algae Pelagic Production Seabirds Euphausiids Arctic Char Sculpins/Zoarcids Marine/Brackish           Fish Sandlance Figure 2.4: Food web linkages in the HB ecosystem with respect to Trophic Level (horizontal lines). Linkages between functional groups were drawn for prey contributing 10% or more to the diet of a predator. For func- tional groups with more than one species, graphical representation of one species within the group was used. Certain functional groups were com- bined to be represented by one image; polar bear (western HB, southern HB, and FB polar bear), beluga (eastern HB, western HB, and JB beluga), walrus (northern and southern walrus), harbour/harp (harbour and harp seals), marine/ brackish sh (Atlantic Salmon, sharks/rays, other marine sh, brackish sh), zooplankton (macro-zooplankton, cephalopods, other meso-zooplankton, and micro-zooplankton). Size of image does not indicate biomass size or individual size. All images c
Megan Bailey, 2010 adapted by permission. 31 2.4. Results Ecosim Fitting Results of time series tting, using the data trends provided in table 2.1, and adjusting the vulnerabilities to obtain the observed trends are presented in gure 2.5. See appendix A for full details of vulnerabilities, details of t- ting each group, and the general model tting process. Primary producer groups ice algae and pelagic production were driven with past sea ice and temperature data. Generally, trends for marine mammal functional groups were more easily t to data, as these time series were created using aerial survey data, and demonstrated gradual changes over time. Data for tting sh groups provided insight as to general trends of abundance, however the model was unable to simulate the extreme increase in capelin and sand- lance populations indicated by their increase in thick-billed murre stomach content, as well as the full decreases in gadiformes and sculpins/zoarcids as suggested by Gaston et al. (2003). This is caused by the high variabil- ity of sh time series as they were compiled from the diets of birds, which demonstrated high annual variability. James Bay beluga abundance was not able to increase to levels as high as survey estimates implied. While migration from the EHB beluga group (de March and Postma, 2003; COSEWIC, 2004b) was included in the model (through biomass accumulation) and improved the t for both EHB and James Bay belugas, the model could not capture the full magnitude of the increase. Conversely, a small decline in EHB belugas was created through hunting mortality and vulnerability settings, but was not fully captured until a negative biomass accumulation component was added to the base Ecopath model, accounting for a loss of this population to the James Bay belugas. Bowhead whales were also unable to increase as rapidly within the model, starting at such a low biomass, and a low P/B, thus a biomass accumulation was added to capture this increase, based on known increases to the population as it recovers from whaling (Higdon and Ferguson, 2010). Rates of biomass accumulation are presented in table 2.2, as annual values (yr1). 32 2.4. Results Model Results Although the tted model cannot fully capture the changes in sh biomass, most notably, increases in capelin and sandlance shifts in sh composition were re
ected. Figure 2.6 identies the changes in sh structure as measured by their percent contribution to the total sh biomass. Since 1970, the model identies declines in Gadiformes and benthic species (sculpins/zoarcids) along with increases in pelagic-based species (capelin and sandlance), as noted in Gaston et al. (2003). Within the model these changes are driven by the decline in sea ice, and subsequent declines in ice algae and benthos, food sources for benthic feeding sh. However the pelagic based sh (capelin and sandlance) fare much better, as pelagic production increases along with SST. This promotes the pelagic production- pelagic detritus- zooplankton- pelagic sh chain allowing increases in capelin and sandlance. Monte Carlo simulations (gure 2.7) indicate that the Ecopath model (for the year 1970) can not support higher marine mammal biomasses than the inputted value for most species groups. Ringed seals have the largest starting biomass of any marine mammal group, and also the highest upper limit or largest biomass, which could be supported by the system, followed by WHB beluga and bowhead whales. Ringed seal biomass had a large uncertainty, as population sizes are not well known. However, the model is able to support a large biomass of these seals. Within the model frame- work, bowheads have the potential to double their biomass while remaining supported by the ecosystem. 33 2.4. Results 1970 1990 2010 0. 01 1 0. 01 4 B (t/k m2 ) Bowhead 1970 1990 2010 0. 00 01 2 0. 00 01 8 FB Polar Bear 1970 1990 20100 .0 00 20 0. 00 04 0 WHB Polar Bear 1970 1990 20100 .0 00 10 0. 00 03 0 B (t/k m2 ) SHB Polar Bear 1970 1990 2010 0. 00 10 0. 00 16 East HB Beluga 1970 1990 2010 0. 00 2 0. 00 4 James Bay Beluga 1970 1990 2010 0. 02 6 0. 03 2 B (t/k m2 ) West HB Beluga 1970 1990 2010 0. 2 0. 6 1. 0 Arctic Cod 1970 1990 2010 0. 1 0. 3 0. 5 Sculpins/Zoarcids 1970 1990 2010 0. 2 0. 8 1. 4 Year B (t/k m2 ) Capelin 1970 1990 2010 0 2 4 6 Year Sandlance Figure 2.5: Biomass trends for functional groups tted to time-series data. Solid lines represent model values, while open circles represent observed data points. Data points for each group were taken from: bowhead (Higdon, 2008; Higdon and Ferguson, 2010), polar bears (Lunn et al., 2002; Stirling and Parkinson, 2006), beluga (Hammill, 2001; DFO, 2002a; Gosselin et al., 2002; COSEWIC, 2004a; Gosselin, 2005; NAMMCO, 2005a; Hammill et al., 2009), sh groups (Gaston et al., 2003). 34 2.4. Results 0 0.1 0.2 0.3 0.4 %  o f T o ta l F is h  B io m a ss Func!onal Group 1970 2009 Figure 2.6: Percent contribution of each sh group to total sh biomass using the Ecopath starting biomass (t  km2), and the Ecosim generated biomass for the 2009 value. 35 Bi o m a s s  ( t  k m − 2 ) R i n g e d  S e a l B e l g u a  W B o w h e a d B e a r d e d  S e a l W a l r u s  N N a r w h a l B e l g u a  E  W a l r u s  S H a r b o u r  S e a l H a r p  s e a l B e l u g a  J a m e s  P o l a r  B e a r  S H K i l l e r  W h a l e 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 A G a d i f o r m e s S a n d l a n c e C a p e l i n A r c t i c  C h a r S c u l p i n s /  Z o a r c i d s O t h e r  M a r i n e  F i s h A t l a n t i c  S a l m o n S e a b i r d s B r a c k i s h  F i s h S h a r k s / R a y s 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 B E c h i n o d e r m s P r i m a r y  P r o d u c t i o n M a c r o Z o o p l . B i v a l v e s M a r i n e  W o r m s C o p e p o d s I c e  A l g a e O t h e r  B e n t h o s M i c r o Z o o p l . E u p h a u s i d s C r u s t a c e a n s O t h e r  M e s o Z o o p l . C e p h a l o p o d s 0 1 2 3 4 5 6 7 8 9 10 11 12 C P o l a r  B e a r  F o x e P o l a r  B e a r  W H B Figure 2.7: Monte Carlo simulation results for Ecopath starting biomass calculated using Eq 2.4. Starting biomass and CV values presented in table 2.2. 36 2.4. Results Compared to the inputted biomasses, the ecosystem is able to support higher sh biomasses than the starting value of 3.42tkm2 for all sh groups within HB. The total zooplankton biomass of 18.91t  km2 falls within the ranges of observed samples, as Harvey et al. (2006) estimated macro and meso-zooplankton from 10-20t  km2 for central HB, while a few samples from Harvey et al. (2001) reached close to 50t  km2 northern HB. Table 2.3: Trophic level of the ecosystem (TLE) and catches (TLC), pre- sented in 10-year increments. Values were calculated annually from 1970- 2009 using Equations 2.6 and 2.7. Year TLE TLC 1970 2.457 3.916 1980 2.512 4.033 1990 2.509 4.037 2000 2.541 4.066 2009 2.512 4.032 Trends for trophic levels (TL) of the ecosystem and catches remain rela- tively stable from 1970-2009 (table 2.3). While catches have a higher trophic level hovering around trophic level 4 (range 3.91-4.07), the ecosystem itself has a much lower trophic level of nearly 2.5 (range 2.45-2.54). This is due to the large proportion of marine mammals being hunted in the system com- pared to small amount of sh at lower trophic levels. The ecosystem TL remains fairly constant even as declines of polar bears, narwhal, and eastern HB beluga are occurring, as increases in killer whales, seals, and western/JB belugas help to keep the ecosystem TL from declining. While eort for sh, seals, and birds increases based on increases in human populations, these contributions to the overall landings and TL of catches are small in relation to marine mammals, therefore allowing the mean TL of catches to remain high. 37 2.4. Results Model Simulations Fitted Model: Past Scenario Starting from the bottom of the food web, shifts caused by forcing func- tions were identied. Figure 2.8 (Past Scenario) identies changes in the ecosystem using the tted model with past sea ice, SST, and hunting data, as presented in % change from the starting 1970 biomass. Declines in ice al- gae and ice detritus of nearly 10% each, and increases in pelagic production (26%), and pelagic detritus (33%). Since both the ice algae and the pelagic production groups were forced, these changes were not surprising. Benthos which rely on energy transported from sinking particles, primarily ice al- gae (Wassmann, 1998; Lavoie et al., 2009), decline under conditions with less ice and ice algae. Zooplankton fare much better, with increases rang- ing from 12% (micro-zooplankton) to 58% (macro-zooplankton). Although zooplankton consume both ice algae and pelagic phytoplankton, biomass for these groups increases from 12% (micro-zooplankton) to 58% (macro- zooplankton), as the increases in pelagic production are high enough to compensate for the loss of ice algae in the diet. 38 DERYH WR WR WR WR EHORZ P o la r B e a r F o x e S H  P o la r B e a r K ill e r W h a le P o la r B e a r W H B H a rp  s e a l R in g e d  S e a l N a rw h a l S h a rk s/ R a y s H a rb o u r S e a l B e lu g a  W B e lu g a  J a m e s B e a rd e d  S e a l S e a b ir d s B e lu g a  E C e p h a lo p o d s W a lr u s S A tl a n ! c S a lm o n B o w h e a d W a lr u s N A rc ! c C h a r G a d if o rm e s B ra ck is h  F is h S cu lp in s/  Z o a rc id s C a p e lin S a n d la n ce O th e r M a ri n e  F is h E u p h a u si d s M a cr o Z o o p la n k to n E ch in o d e rm s C ru st a ce a n s O th e r M e so Z o o p la n k to n M a ri n e  W o rm s B iv a lv e s O th e r B e n th o s C o p e p o d s M ic ro Z o o p la n k to n P ri m a ry  P ro d u c! o n Ic e  A lg a e Ic e  D e tr it u s P e la g ic  D e tr it u s Fi"ed Model # # No Hun!ng # # Constant Climate # # Figure 2.8: Change in biomass from 1970 value under various scenarios: Fitted Model (Fitted model driven using past climate and hunting trends), No Hunting (Model driven with past climate and no hunting), and Constant Climate (Model driven with 1970 climate repeated annually and past hunting). Functional groups are arranged by trophic level, from Foxe Basin polar bears (TL 4.92) to detritus (TL 1) 39 2.4. Results Declines are identied predominantly in sh with benthic or epibenthic diets (Gadiformes: Arctic and Polar cod, Sculpins/Zoarcids: benthic sh, and sharks/rays) due to declines of ice detritus and other benthos. Gad- iformes and sculpins/zoarcids decreased in the diet of thick-billed murres an average of 68% and 57%, respectively, while pelagic-based sh show in- creases, with the largest being capelin and sandlance (gure 2.3). Fitting of time-series data (gure 2.5) from the diet of thick-billed murres appears to be unable to capture the full magnitude of the increase for both capelin and sandlance. Most marine mammal groups were tted to data with the model replicat- ing the trends observed. Polar bears, narwhal, and EHB beluga decline as expected. James Bay beluga, WHB beluga, and bowhead all show increas- ing trends as identied in model tting. However, decreases are identied for southern walrus and bearded seals as hunting mortality impact their relatively small populations throughout the simulation. Northern walrus along with harp, ringed, and harp seals show increases in biomass, as hunt- ing mortality is low relative to the population size, and there are decreases to predators (polar bears). The killer whale functional group biomass was based on sightings data (Higdon, 2007), therefore the biomass was not esti- mated by the model. Other Scenarios: No Hunting and Constant Climate Under the No Hunting scenario, all hunting and shing mortality has been removed, while SST and sea ice were used as environmental drivers as in the past scenario. The biomass of all marine mammal groups increases, with the exception of western and James Bay belugas, which remain the same (gure 2.8). This is due to the relatively low hunting pressure on these specic groups, compared to their biomass. Lower trophic level organisms remain relatively unaected, as climate is still driving the changes to these groups. Gadiformes are the only sh group to decrease further under this scenario indicating the abundance of marine mammals is causing high levels of mortality on this group. 40 2.5. Discussion For the Constant Climate scenario, ice algae and ice detritus show in- creases compared to other scenarios as expected, however the biomass is quite similar to the 1970 value (<5% increase each), while pelagic production and pelagic detritus show slight declines (close to 10% decrease each). With- out the restriction on ice algae, caused by declining sea ice, these changes are propagated through the food web. Increases to benthos are observed as well as declines in zooplankton groups favoring a shift to a more benthic- dominated food web. Fish groups show increases from changes in the lower trophic levels, as well as predator release caused by hunting of marine mam- mals. Biomass for most marine mammal groups remains quite similar to the tted model indicating pressures from hunting are a more important factor in determining biomass than climate change. 2.5 Discussion Fish Biomass and Changes to Fish Composition While past commercial shing endeavors have not been protable (Stewart and Lockhart, 2005), it can be assumed that the region has modest sh biomasses, as Aboriginal communities have harvested sh for thousands of years. This is further corroborated by the ability of the ecosystem to sustain moderate biomasses of sh in the model. Estimates of sh for HB should be considered conservative, as the model only estimates enough sh to satisfy the diets of top predators and shing, with a total sh biomass estimate of 3.42t  km2 for 19701. The contribution of each functional group of sh is based on the diets of predators, and the minimum biomass required of each sh group to satisfy the needs of predators. Compared to other regions at similar latitudes this value is still low, but considering the low productivity of the ecosystem it can be considered a plausible estimate. In comparison, sh biomass estimates for other Ecopath with Ecosim models range from 1This is due to the EE parameter being set to 0.95 for sh species indicating nearly all mortality is caused by shing and food web interactions 41 2.5. Discussion 6.42t  km2 to 49.62t  km2 for other ecosystems at similar latitudes2. As HB is considered oligotrophic (Kuzyk et al., 2011), having a lower cumulative sh biomass than other similar latitude ecosystems is conceivable. Although there is a general lack of data on trends in sh species, the diet of thick billed murres provides insight as to potential changes occurring within the system. Most notably is the shift from a benthic-dominated sys- tem to a pelagic-based ecosystem, demonstrated in the diet of birds as they move from sculpins and zoarcids to pelagic sandlance and capelin (Gaston et al., 2003). Despite the fact that the model ts do not identify the ex- act patterns for the sh functional groups due to dierences in data (gure 2.5)3, changes in composition of sh species are retained (gure 2.6). De- clines in the gadiform group stem from the declines in benthic species as prey items. Although the importance of epibenthic prey has been noted in the literature (Craig et al., 1982), in many regions copepods a predominant dietary staple (Sherwood and Rose, 2005). The model diet re
ects a larger proportion of benthic prey items (see appendix A) facilitating the decline as climate warms. A re-analysis of the tted model identies less severe declines in the gadiform group with increased contribution of copepods and other zooplankton groups to the diet. However, crevasses within sea ice may be important areas for Arctic cod to areas to avoid predators (Gradinger and Blumm, 2004), therefore declines in sea ice would negatively impact Arctic cod. In light of this information, the gadiform group would be ex- pected to decline as demonstrated within the model, albeit possibly with less severity. As the sh data are based on the northern edge of the model region, a greater understanding of sh distribution and diets is important to future modelling. In order to provide more accurate modelling of sh groups, large scale surveys of sh will be necessary for this region. In south- ern HB, sh may be impacted dierently with large freshwater inputs from rivers, causing dierent environmental conditions. 2Fish biomass pertains to the cumulative biomass of all sh groups within the model. Values from other models at similar latitudes include; 1997 Icelandic shelf model (17.1t km2) by Samb (1999), 1980 Bering Sea model (49.62t  km2) by Trites et al. (1999), and a 1964 Ionian Sea model (6.42t  km2) by Piroddi et al. (2010). 3Observed changes of sh groups are inputted as the contribution to bird diets 42 2.5. Discussion Trophic Level of Catches The ecosystem is able to sustain a higher mean TL of catches compared to the mean TL of the ecosystem, throughout the model simulation. Fish catches for 1970 totaled 1.14t for the whole model area, much lower than values reconstructed for the Canadian Arctic by Zeller et al. (2011). Per capita consumption rates used to estimate catches were obtained from the same source (Booth andWatts, 2007), however catch reconstruction included sh fed to dog sled teams, which the model presented does not. It is quite possible in reality that catches are higher than the values used in the model, current shing mortality is low on sh groups indicating they would be able to sustain some increased level of shing. While the constant TL of the ecosystem and catches would imply the ecosystem is stable both in its structure and in the composition of catches, it is uncertain if such a high trophic level of catches can be sustained in- denitely. For example, polar bear populations are shown to be declining within the model, and only under scenarios where harvest is included, in- dicating this level of harvest cannot be sustained. Future reductions of high TL species in the catch composition have the potential to reduce the mean TL of catches over time. This is consistent with ecosystems where sh species have been exploited, a term coined "shing down the food web" by Pauly et al. (1998a). Hudson Bay is one of a small number of ecosystems worldwide where marine mammal catches provide the greatest contribution to landings of all species, re
ected in the high TL of catches. Ultimately, without reductions in catches, populations of marine mammals (such as po- lar bears) have and will likely continue to decline, thus reducing the TL of the ecosystem. Future simulations are necessary to determine if the current hunting and shing pressures are sustainable. Model Simulations Model simulations (No Hunting, Constant Climate) identied expected re- sponses in the model for most functional groups. Removal of hunting pres- sure causes increases in targeted species, with little eects to lower trophic 43 2.5. Discussion levels. The ability of the ecosystem to withstand increases to the starting biomass ranging from 6% to >100% (for narwhal and northern walrus respectively), while not causing declines in sh populations, indicates the ecosystem could support larger biomasses of high trophic level species. Conversely, simulations with constant environmental drivers (sea ice and SST levels) indicate the sensitivity of producers and lower trophic levels to environmental changes. While producers are driven within the model, their responses mimic higher ice cover and lower SST. Although there are a multitude of factors contributing to primary production such as wind, temperature, light, snow cover, ice cover and nutrient input (Legendre et al., 1996; Gregg et al., 2003; Pa