Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Harbour seals, transgenic coho salmon and euphausiids : food dynamics in the Strait of Georgia, British… Li, Lingbo 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_spring_li_lingbo.pdf [ 1.88MB ]
Metadata
JSON: 24-1.0073447.json
JSON-LD: 24-1.0073447-ld.json
RDF/XML (Pretty): 24-1.0073447-rdf.xml
RDF/JSON: 24-1.0073447-rdf.json
Turtle: 24-1.0073447-turtle.txt
N-Triples: 24-1.0073447-rdf-ntriples.txt
Original Record: 24-1.0073447-source.json
Full Text
24-1.0073447-fulltext.txt
Citation
24-1.0073447.ris

Full Text

HARBOUR SEALS, TRANSGENIC COHO SALMON AND EUPHAUSIIDS: FOOD DYNAMICS IN THE STRAIT OF GEORGIA, BRITISH COLUMBIA  by  Lingbo Li B.Sc., Ocean University of China, 1993 M.Sc., University of Bergen, 2006  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES  (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2012  © Lingbo Li, 2012  Abstract  This thesis examines top-down control and a policy of predator culling, the possible impacts of invasive species, and the bottom-up effects of zooplankton on the fish populations in the Strait of Georgia (SoG), British Columbia. In summary, my study highlights indirect interactions and strongly suggests ecosystem-based management in the SoG. For top-down control, I examined interactions between harbour seals and fisheries using Ecopath with Ecosim modelling. Harbour seals feed on herring, hake, and many other commercial fish. Many fish populations in this region have declined in recent decades, while harbour seals increased exponentially after a ban of hunting in 1970, until they reached their carrying capacity in the 1990s. However, model results indicated that a cull of harbour seals may not increase total fisheries catch in the SoG because increased hake would eat more herring. With seals absent, the SoG ecosystem may be dominated by hake. The Ecopath model was then modified to investigate the ecological impacts of invasive species with altered physiology — growth-hormone transgenic (GH) coho. GH coho have the potential to greatly increase the yield of fish farms, but could cause ecological harm should they ever invade natural systems. My model scenarios showed that GH coho may impact the whole ecosystem largely through indirect interactions. Many functional groups were impacted depending on GH coho diet. However, functional groups were more strongly impacted when a bottom-up effect was introduced by changing ocean conditions. Sensitivity analyses showed that the predictions were robust to uncertainty in model parameters although predator-prey vulnerabilities were more sensitive than Ecopath parameters. I assessed the bottom-up effects by analyzing nocturnal zooplankton samples collected from the top 20m of the water column between 1990 and 2007. An abrupt step-like decline occurred in community composition in 1998/1999, especially in euphausiids and copepods. Local environmental factors had low coherence and changes in the SoG zooplankton communities correlated more with large-scale climate forcing than with local factors. The decline in zooplankton communities may be an important factor in the lack of recovery of predatory fish, such as coho salmon and lingcod, even with fishery closures in the 1990s.  ii  Preface A version of Chapter 2 has been published. [Li, L.], Ainsworth, C., and Pitcher T. (2010) Presence of harbour seals (Phoca vitulina) may increase exploitable fish biomass in the Strait of Georgia. Progress in Oceanography 87: 235-241. I developed an Ecopath model for the Strait of Georgia in the year of 2005, conducted all the analyses and wrote the manuscript. Dr. Cameron Ainsworth provided technical assistance in model construction and simulations. Professor Tony Pitcher provided guidance throughout the study. All coauthors provided feedback on the manuscript.  A version of Chapter 3, Potential ecological impacts of escaped transgenic coho salmon in a marine environment, coauthored with Tony Pitcher and Robert Devlin, is currently in review for Marine Ecology Progress Series. I modified the previous Ecopath model of Chapter 2, ran all the scenarios, completed uncertainty analyses, and wrote the manuscript. Professor Tony Pitcher and Dr. Robert Devlin provided advice on model modification and simulations. Professor Tony Pitcher also advised on improving the Monte Carlo function in Ecopath with Ecosim model. All coauthors provided feedback on the manuscript.  A version of Chapter 4 has been submitted with revisions. [Li, L.], Mackas, D., Hunt, B., Schweigert, J., Pakhomov, E., Perry, I.,Galbraith, M., and Pitcher, T. J. accepted. Zooplankton communities in the Strait of Georgia, British Columbia, track large-scale climate forcing over the Pacific Ocean. Progress in Oceanography. pp 53. I conducted all the analyses and wrote the manuscript. Dr. Dave Mackas is the key for data corrections and, together with Professor Tony Pitcher, provided guidance throughout the study. Drs. Brian Hunt and Evgeny Pakhomov assisted in zooplankton grouping and hypothesizing possible mechanisms. Dr. Ian Perry provided constructional advice on the direction of the manuscript. Dr. Jake Schweigert was the principal investigator of sampling program and Moira Galbraith completed most of the lab work. All coauthors provided feedback on the manuscript.  iii  Table of Contents Abstract .................................................................................................................................... ii Preface ..................................................................................................................................... iii Table of Contents ................................................................................................................... iv List of Tables ........................................................................................................................ viii List of Figures ......................................................................................................................... ix Abbreviations ......................................................................................................................... xi Acknowledgements ............................................................................................................... xii Dedication ............................................................................................................................. xiv Chapter 1: Introduction ........................................................................................................ 1 1.1  Research questions .................................................................................................... 1  1.2  The Strait of Georgia ................................................................................................ 4  1.2.1  Physical area ......................................................................................................... 4  1.2.2  Biological system and fisheries ............................................................................ 5  1.3  Methodology ............................................................................................................. 8  1.3.1  Ecopath with Ecosim ............................................................................................ 8  1.3.2  Multi-variate analyses ........................................................................................... 9  1.4  Synopsis of thesis .................................................................................................... 10  Chapter 2: Presence of harbour seals (Phoca vitulina) may increase exploitable fish biomass in the Strait of Georgia .......................................................................................... 13 2.1  Introduction ............................................................................................................. 14  2.2  Methods................................................................................................................... 16  2.2.1  Study area............................................................................................................ 16  iv  2.2.2  The model ........................................................................................................... 16  2.2.3  The Strait of Georgia 2005 Ecopath model ........................................................ 18  2.2.4  Simulations of the ecosystem in 2035 ................................................................ 23  2.3  Results ..................................................................................................................... 24  2.3.1  Scenario 1............................................................................................................ 24  2.3.2  Scenario 2............................................................................................................ 24  2.3.3  Scenario 3............................................................................................................ 26  2.4  Discussion ............................................................................................................... 28  Chapter 3: Investigating ecological impacts of invasive fish with altered physiology using ecosystem modeling..................................................................................................... 31 3.1  Introduction ............................................................................................................. 32  3.2  Materials and methods ............................................................................................ 36  3.2.1  Study area............................................................................................................ 36  3.2.2  Ecopath with Ecosim model parameterization ................................................... 37  3.2.3  Ecosim scenarios ................................................................................................. 40  3.2.4  Model sensitivity................................................................................................. 42  3.3  Results ..................................................................................................................... 44  3.3.1  Model responses to increased salmon ................................................................. 44  3.3.2  Model responses to altered ocean conditions...................................................... 46  3.3.3  Sensitivity analyses ............................................................................................. 49  3.4  Discussion ............................................................................................................... 53  3.4.1  Trophic complex ................................................................................................. 53  3.4.2  Changing ocean conditions ................................................................................. 55  v  3.4.3  Model uncertainty ............................................................................................... 56  Chapter 4: Zooplankton communities in the Strait of Georgia, British Columbia, track large-scale climate forcing over the Pacific Ocean ............................................................ 59 4.1  Introduction ............................................................................................................. 60  4.2  Materials and methods ............................................................................................ 62  4.2.1  Sampling and laboratory processing ................................................................... 62  4.2.2  Large-scale climate indices ................................................................................. 64  4.2.3  Local physical factors ......................................................................................... 66  4.2.4  Data analysis ....................................................................................................... 67  4.3  Results ..................................................................................................................... 71  4.3.1  Large changes in September zooplankton community ....................................... 71  4.3.2  Large changes in June zooplankton communities .............................................. 79  4.3.3  Physical indicators associated with changes in September zooplankton  communities .................................................................................................................... 82 4.3.4 4.4  Correlations with euphausiid biomass ................................................................ 89 Discussion ............................................................................................................... 92  4.4.1  Synchrony of biological responses in Northeast Pacific..................................... 92  4.4.2  Large-scale climate forcing versus local factors ................................................. 93  4.4.3  Possible mechanism for the influence of the previous spring SOIx ................... 95  4.4.4  Implications of this study .................................................................................... 96  Chapter 5: Conclusion ......................................................................................................... 98 5.1  Overview ................................................................................................................. 98  5.2  Ecosystem functioning .......................................................................................... 101  vi  5.3  Ecosystem-based management ............................................................................. 105  5.4  Future studies ........................................................................................................ 106  References ............................................................................................................................ 109 Appendices ........................................................................................................................... 140 Appendix A Basic parameters of the Ecopath with Ecosim model .................................. 140 Appendix B Diet matrix of the Strait of Georgia 2005 Ecopath model............................ 142 Appendix C List of Abbreviations in Appendix A and B ................................................. 146 Appendix D Appendix references ..................................................................................... 147  vii  List of Tables  Table 3.1 Summary of all scenarios. ....................................................................................... 41  Table 4.1 Summary of PCA results of 24 large-scale climate indices.................................... 81 Table 4.2 Summary of PCA results of 34 local measurements. ............................................. 83 Table 4.3 Summary of PCA results of selected local measurements. .................................... 87 Table 4.4 The best environmental indicators of changes in zooplankton communities ......... 88 Table 4.5 Correlation between the environmental indicators of Table 4. . ............................. 88  viii  List of Figures  Figure 1.1 Map of the Strait of Georgia, British Columbia ...................................................... 5  Figure 2.1 Commercial and recreational landings in the Strait of Georgia in 2005. .............. 20 Figure 2.2 Harbour seals (top) and Pacific hake (bottom) diet composition. ......................... 21 Figure 2.3 Annual herring mortality caused by predation and the fishery in 2005. ............... 22 Figure 2.4 Predicted seal and herring biomass with changing herring fishing mortality ....... 25 Figure 2.5 Predicted biomass of the main functional groups with seal absent. ...................... 26 Figure 2.6 Predicted herring biomass relative to 2005 with changing vulnerability of juvenile herring when seals are removed. ............................................................................ 27  Figure 3.1 Four possible diets for GH coho salmon. .............................................................. 38 Figure 3.2 Predicted biomass in the thirtieth year relative to the Ecopath baseline weh the biomass of GH coho was forced to remain10 times for 30 years. ......................... 45 Figure 3.3 Responses of biomass of some functional groups when the biomass of GH coho was forced to be as different times for 30 years..................................................... 46 Figure 3.4 Predicted biomass of some functional groups when the biomass of euphausiid was forced to decrease by 15% of the baseline for 30 years ......................................... 48 Figure 3.5 Relative biomass of some functional groups when euphausiid biomass was kept constant at different levels for thirty years ............................................................. 49 Figure 3.6 Deviation of Monte Carlo runs from the Ecosim baseline. ................................... 50 Figure 3.7 Deviation of Ecosim output from the Ecosim baseline. ........................................ 50  ix  Figure 3.8 The harbour seal-hake-herring trophic triangle in the Strait of Georgia. .............. 55  Figure 4.1 Zooplankton sampling and local factor stations in the Strait of Georgia .............. 64 Figure 4.2 Average proportion of total zooplankton dry biomass in September .................... 72 Figure 4.3 Interannual variation in the September biomass contribution of the four dominant zooplankton groups in the SoG. ............................................................................. 73 Figure 4.4 Cluster analysis and Non-metric Multi-Dimensional Scaling ordination of September zooplankton communities .................................................................... 75 Figure 4.5 The principal component score from principal component analyses of zooplankton communities. .......................................................................................................... 76 Figure 4.6 Average dry biomass in September for the “1990s” and “2000s”. ....................... 76 Figure 4.7 Anomalies for the total biomass in September ...................................................... 78 Figure 4.8 September biomass anomalies of each zooplanton group. .................................... 78 Figure 4.9 Cluster analysis of the zooplankton community in June and average biomass of each zooplankton group in the “1990s” and the “2000s”. ..................................... 80 Figure 4.10 The principal component score of the environmental variables. ......................... 85 Figure 4.11 Normalized 1-year lagged SOIxspring, peak flow day of the Fraser River, and temperature, and their relationship with biomass of euphausiids. ......................... 90  Figure 5.1 Predicted biomass of the main functional groups with a low seal hunting mortality of 2% per year. ....................................................................................................... 99  x  Abbreviations  ANOSIM Analysis of Similarity B  Ecopath biomass parameter  BC  British Columbia  DFO  Department of Fisheries and Oceans, Canada  EE  Ecotrophic Efficiency  EET  Ecotrophic Efficiency Tolerance  EwE  Ecopath with Ecosim  GH  Growth-harmone transgenic  NMDS  Non-metric Multi-Dimensional Scaling  NOI  Extratropical-based Northern Oscillation Index  NPGO  North Pacific Gyre Oscillation  NPI  North Pacific Index  PC  Principal Component  PCA  Principal Component Analysis  PFD  Peak Flow Day of the Fraser River  PDO  Pacific Decadal Oscillation  P/B  Production rate  Q/B  Consumption rate  SIMPER Similarity Percentage SoG  Strait of Georgia  SOI  Southern Oscillation Index  SOIx  Extratropical-based Southern Oscillation Index  SSH  Sea Surface Height  SSS  Sea Surface Salinity  SST  Sea Surface Temperature  STD  Spring Transition Day  TS  Temperature of Surface water at Nanoose station  TI  Temperature of Intermediate water at Nanoose station  TD  Temperature of Deep water at Nanoose station  xi  Acknowledgements  This study could not have been done without the support of many people. First, I offer my enduring gratitude to my supervisor, Dr. Tony Pitcher, for patience, advice and assistance throughout the study. Many thanks to Dr. Ian Perry for many opportunities you gave me, and your advice throughout my study, and to Dr. Evgeny Pakhomov for your consistent encouragement and suggestions. I acknowledge my other committee members Drs. Andrew Trites and Bob Shadwick, who also provided helpful comments and suggestions on my proposal and drafts of this thesis. Many thanks to my unofficial committee members Drs. Cameron Ainsworth and Dave Mackas for advice and support. I owe particular thanks to the scientists I have also enjoyed working closed with, Drs. Bob Devlin and Brian Hunt. Many thanks also to Drs. Villy Christensen, Carl Walters, William Cheung, and Susan Allen for insightful discussions and advice, and to many scientists who participate the Strait of Georgia Ecosystem Research Initiative, particularly, Drs. Diane Masson, Angelica Pena, Jim Irvine, and Caihong Fu. I thank Robert Devlin for financial support and Jeroen Steenbeek and Joe Buszowski for technical support on improving Monte Carle function in Ecopath with Ecosim model. I am grateful to Dr. Chih-hao Hsieh for being my host supervisor during the NSERC Summer Program in Taiwan in 2010. I also thank Drs. Carie Hoover, Divya Varkey, and Mr Rajeev Kumar for technical assistance on ecosystem modelling, and Desiree Tommasi and Leigh Gurney for answering my questions on zooplankton. Thanks to Vicky Lam, Edward Gregr, and Zoraida Alojado for assistance in GIS. I would also like to sincerely thank many friends for English support, especially Dr. Nigel Haggan and Ms. Shannon Obradovich, who provided editorial comments on some thesis chapters, Dr. Mimi Lam for edits and comments  xii  on my funding applications, Ms. Janice Doyle for ongoing help with English pronunciation, and Mr. Shawn Booth for answering my many language questions. Many friends have also provided invaluable support, particularly, Drs. Yan Jiao, Yajie Liu, and Deng Palomares.  I gratefully acknowledge financial support from the National Science and Engineering Research Council (NSERC) three-year Post Graduate Scholarship, the University of British Columbia (UBC) Four-Year Doctoral Scholarship, UBC Biodiversity Research Integrative Training & Education (BRITE) Fifth year PhD Scholarship, and UBC Faculty of Science Graduate Award. I also acknowledge travel funding from NSERC, UBC Zoology, and the North Pacific Marine Science Organization (PICES).  My final and greatest debt is owed to my family. I offer thanks and love to my husband Henry Han, who always encourages me and supports me, emotionally and financially, and to our son Yang, who is always interested in my research and understands when I am busy.  xiii  Dedication  To Yang Han.  xiv  Chapter 1: Introduction  1.1  Research questions  Throughout the world, traditional target species of fish have declined to a fraction of their abundance prior to the industrialization of fisheries; these declines are well-documented (Ward and Myers, 2005; Pauly et al., 2003; e.g. Jackson et al., 2001). Besides an obvious loss of valuable commercial populations, the ecosystem has been altered in more subtle ways. The high value, demersal resources have been replaced by rapid-turnover, low-value pelagic species (Pitcher, 2001). Furthermore, a simplification of the food web and loss of biodiversity has been observed (Jackson, 2008; Worm et al., 2006) and this may have reduced resiliency to perturbation and climate change (Perry et al., 2010b; Stachowicz et al., 2007; Folke et al., 2004).  My study focuses on the Strait of Georgia (SoG) marine ecosystem, an exploited ecosystem (Pauly et al., 1998b) that is like many other ecosystems around the world. Moreover, it is a coastal area surrounded by several large cities with increasing human populations and thus, increasing human impacts (Perry, 2008). In addition to fishing, humans also impact the SoG ecosystem through pollution, marine traffic, habitat destruction, aquaculture, and invasive species (Johannessen and Macdonald, 2009; Perry, 2008) , each of which is associated with economic benefits and environmental costs. Take fish farming as an example. The salmon farming industry was introduced in the 1970s. Now in British Columbia (BC), there are 130 sites covering over 4,000 hectares, primarily in the north SoG and north Vancouver Island.  1  BC is now the fourth largest producer of cultured salmon in the world after Norway, Chile and the United Kingdom (more information at the website http://www.env.gov.bc.ca/omfd/fishstats/aqua/salmon.html). However, there are serious environmental concerns, such as escapes (Naylor et al., 2005; Carss, 1990), transmission of parasites and diseases (Krkošek et al., 2005; Jones et al., 1999), and changes in the sediments near farms (La Rosa et al., 2001; Herwig et al., 1997). Generally, human impacts, like fish farming, are increasingly dominating the once-productive SoG (Perry, 2008).  In addition to human impacts, the SoG and contiguous Puget Sound has also experienced significant environmental changes. Since the arrival of European settlers in the early 1800s, there have been major changes in the shorelines in the large urban areas, such as Vancouver, Bellingham, Seattle, and Victoria, and changes in the estuaries, which have led to differences in sedimentation rates and circulation (Taylor and Horner, 1994). The environmental changes have been accelerating in the last several decades. The peak water flow of the Fraser River has shifted earlier in the year and decreased in volume (Morrison et al., 2002). Average seasurface temperature has increased 1°C since 1970 in the SoG (Masson and Cummins, 2007). More importantly, warming in the SoG subsurface water appears to be faster than in the offshore (Masson and Cummins, 2007), suggesting more striking changes in the SoG organisms.  The SoG is the focus of many resource management and scientific issues within the Pacific Region of Fisheries & Oceans Canada (DFO). My PhD thesis (research) is integrated with DFO’s SoG Ecosystem Research Initiative http://www.pac.dfo-  2  mpo.gc.ca/science/oceans/detroit-Georgia-strait/index-eng.htm. My research is helpful in understanding how this ecosystem works. It focuses on the fish populations and examines three major impacts: top-down control, invasive species, and bottom-up control. Specifically, I asked three research questions:  1. As there are abundant harbour seals and declined fish populations, will a culling of harbour seals (Phoca vitulina) increase the total fish catch in the Strait of Georgia?  2. The growth-hormone (GH) transgenic coho have the potential to greatly increase the yield of fish farms. However, there is major concern about the potential ecological harm transgenic fish may cause should they ever invade natural systems. What would be the ecological impacts on the Strait of Georgia ecosystem should they escape from facilities?  3. Some fish populations, such as coho salmon and lingcod, have not been recovered since fisheries were closed and harbour seals were stable in 1990s. Were there any large changes in zooplankton communities in the last few decades? If so, changes in the SoG zooplankton were correlated more with large-scale climate forcing or local conditions?  3  1.2  1.2.1  The Strait of Georgia  Physical area  The SoG (Figure 1.1) is a large semi-enclosed estuary between Vancouver Island and the mainland coast of BC, Canada (Pawlowicz et al., 2007; LeBlond, 1983). The Fraser River is the largest river entering the SoG, contributing 80% of the fresh water input (LeBlond et al., 1991). It is connected with the Pacific Ocean through southern (Juan De Fuca Strait) and northern (Johnstone Strait) passages, with most of the water mass exchange with the Pacific Ocean occurring through the southern passage (Pawlowicz et al., 2007). The oceanic water is the major source of the SoG deep water (Pawlowicz et al., 2007; Masson, 2002). The outer coast upwelling contributes to salinity, temperature and nutrient variability in the SoG (Masson and Cummins, 2007; Masson, 2002; Mackas and Harrison, 1997).  4  Figure 1.1 Map of the Strait of Georgia, British Columbia  1.2.2  Biological system and fisheries  Harbour seals are important marine mammals in the SoG and have been very abundant since the1990s (DFO, 2010). Harbour seals were harvested in the 1950s and 1960s, but a harvest ban was implemented in 1970 after their numbers declined drastically (DFO, 2010). Their population has since rebounded and is thought to have reached and remained at carrying 5  capacity in the SoG since the 1990s (DFO, 2010). Abundant harbour seals raise an important question on a seal culling policy to save the fish populations, which leads to Chapter 2.  Many species in the SoG have declined greatly in recent decades (Johannessen and Macdonald, 2009; Pauly et al., 1998b). Pacific herring (Clupea pallasii), a forage fish, now dominates the fisheries in the SoG (Li et al., 2010). Salmon are the main subject of the traditional native fisheries; they have been commercially harvested for over 100 years in the SoG. However, the main coho (Oncorhynchus kisutch) troll fishery has been closed since 1995, and the biomass has remained low (DFO, 2008; DFO, 2002). Additionally, chinook (Oncorhynchus tshawytscha) and sockeye (Oncorhynchus nerka) also greatly declined (Irvine et al., 2005). By contrast, the current harvest quantity and value of cultured salmon exceeds that of wild salmon populations. As the GH transgenic coho salmon potentially benefit the industry but with high risks of environmental problems, I examined the possible trophic interactions if they ever escape from facilities (Chapter 3).  Lingcod (Ophiodon elongatus) is a top predator in rocky reef areas along the west coast of North America (Beaudreau and Essington, 2009). The lingcod fishery in the SoG started in the 1860s, peaked in the 1940s and then decreased progressively since the 1950s (Pauly et al. 1998). By the 1980s, both commercial and recreational catches had declined to low levels. A commercial fishery closure was implemented in 1990, followed by a recreational fishery closure implemented in 2002 (DFO, 2005). However, there has been no indication of any rebuilding of lingcod in the SoG (DFO, 2005). Similarly, catches of inshore rockfish have been heavily curtailed since the 1990s, and Rockfish Conservation Areas have been  6  established throughout the British Columbia coast (DFO, 2006). No recovery of depleted fish populations indicates an important question on changes in their prey zooplankton (Chapter 4) as their predators harbour seals have been stable during the time period.  Pacific herring is the most abundant forage fish, and the herring fishery is one of the largest fisheries in the SoG. All herring spawning within the SoG are assumed to belong to a single stock that migrates into the SoG in the late fall and leaves after spawning in March (DFO, 2009). Resident or non-migratory herring appear to be present in a few areas in the SoG and are also considered part of the SoG herring stock (DFO, 2009). From the mid-1940s until the late 1960s, the main herring fishery was a “reduction” fishery that harvested and processed herring into low value products such as fishmeal and oil. After the herring fishery was closed in 1967, herring stocks recovered in the 1970s and a roe fishery has replaced the reduction fishery as the main herring fishery (DFO, 2009). Herring declined again in the mid-1980s, rebuilt to near historical high abundance in 2003 and then declined subsequently (DFO, 2009).  Pacific hake (Merluccius productus) has the highest biomass in the study area, and the SoG resident stock is different from the stock on the west coast of BC (McFarlane and Beamish, 1985). The SoG hake stock was not discovered until 1974 (McFarlane and Beamish, 1985). The SoG hake fishery started at the end of the 1970s and catches peaked in 1995 (King and McFarlane, 2006). Although hake is not an important fishery in the SoG, hake is an important ecological role in the SoG as an important prey of harbour seals and groundfish.  7  1.3  1.3.1  Methodology  Ecopath with Ecosim  I applied Ecopath with Ecosim (EwE) ecosystem modeling to model the interactions between harbour seals and fisheries in Chapter 2 and the trophic impacts of growth-hormone transgenic (GH) coho in Chapter 3. This commonly used food web model was first developed by Polovina (1984) and then advanced by Christensen and Pauly (1993; 1992), Walters et al. (1997), and Christensen and Walters (2004). EwE is used in more than 150 countries, supports more than 300 publications, and has been named one of the top 10 breakthroughs by National Oceanic and Atmospheric Administration, United States (NOAA, 2006). It captures trophic interactions and fishing impacts (see reviews by Plagányi 2007; Christensen and Walters, 2005). In addition, Ecosim offers a comprehensive overview of predator-prey interactions (Jarre-Teichmann, 1998), which allows the user to investigate unintuitive and indirect trophic relationships. EwE is a very useful modeling approach for understanding the dynamics of the upper levels of the food chain (Christensen and Walters, 2004), which is the focus of my two chapters. EwE models have been successfully used to investigate interactions between fisheries and marine mammals (Gerber et al., 2009) and between non-indigenous species and salmon in a reservoir (Harvey and Kareiva, 2005). Thus, EwE should be adequate for the research questions of the two chapters that depend upon food web modelling.  8  1.3.2  Multi-variate analyses  I utilized multi-variate community analyses to examine the large changes in the SoG zooplankton communities and the link with environment in Chapter 4. Due to large correlations among the zooplankton communities, I chose analytical methods based on multivariate analyses rather than focusing on one or two species. A popular software, PRIMER, was extensively used to implement these methods. Before multi-variate analyses, zooplankton group biomass was square-root transformed to balance the weight of dominating groups and rare groups. All the environmental data were normalized to a mean of zero and standard deviation of one to assure each factor has the same weight. To avoid bias in each method, Cluster analysis (Clarke, 1993), Non-metric Multi-Dimensional Scaling (NMDS) (Clarke, 1993) and Principal Component Analysis (PCA) (Chatfield and Collins, 1980) were used to examine the year-to-year change in the biomass of zooplankton communities. The first two methods are based on Bray-Curtis similarity while PCA is based on dissimilarity of Euclidean distance. Cluster analysis was completed using group averages of year similarities. By contrast, the NMDS is a non-metric ordination ranking the similarities between each pair of years, where a shorter distance between the two years means higher similarity. For the physical-biological relationships, linear and nonlinear correlations were used. Linear correlation was conducted between PC scores of zooplankton communities and PC scores of environmental factors. A novel method, Bio-Env procedure (Clarke and Ainsworth, 1993) was applied to investigate the non-linear correlation. I identified the best subsets of physical variables through the maximized Spearman rank correlation coefficient r between the zooplankton similarity matrix and the physical variable similarity matrix.  9  1.4  Synopsis of thesis  Chapter 1 summarizes the goals of my thesis and my research questions. It provides basic background on the SoG study area and describes two major methods I used in my thesis: the EwE ecosystem modeling and multi-variate analyses.  Chapter 2 examines the interactions between harbour seals and fisheries in the SoG using a SoG Ecopath model. It is a new model for SoG in 2005 that is based on a 1950 SoG Ecopath model (Preikshot, 2007) and a northern BC Ecopath model (Ainsworth et al., 2008). I assembled new data and updated parameters, especially some diets for important functional groups such as harbour seal and hake. Three scenarios were run using this new model: 1) I altered annual herring fishing mortality to see how the harbour seal population responds, 2) I forced the harbour seal population to near zero to see how the fish population responds, 3) I reran #2 scenario using different vulnerabilities for juvenile herring to see how the model predictions are influenced. I compare my results to structural changes in other marine ecosystems and with the perception that culling marine mammals will increase fisheries.  Chapter 3 uses ecosystem modeling to examine the long-term impacts on the whole ecosystem should GH coho escape from facilities. I first added one special invasive fish group GH coho, which are not allowed to be used in fish farming, to my Ecopath model. As GH coho are still kept in the labs, I defined four theoretically possible diets and also estimated other model parameters based on experiments and wild salmon population. Secondly, I ran two major types scenarios to examine the changes in biomass of major  10  functional groups: 1) Force GH coho biomass at different levels from 5 - 25 times the biomass baseline of wild salmon for thirty years and 2) Force euphausiids and herring, the major prey for GH coho and many fish, to increase or decline at different levels for thirty years. Finally, considering the huge uncertainty in GH coho parameters, I completed sensitivity analyses of Monte Carlo runs of Ecopath parameters and vulnerabilities in Ecosim. This is the first study to apply EwE’s new Monte Carlo function. A wide range of possible values, uniformly distributed, was included in the Monte Carlo runs of Ecopath parameters, e.g. consumption rate 1-3 times of the wild salmon. In order to investigate predator-prey vulnerabilities to Ecosim performance, prey vulnerabilities to GH coho and GH coho vulnerabilities to their predators were respectively reset from low to high to examine the bottom-up and top-down controls related to GH coho.  Chapter 4 analyzes large changes in SoG zooplankton communities from 1990 -2007 and examine if the changes were more tied to large-scale climate indices or local factors. First, based on night zooplankton samples from the surface 20m, I aggregated zooplankton communities into ten major taxa and investigated interannual changes for September and June separately. Second, I utilized a combination of cluster analysis, NMDS and PCA to investigate year-to-year change in zooplankton communities. Third, I assembled 24 time series of seasonal large-scale climate indices and 34 time series of local environment factors, each with up-to-two year lags. Fourth, I used linear and non-linear methods to examine the link between changes in zooplankton and environmental factors. The linear method was carried out for the large-scale climate indices or local factors, respectively. Principal Component (PC) scores were used to correlate with PC scores of zooplankton communities.  11  For the nonlinear method, I applied the Bio-Env package in PRIMER to identify the combinations of environmental indicators that “best” explain zooplankton community changes. An up-to-two year lag was considered in the physical-biological analyses.  Chapter 5 provides a summary and discussion of my major findings on top-down control, invasive species effects, and bottom-up control. I addressed the ecosystem functioning and make recommendations for the management of the fisheries. Future research directions were presented and comments were made on EwE ecosystem modeling.  12  Chapter 2: Presence of harbour seals (Phoca vitulina) may increase exploitable fish biomass in the Strait of Georgia  I tested the degree to which harbour seal populations compete with fisheries for commercially harvested species, and to what degree seal populations depend on exploited species as prey. In the Strait of Georgia (SoG), harbour seals mainly feed on species targeted by fisheries (Pacific herring and Pacific hake), while herring is also a main prey of hake. Using an Ecopath model constructed based on 2005 conditions, I ran three scenarios: altering herring fishing mortality, removing seal populations and sensitivity analyses of the herring vulnerability parameter. My results showed that with more herring available, the seal population would increase greatly, but with less herring available, the seal populations in the SoG would decrease gradually. My model suggested that the total biomass of commercial fish populations in the SoG may decrease substantially with seals absent. A cull of harbour seals may not increase total fisheries catch in the SoG. Herring benefit from seal predation on herring’s largest predator, hake, so that herring may decline when seals are removed. However, this result was highly dependent on model parameterization. When juvenile herring were considered less vulnerable to hake predation (i.e., when there were many refuges in which to hide), the herring population was less negatively impacted by seal removals. This indicates that survival during this crucial life-stage is important to herring biomass. The model also suggests that, with seals removed, the ecosystem would be dominated by hake.  13  2.1  Introduction  There has long been concern about the interactions between fisheries and marine mammals. Many studies have reported that fisheries pose a serious threat to populations of marine mammals due to bycatch effects and gear entanglements (Read, 2008; Kraus et al., 2005). For example, this has been suspected to be the main reason behind the recent extinction of the baiji (Lipotes vexillifer), a freshwater dolphin endemic to the Yangtze river, China (Turvey et al., 2007). On the other hand, many commercial fish stocks have been seriously depleted worldwide (Worm et al., 2006; Pauly et al., 1998a), and there have been calls to reduce marine mammal populations due to the perception that they compete with fisheries (Gerber et al., 2009; Swartz and Pauly, 2008; Yodzis, 2001). Nevertheless, Gerber et al. (2009) found negligible effects of removing whales on commercial fish biomass relative to the effect of a fishing moratorium in the Northwest African and Caribbean ecosystems. It is noticeable that their study found little trophic overlap between fisheries and whales which was similarly reported by Trites et al., (1997) in the Pacific Ocean where whale prey species compete with commercially targeted fish for plankton and prey at low trophic levels. Swartz and Pauly (2008) reported that only about 1% of all food taken by marine mammals was consumed in areas where they competed with fisheries concluding that there is no evidence to support the suggestion that food competition between marine mammals and fisheries is a global problem. However, pinnipeds can strongly interact with fisheries in local areas and it is still unresolved if culling marine mammals can increase targeted fish stocks when marine mammals feed on commercial target species.  14  I examined the interactions between harbour seals and fisheries in the Strait of Georgia (SoG) marine ecosystem using ecosystem modeling. The SoG is a good ecosystem to investigate this question for two reasons: 1) Harbour seals mainly feed on Pacific herring, Pacific hake and other fishery target species in this ecosystem. Therefore, there is direct competition between seals and fisheries; 2) Harbour seals, which are the dominant top predators, have been increasing in abundance while fisheries have been declining in the SoG. I applied Ecopath with Ecosim (EwE) model because this food web ecosystem model captures trophic interactions and fishing impacts (Plaganyi, 2007; Christensen and Walters, 2005). In addition, Ecosim offers a comprehensive overview of predator-prey interactions (JarreTeichmann, 1998), which allows the user to investigate unintuitive and indirect trophic relationships. EwE models have been successfully used to investigate trophic roles of marine mammals in the Northern Gulf of St. Lawrence (Morissette et al., 2006) and interactions between fisheries and marine mammals (Gerber et al., 2009). In this study, I tested the response of seals to removals of herring by fisheries, and the response of herring and other commercial fish populations to the absence of seals using a present SoG Ecopath model. As harbour seals are removed, the response of the prey populations will depend largely on the form of the feeding functional relationship used in Ecosim. I therefore take herring as an example to perform a sensitivity analysis on the key parameter in this relationship, the vulnerability of juvenile herring to seals.  15  2.2  2.2.1  Methods  Study area  There has been a strong contrast between the abundant harbour seals and declining fisheries in the last few decades. Harbour seal population has rebounded since a ban in 1970 and may have reached carrying capacity in the SoG from the late 1990s (DFO, 2010; Olesiuk, 1999). However, many fisheries in the SoG declined greatly in the last several decades. The commercial coho salmon fishery has been closed since 1995, and a commercial lingcod fishery closure was implemented in 1990. However, there has been no indication of recovering lingcod or coho populations in the SoG. Herring is the most abundant forage fish, and the herring fishery is one of the largest fisheries in the SoG. Hake is the most abundant resident fish in the SoG and important both as predator and prey for a variety of other animals (McFarlane and Beamish, 1985). Pacific herring and Pacific hake now play an important role in the fisheries and also in the ecosystem of the SoG.  2.2.2  The model  A popular food web ecosystem model, EwE, was used in this study. The parameterization of an Ecopath model is based on satisfying two ‘master’ equations (Christensen and Pauly, 1992). The first equation describes how the production terms for each group are divided:  n  Bi  P B i  Yi   B j  Q B  j  DCij  E i  BAi  Bi P B i  1  EE i  j 1  (1) 16  In this equation, Bi and Bj are biomasses of prey (i) and predator (j); P/Bi is the production/biomass ratio; Yi is the total fishery catch rate of group (i); Q/Bj is the consumption/biomass ratio; DCij is the fraction of prey (i) in the average diet of predator (j); Ei is the net migration rate (emigration – immigration); BAi is the biomass accumulation rate for group (i); and EEi is the ecotrophic efficiency, a catch-all term that describes the proportion of mortality that is explicitly accounted for in the model such as predation and fishing mortality. It is the aim of the Ecopath model to describe all mortality factors; hence the ‘other mortality’ should only include generally minor factors such as mortality due to old age, diseases, etc.  The second ‘master’ equation is based on the principle of conservation of matter within a group:  Consumption = production + respiration + unassimilated food  (2)  where respiration is estimated based on consumption, production, and unassimilated food.  Ecosim (Walters et al., 1997) provides a dynamic simulation capability at the ecosystem level, with key initial parameters inherited from the base Ecopath model. Biomass dynamics are described as follows:  n n dBi  gi  f ( B j , Bi )   f ( Bi , B j )  I i  ( M i  Fi  ei )  Bi dt j 1 j 1  (3)  17  where dBi/dt represents biomass growth rate of group (i) during the interval dt; gi represents the net growth efficiency (production/consumption ratio); Ii is the immigration rate; Mi and Fi are natural and fishing mortality rates of group (i), respectively; ei is the emigration rate; and f(Bi, Bj) is a function used to predict consumption rates of predator (j) on prey (i) according to the assumptions of foraging arena theory (Ahrens et al., 2012; Walters and Martell, 2004; Walters and Korman, 1999; Walters and Juanes, 1993). This predator-prey functional response is dependent on Ecosim’s vulnerability parameter, which defines the maximum allowable increase in predation mortality under conditions of high predator biomass (Ahrens et al., 2012; Christensen and Walters, 2004) (i.e., this parameter defines the current position on the feeding functional response in relation to the predation mortality asymptote). Depending on the value used for this parameter, it defines a predator controlled or donor driven feeding relationship.  2.2.3  The Strait of Georgia 2005 Ecopath model  I built an Ecopath model for the SoG based on conditions in 2005. There have been several previous SoG models (Preikshot, 2007; Martell et al., 2002) for different research questions. I chose the year 2005 because more data are available for the ecosystem for this period and the seal population had been stabilized for about one decade by that year (DFO, 2010). My Ecopath model has 39 functional groups with multi-stanza groups for its most important and abundant prey: herring, hake and pollock. Production (P/B) and consumption (Q/B) parameters were mainly based on a SoG 1950 model by Preikshot (2007) and a Northern British Columbia 2000 model by Ainsworth (2006). Biomass data are compiled from  18  Department of Fisheries and Oceans Canada (DFO) Pacific Scientific Advice Review Committee stock status reports available at http://www.pac.dfompo.gc.ca/SCI/psarc/Default_e.htm and Canadian Science Advisory Secretariat stock status reports available at: http://www.meds-sdmm.dfo-mpo.gc.ca/csas/applications/Publications/ publicationIndex_e.asp. When stock assessment reports were unavailable, biomass was estimated for dogfish and groundfish taxa (Gordon McFarlane, Pacific Biological Station, Nanaimo, BC Canada, Pers. Comm.) and salmon (James Irvine, Pacific Biological Station, Nanaimo, BC Canada, Pers. Comm.) based on expert opinions. Commercial and recreational landings data are from DFO with herring dominating the fishery in the SoG (Figure 2.1). A full list of taxa in the model and their parameter values can be found in the Appendix A and B.  Detailed diet data were available for this study (Diet matrix in Appendix B). Harbour seals are opportunistic predators and their diets vary regionally and seasonally depending on the local availability of prey (Lance and Jeffries, 2007). Their diet in the SoG is dominated by Pacific hake and herring, which comprised 75% of the overall diet throughout the year based on about 3,000 scat samples collected at fifty-eight sites in the SoG in all months (Olesiuk et al., 1990). The seal diet in the balanced Ecopath model (Figure 2.2) is based on Olesiuk et al. (1990). Olesiuk et al. (1990) reported a strong seasonal and regional pattern in harbour seal diets in which hake dominated the diet in summer and herring dominated in winter, in all non-estuaries. The seal diet appears consistent with the timing of when spawning herring return from the offshore (Hay et al., 2001) and when the herring roe fishery occurs (DFO, 2009). It is also consistent with regards to when the post-spawning school of hake is thought  19  20 15 10 5  Krill  Shrimps  Clams  Crabs  Soles  Pollock  Dogfish  Hake  Salmon  Herring  0  Landings in 2005 ('000 tonns)  Figure 2.1 Commercial and recreational landings in the Strait of Georgia in 2005.  to be present in shallow waters (McFarlane and Beamish, 1985), which occurs after the peak of the hake fishery (King and McFarlane, 2006). I therefore expect some overlap in the size of herring and hake preyed upon by seals and targeted by fisheries. Notably, both herring and hake have been shown to be present in the diets of seals year-round (Oleskiuk et al., 1990). I therefore assume that a part of the consumption occurring during non-peak times of the year is composed of juvenile hake and juvenile herring. This is consistent with the observation that both juvenile hake and juvenile herring are present in the shallow waters of the SoG (McFarlane and Beamish, 1985; Hourston, 1956). Additionally, what hake feed on is also important in this study because of the high abundance of hake and its high predation on 20  Figure 2.2 Harbour seals (top) and Pacific hake (bottom) diet composition.  21  0.30 0.25 0.20 0.15 0.10  Others  Res. Salmon  sea lions  Lingcod  Sea birds  Rockfish  Seals  Dogfish  Fishery  Hake  0.05 0.00  Annual Herring Mortality  Figure 2.3 Annual herring mortality caused by predation and the fishery in 2005.  herring. Herring, krill and shrimp constitute the main prey of hake (Buckley and Livingston, 1997; McFarlane and Beamish, 1985) (Figure 2.2). Comparing the fishery and predation among groups, the largest source of mortality for herring is predation from hake followed by the fishery based on the balanced Ecopath model (Figure 2.3). The model also shows the predation from dogfish is larger than that of seals due to the high abundance of dogfish.  Ecosim has a limited ability to represent migratory species (Martell, 2004) but one facility that does exist is to assume that some fraction of diet originates outside of the model domain.  22  I used this 'diet import' term (Christensen and Walters, 2004) for the migratory groups such as orcas, sea lions, transient salmon and herring. Although this imparts realism, the diet import is not subject to the same systemic fluctuations in productivity that affect local diet components. Thus, highly migratory groups can show hyper-stability during production minima. Detailed diet composition data are in the Appendix B.  2.2.4  Simulations of the ecosystem in 2035  Vulnerabilities, normally determined by fitting to time series data, are a key parameter in Ecosim describing the flux of prey from safe refuges to feeding areas, where they are subject to predation (Walters et al., 1997). In a recent model with no time series data available to fit the model, Cheung et al. (2002) first reported that scaling vulnerabilities proportional to predator trophic levels was more realistic than assuming a global vulnerability parameter, the default setting in Ecosim, in which every interaction is parameterized the same (i.e., as a predator-controlled interaction, a donor-controlled interaction or a mixed trophic control interaction). Chen et al. (2008) used this approach to evaluate different fisheries management strategies. However, Ainsworth and Pitcher (2004) and Ainsworth et al. (2008) reported that scaling vulnerabilities to prey trophic levels produced better predictions than scaling to predator trophic levels after comparing the predictions with time series data. I therefore set vulnerabilities proportional to prey trophic level in the range of 1-15 (see values in Appendix A). This range was chosen because it encompasses both predator-controlled and donordriven feeding relationships.  23  The SoG ecosystem was simulated from 2005 to 2035 using three scenarios. In the first scenario, I altered annual herring fishing mortality from 0.5 to 3 times the 2005 level to see how the harbour seal population responded. In the second scenario, I removed the harbour seal population by forcing their abundance to near zero to see how the fish population responded. In the third scenario, I tested the sensitivity of the model to the vulnerability parameter. I re-ran scenario 2 using different juvenile herring vulnerabilities to see how my assumptions on the seal-herring predation functional response affected my predictions regarding herring population changes in the absence of seals.  2.3  2.3.1  Results  Scenario 1  The herring fishery has a strong impact on the seal population through trophic effects (Figure 2.4). Harbour seal abundance increases greatly when more herring are available due to reduced herring fishing efforts. When more herring are removed from the system by fisheries, the population of seals decreases gradually even though seals increase their predation on hake, salmon and other prey.  2.3.2  Scenario 2  When the harbour seal population is reduced, hake and some groundfish increase in biomass, and herring and sand lance biomass decreases (Figure 2.5). Herring abundance depends on  24  Figure 2.4 Predicted harbour seal biomass (filled circle) and herring biomass (open circle) in 2035 relative to 2005 when altering the fishing mortality of herring relative to 2005.  seal abundance because seals predate other predators of herring. When seals are removed from the ecosystem, most fish increase immediately due to lack of seal predation. The forage fish then decrease due to increased predation from hake and groundfish. Some groundfish also decrease as a result of reduced food availability and increased food competition. The total fish biomass decreases by 35% and the total commercial fish biomass decreases by 33% after seals are removed. When seals are removed, the SoG shifts from a seal dominated ecosystem to a hake dominated ecosystem.  25  Figure 2.5 Predicted biomass of the main functional groups relative to their biomass in 2005 after removing seal population in the Strait of Georgia. Increasing groups in 2035 are: soles (open circle), hake (filled circle), lingcod (filled square). Decreasing groups are: dogfish (x), pollock (open inverted triangle), sand lance (filled triangle) and herring (open triangle).  2.3.3  Scenario 3  Herring biomass is very sensitive to the vulnerabilities of juvenile herring. The dynamics change greatly when vulnerabilities are set low (Figure 2.6). In this situation, juveniles are assumed to have more predation refuges and the rate of predation mortality on juveniles is controlled by prey biomass (i.e., donor control) rather than predator biomass. The total herring biomass increases when seals are removed. However, when the vulnerabilities 26  2.0 1.5 1.0 0.5 0.0  Predicted Herring Biomass / Biomass in 2005  2005  2010  2015  2020  2025  2030  2035  Year Figure 2.6 Predicted herring biomass relative to 2005 with changing vulnerability of juvenile herring when seals are removed. Straight line shows prediction with vulnerability of juvenile herring (8.4) proportional to its trophic level. Symbols show prediction with vulnerabilities of juvenile herring at 1.1 (open triangle), 1.5 (plus), 2 (filled circle), 2.5 (open inverted triangle), 3 (filled square), 4 (filled triangle), 6 (open circle), 100 (x), respectively.  increase slightly, the herring biomass decreases dramatically. The changes in the response of herring biomass are increasingly smaller when the vulnerabilities of juvenile herring are larger than 2.  27  2.4  Discussion  This model suggests that removing seals from the SoG may not result in better fisheries or higher abundances of fish populations. This finding is surprising compared to Gerber et al. (2009) who found that fish stocks would increase in biomass after whales are removed from the ecosystem with little overlap between fisheries and whale prey. However, my model shows that the total fish biomass or the total exploitable fish biomass may decrease by about one third after seals are removed. Opposite to the perception of culling marine mammals to increase fisheries, the presence of harbour seals may increase the overall abundance of exploitable fish populations in the SoG ecosystem where seals and fisheries compete directly. In exploited marine ecosystems, fisheries do not necessarily increase with a cull of top predators because of the complexity of the food web and other issues (Yodzis, 2001). The critical reason is that herring benefits from predation by seals on its predators and competitors, leading to an overall positive impact of seals on herring. This is consistent with the northern Gulf of St. Lawrence ecosystem where apex predators had a negative impact on their dominant prey, the high trophic level fish, but an indirect positive effect on the prey of their preferred prey such as plaice (Morissette et al., 2006). Lessard et al. (2005) arrived at a similar conclusion regarding the effect of a seal cull in the SoG. While they found that it would increase abundance of chinook and coho salmon in the short-term, long-term abundance may not increase because of compensatory mortality from increased hake predation. They cautioned that this result is highly dependent on the assumed contribution of salmon to the diet of seals and hake. Moreover, such a 'symptomatic' policy option, they suggest, must also be accompanied by strict monitoring and tested in an adaptive  28  management design until risks inherent in manipulating species abundances are better understood.  This study demonstrates that removing seals may result in a severe decline for herring and that Pacific hake may dominate the ecosystem. The decline of herring is highly dependent on the assumptions used to model the seal-herring predation functional response, as indicated by scenario 3. Nevertheless, this finding is corroborated by circumstantial evidence. The herring collapse in the late 1960s occurred at a time when the seal population was at low abundance following two decades of hunting. The collapse was attributed to overfishing and unfavorable ocean conditions (DFO, 2009) but my study suggests an additional possible contribution of increased hake predation due to the low abundance of seals in the late 1960s. This is consistent with previous findings in different ecosystems. Merrick (1997) reported that the Bering Sea ecosystem became dominated by walleye pollock after a dramatic decline of pinnipeds and whales since the early-1970s. It was in this period that fur seals were listed as depleted under the US Marine Mammal Protection Act. In the opposite condition, Bundy et al. (2009) reported forage fish increased when seals replaced cod as a top predator in the four Northwest Atlantic ecosystems after the cod stock collapsed 15 years ago. Marine mammals are consumers of production at most trophic levels and have a major influence on the structure and function of aquatic communities (Bowen, 1997). It is therefore likely that with the top predators greatly depressed or removed, the whole ecosystem assemblage tends to shift.  29  Notably, this study is based on the assumption of fixed primary production. However, longterm average sea surface temperature has increased 1°C since 1970 (Masson and Cummins, 2007) while the timing of the peak water flow of the Fraser River has also become earlier in the year and the amount of flow has diminished (Morrison et al., 2002). Both of these factors are likely to impact the overall productivity of the system, and therefore the exploitable biomass of fish. The ecosystem responses will be further complicated with the addition of other random variations. Although these are out of the scope of the current study, they would be necessary to include when considering future climate change impacts.  30  Chapter 3: Investigating ecological impacts of invasive fish with altered physiology using ecosystem modeling  Growth hormone (GH) transgenic fish have the potential to greatly increase the yield of fish farms but are kept in captivity due to serious ecological concerns. This study is the first attempt to examine the possible trophic impacts of GH coho on the Strait of Georgia (SoG) ecosystem should they ever escape from facilities. Ecopath with Ecosim (EwE) modelling using lab-determined parameters showed that GH coho would impact the whole ecosystem largely through indirect interactions. Many functional groups were impacted depending on assumptions about GH coho diet. However, functional groups were more strongly impacted when a bottom-up effect was introduced by changing ocean conditions. Different assumptions about GH coho diets only influence GH coho and the wild salmon biomass. The harbour seal-hake-herring trophic triangle remained stable across many scenarios despite the direct or indirect interactions of GH coho with each component. A new EwE’s Monte Carlo approach showed that the model predictions were robust to uncertainty in GH coho parameters, especially Ecopath parameters such as production rate and consumption rate. High prey vulnerabilities to GH coho, a top-down control by GH coho, produced the most different predictions. The EwE ecosystem modelling framework proves an excellent tool to try to forecast complicated trophic interactions.  31  3.1  Introduction  Growth hormone (GH) transgenic fish have the potential to greatly increase the yield of fish farms. However, there is major concern about the potential ecological harm transgenic fish may cause should they ever invade natural systems. Since the first transgenic fish were produced in China (Zhu et al., 1985), there have been more than 30 species of genetically engineered fish, including many of the major world aquaculture species (Devlin et al., 2006). GH fish grow substantially faster than their unmodified counterparts: for example, GH transgenic coho salmon (Oncorhynchus kisutch, hereafter GH coho) at 14 months of age are on average 11 times larger than control salmon (Devlin et al., 2004a; Devlin et al., 1994). This potential gain must however be balanced against the risk of escape from aquaculture facilities.  Aquaculture is considered as a gateway for exotic species (Naylor et al., 2001). For example, escaped farmed Atlantic salmon have been caught in the North Pacific salmon fishery (Fiske et al., 2006; Hansen et al., 1999) and have spawned in wild Pacific salmon rivers (Crozier, 1993; Lura and Sægrov, 1991). Farmed salmon escapes are inevitable so long as fish are reared in an open aquatic environment: accidental escape can happen during fish transfer and harvesting, extreme weather and breaches in pen integrity caused by marine mammals. (http://www.al.gov.bc.ca/fisheries/cabinet/Escape_Stats.PDF). Thus it is not surprising that environmental impact tops the list of fears about transgenic animals (Check, 2002).  32  Several studies have reviewed the ecological effects of transgenic fish and defined what needs to be assessed (Kapuscinski et al., 2007; Aerni, 2004; Muir and Howard, 2004; Devlin and Donaldson, 1992; Kapuscinski and Hallerman, 1990). The overall environmental concerns posed by GH animals is associated with impacts on ecosystem components, and is influenced by (1) the probability of escape and dispersal into diverse communities, (2) the fitness of transgenic fish in the wild and (3) the stability and resilience of the receiving community (National Research Council, 2002). Some studies have examined the fitness of transgenic fish with an emphasis on the spread of the transgenes should they escape from culture facilities. In one scenario, Trojan gene that confer combinations of enhanced and reduced fitness characteristics have been modeled to indicate populations could be driven to extinction by certain transgenes (Muir and Howard, 1999). For example, when a transgene enhances reproduction, such as higher fecundity (Pennington and Kapuscinski, 2011; Bessey et al., 2004; Muir and Howard, 2001), while reducing the viability of their young, the reproduction advantage could make the transgene spread quickly into a wild population but the low viability of its offspring may increase the probability of the extinction of the natural and transgenic population (Hedrick, 2001; Muir and Howard, 1999). Ahrens and Devlin (2011) found background genetics had the potential to play an important role in mediating the transgene persistence in populations and that evolution in the transgenic sub-population may affect non-transgenic conspecifics as well. To reduce environmental risks, some scientists have suggested applying sterilization technology such as triploidy which impairs develop of the gonad in many fish species (Devlin and Donaldson, 1992). Coupling triploidization with mono-sex female technologies can produce sterility in large numbers of  33  production fish and could be an attractive approach despite the impairment of growth and reduced disease-resistance (Devlin et al., 2004a; Razak et al., 1999).  In contrast to many studies focusing on transgene dispersal and control, the resilience of the receiving ecosystems has rarely been studied. Even with a 100% sterilization success rate, the ecosystem still carries a high risk from trophic interactions because GH transgenic fish are vigorous foragers and competitors for prey compared to the native population (Sundström et al., 2004; Devlin et al., 1999). There are concerns about increased predation on the prey populations and competition with wild populations (Kapuscinski et al., 2007; Aerni, 2004; Muir and Howard, 2004). Nevertheless, to date and to my knowledge, no quantitative studies have examined the trophic impacts on the whole ecosystem. Even ecological studies on escaped Atlantic salmon have focused on the wild population only (Liu et al., 2012; Hutchings and and Fraser, 2008; Hindar et al., 2006).  In general, experiments with GH coho in laboratory or semi-natural conditions have found that their survival, growth, and impacts on other species in the system, are highly influenced by environmental conditions (e.g. simple vs. complex environments, food availability, predator load), and that these responses differ from those see with non-transgenic coho (Sundström et al., 2007; Bessey et al., 2004; Devlin et al., 2004b). Such genotype x environment interaction makes it difficult to predict outcomes to nature based on laboratory data alone (Devlin et al., 2007; Devlin et al., 2006). Further, unlike the situation for genetically modified plants, release of transgenic fish to nature to conduct risk assessment impact evaluations is not currently conducted as a precaution in case such fish established  34  themselves and begin inducing ecosystem impacts (eradicating fish from large complex ecosystems such as are inhabited by salmon is essentially impossible). Thus, risk assessment impacts for transgenic fish in large oceanic and riverine environments are not supported by a large body of empirical data, and what information that does exist possesses significant uncertainty. The application of complex ecosystem modelling to risk assessment research can provide estimates of potential influences of genetically modified fish under different scenarios that cannot otherwise be obtained through empirical experimentation in nature.  In this study, I concentrated on the Strait of Georgia (SoG) marine ecosystem’s resilience to escaped GH coho and evaluate the potential long-term changes to the ecosystem using the Ecopath with Ecosim (EwE) modelling approach. Transgenic fish cannot be released to nature to experimentally study ecological interactions, and thus environmental risk assessment studies on transgenic fish are currently restricted to the laboratory and the data derived are limited due to environmental complexities that arise therein. Therefore, experimental studies of ecological effects of transgenic fish have to rely on laboratory and modeling studies, rather than experimentation in the field (Devlin et al., 2007). Ecosystem modeling has been widely used in ecosystem function and predictions (e.g. Perry et al., 2010a). For example, EwE models have been successfully used to investigate interactions between fisheries and marine mammals (Li et al., 2010; Gerber et al., 2009) and between non-indigenous species and endangered salmon (Harvey and Kareiva, 2005). Here I applied a recently published SoG EwE model (Chapter 2; Li et al., 2010) to examine how the GH coho might influence the whole food web in the context of environmental variability and fishing should they enter the SoG.  35  Unlike many invasion biology studies trying to predict the establishment success of alien species (e.g. Hayes and Barry, 2008; Kolar and Lodge, 2002), I assumed certain biomass levels of escaped GH coho living in the SoG marine ecosystem and examined the consequences of trophic interactions on the organisms of this ecosystem. Instead of covering their complex life history, I focused on their marine life and their trophic impacts on the SoG marine ecosystem. My aim was not to carry out accurate and precise predictions in changes in biomass of marine organisms, which are impossible, but to stress the general trends, most sensitive trophic linkages, and seek to identify important factors in potential trophic interactions. I also consider a wide range for each GH coho parameter to assess the uncertainty of my predictions.  3.2  3.2.1  Materials and methods  Study area  The SoG is a key area for foraging of Pacific salmon species, herring and many groundfish species and is therefore a key marine ecosystem in which to study the potential impacts of GH coho. Harbour Seals are thought to have reached carrying capacity in the SoG since late 1990s (DFO, 2010; Olesiuk, 1999). Due to the decline in groundfish populations in recent decades (Johannessen and Macdonald, 2009; Akenhead et al., In review), Pacific herring, the main forage fish in the SoG, has become the major fishery (Li et al., 2010). Hake and its major prey herring are both important prey for harbour seals, which together compose an important food links in the SoG (Li et al., 2010).  36  3.2.2  Ecopath with Ecosim model parameterization  EwE version 6 was used in this study. The basic theory and the SoG model for the year of 2005 have been described in Chapter 2 (Li et al. 2010). There are two salmon groups in the model: transient salmon (including sockeye, chum (Oncorhynchus keta), and pink (O. gorbuscha)) and resident salmon (including coho salmon and chinook salmon (Oncorhynchus tshawytscha)). As transient salmon only pass the SoG and migrate to the Pacific Ocean, I pay more attention to the resident salmon group. I added one GH coho group to this SoG Ecopath model with 39 functional groups (Li et al. 2010). I set its initial biomass the same as the wild resident salmon. GH coho have inherently increased appetite and higher feed intakes. They consume about 3 times as much food as non-transgenic coho of the same size when excess food is provided in freshwater (Devlin et al., 1999) and seawater experiments (Devlin, unpublished data). However, in natural streams, GH coho could not realize their full growth potential (Sundström et al., 2007) due to presence of predators and complex habitat. I thus set Q/B of CH coho to 14.6, twice that of resident salmon. GH coho have enhanced ability to effectively utilize digestive energy and have about 10-20% higher food conversion efficiency (Higgs et al., 2009; Raven et al., 2006; Devlin et al., 2004a). Thus I assume a P/Q 20% higher than the resident salmon and estimated P/B 5.26 for GH coho. EE of GH coho is as low as 0.15, suggesting fishing and predation only contribute a small proportion to the total mortality in the model. This allows a large amount of GH coho to die of disease – indeed, they have been observed to have reduced resistance to bacterial pathogen (Jhingan et al., 2003). Uncertainties in these GH coho parameter values were later employed in Monte Carlo simulations.  37  Figure 3.1 Four possible diets for GH coho salmon.  In the absence of field studies, I hedged uncertainty by creating four different possible diets for GH coho, including a baseline assumption of the same diet as resident salmon (Figure 3.1). For the other diets, I assumed that GH coho are pure opportunistic predators, i.e., diet is directly proportional to the biomass of organisms selected as prey (this is automatically driven by the Ecosim model as selected prey biomass changes). The three possible diets differ in the selected potential prey they may feed on. For the forage diet, I assume that GH coho feed on forage fish and invertebrates. For the forage hake diet, I assume that GH coho also feed on hake in addition to the forage diet. For the broad diet, I assume only 20% of  38  demersal fish, given that GH coho, like their wild counterparts are more likely to encounter pelagic organisms and invertebrates.  I assume GH coho to have the same predators as resident salmon, but with increased risk of predation because they tend to position themselves closer to the surface (Sundström et al., 2003), have increased predation mortality (Sundström et al., 2004), have reduced swimming ability (Lee et al., 2003; Farrell et al., 1997), and brain structure changes ((Kotrschal et al., 2012). Some lab-based semi-natural habitat experiments have revealed higher predation mortality in GH coho than in their counterparts (Sundström et al., 2005; Sundström et al., 2004). Therefore, I assume GH coho have 50% higher predation mortality than resident salmon to bigger predators such as harbour seals. I also apply lower predation mortality for GH coho to small predators such as birds due to their difficulty in handling big prey. I found a small proportion of resident salmon contribute to adult pollock diet (Appendix B) in Li et al (2010) based on Preikshot (2007). However, salmon do not contribute to the diet of adult pollock in Puget Sound (Harvey et al., 2010), North BC (Ainsworth 2006), or Alaska (Yang and Nelson, 1999). I thus corrected it by replacing salmon with euphausiids. Additionally, I applied the same low fishing mortality (recreational fishery) for GH coho as the resident salmon (Li et al., 2010).  The Ecopath model was automatically balanced after adding GH coho group, suggesting an ability of receiving alien animals without breaking the balance. For Ecosim, the same vulnerabilities were applied as in Li et al. (2010) for all groups. The GH coho vulnerabilities  39  to predators were set the same as the wild population at 9.9 (top down) and the prey vulnerabilities to GH coho were set to the default value of 2 (intermediate).  3.2.3  Ecosim scenarios  I ran two major types of scenarios. The first type of scenarios examined the impacts of GH coho at different biomass levels under fixed ocean conditions (no changes in primary and secondary productions caused by ocean conditions) (Table 3.1, Scenario 1-3). As Chapter 4 shows dramatic changes could happen in zooplankton communities, I then added uncertainty in euphausiids and herring, assumingly caused by altered ocean conditions, in the second type of scenarios (Table 3.1, Scenario 4-8). For the both types of scenarios, I also examined influences of four different GH coho diets and compared the impacts of GH coho and resident salmon populations at similar biomass level. GH coho were assumed to have no breeding with the resident salmon. The forced high biomass of GH coho can result from direct escape and self-reproduction.  For the first types of scenarios, I first compared the impacts of different diets of GH coho at a fixed biomass level (Table 3.1, Scenario 1). I held GH coho biomass constant at 10 times its initial biomass, with four different diets respectively, for thirty simulated years (the model runs 12 time steps per simulated year). Secondly, to compare the responses of the functional groups to the increased GH coho and wild population, I ran the model of Li et al. (2010) keeping the resident salmon population (with no GH coho present) constant at 11 times its initial biomass (which is equivalent to the total biomass of resident salmon and GH salmon in  40  Table 3.1 Summary of all scenarios. Scenario 2 and 5 were carried out using the model of Li et al. (2010) with no GH coho present. The same diet of Scenario 1 and 6 were used for model uncertainty analyses. Forced GH coho Code biomass above baseline 1 10 times 2 3 5, 15, 20, and 25 times 4 10 times 5 6 7  10 times  8  10 times  Forced increase/decrease biomass of other groups from baseline Resident salmon 11 times Euphausiid decline 15% Euphausiid decline 15% Euphausiid decline 15% Euphausiid decline 30%, increase 15% and 30%, respectively Herring declining 30% and 15%, increase 15% and 30%, respectively  GH coho diet 4 diets The same diet 4 diets 4 diets  Figure 3.2 3.3 3.4 a 3.4 a 3.4 b 3.5  The same diet The same diet  the previous scenario) for thirty years (Table 3.1, Scenario 2). Last, I took the same diet of GH coho as an example and used biomass forcing function to hold GH coho biomass constant at 5, 15, 20 and 25 times its initial biomass for thirty simulated years to investigate the impacts of GH coho at different biomass levels (Table 3.1, Scenario 3). Historic SoG salmon farm production was obtained from Ministry of Agriculture, BC, http://www.env.gov.bc.ca/omfd/fishstats/index.html. A biomass of 25 times of the baseline is equivalent to a half the average annual total production of farmed salmon (all species including Atlantic salmon (Salmo salar L.), chinook salmon, coho salmon, and steelhead (Oncorhynchus mykiss)) from 2000 to 2010 in the SoG.  For the second types of scenario, the responses of functional groups were compared to a combination of declined euphausiids (or herring) and different GH coho conditions. A simulation was first run keeping euphausiid biomass constantly 15% lower than its baseline,  41  for thirty years, with different 4-diet GH coho conditions respectively: 10 times biomass forcing; not present; present but with no forcing (Table 3.1, Scenario 4-6). I then held the same diet of GH coho at 10 times its baseline biomass and kept euphausiid or herring biomass at a constant 30% and 15% below, and 15% and 30% above baseline levels (Table 3.1, Scenario 7-8).  3.2.4  Model sensitivity  Considering the huge uncertainty associated with GH coho parameters based on currently available empirical data, I performed two separate analyses to examine the sensitivity of the Ecosim predictions: (1) on the initial Ecopath parameters and (2) on the Ecosim parameters. I completed the two sensitivity analyses keeping the biomass of GH coho constant at 10 times that of resident salmon and with the same diet as resident salmon (same diet in Scenario 1). I compared the perturbed Ecosim predictions with the Ecosim baseline. However, the biomass forcing may bring about changes in the P/B ratio. I thus repeated the two sensitivity analyses with no forcing of the biomass of GH coho, with GH coho eating the same diet as resident salmon and a fixed 15% reduction in the biomass of euphausids (the same diet in Scenario 6).  For the Ecopath parameter sensitivity analyses, I used the Monte Carlo approach (Christensen et al., 2008; Christensen and Walters, 2004). I first improved Monte Carlo function in EwE 6 by adding Q/B and a new parameter EE tolerance (EET) to the list of parameters that can be varied on the main Monte Carlo interface. First, I set a small number  42  0.00005 for EET to allow close-to-exactly-balanced Ecopath models to run, rather than requiring an exact balance, and thus minimize the difference between the range of the parameter values actually used and set by the user. I set a wide range (upper and lower limit) for Q/B of GH coho at 7.3-21.9, which is 1-3 times the value for resident salmon, based on experiments under different conditions (Sundström et al., 2005; Devlin et al., 2004b; Sundström et al., 2004; Devlin et al., 1999). For P/B of GH transgenic coho, I also consider a wide range 1.2-3.6 times as the resident salmon, which is 2.6-7.9 for GH coho. Additionally, I set the coefficient of variation to a large value such as 1000 to emulate a uniform distribution within the upper and lower limit so that new sets of Ecopath parameters are randomly drawn within the range. If EE for any group is greater than 1.0 + EET or less than 0 - EET, then the input parameters will be rejected and another attempt will be made at varying the Ecopath parameters. Ecosim simulates the ecosystem only when a new balanced Ecopath model (within EET) is achieved. I completed 200 successful Monte Carlo runs with altered P/B and Q/B, single or combined, to compare with the Ecosim baseline.  For the Ecosim parameter sensitivity analyses, I varied the vulnerabilities of GH coho in Ecosim, which represents the degree to which a large increase in predator biomass will cause predation mortality for a given prey. A low vulnerability (1<V<2) indicates a bottom-up control, so that an increase in predator biomass will not cause any significant increase in the predation mortality on the given prey. In contrast, a higher vulnerability (V>2) indicates a top-down control, e.g., if the predator biomass is doubled, it will come close to doubling the predation mortality (Christensen et al., 2008). I re-ran the same diet of Scenario 1 and 6 with varying prey vulnerabilities to GH coho and GH coho vulnerabilities to predators. For all  43  predation on GH coho, I increased GH coho vulnerabilities to 100, causing strong top-down control on GH coho, and then decreased vulnerabilities to 1.5, resulting in strong bottom-up control by GH coho. Similarly, I altered the prey vulnerabilities to GH coho from strong topdown control to strong bottom-up control. I compared the Ecosim runs using these different vulnerability settings with the Ecosim baseline.  3.3  3.3.1  Results  Model responses to increased salmon  When GH coho biomass is maintained at 10 times the baseline, the impacts on different groups vary with GH coho diet (Figure 3.2). Seal biomass increases at all levels due to extra food from the GH coho, and this higher seal population also increases predation on other prey such as hake and groundfish that compete for food with GH coho. Herring biomass declines greatly when GH coho have the forage diet, preying heavily on herring. In contrast, herring biomass increases even when GH coho have a diet of forage and hake, where herring contribute a similar proportion. This effect arises because with a diet of forage and hake, GH coho also prey on herring’s largest predator, hake, and thus reduce the predation from hake. Under the forage diet, hake decline least due to the smallest increase in the population of seals. Resident salmon decline most when GH coho have the same diet, whereas little changes are observed when GH coho have a forage diet. With any GH coho diet, the biomass of hake, resident salmon and lingcod decrease and seal biomass increases while changes in the biomass of herring, pollock and dogfish and euphausiid depend on the diet of GH coho.  44  Figure 3.2 Predicted biomass in the thirtieth year relative to the Ecopath baseline when the biomass of GH coho, with four different diets, was forced to remain10 times for 30 years.  Changes in the functional groups increase with increasing biomass of GH coho with the same diet (Figure 3.3). When the biomass of resident (non-transgenic) salmon is held at a level 11 times baseline, with no GH coho present, the ecosystem displayed the largest changes in the biomass of seals, hake and herring but little changes in the biomass of other functional groups. This is due to an assumption that GH coho share the resident salmon’s proportion to seal’s diet and contribute more to seals’ diet than their counterpart at the same biomass. Resident salmon suffered higher mortality by harbour seals when GH coho were absent than when they were present. Consequently, an increased biomass of resident salmon in the model with GH coho absent caused an increase in the seal population, which depressed the biomass of hake and allowed the biomass of herring to increase.  45  Figure 3.3 Responses of the biomass in seals, pollock, lingcod, hake, resident salmon and herring when the biomass of GH coho, with the same diet as resident salmon, was forced to a level 5, 10, 15, 20 and 25 times the baseline biomass of resident salmon for 30 years.  3.3.2  Model responses to altered ocean conditions  Reducing euphausiid biomass by 15% of the baseline in addition increasing the biomass of GH coho 10 times, decreases the biomass of hake, resident salmon and most groundfish and greatly increases the biomass of herring and seals (Figure 3.4 a). Hake decrease due to decrease of its major euphausiid prey and higher predation by seals. Herring biomass increases due to reduced predation from hake while they acquire enough food from copepods. Seals benefit from increased herring and GH coho. Notably, with different GH coho diets, there are no large differences in the responses of other functional groups except  46  salmon groups (Figure 3.4 a). This indicates that bottom-up effects play an important role in this ecosystem's functioning.  When euphausiids are forced to decrease by 15% with GH coho absent, functional groups show similar trends with smaller changes in some groups (Figure 3.4 a). With no extra food from GH coho, harbour seals increased less and thus hake declined less. Harbour seal biomass increases due to increased prey herring resulting from decreased hake. When euphausiids are decreased by 15% with GH coho present but no forcing, responses of functional groups are similar to the situation with GH coho absent, again suggesting a strong bottom-up effect. The different GH coho diets only make a difference for resident salmon and GH coho (Figure 3.4 b): GH coho biomass declines only under the same diet as resident salmon, in which they feed heavily on euphausiids. GH coho benefit from increased herring and thus increase greatly in biomass under the forage diet and the forage and hake diet.  The whole ecosystem changes greatly with 10 times GH coho biomass with the same diet as resident salmon under altered the euphausiid biomass (Figure 3.5). More euphausiids produces more hake and pollock and, through predation, lower biomass of herring and harbour seals. This indicates strong bottom-up effects of euphausiids on hake and pollock and hereafter a trophic cascade influence on herring and seals. I do not present the scenarios of altered herring biomass in this study because the results of decreased euphausiids are similar to the results of increased herring and opposite to the results of decreased herring. This correlation of altered euphausiid and herring biomass indicates two major pathways in the SoG food web: one via euphausiids the other via herring.  47  (a)  (b)  Figure 3.4 Predicted biomass relative to the baseline when the biomass of euphausiid was forced to decrease by 15% of the baseline for 30 years with three different conditions of the biomass of 4-diet GH coho: 10 times forcing and with GH coho absent, respectively (a), and GH coho present but no forcing (b).  48  Figure 3.5 Relative biomass in seals, pollock, lingcod, hake, resident salmon and herring when euphausiid biomass was kept at a constant 30% and 15% below and 15% and 30% above the background levels for thirty years with GH coho at 10 times the baseline biomass of resident salmon and eating the same diet.  3.3.3  Sensitivity analyses  In the first set of sensitivity analyses, when the biomass of GH coho was forced to 10 times above background levels for resident salmon and they both had the same diet (same diet in Scenario 1), Ecosim estimates of biomass in functional groups are more sensitive to P/B than Q/B of GH coho (Figure 3.6). With variable P/B or Q/B, the mean biomass of each functional-group is very close to the Ecosim baseline with the maximum difference of 1.9% in average seal biomass. However, the variability in predicted biomass is prominently larger with variable P/B than with variable Q/B, except for wild resident salmon, which is impacted  49  P/B Q/B P/B & Q/B  0.1  0  Herring  -0.2  Res salmon  Soles  Lingcod  Hake  Pollock  -0.1  Seals  Change from Ecosim baseline  0.2  Figure 3.6 Deviation of Monte Carlo runs from the Ecosim baseline (mean ± standard deviation), keeping GH coho with the same diet as resident salmon and at 10 times the resident salmon biomass, with the following altered Ecopath parameters for GH coho: uniform-distributed P/B (from 2.6-7.9), uniform-distributed Q/B (from 7.3-21.9) and combined.  As prey 1.5 As prey 100 As predator 1.5 As predator 100  0.3  0  herring  Res salmon  Soles  Lingcod  hake  -0.6  pollock  -0.3  Seals  Change from Ecosim baseline  0.6  Figure 3.7 Deviation of Ecosim output from the Ecosim baseline, keeping the biomass of GH coho with the same diet 10 times that of the resident salmon, with varying predator-prey interactions: GH  50  coho vulnerability 1.5 and 100 (GH coho as prey) to predators and the prey vulnerability 1.5 and 100 (GH coho as predator) to GH coho. Vulnerabilities >2 indicate top-down control from predators and <2 indicate bottom-up control from preys.  strongly by Q/B. A combined effect also produces similar variability with slightly larger deviation in the mean functional-group biomass than seen with P/B or Q/B alone.  By contrast, Ecosim estimates of biomass in functional groups were more sensitive to the predation-prey interactions (Figure 3.7). Among the different vulnerabilities tested, GH coho vulnerability 100 to predators and prey vulnerability 1.5 to GH coho produced very similar results to the Ecosim baseline. A prey vulnerability of 100 to GH coho produced a greater difference from the baseline than the other three vulnerabilities and ranked top among all the parameters considered in the sensitivity analyses. Under this strong top-down scenario, the average deviation of all groups decreased 44% with the largest difference in resident salmon (Figure 3.7). Resident salmon decreased 60% from the baseline and became almost extirpated. Additionally, pollock and hake declined about 40% from the baseline and remained at low biomass levels. The impacts on the whole ecosystem were the largest when GH coho had top-down control on their prey.  In the second set of sensitivity analyses, when the biomass of euphausiids was forced 15% below background levels, and both GH coho and resident salmon had the same diet (Scenario 6),  GH coho Ecopath parameters have only a minor influence on Ecosim estimates of  biomass probably due to low biomass of GH coho. Variability of predicted biomass was very  51  small for the P/B perturbation (GH coho 0.017 and other groups < 0.01), Q/B perturbation (<0.01), and combined perturbation (GH coho 0.018 and other groups < 0.01). Ecosim estimates of biomass of GH coho had a mean deviation of only 5% for P/B perturbation, a negligible deviation for Q/B perturbation (<0.001) and a mean deviation of 7% for the combined perturbation to the baseline. Ecosim predictions of other functional groups were close to the Ecosim baseline with the largest mean deviation of 2% for P/B perturbation, 2.6% for Q/B perturbation, and 2.6% for the combined.  Unlike Scenario 1, setting prey vulnerabilities of 100 to GH coho had the largest effects only on GH coho and resident salmon groups, again suggesting a weak impact of GH coho at low biomass in Scenario 6. Under these vulnerability settings, GH coho decline 73% and resident salmon increase 27% compared to the Ecosim baseline. Thus, contrary to the Ecosim baseline (Figure 3.4 b), GH coho decline much more than resident salmon (prey vulnerabilities to resident salmon have a range from 5.3-9.8) when euphausiids decline. As the two groups have exactly the same diet, the biomass of the group with higher vulnerability declines more than the other. For the other three sets of vulnerabilities, Gh coho vulnerabilities 1.5 and 100 to predators and prey vulnerability 1.5 to GH coho, the deviation of GH coho is less than 3%, while other functional groups are less than 1%.  52  3.4  3.4.1  Discussion  Trophic complex  The present analyses highlight the potential for complicated indirect trophic interactions through trophic cascades and competition arising from introductions of a fish strain with altered physiology into an ecosystem. Unlike other invasive species such as Nile perch in Lake Victoria, where direct predation is a major impact on the native species and has caused extirpation of many fish (e.g. Pitcher and Hart, 1995), my model suggests that GH coho living in the wild could impact the whole ecosystem largely through indirect interactions. GH coho feed on euphausiids and herring, which are the major prey of resident salmon, hake, and many ground fish. Resident salmon and GH coho are also preyed upon by their mutual predator, harbour seal. Thus, resident salmon and groundfish not only compete with GH coho for food but would also suffer from increased predation by seals due to artificially increased GH coho. Hake is also a potential prey of GH coho (in the forage and hake diet and the broad diet) in addition to the indirect effects. Herring is even more complicated as it is a prey of GH coho and hake and is also eaten by their predator harbour seals. Both herring and its largest predator hake are major prey of seals in the SoG (Figure 3.8). Therefore, each functional group in the model has been modelled as a consequence of all trophic interactions mentioned above. The indirect trophic interactions (DeCesare et al., 2010; Libralato et al., 2006; Lessard et al., 2005) cannot be investigated with many traditional methods such as bioenergetics model. Here, EwE ecosystem modeling demonstrates an excellent tool to capture all the trophic dynamics, especially the complicated indirect effects.  53  This study also revealed the importance of including minor diet items in the Ecopath model. Resident salmon and GH coho are initially at low biomass and only contribute a small proportion to the seal diet. Overlooking minor diet items in the diet composition for predators can result in incorrect assessments of predation impacts on minor prey species (Christensen and Walters, 2004; Walters and Kitchell, 2001). For example, large predatory adult fish crop down forage species that are predators of their own juveniles. However, when adult fish decline due to fishing, increased forage fish may cause serious decrease in the juvenile survival (although juvenile predatory fish are only a minor diet item in the forage fish) and thus delay the recovery of the large predatory fish even with the closure of the fishery (Walters and Kitchell, 2001). By contrast, I show that the minor prey (either resident salmon or GH coho), with increased biomass, is able to produce an increased predator harbour seal population. Therefore, including the minor diet items in the model is important in order to investigate the impacts on the predator as well.  I have predicted many different results depending on different conditions. However, my modeling results confirm the crucial role of the harbour seal-hake-herring trophic triangle in the food web (Figure 3.8) in the SoG and the consistent relationship between them. This triangle was fundamental in the major finding in Li et al. (2010) that removing seals will threaten herring due to increased predation from hake. Except when harbour seals increase less than 10% (the forage diet in Figure 3.2), all my scenarios here agree with Li et al. (2010) in that harbour seals show an inverse relationship with hake and the same trend as herring. My modeling suggests that invasive GH coho are unlikely to alter this trophic triangle relationship in spite of many direct or indirect trophic interactions with each component of  54  this triangle. My results indicate a powerful role of this trophic triangle in the SoG as this triangle incorporates influential species: an abundant marine mammal harbour seal, a major forage fish herring and the most abundant resident fish hake. Some studies have highlighted the keystone species that are at low biomass organisms but play an important role in structuring the ecosystems (Libralato et al., 2006; Power et al., 1996). However, my study suggests more attention to fixed links in the ecosystems such as the trophic triangle.  Figure 3.8 The harbour seal-hake-herring trophic triangle in the Strait of Georgia where hake feed on herring and both of them are the major prey of harbour seals. Arrows represent the energy flows.  3.4.2  Changing ocean conditions  Unlike responses under the fixed ocean conditions, the impacts on most functional groups are independent on the GH coho’s diet and biomass level under the altered ocean conditions. This indicates strong bottom-up effects on the SoG marine ecosystem. When euphausiid biomass declines, the biomass of GH coho may increase depending on its diet. Although GH coho hatch sooner (Lõhmus et al., 2010) and grow much faster than wild type salmon (Devlin et al., 2004a), stream experiments have showed that seasonal timing of seaward migration in GH coho occurs at approximately normal times (Sundström et al., 2010)  55  suggesting strong food competition with resident salmon in seawater. Because GH coho are more willing to feed on novel prey (Sundström et al., 2004), GH coho may easily feed on whatever is available and adapt to new prey in the changed ocean in the case of the large changes of ecosystem, such as regime shifts (Jiao, 2009; Hare and Mantua, 2000). Future experiments are needed to investigate what prey GH coho eat in seawater.  3.4.3  Model uncertainty  My sensitivity analyses reveal that my Ecosim predictions are robust to uncertainty in both Ecopath and Ecosim parameters of GH coho. All the trends (increase or decrease) in all functional groups are the same as in the Ecosim baseline with the forcing of increasing GH coho or declining euphausiid biomass. EwE involves numerous parameters, and model uncertainty has long been a concern. The Monte Carlo simulation routine has been used to improve Ecopath inputs to fit Ecosim to historic time series data (e.g. Hoover, 2012). Different vulnerabilities have been applied to examine how the Ecosim predictions changed (Li et al., 2010; Harvey and Kareiva, 2005). To date and to my knowledge, there have been no comprehensive studies on model uncertainty and my study is the first attempt to apply both Monte Carlo and Ecosim sensitivity analyses.  My study also shows Ecosim predications are more sensitive to vulnerabilities than Ecopath parameters. The top-down control by GH coho in the predator-prey interactions amplifies the trends in Ecosim predictions. By contrast, Q/B perturbations do not make a large difference despite the wide range of possible consumption rates considered. I find that GH coho’s  56  consumption rate remains at only 4 in the Ecosim baseline (14.6 in the Ecopath model baseline) due to a density-dependent mechanism when I force the biomass of GH coho to high levels. This means GH coho might be restricted by food availability and could not achieve their full growth potential. In this case, it is not meaningful to adjust the consumption rate because doing so would not mean that the fish could obtain more food. However, vulnerability can determine the predation through the exchange of prey between vulnerable and non-vulnerable states and thus is thought to be the most sensitive parameter in EwE model (e.g. Christensen and Walters, 2004). A high vulnerability represents a high exchange rate and thus allows the depleted vulnerable prey quickly replaced with previously invulnerable prey. The consumption with high vulnerability is therefore determined by the products of the prey biomass and predator biomass (Christensen and Walters, 2004). A strong top-down control in the predator-prey interaction, by setting a high vulnerability of GH coho as predator, improves GH coho’s actual consumption rate to 11 (although they still starve to some extent) when the biomass of GH coho is forced to constantly remain high. This confirms that impacts of GH coho are indeed larger (Figure 3.7) when they actually consume more food, although Ecosim predictions do not change much with directly altered Q/B. This is actually supported by EwE theory in that the calculation of consumption rate is predicted from vulnerability and other factors (Christensen and Walters, 2004). Similarly, the high vulnerability can make the consumption more sensitive to the decline of prey biomass. With no density-dependence, this situation will accelerate the decline of the predator biomass as shown in the sensitivity analyses with declining euphausiids. The strong sensitivity of vulnerability indicates that future studies should focus on more detailed sensitivity analyses on vulnerabilities.  57  Although I have considered the wide range of uncertainty in GH coho, many physiological and ecological factors, that may play an important role, are not included in my model. For example, I assume GH coho predators are the same as resident salmon’s. I may not have included other predators that may specifically feed on GH coho. Experiments have found a positive effect of temperature on growth rates of GH coho, suggesting higher optimal thermal conditions for GH coho than for the wild-type fish (Lõhmus et al., 2010). Additionally, GH transgenic fish have altered standard oxygen consumption than transgenic fish (Deitch et al., 2006; Leggatt et al., 2003). Thus, GH coho may differentially respond and adapt to shifts in environmental conditions compared with wild-type, influencing their ability to survive and interact in ecosystems (Lõhmus et al., 2010). Furthermore, this Ecopath model does not explicitly simulate salmon’s freshwater life. Thus uncertainties in their early life stage are ignored in this study.  Despite the subjectivity involved, EwE is still an extremely valuable tool to investigate the complicated trophic interactions. Even with extreme values considered, my model consistently predicts the basic trends in functional groups. It is possible to investigate scientific questions of interest and make meaningful predictions concerning ecosystem response (Ainsworth et al., 2008), specifically, for “strategic” rather than “tactical” application (Essington, 2007; Christensen and Walters, 2004). My study is expected to contribute to evaluating risk assessment methodologies that may be applied for the regulation of transgenic fish should such technology be adopted in the future.  58  Chapter 4: Zooplankton communities in the Strait of Georgia, British Columbia, track large-scale climate forcing over the Pacific Ocean  Many studies have reported a correlation between offshore biological changes and largescale climate forcing. However, it remains unclear whether the biological changes in spatially confined estuaries are associated more with large-scale climate forcing or with local environmental changes. The Strait of Georgia (SoG) was the location used to investigate this question. Based on night zooplankton samples collected from the surface 20m between 1990 and 2007, I identified ten major taxa and investigated interannual changes for September and June. A combination of cluster analysis, NMDS and PCA showed an abrupt shift in the September community composition in 1998/1999 associated with a decline in the biomass of euphausiids and calanoid copepods in particular. Although the June samples did not have complete coverage throughout the time period of interest, they confirmed a similar pattern of community change. The PCA analysis showed a phase shift in the large-scale climate indices in 1999. By contrast, local environmental factors were less coherent. The best single indicator of zooplankton community change was the spring extratropical-based Southern Oscillation Index (SOIx), with a one-year lag. The best single local indicator of zooplankton change was the day of the peak flow of the Fraser River. Both linear and non-linear correlations showed that local environmental factors were weakly correlated with zooplankton community changes in the Strait, highlighting this as an important area for future research to understand long-term ecosystem changes. As secondary production plays a  59  key role in food webs, these large changes in the zooplankton community may have had a substantial impact on higher trophic levels and the ecosystem as a whole.  4.1  Introduction  There has been increasing interest in the impacts of environmental variability on marine ecosystems at interannual to decadal scales. Changes in large-scale climate indices (Drinkwater et al., 2010a; Stenseth et al., 2003) have been demonstrated to be correlated with changes in offshore ecosystems, in particular, the North Atlantic Oscillation (NAO) in the Atlantic (Stenseth et al., 2002), the Pacific Decadal Oscillation (PDO) in the North Pacific (Hare and Mantua, 2000), and the influence of the North Pacific Gyre Oscillation (NPGO) on southern California (Di Lorenzo et al., 2008).  However, it remains unclear if biological changes in spatially confined estuaries are also tied to these large-scale climate forcings. Estuaries are frequently highly productive and support important fisheries resources. They may be subject to a variety of local influences, such as local winds and fresh water inputs into the estuary (Eloire et al., 2010). Links between ecosystem change in these systems and large-scale climate indices may therefore not be as clear as for offshore regions, where climate and large-scale ocean dynamics are likely to be dominant environmental parameters affecting biological communities (Cloern and Jassby, 2008). In the largest estuary on the US east coast, the Chesapeake Bay, river flow is the primary driver of ecological variability (Kimmel et al., 2009). In this system, river flow is driven by regional-scale weather variability, and is therefore not strongly controlled by largescale climate indices. In San Francisco Bay, biological communities have been shown to 60  track large-scale climate forcing over the North Pacific (Cloern et al., 2010). However, due to limited data on local parameters such as water circulation, the way in which these largescale climate patterns interact with local drivers to structure the estuarine biological communities remains unclear.  The Strait of Georgia (SoG) is a large semi-enclosed estuary between Vancouver Island and the mainland coast of BC, Canada (Pawlowicz et al., 2007; LeBlond, 1983). It is a key area for foraging of salmon, herring and many groundfish species. However, groundfish populations have declined in recent decades (Johannessen and Macdonald, 2009; Akenhead et al., In review). In the 1990s a series of measures, such as fishery closures and restricted fishing areas, were implemented by the Department of Fisheries and Oceans (DFO), Canada, to promote the recovery of depleted fish populations such as coho salmon, lingcod and rockfish. Thus far, no significant recoveries of those fish populations have occurred which suggests a role for bottom-up food web control of their populations.  A regime shift (e.g. Jiao, 2009) in 1999 was reported in both offshore and inshore regions of the Northeast Pacific. In the southern and central parts of the California Current, an abrupt switch to positive anomalies in 1999 was observed in many of the dominant zooplankton taxa, including copepods, euphausiids, chaetognaths and tunicates (Lavaniegos and Ohman, 2007). In San Francisco Bay, a step-like increase was detected in 1999 in phytoplankton, shrimp, crab and demersal fish biomass, which tracked large-scale climate indices (Cloern et al., 2010). In the northern California Current and off southern Vancouver Island, the abundance of various cool water high latitude species increased abruptly in 1999, with a  61  corresponding decrease in the abundance of various warm water low latitude species (Mackas et al., 2007; Mackas et al., 2004; Peterson and Schwing, 2003; Mackas et al., 2001). Here I use a time series of zooplankton data from the SoG, spanning the 1999 regime shift (1990-2007), as a case study to examine the interaction between large-scale climate patterns and the local environment, and their impact on the pelagic ecosystem. Specifically, my objective is to determine the response of this important coastal estuarine ecosystem to the oceanic 1999 regime shift.  4.2  4.2.1  Materials and methods  Sampling and laboratory processing  Zooplankton sampling was conducted during routine surveys by the DFO in support of research on herring recruitment. All sampling was done using a 0.19 m diameter Bongo net fitted with 350 µm mesh, and towed as an Oblique Net Haul (ONH) between 20 m and the surface. Sampling was conducted after sunset at nearshore stations all around the SoG (Figure 4.1) between the months of May and October, from 1990 to 2007 with a single year gap in 1995. Zooplankton were enumerated to variable taxonomic resolution between surveys, and taxa were therefore consolidated into ten major groups: euphausiids, calanoid copepods, amphipods, gelatinous zooplankton (including siphonophores, ctenophores and medusae), pteropods, copepod nauplii, larvacea, chaetognaths, decapods, and cladocerans. An additional “other” category was included for taxa not falling within these groups. Zooplankton dry weight biomass was estimated based on abundance and average individual  62  weights, and converted to g dry weight/m2 for each of the ten major groups in this study (Figure 4.2).  Potential caveats in the data need to be considered prior to proceeding with the analysis. The Bongo net used in this study may have underestimated some mesozooplankton groups, including larger more active species due to net avoidance and small copepods and copepodite stages (≤ 1 mm long) due to the 350 µm mesh used. However, the consistency of the sampling gear throughout the period of observation made this data set comparable between years and suitable for detecting interannual change. Similarly, the sampling depth of 20 m to surface is not expected to have provided samples representative of the entire water column mesozooplankton. However, the maximum in the mean chlorophyll distribution is typically located at ~10 m depth in the SoG (Masson and Peña, 2009) and fish and invertebrate predators also migrate to the surface layer at night to forage. A consistent sampling of 0-20 m depth at night should therefore be adequate to examine changes in zooplankton, especially in the context of trophic dynamics affecting juvenile fish and invertebrates. Finally, many zooplankton species may have changes in their phenology (seasonal timing) associated with climate changes (Drinkwater et al., 2010a; Mackas and Beaugrand, 2010). As June and September community patterns were similar at the level of the zooplankton groups used in this study, I considered that phenology changes were not an important aspect in this study.  63  Figure 4.1 Map of the Strait of Georgia with zooplankton sampling stations (unfilled circles, samples from deeper stations in oval areas were selected for multivariate analyses) and local factor stations: temperature and salinity stations (filled triangle, from left: Chrome Island, Entrance Island, and Nanoose Station), sea surface height in Victoria (filled circle), wind speed in Sandhead Island (square), and upwelling index in west La Push (star).  4.2.2  Large-scale climate indices  The large-scale climate indices included the extratropical-based Northern Oscillation Index (NOI), Southern Oscillation Index (SOI), extratropical-based Southern Oscillation Index (SOIx), North Pacific Index (NPI), Pacific Decadal Oscillation (PDO) and North Pacific Gyre Oscillation (NPGO) (Table 4.1). The NOI, SOIx and SOI are the differences of sea 64  level pressure between Darwin (Australia) and the North Pacific High in the northeast Pacific (NOI), the South Pacific High in the southeast Pacific (SOIx), and the Tahiti station in the tropical Pacific (SOI) (Schwing et al., 2002). Time series data for these indices were downloaded from the website http://www.pfeg.noaa.gov/. The NPI is the area-weighted sea level pressure over the region 30°N-65°N, 160°E-140°W, which depicts the Pacific-North American teleconnection pattern (Wallace and Gutzler, 1981) and changes in the intensity of the Aleutian low in winter (Trenberth and Hurrell, 1994). Monthly time series of the NPI were downloaded from http://www.cgd.ucar.edu/ . In addition to these atmospheric indices, I also included two major oceanic indices in the North Pacific, the Pacific decadal Oscillation (PDO) and North Pacific Gyre Oscillation (NPGO). The PDO is the leading principal component (PC) of monthly SST anomalies in the North Pacific Ocean (Mantua et al., 1997) and is available at http://www.atmos.washington.edu. The NPGO is defined as the second PC (the first PC is highly correlated with the PDO) of sea surface height (SSH) anomalies over the region (180°W – 110°W; 25°N-62°N) of the Northeast Pacific (Di Lorenzo et al., 2008). Time series of the NPGO are available at http://www.o3d.org/.I calculated the average value for each season per year for each of the large-scale climate indices, providing a total of 24 climate index time series. The winter index was an average from the previous December to February, spring from March to May, summer from June to August, and autumn from September to November. A one-month later seasonal classification (e.g. winter from January to March) was also applied but had weaker correlations with zooplankton changes.  65  4.2.3  Local physical factors  The local physical variables included temperature, sea surface salinity (SSS), wind speed, water flow of the Fraser River, sea surface height (SSH), Bakun upwelling index and spring transition day (see Table 4.2 and stations in Figure 4.1). Monthly time series of sea surface temperature (SST) and SSS from Chrome Island and Entrance Island lighthouses were obtained from the DFO website http://www.pac.dfo-mpo.gc.ca/science/oceans/datadonnees/lighthouses-phares/index-eng.htm . I used an average of the two lighthouses to represent SST and SSS in the SoG. In addition to SST, average water column temperature for three depth strata were included from semi-monthly observations at the Nanoose station (Masson and Cummins 2007): surface layer (TS, 0-51m), intermediate layer (TI, 51-201m) and deep water (TD, below 201m). Similar depth layers were classified in Pawlowicz et al (2007). Hourly wind speed was obtained at Sandheads (49° 6' N, 123° 18' W) (Allen and Wolfe, accepted). I calculated the cubic speed as a proxy for the strength of wind mixing and the relative mixed layer depth (Collins et al., 2009; Hsieh et al., 2009; Bakun and Parrish, 1991). As with the climate indices, I calculated seasonal averages for temperature, SSS and wind speed cubed each year. Time series of daily water flow of the Fraser River at Hope (49°22' 50"N, 121°27'5"W) were obtained from Environment Canada at http://www.wsc.ec.gc.ca/. In addition to average seasonal water flow, I identified the largest flow volume as the peak flow and the Julian Day of the peak flow as the peak flow day (PFD) for each year, giving a total of six indices related to water flow. Water flow peaked usually in May and June during my study period. When exactly the same peak flow occurred on two different days within one year, I used the first day as the PFD. The annual-averaged  66  SSH at Victoria (Crawford and Irvine, 2010) was included because of a correlation found with coho salmon production (Beamish et al., 1999). As the outer coast upwelling contributes to the variability of salinity, temperature and nutrients in the SoG (Masson and Cummins, 2007; Masson, 2002; Mackas and Harrison, 1997), an average of the Bakun upwelling index in spring and summer was included as a local environmental parameter. This index is derived from 48°N 125°W, west of La Push, Washington State (USA), and monthly values were downloaded from the Pacific Fisheries Environmental Laboratory (PFEL), National Marine Fisheries Service's Southwest Fisheries Science Center, http://www.pfeg.noaa.gov/products/PFEL/ . The beginning of upwelling, the spring transition day (STD) at the same location, was updated by Van Holmes with additional information in (Van Holmes, 2007). Thus, a total of 34 local time series were included in this study.  4.2.4  Data analysis  Prior to multivariate analysis, the zooplankton biomass data were square root transformed. The log-transformation strongly decreases the influence of the dominating groups and hence increases the relative influences of rare groups of zooplankton. Thus, this commonly used transformation introduces bias from the perspective of the trophic interactions due to the important roles of the dominating groups in the food web. A square-root transformation was therefore used to better balance the weighting of both dominating and rare groups of zooplankton on the similarity of samples. Similarities between years were then calculated  67  based on the Bray-Curtis similarity index. The Bray-Curtis similarity (Sjk), between the jth and kth year, is defined as:  ∑ ∑  (4)  where yjk represents the biomass in the ith taxa and jth year of the data matrix (i=1,2,…,p; j= 1,2,…, 9 (for June samples),…,15 (for September samples)). Similarly, yjk is the biomass for the ith taxa in the kth sample.  Analysis of Similarity (ANOSIM) was used to assist the selection of data to be used for the time series analysis. ANOSIM is analogous to an ANOVA, but instead of being computed on a univariate index, it contrasts community composition within and between groups of samples by comparing their rank similarities (Clarke and Green, 1988), in this case based on the Bray-Curtis metric. I investigated factors that may influence the community structure of samples, including stations, region (north and south SoG), bottom depth, sampling time (nautical twilight and full darkness), and month. A significant month effect indicated that samples could not be combined across seasons, and further analysis therefore focused on the September and June samples which had the most consistent coverage across years. ANOSIM results showed no significant difference in any of the above mentioned factors after omitting shallow stations and occasionally sampled stations, reducing the data set to 167 September samples and 151 June samples. Zooplankton biomass was averaged for each month of each year. As only nine years (1990-94, 1996-97, 2000, and 2001) were sampled in June, these samples were primarily used to investigate if the changes in zooplankton communities in 68  early summer were similar to those in early autumn. The September samples were further used to investigate long-term (1992-2007) physical-biological connections.  Year-to-year changes in the zooplankton community were investigated using three multivariate methods to avoid bias in each method: Cluster analysis, NMDS and PCA. The former two methods (Clarke, 1993) are based on the Bray-Curtis similarity while PCA (Chatfield and Collins, 1980) is based on dissimilarity (Euclidean distance). Cluster analysis was completed using group averages of year similarities. By contrast, NMDS is a non-metric ordination which ranks the similarities between each pair of years, where a shorter distance between the two years means higher similarity. Clarke (1993) argued that cluster and NMDS are more suitable for biological analyses because neither requires assumptions of linearity. However, PCA is extensively used in biological studies (e.g. Mackas et al., 2001; Hare and Mantua, 2000). I used these three methods to identify results which are robust to their different assumptions. After the year-clusters were identified, a post hoc Similarity Percentage (SIMPER) analysis was carried out to assess the percent contribution of zooplankton groups to the year-clusters.  In the case of physical variables, time series were normalized to a mean of 0 and standard deviation of 1 for the time period 1992-2007, with 1995 removed (due to no zooplankton sampling in 1995) before further analyses. Euclidean distance (rather than Bray-Curtis) was used as the resemblance measure for the physical variable time series.  69  Lag correlations were calculated between physical conditions and biological responses up to a maximum of two years. The lag was based on the life cycle duration of plankton groups. Most mesozooplankton have a one year life cycle. Euphausia pacifica, the dominant species of euphausiids in the region, has a maximum life span of two years in the North Pacific (Siegel, 2000). Additionally, Cloern et al. (2010) reported a time lag of two years from atmospheric forcing to the PC1 of biological changes of multiple trophic levels in San Francisco Bay. A maximum of a two-year lag was therefore adequate in this study.  Physical-biological relationships were further investigated using two methods: linear correlation between PC scores of physical factors and PC scores of zooplankton communities, and the non-linear Bio-Env procedure. For the former method, PCA was performed for 24 large-scale climate indices and 34 local environmental variables, separately, both with 0, 1 and 2 year lags. When a PC explained a small amount of variance in the large-scale climate indices or local environment but was strongly correlated with the PC scores of zooplankton communities, a subset of variables was selected, depending on the loadings to that PC, and used in an additional PCA. The non-linear Bio-Env procedure in PRIMER (Clarke and Ainsworth, 1993) was used to identify the “best” combinations of environmental indicator(s) explaining zooplankton community changes. I used the combination of the 24 large-scale and 34 local physical variables with up to two-year lags, giving a total of 174 variables as my physical factor dataset. First, a similarity matrix of each combination of k-variable(s) (k can be a natural number between 1 and 173 inclusive; k=1 represents the single variable) out of 174 physical factors was calculated based on Euclidean distance estimates. The k-variable combinations of physical factors were then ranked through  70  the Spearman correlation coefficient (r) between the zooplankton Bray-Curtis similarity matrix and the k-variable physical factor similarity matrix. Starting from 1, I increased k by 1 at a time. Further analysis was stopped when r dropped at an increased k. An ordination based on the best environmental indicators would group years in the same way as for the zooplankton and thus produce the highest r. However, the match will also worsen if abiotic factors irrelevant to the zooplankton community structure are included (Clarke and Warwick, 2001). Finally, the role of the best single environmental indicator was further investigated through direct correlation with the dominant zooplankton group.  4.3  4.3.1  Results  Large changes in September zooplankton community  The ten zooplankton groups used in this analysis comprised 89% of the total zooplankton square-rooted biomass over the study period and were thus representative of the bulk of the early autumn (September) zooplankton community (Figure 4.2). Euphausiids, amphipods, calanoid copepods and gelatinous plankton were the dominant contributors to average total zooplankton biomass, and also played important roles in inter-annual variation in zooplankton community structure (Figure 4.3).  71  DECAP CLADO 1% CHAET 2% Others 2% 11% LARVACEA  EUPH 22%  COPNAU 4% 4% PTER 5% GELA 15%  AMPH 18% CALA 16%  Figure 4.2 Average proportion of total zooplankton dry biomass (square-root transformed) contribution by ten zooplankton taxa and “other” in September over the study period 1992-2007, with a year gap in 1995: euphausiids (EUPH), calanoid copepods (CALA), amphipods (AMPH), gelatinous zooplankton (GELA), chaetognaths (CHAET), decapods (DECAP), copepods nauplii (COPNAU), larvacea, pteropods (PTER), cladocerans (CLADO).  72  Figure 4.3 Interannual variation in the September biomass (%) contribution of the four dominant zooplankton groups in the SoG. Group names are the same as in Figure 4.2.  Cluster analysis identified two groups of years based on the biomass distribution of zooplankton taxa, separated at the 56 % level of similarity (Figure 4.4 a). The “1990s” cluster comprised the years 1992-1998, 2002 and 2005, while the “2000s” cluster comprised the years 1999-2001, 2003, 2004, 2006 and 2007. These year-clusters were replicated by the NMDS, the separation occurring largely along the vertical direction of the diagram (Figure 4.4b). In the PC analysis, PC1 showed a shift from negative scores before 1999 to predominantly positive scores after 1999 (Figure 4.5), accounting for 63% of the interannual variations with high loadings by euphausiids (Eigenvector value 0.97). PC2 showed a shift in 1998 from negative scores before 1998 to mostly positive scores after 1998 (Figure 4.5), accounting for 19% of the interannual variation, with strong negative loadings by calanoid copepods (Eigenvector value -0.72) and positive loadings by gelatinous zooplankton 73  (Eigenvector value 0.63). Negative loadings on the PC2 scores in 2002 and 2005 supported the grouping of these years with pre-1999 samples in the cluster and NMDS analyses. Furthermore, I redid PCA using a log-transformation of biomass for each zooplankton group, with similar results (PC1 shifted in 1999 and explained 66% of variability; PC2 shifted in 1998 and explained 16% of variability). This confirmed that the shifts in zooplankton communities are independent of the biomass transformation. Therefore, the results of cluster, NMDS and PCA, with different biomass transformations, all supported a shift in the autumn SoG zooplankton community structure in (or near) 1999.  Substantial differences were observed in the zooplankton community structure between the two year-clusters (Figure 4.6). Euphausiids dominated the zooplankton community in the “1990s” group, followed by calanoid copepods and amphipods. The decline in the biomass of these three zooplankton groups and a slight increase in the contribution of gelatinous zooplankton biomass were the most prominent features of the shift to the “2000s” grouping. SIMPER analyses found that euphausiids, calanoid copepods, amphipods and gelatinous zooplankton contributed 80% to the dissimilarity between the two year-clusters.  74  (a)  (b)  Figure 4.4 Between year (a) cluster analysis and (b) Non-metric Multi-Dimensional Scaling ordination of September zooplankton communities, based on Bray-Curtis similarities of square-root transformed biomass data.  75  Figure 4.5 PC1 (a) and PC2 (b) score from principal component analyses of the ten zooplankton  1 "1990s" "2000s" Contribution 0.5  0 CLADO  DECAP  CHAET  LARVACEA  PTER  COPNAU  GELA  AMPH  CALA  EUPH  Biomass in September (√g/m2); Contribution to the disimilarity  groups time series.  Figure 4.6 Average dry biomass (Square-root transformed) of each zooplankton group in September for the two year-clusters identified by cluster analyses (Figure 4.4 a), i.e., “1990s” and “2000s”. Triangles indicate the results of SIMPER analysis, the cumulated contribution of each zooplankton group to the dissimilarity between the two year-clusters. Group names are the same as in Figure 4.2.  76  The anomalies of the total zooplankton biomass, and of each zooplankton group, showed different amounts of shift in 1999 (Figures 4.7 and 4.8). The total zooplankton biomass showed largely positive anomalies before 2000 and negative anomalies in the 2000s (Figure 4.7). Considering individual zooplankton groups, euphausiid and calanoid copepod biomass shifted from primarily positive to primarily negative anomalies in 1999 and 1998 respectively (Figure 4.8 a and b). For both of these zooplankton groups, a student’s t-test showed that the decline in average biomass after the shift year was significant (P=0.01 for euphausiids, P<0.001 for calanoid copepods). This was consistent with the findings of PC1 and PC2 respectively. It is worth noting that calanoid copepods showed a strong positive biomass anomaly in 2002 (Figure 4.8 b) which contributed substantially to the clustering of that year in the “1990s” group. Amphipod biomass had positive anomalies in most years through 1999 and negative values thereafter, with the exception of 2005 (Figure 4.8 c). Strong positive biomass anomalies of amphipods in 2005 contributed to clustering of that year with the “1990s” group. Gelatinous zooplankton showed strong interannual variability with peaks in 1998, 1999 and 2003 contributing to the increase of this group in the cluster of the “2000s” (Figure 4.8 d and Figure 4.6). Among the six rare groups, larvacea and decapods showed clearer shifts than the other groups (Figure 4.8). Notably, larvacea is the only group that had positive anomalies of biomass in most years of the 2000s (Figure 4.8 j).  77  Figure 4.7 Anomalies for the total biomass of ten dominant zooplankton groups in September. The sum of the ten square-rooted biomass zooplankton groups was normalized to a mean of 0 and standard deviation of 1.  Figure 4.8 September biomass anomalies of euphausiids (a), calanoid copepods (b), amphipods (c), gelatinous zooplankton (d), chaetognaths (e), decapods (f), cladocerans (g), copepods nauplii (h), pteropods (i), larvacea (j). The square-rooted biomass of each group was normalized to a mean of 0 and standard deviation of 1.  78  4.3.2  Large changes in June zooplankton communities  Early summer (June) zooplankton communities were also different between the 1990s and 2000s, separating into two clusters of years, the “1990s” (1990-1994, and 1996) and “2000s” (1997, 2000 and 2001) (Figure 4.9 a). The four major zooplankton groups, euphausiids, calanoids, amphipods and gelatinous zooplankton, dominated the water column in early summer, as they did in autumn (Figure 4.9 b). The biomass of the four dominating zooplankton groups decreased considerably between the “1990s” to the “2000s” cluster, contributing 81% to the dissimilarity between the two year-clusters. The consistency of the interannual changes observed in June and September samples supports the conclusion that these changes were not an artifact of seasonality but a reflection of a long-term persistent shift in zooplankton structure.  79  Biomass in June (√g/m2); Contribution to the disimilarity  (a)  1  (b)  "1990s" "2000s" Contribution  0.5  0 CLADO  PTER  LARVACEA  COPNAU  DECAP  CHAET  GELA  AMPH  CALA  EUPH  Figure 4.9 Cluster analysis of the zooplankton community in June (a) and average biomass of each zooplankton group in June in the “1990s” and the “2000s” (b). Triangles in (b) indicate the SIMPER results, the cumulated contribution of each zooplankton group to the dissimilarity between the two year-clusters. Group names are the same as in Figure 4.2.  80  Table 4.1 Summary of PCA results of 24 large-scale climate indices with up to two-year lags including the total variance explained by each PC, the correlation and P-value of each PC with PC1 score of zooplankton communities, and eigenvalue of each index. P≤0.05 indicated in boldface.  Variance Correlation P-value Eigenvalue SOIspring SOIsummer SOIautumn SOIwinter PDOspring PDOsummer PDOautumn PDOwinter NOIspring NOIsummer NOIautumn NOIwinter SOIxspring SOIxsummer SOIxautumn SOIxwinter NPIspring NPIsummer NPIautumn NPIwinter NPGOspring NPGOsummer NPGOautumn NPGOwinter  PC1 0.41 -0.6 0.01  Lag 0 PC2 PC3 0.21 0.11 -0.3 0.0 0.30 0.92  Lag 1 PC1 PC2 PC3 0.39 0.24 0.12 -0.8 0.2 0.2 <0.001 0.37 0.47  PC1 0.34 -0.4 0.20  Lag 2 PC2 0.25 -0.1 0.74  PC3 0.12 -0.2 0.55  -0.2 -0.2 -0.2 -0.2 0.2 0.3 0.2 0.1 -0.3 -0.2 -0.1 -0.2 -0.3 -0.1 -0.2 -0.2 -0.2 0 0.1 0 -0.3 -0.3 -0.3  0 0.4 0.3 -0.3 0.2 -0.1 -0.2 0.1 -0.1 0.3 0.3 -0.2 0.1 0.2 0.3 -0.2 -0.2 0.2 0.1 -0.2 -0.1 -0.1 -0.1  -0.3 0 -0.1 -0.2 0 0 0.2 0.1 -0.1 0 -0.2 -0.2 -0.2 0.2 0 -0.1 0.2 0 -0.4 -0.4 0.2 0.2 0.2  -0.2 -0 .1 -0.2 -0.2 0.1 0.2 0.3 0.3 -0.2 -0.3 -0.2 -0.1 -0.2 -0.3 -0.1 -0.1 -0.2 -0.1 0.1 -0.1 -0.3 -0.3 -0.3  0 0.3 0.3 -0.2 0.2 0.2 0 -0.2 -0.3 -0.1 0.3 0.3 -0.1 0.1 0.2 0.3 -0.2 0.2 0.2 -0.2 -0.1 -0.1 -0.1  0.4 -0.1 0.1 0.2 0 0.1 0.1 -0.1 0.2 0.1 -0.1 0.2 0.2 0.3 -0.4 0.1 -0.2 -0.1 0.4 0.3 -0.1 -0.2 -0.2  -0.3 -0.1 -0.2 -0.2 0.1 0.2 0.3 0.2 -0.2 -0.3 -0.2 -0.1 -0.2 -0.3 0 -0.1 -0.1 0 0 -0.1 -0.3 -0.3 -0.3  -0.1 -0.4 -0.3 0.2 -0.1 -0.2 0.1 0.2 0.2 0.1 -0.3 -0.3 0.1 -0.2 -0.2 -0.3 0.3 -0.2 -0.2 0.1 0 0.1 0.1  0.3 0 0 0.2 -0.1 0 0 -0.2 0.2 0.2 -0.1 0.2 0.1 0.1 -0.3 0 -0.3 -0.1 0.3 0.3 -0.2 -0.2 -0.2  -0.2  0  0.4  -0.3  0  -0.3  -0.3  0  -0.3  81  4.3.3  Physical indicators associated with changes in September zooplankton  communities  PCA was used to identify the most important patterns of shared variability in the 24 largescale climate indices and 34 local physical variables, at 0, 1 and 2-year lags. Like the zooplankton community, PC1 scores of the large-scale climate indices with no lag and 1-year lag showed a shift in 1999, accounting for 41% and 39% of the total variance in large-scale indices, respectively (Figure 4.10 a and b, Table 4.1). However, among the PCs of all the climate indices and local physical variables with different lags, one-year lag PC1 had the highest correlation (r = 0.8, P<0.001) with PC1 of zooplankton samples (Table 4.1). A student’s t-test confirmed significant differences (p= 0.001) in PC1 scores of one-year lagged climate indices before and after 1999. Strong negative loadings, with a one-year lagged PC1, were observed on the NOIsummer, SOIxsummer, and the four seasons of NPGO, while strong positive loadings were observed for autumn and winter PDO (Table 4.1).  82  Table 4.2 Summary of PCA results of 34 local measurements with up to two-year lags including the variance explained by each PC, the correlation and P-value of each PC with PC1 score of zooplankton communities, and eigenvalue of each variable. P≤0.05 indicated in boldface. PC1 0.33 Variance Correlation -0.2 P-value 0.41 Eigenvalue Windspring -0.1 Windsummer 0.1 Windautumn 0.1 Windwinter -0.1 SSTspring 0.2 SSTsummer 0.2 SSTautumn 0.1 SSTwinter 0.2 SSSspring 0 SSSsummer 0.2 SSSautumn 0.2 SSSwinter 0 Upwellspring 0.1 Upwellsummer 0.2 STD -0.1 Flowspring 0 Flowsummer -0.1 Flowautumn -0.1 Flowwinter 0.1 Peak Flow -0.3 PFD -0.1 SSH 0.1 TSspring 0.3 TIspring 0.2 TDspring 0 TSsummer 0.3 TIsummer 0.3 TDsummer 0.2 TDautumn 0.3 TSwinter 0.2 TIwinter 0.1 TDwinter 0  Lag 0 PC2 PC3 0.15 0.12 0.0 0.4 0.97 0.14 0.2 -0.1 0.1 0.3 0 0 0 0.1 -0.3 -0.3 -0.2 -0.2 -0.1 0 0 0.2 0.3 0.1 0 0.1 -0.2 0.1 0 0.2 0.4 0.1 0.1 0.2 0 0.2 0.3 0.3  0.2 0.2 0.1 0.3 -0.2 -0.1 -0.2 0.2 0.3 0.1 0.2 0.2 0.1 0.2 -0.1 0.2 -0.1 -0.1 0.1 0 0.2 -0.3 0 0.1 0.1 0.1 -0.1 -0.1 -0.1 0.2 0.2 0.1  Lag 1 PC1 PC2 PC3 0.33 0.15 0.12 0.1 0.2 0.5 0.84 0.44 0.03  Lag 2 PC1 PC2 0.34 0.18 0.2 0.3 0.47 0.34  PC3 0.11 -0.2 0.53  -0.1 -0.1 0.2 0.1 0.2 0.2 0.1 0.2 0 0.2 0.2 0 0.1 0.2 0 -0.1 0 -0.2 -0.1 -0.2 -0.1 0.1 0.3 0.2 0.1 0.3 0.3 0.2 0.3 0.2 0.1 0  -0.1 -0.1 0.1 0.1 0.2 0.2 0.1 0.2 0 0.2 0.2 0 0.1 0.1 0 -0.1 0 -0.1 -0.1 -0.2 -0.2 0.1 0.3 0.2 0.1 0.3 0.3 0.3 0.3 0.2 0.1 0  -0.3 -0.1 -0.1 0.2 0.2 0.1 0.1 -0.2 -0.3 -0.1 -0.3 -0.2 0 -0.2 0 0 -0.2 0.1 0 0 -0.2 0.4 0 -0.1 -0.1 -0.1 0 0.1 0.1 -0.3 -0.3 -0.2  0.2 0.2 -0.1 0 0 0 0 0.1 -0.2 -0.3 -0.2 -0.2 -0.2 0 0.1 -0.1 0.2 0.3 0.1 0.2 -0.2 0.2 0 0.2 0.4 0.1 0.1 0.2 0 0.2 0.3 0.3  0.3 0.1 0.1 0.1 -0.2 -0.1 -0.2 0.2 0.2 0 0.2 0.1 0.2 0.3 -0.1 0.1 0.2 0 0 0 0.2 -0.3 0 0.1 0.1 0.1 -0.1 -0.1 -0.1 0.2 0.2 0  0.2 0.2 -0.1 -0.1 0 0 0.1 0 -0.2 -0.2 -0.2 -0.2 -0.3 -0.1 0.1 -0.1 0.2 0.3 0.1 0.1 -0.2 0.2 0 0.2 0.4 0 0.1 0.2 0 0.1 0.2 0.3  83  A weaker correlation with zooplankton changes was detected with the PC scores of the 34 local physical factors. The highest correlation was seen in the PC3 scores at 0 and 1- year time lags (r= 0.4 and 0.5 and p= 0.14 and 0.03, respectively), each accounting for 12% of the variance (Table 4.2). There was no significant correlation between the first two PC scores and the zooplankton community although these two PCs accounted for 48% of the variability in the physical data. Both the 0 and 1-year time lag PC3s had their strongest negative loadings on SSH and the TSautumn. The strongest positive loadings were on Windwinter and SSSspring for 0-year lagged PC3 and Windspring and Upwellsummer for 1-year lagged PC3. I redid the PCA selecting a subset of local measurements with absolute eigenvalues in the upper quartile of PC3 at both 0 and 1-year time lags respectively. However, none of the first three PCs of the selected local measurements, at both 0 and 1-year time lags, had significant correlation with zooplankton changes (Table 4.3 a) indicating poor covariance among these factors. A further PCA was conducted using a subset of local measurements with absolute eigenvalues larger than the median of absolute eigenvalues in the PC3 of both 0 and 1-year time lags respectively. The PC2 scores (Figure 4.10 c and d), at both time lags, had significant correlation with zooplankton community changes, explaining 22% of the variance (r=-0.55 and 0.51, P= 0.03 and 0.05, respectively, Table 4.3), but were not as strongly correlated as the large-scale climate indices.  84  Figure 4.10 The first principal component score from a principal component analysis of the 24 largescale climate indices with no lag (a), a 1-year lag (b), and the second principal component of selected local measurements (please see a list in table 4.3 b) with no lag (c) and 1-year lag (d).  The Bio-Env procedure identified 1-year lagged SOIxspring (r = 0.43) (Figure 4.11 a) as the best single indicator of zooplankton community change, followed by the 1-year lagged SOIspring (r = 0.34), and the previous PDOautumn (r = 0.31) (Table 4.4). Only 6 local factors (PFD0, 1-year lagged TIwinter, Windautumn, 1-year lagged TSwinter, 1-year lagged Flowsummer, 1year lagged TDspring) were selected in the best twenty single indicators of zooplankton community change, the highest correlated being PFD (r = 0.27, Table 4.4), which was ranked fifth overall. This was consistent with the PCA analyses which demonstrated that large-scale climate indices were more strongly correlated with zooplankton changes. The combination of 1-year lagged SOIxspring and PFD was the best two-variable indicator of zooplankton community change (r = 0.54). The best 3-variable combination was the 1-year lagged SOIxspring, NPGOsummer and windautumn (r = 0.59). The correlation coefficient dropped slightly 85  to 0.58 for the best 4-variable combinations. In addition to windautumn and PFD, TSwinter was another local factor contributing to the best five 2 (or 3)-variable combinations. The commonly used local factor SST was not a good single or combined indicator in this study.  Covariance among the local physical factors appeared to be weaker than among the largescale climate indices. For example, in the 0-year lag PCA (Tables 4.1, 4.2 and 4.3), PC1 of the large-scale climate indices explained more than 40% of the variability while the PC1 of the 34 local factors explained 33% of the variability. With fewer numbers of selected factors, PC1 of the selected local measurement did not explain higher variability. This indicates that interannual changes in the local environment were less coherent than largescale climate indices. The high coherence amongst large-scale climate indices was apparent in the high correlation between large-scale climate indices (Table 4.5).  I note that the environmental factors were correlated at different levels (Table 4.5). The high correlation between PFD and seasonal NPGO indices supports the idea that both the NPGO and the PFD are correlated with the SOI (Di Lorenzo et al., 2010; Foreman et al., 2001). Unlike Chesapeake Bay, where water flow is dominated by precipitation resulting from the regional weather variability (Kimmel et al., 2009), the peak flow of the Fraser River is controlled by snow melt driven by the large-scale atmospheric forcing (Morrison et al., 2002; Foreman et al., 2001).The correlation was also very high between the 1-year lagged SOIxspring and the 1-year lagged SOIspring (Schwing et al., 2002). This may explain why SOIxspring is absent in the 2 or 3-variable combinations. In cases with similar values of rand correlation  86  Table 4.3 Summary of PCA results of selected local measurements with 0 or 1 lag, whose absolute eigenvalues are in the upper quartile (a) or more than the median (b) of the PC3 of 34 local measurements with lag 0 and 1. The values in the table present the variance explained by each PC, the correlation and P-value of each PC with PC1 score of zooplankton communities, and eigenvalues of each variable. P≤0.05 indicated in boldface. (a) Variance Correlation P-value Eigenvalue Windwinter SSTspring SSTwinter SSSspring SSSautumn Upwellsummer PFD SSH TSautumn  PC1 0.32 0.39 0.15  Lag 0 PC2 0.29 0.29 0.30  PC3 0.18 -0.20 0.47  0.19 -0.37 0.09 0.40 0.23 0.20 0.39 -0.46 -0.45  0.11 -0.36 -0.52 -0.23 -0.51 -0.43 0.20 -0.03 -0.22  0.61 -0.21 0.17 -0.42 0.00 0.36 -0.26 0.31 -0.29  PC1 0.23 0.14 0.63  Lag 0 PC2 0.22 -0.55 0.03  PC3 0.17 0.09 0.76  -0.04 -0.24 -0.13 -0.12 -0.10 -0.46 -0.18 -0.41 -0.03 -0.42 -0.14 0.11 0.05 -0.01 -0.43 -0.30  -0.16 -0.07 -0.20 0.38 0.30 0.03 -0.32 -0.04 -0.26 -0.03 -0.12 -0.42 0.40 0.41 0.08 0.02  0.34 -0.19 0.45 -0.19 -0.05 -0.05 -0.33 -0.26 -0.27 -0.05 0.37 -0.16 0.15 -0.17 0.14 0.35  (b) Variance Correlation P-value Eigenvalue Windspring Windsummer Windwinter SSTspring SSTautumn SSTwinter SSSspring SSSautumn SSSwinter Upwellsummer Flowspring PFD SSH TSautumn TSwinter TIwinter  Windspring SSSspring SSSautumn Upwellsummer Flowspring SSH TSautumn TSwinter  Windwinter Windsummer SSTspring SSTautumn SSTwinter SSSspring SSSautumn Upwellspring Upwellsummer STD Flowspring PFD SSH TSautumn TIautumn TSwinter TIwinter  PC1 0.34 0.46 0.09  Lag 1 PC2 0.23 0.19 0.51  PC3 0.20 -0.22 0.42  0.20 0.38 0.49 0.46 0.16 -0.38 -0.28 0.35  0.62 -0.39 -0.32 0.03 0.52 0.11 -0.25 0.15  0.01 -0.31 0.20 0.42 -0.20 0.54 0.37 0.47  PC1 0.26 0.04 0.88  Lag 1 PC2 0.22 0.51 0.05  PC3 0.15 -0.28 0.31  0.06 -0.25 -0.32 -0.18 -0.39 -0.07 -0.30 -0.04 -0.29 0.03 0.06 0.28 -0.13 -0.26 -0.44 -0.30 -0.14  0.25 0.20 -0.23 -0.19 0.20 0.27 0.27 0.20 0.30 -0.15 0.21 0.27 -0.37 -0.35 -0.16 0.23 0.15  -0.40 0.19 0.22 -0.10 -0.06 0.28 0.11 0.34 -0.09 -0.28 -0.21 0.30 -0.17 0.10 0.13 -0.25 -0.44  87  Table 4.4 The K (one, two or three)-variable combinations of environmental variables, taken K at a time, yielding the best five matches of biotic and abiotic similarity matrices for each K, as measured by Spearman rank correlation r with the value in the (). Subscripted numbers represent the year time lag. Bold font indicates the combination with the overall highest r value. K Best variable combinations 1 SOIxspring1 (0.43), SOI spring1 (0.34), PDOautumn1 (0.31), SOIautumn1 (0.31), PFD0 (0.27) SOIxspring1 and PFD0 (0.54), SOIxspring1 and NPGOsummer0 (0.51), SOIxspring1 and NPGOspring0 2 (0.51), PFD0 and TSwinter1 (0.49), SOIxspring1 and windautumn0 (0.49) SOIxspring1, NPGO summer0 and windautumn0 (0.59); SOIxspring1, NPGO spring0 and windautumn0 (0.57); SOIxspring1, PFD0 and TSwinter1 (0.57); SOIxspring1, NPGOspring0 and TSwinter1 (0.55); 3 SOIxspring1, PFD0 and TSwinter2 (0.55);  Table 4.5 Correlation between the environmental indicators of Table 4.4. Absolute values >0.7 indicated in boldface.  PDOautumn0 PDOautumn1 SOIxspring1 NPGOspring0 NPGOsummer0 NPGOautumn0 Windautumn0 PFD0 TSwinter1 TSwinter2  SOI  SOI  PDO  PDO  SOIx  NPGO  NPGO  NPGO  Wind  spring1  autumn1  autumn0  autumn1  spring1  spring0  summer0  autumn0  autumn0  -0 -0.6 0.89 0.3 0.34 0.44 0.08 0.41 0.44 0.63  -0.3 -0.8 0.72 0.55 0.51 0.5 -0.1 0.56 0.48 -0  0.19 -0 -0.4 -0.5 -0.4 -0.3 -0.2 -0.2 -0.2  -0.7 -0.6 -0.6 -0.6 0.14 -0.6 -0.5 -0.1  0.42 0.44 0.52 0.02 0.42 0.41 0.36  0.96 0.05 0.75 0.11 0.08  0.12 0.78 0.21 0.16  0.96 0.93 -0 0.79 0.07 0.04  -0.1 0.05 0.07  PFD0  TS winter1  0.05 0.11  0.14  88  coefficient (r) and high correlation among different variables, it might be very important to identify which combination/single indicator is better than the other. On the other hand, the cubed autumn wind speed (windautumn) had the weakest correlation with most factors (Table 4.5). However, combined with the 1-year lagged SOIxspring, it also provided a good indicator to identify which combination/single indicator is better than the other. On the other hand, the cubed autumn wind speed (windautumn) had the weakest correlation with most factors (Table 4.5). However, combined with the 1-year lagged SOIxspring, it also provided a good indicator of zooplankton community change, as was the case with the NPGO and PFD. Therefore, several indices may have driven zooplankton community structure, indicating multidimensional changes in the environment, multi-dimensional responses in the zooplankton communities and multiple mechanisms coupling environmental changes and zooplankton communities.  4.3.4  Correlations with euphausiid biomass  Zooplankton groups have variable associations with the 1-year lagged SOIxspring due to different life history and different ways of interacting with environmental factors. I used euphausiids, a dominant group in the zooplankton communities, as an example to examine the possible mechanisms responsible for zooplankton compositional changes. Euphausiids showed the strongest negative correlation with 1-year lagged SOIxspring (r=-0.78 and P<0.001, Figure 4.11 b). None of the local environmental parameters had as strong a statistical relationship with zooplankton as the SOIx. Nevertheless, among the time series available, there were two local physical factors that were correlated with the SOI (which is  89  Figure 4.11 Normalized environmental variables: 1-year lagged SOIxspring (a), peak flow day of the Fraser River (c), and average temperature from the Nanoose Station from January to March at depth 81-99m (e). Panels b, d, and f present these variables versus the square-rooted biomass of euphausiids.  strongly correlated with SOIx). These included Fraser River PFD (Morrison et al., 2002; Foreman et al., 2001) and the SoG deep water temperature (Masson and Cummins, 2007). Although time lags between the SOI and PFD were not studied, the earlier snow melt in warm years of negative SOI (Masson and Cummins, 2007) generally result in peak flows occurring earlier in the year (Morrison et al., 2002; Foreman et al., 2001). Unlike the consistent pattern in the 1-year lagged SOIxspring, the anomalies of PFD turned positive in  90  1999 but became negative again after 2004 (Figure 4.11 c). I examined the subsurface SoG temperature from January to March, corresponding to the 1-year lagged SOIxspring based on the 10-month lag between SOI and subsurface temperature (Masson and Cummins, 2007), at the 81-99 m depth range (the major depth layer for E. pacifica in daytime in the SoG). The temperature generally showed anomalies of opposite sign to the PFD, with the exception of a large negative year in 1997 (Figure 4.11 e).  The highest euphausiid biomass values tended to be associated with earlier PFDs, but there were also two early PFD years of low euphausiid biomass (r=0.56 and P=0.03, Figure 4.11 d). I examined three potential scenarios that could have made a difference to euphausiid biomass during early PFD years: (a) timing of the spring phytoplankton bloom (Allen and Wolfe, This issue), (b) the time lag between the timing of the spring phytoplankton bloom and the PFD, and (c) presence and timing of the leading edge of water flow (see daily water flow data in Section 2.3) that may input highly turbid (sediment loaded) water and thus collapse the spring bloom by reducing the light availability in the euphotic layer. However, none of the above indices were able to separate the high and low euphausiid biomass during the early PFD years. This indicated that PFD effects were independent of the spring phytoplankton bloom. Similar to the correlation of PFD with euphausiids, I found large variability in the biomass of euphausiids among warm years (r=0.25 and P=0.36, Figure 4.11 f). A combination of the temperature and PFD also failed to explain the large variability in euphausiid biomass during warm years.  91  4.4  4.4.1  Discussion  Synchrony of biological responses in Northeast Pacific  The changes in the SoG zooplankton community demonstrated by this study appeared to be in synchrony with changes observed in the other ecosystems of the Northeast Pacific. The SoG zooplankton community shifted in 1999 the same year as in the south and central California Current (Lavaniegos and Ohman, 2007), northern (Mackas et al., 2007; Peterson and Schwing, 2003; Mackas et al., 2001), and San Francisco Bay (Cloern et al., 2010). This biological shift in the North Pacific, in both offshore and coastal sites, coincided with a shift in large-scale climate indices as highlighted by PCA in this study. This indicates that a strong basin-scale forcing influenced all North Pacific ecosystems.  It should be noted that the shift in the SoG zooplankton communities was opposite to the shift in the California Current system, the latter experiencing an increase in the biomass of euphausiids and copepods (Lavaniegos and Ohman, 2007). This may indicate the same forcing but different mechanisms involved in these systems. Unlike the SoG (see Section 4.2 and 4.3, below), the California Current is primarily a wind driven system. Wind determines the strength of the summer upwelling, which in turn drives ecosystem dynamics (Peterson and Schwing, 2003). The warm phase of 1990-1998, characterized by high stratification and low nutrients, was followed by a sharp transition to strong upwelling induced by the 1999 La Nina, highly mixed surface waters, and increased nutrient levels (Peterson and Schwing, 2003; Whitney and Welch, 2002). My PCA results show that this large-scale climate forcing  92  pattern has persisted over the Pacific Ocean since 1999 and thus may have maintained these conditions in the California Current (Cloern et al., 2010; Lavaniegos and Ohman, 2007).  4.4.2  Large-scale climate forcing versus local factors  My results from both PCA and Bio-Env showed that changes in zooplankton communities were correlated more strongly with climate forcing over the Pacific Ocean than with local physical factors. A strong shift in 1999 was pervasive amongst the large-scale climate indices. By contrast, observation of a shift in local environmental factors was dependent on the suite of local factors considered. Overall, unlike large-scale climate indices, local factors showed low covariance. It is expected that animals will respond to local conditions (Drinkwater et al., 2010a). It is possible that by aggregating zooplankton species into groups, species specific changes were masked by compensatory changes in different species (Lavaniegos and Ohman, 2007). The scale itself can constrain recognition of the drivers (Perry and Ommer, 2003). Likely, the large scale I chose for the zooplankton communities determined the large scale of the best environmental indicators.  Large-scale atmospheric forcing incorporates a wide range of temporal and spatial scales of several variables, including temperature, wind, and precipitation, and thus can be more representative of the cumulative environmental changes rather than forcing of single local variables (Drinkwater et al., 2010a; Stenseth and Mysterud, 2005; Hallett et al., 2004; Stenseth et al., 2003; Stenseth et al., 2002). This is consistent with the best single indicator in this study being the 1-year lagged SOIxspring followed by the previous SOIspring. The eastern  93  Pacific is teleconnected across the equator with the California Current and Humboldt Current Systems that experience correlated decadal climate forcing by the atmosphere (Schwing et al., 2010).  The SOIx is a newly developed index but many studies have documented the extensive and intensive impacts of El Nino on ecosystems (Chavez et al., 2002; Stenseth et al., 2002). Considerable environmental changes have been reported in the SoG during El Nino years. For example, it has been documented that the Fraser River flows were smaller and the water temperatures were higher during El Nino years (Foreman et al., 2001). Furthermore, the late winter deep water renewal from the Pacific Ocean was shut down during El Nino years leading to warmer, oxygen depleted water in the SoG (Masson, 2002). These conditions appeared to persist for one to three years after the 1997 El Nino event, as inferred from the SoG deep water residence time (Pawlowicz et al., 2007; Masson, 2002). As oceanic water represents an important nutrient source for the SoG (Mackas and Harrison, 1997), the El Nino may also have negatively influenced the SoG nutrient levels during the following year(s). Additionally, SoG biological communities, e.g. dinoflagellates (Pospelova et al., 2010) and zooplankton (Haro-Garay and Soberanis, 2008) changed during the 1997 El Nino. Decadal scale SoG salmon production has also been linked to the SOI (Beamish et al., 1997). The observed biological changes could have been in response to any (or any combination) of the changes in the SoG local environment driven by SOIx / SOI, and it is difficult to single out any one local driver. This was statistically supported by the correlation analyses of euphausiids. I show that SOIxspring may have impacted the SoG zooplankton communities through the subsurface temperature and PFD to some extent. However, the strongest  94  correlation between the 1-year lagged SOIxspring itself and euphausiid biomass supports the conclusion that large-scale climate indices better explain animal performances than the local variables (Stenseth and Mysterud, 2005; Hallett et al., 2004).  4.4.3  Possible mechanism for the influence of the previous spring SOIx  The underlying mechanism for the influence of the previous spring SOIx can only be revealed through investigations into the role of the local environment and its connection to the large-scale atmospheric indices (Drinkwater et al., 2010b). In the case of euphausiids, hatching and larval development times are significantly reduced with increased temperature (Pinchuk and Hopcroft, 2006; Iguchi and Ikeda, 1995; Iguchi and Ikeda, 1994), although E. pacifica can survive a wide range of temperatures (Sun et al., 2011; Iguchi et al., 1993). As a consequence, E pacifica biomass may be higher in warm years because shorter development times reduce larval mortality and can lead to overall recruitment success (Rumsey and Franks, 1999), and hence enhance population biomass (Tanasichuk, 1998).  However, euphausiid larval development and growth is also dependent on the food supply (Gómez-Gutiérrez et al., 2007; Pinchuk and Hopcroft, 2006; Daly, 1990; Ross and Quetin, 1989). With no correlation found between the spring phytoplankton bloom and euphausiid biomass, I hypothesize that the second phytoplankton bloom, which is in early summer, may be a major determinant of SoG zooplankton production. Comparison of the timing of the peak of the second seasonal SoG phytoplankton bloom (Johannessen and Macdonald, 2009; Johannessen et al., 2005; Bornhold, 2000) with the PFD indicated a strong correspondence  95  between them. It is suggested that the timing and magnitude of the May-June freshet may control the entrainment of nutrients and maintain high primary productivity in late spring and early summer (Yin et al., 1997) and thus influence the growth and composition of mesozooplankton. Additionally, nutrients entrained by the peak flow can also modify the composition of the phytoplankton community, further influencing zooplankton growth through an imbalanced diet (El-Sabaawi et al., 2009). An early PFD may thus create favorable feeding conditions for the offspring of euphausiids which may spawn after the spring bloom, leading to enhanced autumn populations.  4.4.4  Implications of this study  The ten aggregated zooplankton groups used in this study represent important lower trophic levels in the foraging layer at night, and the changes observed therefore have significant implications for the SoG ecosystem. As bottom-up effects are critical for higher trophic levels (Cloern et al., 2010; Perry and Schweigert, 2008; Trites et al., 2007), the decline in prey biomass post-1999 may be an important contributing factor to the lack of recovery of predators such as coho salmon, lingcod and rockfish. My analysis indicates that the SOIx/SOI was the major overarching driver of zooplankton community change in the SoG. However, the manner in which these large-scale climate indices were transmitted through local environmental parameters to the zooplankton remains unclear. Studies on how largescale climate forcing impacts the local environment is a critical area of future research to understand the mechanisms controlling long-term ecosystem change. With ongoing observations, this dataset will soon include additional years which may help address present  96  uncertainties and test whether the previous year’s spring SOIx can be used to predict zooplankton community and ecosystem changes in the Strait of Georgia.  97  Chapter 5: Conclusion  5.1  Overview  In this thesis, I studied top-down control, invasive species and bottom-up effects on fish populations. I developed an Ecopath model for the recent SoG ecosystem and used it to examine interactions between harbour seals and fisheries (Chapter 2) and to investigate the ecological impacts of invasive species with altered physiology on the whole ecosystem (Chapter 3). In both cases, dynamics were dominated by a seals-hake-herring trophic triangle. For the bottom-up effects (Chapter 4), a series of multi-variate analyses were applied to examine shifts in zooplankton communities and their links with physical factors.  Chapter 2 reveals that a cull of harbour seals may not increase total fisheries catch in the SoG, where harbour seals and fisheries have strong interactions. An Ecopath model based on 2005 conditions showed that, in the absence of harbour seals, the total biomass of commercial fish populations would substantially decrease. Herring, especially the major fishery in the SoG, showed the greatest decline because seals also prey on hake, herring’s largest predator. With seals absent, the SoG ecosystem may be dominated by hake. In order to test a more realistic policy for seal control, I further ran the model with a small seal hunting mortality of 2% per year. In the results, herring still showed large decline in biomass and hake still increased in biomass (Figure 5.1). Opposite to the increase in lingcod in the seal removal scenario (Figure 2.5), lingcod biomass declined gradually (Figure 5.1). Like the  98  2.0 1.5 1.0 0.5 0.0  Predicted biomass in 2035 relative to 2005  2005  2010  2015  2020  2025  2030  2035  Year  Figure 5. 1 Predicted biomass of the main functional groups relative to their biomass in 2005 after a fixed seal hunting rate of 2% per year in the Strait of Georgia. Increasing groups in 2035 are: soles (orange), hake (green), dogfish (blue) and pollock (light blue). Decreasing groups in 2035 are: seal (black), lingcod (purple), sand lance (grey), and herring (red).  harbour seals complete removal scenario, a partial culling of harbour seals still may not increase the commercial catch in the SoG. However, this conclusion may not be robust against very low values for juvenile herring vulnerabilities to predators. Ecosim sensitivity analyses showed that the herring population was less negatively impacted by seal removals when juvenile herring were less vulnerable to hake predation (i.e., when we assume there are many refuges in which to hide). This indicated that juvenile survival is important to herring  99  abundance. My model also suggested that the seal biomass was positively correlated with herring availability.  Chapter 3 suggests that GH coho, an invasive species with altered physiology, would impact the whole ecosystem largely through indirect interactions. A modified SoG Ecopath model (Li et al. 2010) showed that increases in GH coho would impact many functional groups depending on GH coho diet. However, these impacts were weaker when a bottom-up effect was introduced by changing ocean conditions. Changes in the functional groups followed the trends forced by this bottom-up effect. Different assumptions about GH coho diets only influence GH coho and the wild salmon biomass. Under a variety of scenarios, harbour seal biomass was positively correlated with herring biomass and negatively correlated with hake biomass. This is consistent with the results seen in Chapter 2. The harbour seal-hake-herring trophic triangle remained stable despite the direct or indirect interactions of GH coho with each component. Model sensitivity analyses showed that the model predictions were robust to uncertainty in GH coho parameters. In spite of the wide range of production rates and consumption rates considered, they were less sensitive than predator-prey vulnerabilities. High prey vulnerabilities to GH coho, which is a top-down control by GH coho, produced the most different predictions. When the GH coho were forced to keep high biomass, the parameter uncertainty impacted the changes in the biomass of most functional groups. When the GH coho were present at low biomass with no forcing, the parameters uncertainty mainly influenced the biomass of GH coho and wild salmon.  100  Chapter 4 highlights an abrupt shift in September zooplankton composition in 1998/1999 and its strong correlation with large-scale climate indices. The biological shift was based on night zooplankton samples collected from the top 20m between 1990 and 2007 using three different methods of cluster analysis, and NMDS and PCA. Dramatic declines were observed particularly in euphausiids and copepods, the major preys of many fish and invertebrates. The PCA analysis showed a dominating phase shift in the large-scale climate indices in 1999. By contrast, the local environmental factors were less coherent. The best single indicator of zooplankton community change was the spring extratropical-based Southern Oscillation Index (SOIx) with a one-year lag. The best single local indicator was the peak flow day of the Fraser River despite explaining less variability in zooplankton than SOIx. Both linear and non-linear correlation shows that changes in zooplankton communities were correlated more with large-scale climate forcing than with local factors. As secondary production plays a key role in food webs, these large changes in the zooplankton community may have had a substantial impact on higher trophic levels and the ecosystem as a whole.  5.2  Ecosystem functioning  My study highlights the important role of indirect trophic interactions and trophic cascades in structuring the ecosystem. Whether altering the euphausiid biomass, removing harbour seals, or introducing an invasive species, changes in biomass of one species or group may change the whole ecosystem. Top-down effects have also been observed in many ecosystems. For example, small pelagic fish (and sometimes decrease in zooplankton) increase after removing large predatory cod (Cloern et al., 2010; Bundy et al., 2009; Casini et al., 2009; Frank et al.,  101  2005). Bottom-up effects can cascade up to fish and even to marine mammals (Cloern et al., 2010; Trites et al., 2007). The importance of resource limitation by bottom-up control and predation by top-down control has been assessed in regulating ecosystems (Sinclair et al., 2003; Shurin et al., 2002; Hunter and Price, 1992). The conception of top-down, bottom-up and indirect interactions has become increasingly popular in marine science and ecosystem studies.  Both chapter 2 and 3 reveal a powerful seal-hake-herring trophic triangle in the SoG ecosystem. Each species of the triangle is influential in the SoG ecosystem: seals are abundant marine mammals heavily preying on fish; herring is a major forage fish and an important fishery; and hake is the most abundant resident fish, contributing to the diet of groundfish and competing with them for herring and euphausiids. This triangle is extremely powerful, as changes in any one of the three will greatly change the whole ecosystem. Recent studies highlighted keystone species, i.e., those species which have a strong role in the structure and function of ecosystems despite having a relatively low biomass and low food intake (Libralato et al., 2006; Power et al., 1996). Relative to their abundance, they disproportionately impact the ecosystems that they inhabit, such as seals in the east coast of Canada (Bundy et al., 2009). However, the trophic triangle may have a more powerful role in ecosystem structure and function than a single keystone species.  Many of the scenarios of EwE modelling in Chapter 2 and 3 show that harbour seals and herring have the same trend in biomass change, which is opposite to hake. The trends within the trophic triangle are corroborated by circumstantial evidence. The herring collapse in the  102  late 1960s occurred at a time when the seal population was low after two-decade hunting. A comparison of harbour seals (DFO, 2010) and herring (DFO, 2009) abundance shows that both increased greatly from the 1970s through the 1980s. Furthermore, this trophic triangle is stable in spite of the high biomass of invasive GH coho, which have interactions with each species.  The dramatic decline in SoG copepods and euphausiids in the 2000s (revealed in Chapter 4) may suggest a strong cascade up to harbour seals and influence the stability of the trophic triangle. The EwE model showed that a decline in SoG copepods would depress herring abundance, while a decline in euphausiids would reduce hake populations. This is consistent with the positive relationship between copepod biomass and herring recruitment in the SoG (Schweigert et al., accepted) and the recent decline in hake biomass seen in the 2011 hydroacoustics survey, compared to surveys in the 1990s (Akenhead et al., In review). Concurrently, the seal population has been observed to have a lower birth rate and a wider foraging range, e.g., far up the Fraser River (Peter Olesiuk, Pacific Biological Station, Nanaimo, personal communication). This bottom-up mechanism, from zooplankton to high trophic levels, can be better understood through an ecosystem simulation from1990 to 2007 using zooplankton forcing. This makes it possible to explore the trophic pathways from low trophic levels to high trophic levels, as well as investigate the stability of the trophic triangle.  EwE has proved an extremely valuable tool to capture all the complicated indirect interactions in an ecosystem. In addition to fishing, we also need to consider changes in environment, predators, prey, and competitors. All of these can be included in an EwE  103  model. For example, a combination of fishing, trophic interactions, and bottom-up effects has been applied in an EwE model to investigate the decline of Steller sea lions in Alaska (Guenette et al., 2006) and to predict the potential impacts of climate change on the Northeast Pacific (Ainsworth et al., 2011). However, uncertainty in the many model parameters needs to be addressed. I completed model sensitivity analyses in both Chapter 2 and Chapter 3 to address how the parameter errors influence the model performance. Ecopath’s new Monte Carlo function was used for the first time to investigate the sensitivity of Ecopath parameters to Ecosim performance. The analyses showed that the ecosystem simulations are insensitive to production and consumption rates over a wide range of uniformly-distributed values. Predator-prey vulnerabilities seem largely able to influence the ecosystem performance. However, many other uncertainties (Plaganyi and Butterworth, 2004; Fulton et al., 2003) are not included. For example, one weakness of EwE model structure (Plaganyi and Butterworth, 2004) is that, Ecosim only uses biomass, while, nutritional content may be more important than quantity (Trites et al., 2007; Trites and Donnelly, 2003). This may not be an issue in the current model where the dramatic decline in the biomass of SoG zooplankton communities seems adequate to explain the changes at high trophic levels. Although there is subjectivity involved in the choice of model parameters and structure, EwE models are very useful for “strategic” applications (Essington, 2007; Christensen and Walters, 2004).  104  5.3  Ecosystem-based management  The results of my study highly recommend ecosystem-based management due to complicated indirect interactions. Focusing solely on managing one species or fish stock at a time has become less of a viable option (Link, 2010). Fisheries management should be focused on the ecosystem as a whole, rather than just target species. Ecosystem-based management has emerged as a useful tool to improve management techniques and gain a better understand of the interactions between human beings and their environment (Link, 2010; Pitcher et al., 2009; Link et al., 2002).  My study supports a new way to predict future ocean ecosystems for ecosystem-based management. Perry et al. (2004) suggested using zooplankton as an indicator of ecosystem change because they are a critical link from the environment to fish. In Chapter 4, I found that SoG zooplankton community changes were more correlated with large-scale climate forcing than local factors. We can thus use a large-scale climate index as a predictor for the zooplankton communities and then use an ecosystem model to predict the responses of high trophic levels in the ecosystem. As there is a one-year lag from the best indicator spring SOIx to the SoG zooplankton communities and another extra few years for fish recruitment, the prediction is helpful to take actions in advance.  105  5.4  Future studies  Vulnerabilities, based on foraging arena theory (Ahrens et al., 2012; Walters and Martell, 2004; Walters and Korman, 1999; Walters and Juanes, 1993), are one of the most important parameters in EwE (Christensen and Walters, 2004). Vulnerabilities are commonly estimated by fitting to time series data as suggested by Christensen and Walters (2004). However, numerous combinations of vulnerabilities can potentially fit model predictions to time series and normally only one of the combinations is used in Ecosim simulation. Very few studies have examined the effects of vulnerability settings on Ecosim simulations. Ainsworth et al. (2008) compared different vulnerability methods including fitting to time series (by predator, by prey, and by individual predation-prey interaction), flat vulnerability (fixed vulnerability at 1.2, 2 and 3.7, respectively), and scaling vulnerability to trophic levels of predators and prey. Scaling vulnerabilities to the trophic levels of prey produced reasonable predictions compared to fitting to time series method. I applied this method for a present SoG Ecopath model in both Chapter 2 and 3. My uncertainty analyses in both Chapter 2 and 3 showed large changes in Ecosim simulations with a-key-group related vulnerabilities altered widely from top-down control to bottom-up control. Considering the important role and large uncertainty of vulnerabilities, there is a strong need to quantitatively study vulnerabilities for future studies in EwE modelling.  Future EwE models should incorporate seasonality in Ecopath modelling. Incorporating seasonality will allow us to better represent the dynamics at low trophic levels. An ecosystem model normally includes organisms from phytoplankton to marine mammals,  106  which have very different life spans and time scales. Using the unit of “per year” in the Ecopath model makes phytoplankton biomass largely driven by the peak biomass of the spring bloom. Therefore, higher phytoplankton biomass in the model produces more energy flow from phytoplankton directly to the zooplankton. By contrast, modeling seasonal phytoplankton biomass (e.g. at monthly level) allows for much phytoplankton to sink to the sea floor during the spring bloom, just as in reality. Additionally, incorporating seasonality will better represent higher tophic level dynamics, such as migration and diet. For example, seal diet is dominated by spawning herring in winter and hake in summer (Olesiuk et al., 1990b). Including the seasonality in harbour seal diet may reveal a clearer top-down effect, especially on the herring and hake populations.  My study also highlights the importance of regular monitoring of the ecosystem components, such as sampling water, phytoplankton and zooplankton, and carrying out surveys for fish and marine mammals. Without detailed field data, an Ecopath model cannot accurately represent the ecosystem and make further correct predictions. With a longer time series of zooplankton samples, my statistical results, such as physical-biological correlation, would be more powerful. In addition to biomass or abundance of organisms, updating diets is also very important. For example, seal diet is a key factor in the SoG ecosystem model. However, the most comprehensive diet data is from Olesiuk et al. (1990b), which analyzed 3,000 scats collected at 58 sites in the SoG in all months. It is probably still the best estimate of seal diet in the SoG, although fish populations have changed to some extent over the past two decades. Because seal diet is impacted by locality and seasonality (Lance and Jeffries, 2007; Olesiuk et al., 1990a), recent studies (e.g. Lance and Jeffries, 2007) based only on several  107  months of data in a small area, and are not representative of the annual seal diet in the SoG. Additionally, many diet papers focus only on the major prey of a predator. However, the rare items in the diet are also important. When the biomass of rare prey increases, the predator will likely increase consumption of them. Chapter 3 shows that a large increase in any rare prey could potentially result in an increase in predator biomass. Fortunately, the Ecosim simulation model has this facility of adjusting the actual diet composition according to changes in biomass of prey (Christensen and Walters, 2004). However, this could not help if the rare diet items are not included in the model.  Heavily exploited ecosystems are more vulnerable to climate change (Perry et al., 2010b) and there will probably be more accelerated climate change in the future, e.g. increased temperature, hypoxia, ocean acidification, and changes in the amount and/or timing of freshwater run-off. On the other hand, ecosystem health might be increasingly impacted by human activities, due to more coastal development and increasing human populations. There is thus strong need to consider more environmental and anthropogenic factors and use holistic and broad approaches for ecosystem-based management.  108  References Aerni, P. 2004. Risk, regulation and innovation: The case of aquaculture and transgenic fish. Aquatic Sciences, 66, 327-341.  Ahrens, R.N.M., Devlin, R.H. 2011. Standing genetic variation and compensatory evolution in transgenic organisms: a growth-enhanced salmon simulation. Transgenic Research, 20, 583-597.  Ahrens, R.N.M., Walters, C.J., Christensen, V. 2012. Foraging arena theory. Fish and Fisheries, 13, 41-59.  Ainsworth, C. ,2006. Strategic marine ecosystem restoration in Northern British Columbia. Ph.D Thesis, the University of British Columbia. Vancouver, Canada. .  Ainsworth, C., Pitcher, T.J. 2004. Estimating the effects of predator-prey vulnerability settings on Ecosim’s dynamic forecasts. In T.J. Pitcher. Back to the Future: Advances in Methodology for Modeling and Evaluating Past Ecosystems as Future Ploicy Goals, Vol. 12(1). (pp. 45-47). Fisheries Centre Research Reports.  Ainsworth, C., Pitcher, T.J., Heymans, J.J., Vasconcellos, M. 2008. Historical Reconstruction of the Marine Ecosystem of Northern British Columbia from Pre-European Contact to the Present. Ecological Modeling, 216, 354-368.  Ainsworth, C.H., Samhouri, J.F., Busch, D.S., Cheung, W.W.L., Dunne, J., Okey, T.A. 2011. Potential impacts of climate change on Northeast Pacific marine foodwebs and fisheries. ICES Journal of Marine Science, 68, 1217-1229. 109  Ainsworth, C.H., Pitcher, T.J., Heymans, J.J., Vasconcellos, M. 2008. Reconstructing historical marine ecosystems using food web models: Northern British Columbia from Pre-European contact to present. Ecological Modelling, 216, 354-368.  Akenhead, S.A., Irvine, J.R., Perry, I.R., Preikshot, D., Li, L., Fu, C. In review. Responses of the fisheries food web within the Strait of Georgia to 150 years of management experiments. Progress in Oceanography.  Allen, S.E., Wolfe, M.A. Accepted. Hindcast of the timing of the spring phytoplankton bloom in the Strait of Georgia, 1968-2010.  Bakun, A., Parrish, R.H. 1991. Comparative studies of coastal pelagic fish reproductive habitats: the anchovy (Engraulis anchoita) of the southwestern Atlantic. ICES Journal of Marine Science, 48, 343-361.  Beamish, R.J., McFarlane, G.A., Thomson, R.E. 1999. Recent declines in the recreational catch of coho salmon (Oncorhynchus kisutch) in the Strait of Georgia are related to climate. Canadian Journal of Fisheries and Aquatic Sciences, 56, 506-515.  Beamish, R.J., Mahnken, C., Neville, C.M. ,1997. Hatchery and wild production of Pacific salmon in relation to large-scale, natural shifts in the productivity of the marine environment. ICES Journal of Marine Science, 54, 1200-1215.  Beaudreau, A.H., Essington, T.E. 2009. Development of a new field-based approach for estimating consumption rates of fishes and comparison with a bioenergetics model for  110  lingcod (Ophiodon elongatus). Canadian Journal of Fisheries and Aquatic Sciences, 66, 565-578.  Bessey, C., Devlin, R.H., Liley, N.R., Biagi, C.A. 2004. Reproductive Performance of Growth-Enhanced Transgenic Coho Salmon. Transactions of the American Fisheries Society, 133, 1205-1220.  Bornhold, E.A. ,2000. Interannual and interdecadal patterns in timing and abundance of phytoplankton and zooplankton in the central Strait of Georgia, BC: with special reference to Neocalanus plumchrus. MSc Thesis. Department of Earth and Ocean Sciences. The University of British Columbia., 132 p.  Bowen, W.D. 1997. Role of marine mammals in aquatic ecosystems. Marine Ecology Progress Series, 158, 267-274.  Buckley, T.W., Livingston, P.A. 1997. Geographic variation in the diet of Pacific hake, with a note on cannibalism. California Cooperative Oceanic Fisheries Investigations Report, 38, 53-62.  Bundy, A., Heymans, J.J., Morissette, L., Savenkoff, C. 2009. Seals, cod and forage fish: a comparative exploration of variations in the theme of stock collapse and ecosystem change in four Northwest Atlantic ecosystems. Progress in Oceanography, 81, 188-206.  Carss, D.N. 1990. Concentrations of wild and escaped fishes immediately adjacent to fish farm cages. Aquaculture, 90, 29-40.  111  Casini, M., Hjelm, J., Molinero, J., Lovgren, J., Cardinale, M., Bartolino, V., Belgrano, A., Kornilovs, G. 2009. Trophic cascades promote threshold-like shifts in pelagic marine ecosystems. Proceedings of the National Academy of Sciences of the United States of America, 106, 197-202.  Chatfield, C., Collins, A.J. 1980. Introduction to multivariate analysis. Chapman and Hall, London.  Chavez, F.P., Collins, C.A., Huyer, A., Mackas, D.L. ,2002. El Nino along the west coast of North America. Progress in oceanography, 54, 1-5.  Check, E. 2002. Environmental impact tops list of fears about transgenic animals. Nature, 418, 805-805.  Chen, Z., Qiu, Y., Jia, X., Xu, S. 2008. Simulating fisheries management options for the Beibu Gulf by means of an ecological modelling optimization routine. Fisheries Research, 89, 257-265.  Cheung, W.L., Watson, R., Pitcher, T. 2002. Policy simulation of fisheries in the Hong Kong marine ecosystem. Fisheries Centre Research Reports, 10, 46-53.  Christensen, V., Pauly, D. 1992. Ecopath II - a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling, 61, 169-185.  Christensen, V., Pauly, D. 1993. Flow characteristics of aquatic ecosystems. Bioscience, 36, 251-257.  112  Christensen, V., Walters, C.J. 2004. Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling, 172, 109-139.  Christensen, V., Walters, C.J. 2005. Using ecosystem modeling for fisheries management: Where are we? ICES CM, /M:19.  Christensen, V., Walters, C.J., Pauly, D., and Forrest, R. 2008. Ecopath with Ecosim version 6 user guide. Available: www.ecopath.org.  Clarke, K.R. 1993. Non‐parametric multivariate analyses of changes in community structure. Australian Journal of Ecology, 18, 117-143.  Clarke, K.R., Ainsworth, M. 1993. A method of linking multivariate community structure to environmental variables. Marine Ecology Progress Series, 92, 205-219.  Clarke, K.R., Green, R.H. 1988. Statistical design and analysis for a" biological effects" study. Marine Ecology Progress Series, 46, 213-226.  Clarke, K.R., Warwick, R.M. 2001. Changes in marine communities: an approach to statistical analysis and interpretation, 2nd edition. PRIMER-E: PLymouth.  Cloern, J.E., Hieb, K.A., Jacobson, T., Sansó, B., Di Lorenzo, E., Stacey, M.T., Largier, J.L., Meiring, W., Peterson, W.T., Powell, T.M. 2010. Biological communities in San Francisco Bay track large‐scale climate forcing over the North Pacific. Geophysical Research Letters, 37, L21602.  113  Cloern, J.E., Jassby, A.D. 2008. Complex seasonal patterns of primary producers at the land– sea interface. Ecology Letters, 11, 1294-1303.  Cloern, J.E., Jassby, A.D., Thompson, J.K., Hieb, K.A. 2007. A cold phase of the East Pacific triggers new phytoplankton blooms in San Francisco Bay. Proceedings of the National Academy of Sciences, 104, 18561-18565.  Collins, A.K., Allen, S.E., Pawlowicz, R. 2009. The role of wind in determining the timing of the spring bloom in the Strait of Georgia. Canadian Journal of Fisheries and Aquatic Sciences, 66, 1597-1616.  Crawford, W.R., Irvine, J.R. 2010. State of physical, biological, and selected fishery resources of Pacific Canadian marine ecosystems in 2009. DFO Can. Sci. Advis. Sec. Res. Doc. 2010/053. 137 p.  Crozier, W.W. 1993. Evidence of genetic interaction between escaped farmed salmon and wild Atlantic salmon (Salmo salar L.) in a Northern Irish river. Aquaculture, 113, 19-29.  Daly, K.L. 1990. Overwintering development, growth, and feeding of larval Euphausia superba in the Antarctic marginal ice zone. Limnology and Oceanography, 1564-1576.  DeCesare, N.J., Hebblewhite, M., Robinson, H.S., Musiani, M. 2010. Endangered, apparently: the role of apparent competition in endangered species conservation. Animal Conservation, 13, 353-362.  Deitch, E.J., Fletcher, G.L., Petersen, L.H., Costa, I.A.S.F., Shears, M.A., Driedzic, W.R., Gamperl, A.K. 2006. Cardiorespiratory modifications, and limitations, in post-smolt 114  growth hormone transgenic Atlantic salmon Salmo salar. Journal of Experimental Biology, 209, 1310-1325.  Devlin, R.H., Sundström, L.F., Johnsson, J.I., Fleming, I.A., Hayes, K.R., Ojwang, W.O., Bambaradeniya, C., Zakaraia-Ismail, M. 2007. Assessing ecological effects of transgenic fish prior to entry into nature. In A.R. Kapuscinski, K.R. Hayes, S. Li, G. and Dana. Environmental risk assessment of genetically modified organisms, Vol. 3. (pp. 151-187). CABI. UK.  Devlin, R.H., Biagi, C.A., Yesaki, T.Y. 2004a. Growth, viability and genetic characteristics of GH transgenic coho salmon strains. Aquaculture, 236, 607-632.  Devlin, R.H., D'Andrade, M., Uh, M., Biagi, C.A. 2004b. Population effects of growth hormone transgenic coho salmon depend on food availability and genotype by environment interactions. Proceedings of the National Academy of Sciences, USA, 101, 9303-9308.  Devlin, R.H., Donaldson, E.M. 1992. Containment of genetically altered fish with emphasis on salmonids. In C.L. Hew, G.L. Fletcher. Transgenic Fish. (pp. 229-266). World Scientific. Singapore.  Devlin, R.H., Johnsson, J.I., Smailus, D.E., Biagi, C.A., Joensson, E., Bjoernsson, B.T. 1999. Increased ability to compete for food by growth hormone-transgenic coho salmon Oncorhynchus kisutch (Walbaum). Aquaculture Research, 30, 479-482.  115  Devlin, R.H., Sundström, L.F., Muir, W.M. 2006. Interface of biotechnology and ecology for environmental risk assessments of transgenic fish. Trends in Biotechnology, 24, 89-97.  Devlin, R.H., Yesaki, T.Y., Biagl, C.A., Donaldson, E.M., Swanson, P., Chan, W. 1994. Extraordinary salmon growth. Nature, 371, 209-210.  DFO. ,2002. Coho salmon in Georgia Basin. , DFO Science Stock Status Report D6-07 (2002).  DFO. ,2005. Strait of Georgia lingcod (Ophiodon elongatus) assessment and advice for fishery management. , DFO Can. Advis. Sec. Sci. Advis. Rep. 2005/042.  DFO. ,2006. Rockfish conservation areas: protecting British Columbia's rockfish. , Fisheries and Oceans Canada, Pacific Region, 174 pp.  DFO. ,2008. 2007 Marine survival forecast of southern British Columbia coho. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2008/032.  DFO. ,2009. Stock assessment report on Pacific herring in British Columbia. , DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2009/059.  DFO. ,2010. Population assessment Pacific harbour seal (Phoca vitulina richardsi). , DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2009/011.  Di Lorenzo, E., Cobb, K.M., Furtado, J.C., Schneider, N., Anderson, B.T., Bracco, A., Alexander, M.A., Vimont, D.J. 2010. Central Pacific El Niño and decadal climate change in the North Pacific Ocean. Nature Geoscience, 3, 762-765.  116  Di Lorenzo, E., Schneider, N., Cobb, K.M., Franks, P.J.S., Chhak, K., Miller, A.J., McWilliams, J.C., Bograd, S.J., Arango, H., Curchitser, E. 2008. North Pacific Gyre Oscillation links ocean climate and ecosystem change. Geophysical Research Letters, 35, L08607.  Drinkwater, K., Hunt, G., Lehodey, P., Lluch-Cota, S., Murphy, E.J., Sakurai, Y., Schwing, F., Beaugrand, G., and Sunby, S. 2010a. Climate forcing on marine ecosystems. In M. Barange, J.G. Field, R.P. Harris, E.E. Hofmann, R.I. Perry, F.E. Werner. Marine ecosystems and global change. (pp. 11-39). Oxford University Press. New York, USA.  Drinkwater, K.F., Beaugrand, G., Kaeriyama, M., Kim, S., Ottersen, G., Perry, R.I., Pörtner, H.O., Polovina, J.J., Takasuka, A. 2010b. On the processes linking climate to ecosystem changes. Journal of Marine Systems, 79, 374-388.  Eloire, D., Somerfield, P.J., Conway, D.V.P., Halsband-Lenk, C., Harris, R., Bonnet, D. 2010. Temporal variability and community composition of zooplankton at station L4 in the Western Channel: 20 years of sampling. Journal of Plankton Research, 32, 657-679.  El-Sabaawi, R., Dower, J.F., Kainz, M., Mazumder, A. 2009. Interannual variability in fatty acid composition of the copepod Neocalanus plumchrus in the Strait of Georgia, British Columbia. Marine Ecology Progress Series, 382, 151-161.  Essington, T.E. 2007. Evaluating the sensitivity of a trophic mass-balance model (Ecopath) to imprecise data inputs. Canadian Journal of Fisheries and Aquatic Sciences, 64, 628637.  117  Farrell, A.P., Bennett, W., Devlin, R.H. 1997. Growth-enhanced transgenic salmon can be inferior swimmers. Canadian Journal of Zoology, 75, 335-337.  Fiske, P., Lund, R.A., Hansen, L.P. 2006. Relationships between the frequency of farmed Atlantic salmon, Salmo salar L., in wild salmon populations and fish farming activity in Norway, 1989–2004. ICES Journal of Marine Science, 63, 1182-1189.  Folke, C., Carpenter, S., Walker, B., Scheffer, M., Elmqvist, T., Gunderson, L., Holling, C.S. 2004. Regime shifts, resilience, and biodiversity in ecosystem management. Annual Review of Ecology Evolution and Systematics, 35, 557-581.  Foreman, M.G.G., Lee, D.K., Morrison, J., Macdonald, S., Barnes, D., and Williams, I.V. 2001. Simulations and retrospective analyses of Fraser watershed flows and temperatures. Atmosphere-Ocean, 39, 89-105.  Frank, K.T., Petrie, B., Choi, J.S., Leggett, W.C. 2005. Trophic cascades in a formerly coddominated ecosystem. Science, 308, 1621-1623.  Fulton, E.A., Smith, A.D.M., Johnson, C.R. 2003. Effect of complexity on marine ecosystem models. Marine Ecology Progress Series, 253, 1-16.  Gerber, L.R., Morissette, L., Kaschner, K., Pauly, D. 2009. Should whales be culled to increase fishery yield? Science, 323, 880-881.  Gómez-Gutiérrez, J., Feinberg, L.R., Shaw, T.C., Peterson, W.T. 2007. Interannual and geographical variability of the brood size of the euphausiids Euphausia pacifica and  118  Thysanoessa spinifera along the Oregon coast (1999-2004). Deep Sea Research Part I: Oceanographic Research Papers, 54, 2145-2169.  Guenette, S., Heymans, S.J.J., Christensen, V., Trites, A.W. 2006. Ecosystem models show combined effects of fishing, predation, competition, and ocean productivity on Steller sea lions (Eumetopias jubatus) in Alaska. Canadian Journal of Fisheries and Aquatic Sciences, 63, 2495-2517.  Hallett, T., Coulson, T., Pilkington, J., Clutton-Brock, T., Pemberton, J., Grenfell, B. 2004. Why large-scale climate indices seem to predict ecological processes better than local weather. Nature, 430, 71-75.  Hansen, L., Jacobsen, J., Lund, R. 1999. The incidence of escaped farmed Atlantic salmon, Salmo salar L., in the Faroese fishery and estimates of catches of wild salmon. ICES Journal of Marine Science, 56, 200-206.  Hare, S.R., Mantua, N.J. 2000. Empirical evidence for North Pacific regime shifts in 1977 and 1989. Progress in Oceanography, 47, 103-145.  Haro-Garay, M.J., Soberanis, L.H. 2008. Dominance shift of zooplankton species composition in the central Strait of Georgia, British Columbia during 1997. Hidrobiológica, 18, 53-60.  Harvey, C.J., Bartz, K.K., Davies, J.R., Francis, T.B., Good, T.P., Guerry, A.D., Hanson, M.B., Holsman, K.K., Miller, J., Plummer, M., Reum, J.C.P., Rhodes, L.D., Rice, C.A., Samhouri, J.F., Williams, G.D., Yoder, N.J., Levin, P.S., and Ruckelshaus, M.H. ,2010.  119  A mass-balance model for evaluating food web structure and community-scale indicators in the central basin of Puget Sound. , NOAA Tech. Memo. NMFS-NWFSC106.  Harvey, C.J., Kareiva, P.M. 2005. Community context and the influence of non-indigenous species on juvenile salmon survival in a Columbia River reservoir. Biological Invasions, 7, 651-663.  Hay, D.E., McCarter, P.B., Daniel, K.S. 2001. Tagging of Pacific herring Clupea pallasi from 1936–1992: a review with comments on homing, geographic fidelity, and straying. Canadian Journal of Fisheries and Aquatic Sciences, 58, 1356-1370.  Hayes, K.R., Barry, S.C. 2008. Are there any consistent predictors of invasion success? Biological Invasions, 10, 483-506.  Hedrick, P.W. 2001. Invasion of transgenes from salmon or other genetically modified organisms into natural populations. Canadian Journal of Fisheries and Aquatic Sciences, 58, 841-844.  Herwig, R.P., Gray, J.P., Weston, D.P. 1997. Antibacterial resistant bacteria in surficial sediments near salmon net-cage farms in Puget Sound, Washington. Aquaculture, 149, 263-283.  Higgs, D.A., Sutton, J.N., Kim, H., Oakes, J.D., Smith, J., Biagi, C., Rowshandeli, M., Devlin, R.H. 2009. Influence of dietary concentrations of protein, lipid and carbohydrate on growth, protein and energy utilization, body composition, and plasma titres of growth  120  hormone and insulin-like growth factor-1 in non-transgenic and growth hormone transgenic coho salmon, Oncorhynchus kisutch (Walbaum). Aquaculture, 286, 127-137.  Hindar, K., Fleming, I.A., McGinnity, P., Diserud, O. 2006. Genetic and ecological effects of salmon farming on wild salmon: modelling from experimental results. ICES Journal of Marine Science, 63, 1234-1247.  Hoover, C. ,2012. Effects of harvest and climate change on polar marine ecosystems. PhD thesis. The University of British Columbia, Vancouver, pp: 490.  Hourston, A.S. 1956. Investigations of the juvenile herring stocks in the Strait of Georgia region, 1955. Manuscript Report No.BQ114, Fisheries Research Board of Canada, Pacific Biological Station, Nanaimo, BC. 59 p.  Hsieh, C.H., Chen, C.S., Chiu, T.S., Lee, K.T., Shieh, F.J., Pan, J.Y., Lee, M.A. 2009. Time series analyses reveal transient relationships between abundance of larval anchovy and environmental variables in the coastal waters southwest of Taiwan. Fisheries Oceanography, 18, 102-117.  Hunter, M.D., Price, P.W. 1992. Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology, 73, 724-732.  Hutchings, J.A., and Fraser, D.J. 2008. The nature of fisheries-and farming-induced evolution. Molecular Ecology, 17, 294-313.  121  Iguchi, N., Ikeda, T. 1994. Experimental study on brood size, egg hatchability and early development of a euphausiid Euphausia pacifica from Toyama Bay, southern Japan Sea. Bulletin of the Japan Sea National Fisheries Research Institute, 44, 49-57.  Iguchi, N., Ikeda, T. 1995. Growth, metabolism and growth efficiency of a euphausiid crustacean Euphausia pacifica in the southern Japan Sea, as influenced by temperature. Journal of Plankton Research, 17, 1757-1769.  Iguchi, N., Ikeda, T., Imamura, A. 1993. Growth and life cycle of an euphausiid crustacean (Euphausia pacifica Hansen) in Toyama bey, southern Japan sea. Bulletin of the Japan Sea National Fisheries Research Institute, 43, 69-81.  Irvine, J.R., Gross, M.R., Wood, C.C., Holtby, L.B., Schubert, N.D., Amiro, P.G. 2005. Canada's Species at Risk Act. Fisheries, 30, 11-19.  Jackson, J.B.C. 2008. Ecological extinction and evolution in the brave new ocean. Proceedings of the National Academy of Sciences, 105, 11458.  Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J., Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science, 293, 629.  Jarre-Teichmann, A. 1998. The potential role of mass balance models for the management of upwelling ecosystems. Ecological Applications, 8, 93-103.  122  Jhingan, E., Devlin, R.H., Iwama, G.K. 2003. Disease resistance, stress response and effects of triploidy in growth hormone transgenic coho salmon. Journal of fish biology, 63, 806823.  Jiao, Y. 2009. Regime shift in marine ecosystems and implications for fisheries management, a review. Reviews in Fish Biology and Fisheries, 19, 177-191.  Johannessen, S.C., Macdonald, R.W. 2009. Effects of local and global change on an inland sea: the Strait of Georgia, British Columbia, Canada. Climate Research, 40, 1-21.  Johannessen, S.C., O'Brien, M.C., Denman, K.L., Macdonald, R.W. 2005. Seasonal and spatial variations in the source and transport of sinking particles in the Strait of Georgia, British Columbia, Canada. Marine Geology, 216, 59-77.  Jones, S.R.M., MacKinnon, A.M., Groman, D.B. 1999. Virulence and pathogenicity of infectious salmon anemia virus isolated from farmed salmon in Atlantic Canada. Journal of Aquatic Animal Health, 11, 400-405.  Kapuscinski, A.R., Hallerman, E.M. 1990. Transgenic fish and public policy: anticipating environmental impacts of transgenic fish. Fisheries, 15, 2-11.  Kapuscinski, A.R., Hayes, K.R., Li, S., Dana, G., Hallerman, E.M., Schei, P.J. 2007. Environmental risk assessment of genetically modified organisms. Volume 3. Methodologies for Transgenic Fish. CAB International, Oxfordshire.  123  Kimmel, D.G., Miller, W.D., Harding, L.W., Houde, E.D., Roman, M.R. 2009. Estuarine ecosystem response captured using a synoptic climatology. Estuaries and Coasts, 32, 403-409.  King, J.R., McFarlane, G.A. 2006. Shift in Size-at-age of the Strait of Georgia Population of Pacific Hake (Merluccius productus). California Cooperative Oceanic Fisheries Investigations Report, 47, 111-118.  Kolar, C.S., Lodge, D.M. 2002. Ecological predictions and risk assessment for alien fishes in North America. Science, 298, 1233-1236.  Kotrschal, A., Sundström, L., Brelin, D., Devlin, R., Kolm, N. 2012. Inside the heads of David and Goliath: environmental effects on brain morphology among wild and growth‐ enhanced coho salmon Oncorhynchus kisutch. Journal of Fish Biology.  Kraus, S.D., Brown, M.W., Caswell, H., Clark, C.W., Fujiwara, M., Hamilton, P.K., Kenney, R.D., Knowlton, A.R., Landry, S., Mayo, C.A., McLellan, W.A., Moore, M.J., Nowacek, D.P., Pabst, D.A., Read, A.J., Rolland, R.M. 2005. North Atlantic right whales in crisis. Science, 309, 561-562. Krkošek, M., Lewis, M.A., Volpe, J.P. 2005. Transmission dynamics of parasitic sea lice from farm to wild salmon. Proceedings of the Royal Society B: Biological Sciences, 272, 689-696.  124  La Rosa, T., Mirto, S., Mazzola, A., Danovaro, R. 2001. Differential responses of benthic microbes and meiofauna to fish-farm disturbance in coastal sediments. Environmental Pollution, 112, 427-434.  Lance, M.M., and Jeffries, S.J., 2007. Temporal and spatial variability of harbor seal diet in the San Juan Island archipelago. Contract Report to SeaDoc Society Research Agreement No. K004431 - 25. Washington Department of Fish and Wildlife, Olympia, WA. 21 pp.  Lavaniegos, B.E., Ohman, M.D. 2007. Coherence of long-term variations of zooplankton in two sectors of the California Current System. Progress in Oceanography, 75, 42-69.  LeBlond, P.H., Ma, H., Doherty, F., Pond, S. 1991. Deep and Intermediate Water Replacement h the Strait of Georgia. Atmosphere-Ocean, 29, 288-312.  LeBlond, P.H. ,1983. The Strait of Georgia: functional anatomy of a coastal sea. Canadian Journal of Fisheries and Aquatic Sciences, 40, 1033-1063.  Lee, C.G., Devlin, R.H., Farrell, A.P. 2003. Swimming performance, oxygen consumption and excess post-exercise oxygen consumption in adult transgenic and ocean-ranched coho salmon. Journal of Fish Biology, 62, 753-766.  Leggatt, R.A., Devlin, R.H., Farrell, A.P., Randall, D.J. 2003. Oxygen uptake of growth hormone transgenic coho salmon during starvation and feeding. Journal of fish biology, 62, 1053-1066.  125  Lessard, R.B., Martell, S.J.D., Walters, C.J., Essington, T.E., Kitchell, J.F. 2005. Should ecosystem management involve active control of species abundances? Ecology and Society, 10, 1.  Li, L., Ainsworth, C., Pitcher, T. 2010. Presence of harbour seals (Phoca vitulina) may increase exploitable fish biomass in the Strait of Georgia. Progress in Oceanography, 87, 235-241.  Libralato, S., Christensen, V., Pauly, D. 2006. A method for identifying keystone species in food web models. Ecological Modelling, 195, 153-171.  Link, J.S. 2010. Ecosystem-based fisheries management: confronting tradeoffs. Cambridge University Press. Cambridge, UK.  Link, J.S., Brodziak, J.K.T., Edwards, S.F., Overholtz, W.J., Mountain, D., Jossi, J.W., Smith, T.D., Fogarty, M.J. 2002. Marine ecosystem assessment in a fisheries management context. Canadian Journal of Fisheries and Aquatic Sciences, 59, 14291440.  Liu, Y., Diserud, O.H., Hindar, K., Skonhoft, A. 2012. An ecological–economic model on the effects of interactions between escaped farmed and wild salmon (Salmo salar). Fish and Fisheries, 16p.  Lõhmus, M., Sundström, L.F., Björklund, M., Devlin, R.H. 2010. Genotype-temperature interaction in the regulation of development, growth, and morphometrics in wild-type, and growth-hormone transgenic Coho Salmon. PloS ONE, 5(4), e9980.  126  Lura, H., Sægrov, H. 1991. Documentation of successful spawning of escaped farmed female Atlantic salmon, Salmo salar, in Norwegian rivers. Aquaculture, 98, 151-159.  Mackas, D.L., Batten, S., Trudel, M. 2007. Effects on zooplankton of a warmer ocean: recent evidence from the Northeast Pacific. Progress in Oceanography, 75, 223-252.  Mackas, D.L., Beaugrand, G. 2010. Comparisons of zooplankton time series. Journal of Marine Systems, 79, 286-304.  Mackas, D.L., Harrison, P.J. 1997. Nitrogenous nutrient sources and sinks in the Juan de Fuca Strait/Strait of Georgia/Puget Sound estuarine system: assessing the potential for eutrophication. Estuarine, Coastal and Shelf Science, 44, 1-21.  Mackas, D.L., Thomson, R.E., Galbraith, M. 2001. Changes in the zooplankton community of the British Columbia continental margin, 1985-1999, and their covariation with oceanographic conditions. Canadian Journal of Fisheries and Aquatic Sciences, 58, 685702.  Mackas, D.L., Harrison, P.J. 1997. Nitrogenous nutrient sources and sinks in the Juan de Fuca Strait/Strait of Georgia/Puget Sound Estuarine System: assessing the potential for eutrophication. Estuarine, Coastal and Shelf Science, 44, 1-21.  Mackas, D.L., Peterson, W.T., Zamon, J.E. ,2004. Comparisons of interannual biomass anomalies of zooplankton communities along the continental margins of British Columbia and Oregon. Deep-sea research. Part II, Topical studies in oceanography, 51, 875.  127  Mantua, N.J., Hare, S.R., Zhang, Y., Wallace, J.M., Francis, R.C. 1997. A Pacific interdecadal climate oscillation with impacts on salmon production. Bulletin of the American Meteorological Society, 78, 1069-1080.  Martell, S. 2004. Dealing with migratory species in ecosystem models. Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports, 12, 158.  Martell, S.J.D., Beattie, A.I., Walters, C.J., Nayar, T., Briese, R. 2002. Simulating fisheries management strategies in the Strait of Georgia ecosystem using Ecopath and Ecosim. Fisheries Centre Research Reports, 10, 16-23.  Masson, D., Peña, A. 2009. Chlorophyll distribution in a temperate estuary: the Strait of Georgia and Juan de Fuca Strait. Estuarine, Coastal and Shelf Science, 82, 19-28.  Masson, D. 2002. Deep water renewal in the Strait of Georgia. Estuarine, Coastal and Shelf Science, 54, 115-126.  Masson, D., Cummins, P.F. 2007. Temperature trends and interannual variability in the Strait of Georgia, British Columbia. Continental Shelf Research, 27, 634-649.  McFarlane, G.A., Beamish, R.J. 1985. Biology and fishery of Pacific whiting, Merluccius productus, in the Strait of Georgia. Marine Fisheries Review, 47, 23-34.  Merrick, R.L. 1997. Current and historical roles of apex predators in the Bering Sea ecosystem. Journal of Northwest Atlantic Fishery Science, 22, 343-356.  128  Morissette, L., Hammill, M.O., Savenkoff, C. 2006. The trophic role of marine mammals in the northern Gulf of St. Lawrence. Marine Mammal Science, 22, 74-103.  Morrison, J., Quick, M.C., Foreman, M.G.G. 2002. Climate change in the Fraser River watershed: flow and temperature projections. Journal of Hydrology, 263, 230-244.  Muir, W.M., Howard, R.D. 1999. Possible ecological risks of transgenic organism release when transgenes affect mating success: Sexual selection and the Trojan gene hypothesis. Proceedings of the National Academy of Sciences, USA, 96, 13853-13856.  Muir, W.M., Howard, R.D. 2001. Fitness components and ecological risk of transgenic release: a model using Japanese medaka (Oryzias latipes). American Naturalist, 158, 116.  Muir, W.M., Howard, R.D. 2004. Characterization of environmental risk of genetically engineered (GE) organisms and their potential to control exotic invasive species. Aquatic Sciences, 66, 414-420.  National Research Council. 2002. Animal biotechnology: science-based concerns. National Academy Press.  Naylor, R., Hindar, K., Fleming, I.A., Goldburg, R., Williams, S., Volpe, J., Whoriskey, F., Eagle, J., Kelso, D., Mangel, M. 2005. Fugitive salmon: assessing the risks of escaped fish from net-pen aquaculture. Bioscience, 55, 427-437.  Naylor, R.L., Williams, S.L., Strong, D.R. 2001. Aquaculture--A gateway for exotic species. Science, 294, 1655-1656. 129  NOAA. 2006. Ecopath modeling: Precursor to an ecosystem approach to fisheries management. National Oceanic and Atmospheric Administration, http://celebrating200years.noaa.gov/breakthroughs/ecopath/welcome.html.  Olesiuk, P.F. 1999. An assessment of the status of harbour seals (Phoca vitulina) in British Columbia. Canadian Stock Assessment Secretariat Research Document 1999/33, 130.  Olesiuk, P.F., Bigg, M.A., Ellis, G.M. 1990a. Recent trends in the abundance of harbour seals, Phoca vitulina, in British Columbia. Canadian Journal of Fisheries and Aquatic Sciences, 47, 992-1003.  Olesiuk, P.F., Bigg, M.A., Ellis, G.M., Crockford, S.J., Wigen, R.J. 1990b. An assessment of the feeding habits of harbour seals (Phoca vitulina) in the Strait of Georgia, British Columbia, based on scat analysis. Canadian Technical Report of Fisheries and Aquatic Sciences, 1730, 1-135.  Pauly, D., Alder, J., Bennett, E., Christensen, V., Tyedmers, P., Watson, R. 2003. The future for fisheries. Science, 302, 1359-1361.  Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres Jr, F. 1998a. Fishing down marine food webs. Science, 279, 860-863.  Pauly, D., Pitcher, T.J., Preikshot, D., Hearne, J. 1998b. Back to the future: reconstructing the Strait of Georgia ecosystem. Fisheries Centre, University of British Columbia.  Pawlowicz, R., Riche, O., Halverson, M. 2007. The circulation and residence time of the Strait of Georgia using a simple mixing-box approach. Atmosphere-Ocean, 45, 173-193. 130  Pennington, K.M., Kapuscinski, A.R. 2011. Predation and food limitation influence fitness traits of growth-enhanced transgenic and wild-type fish. Transactions of the American Fisheries Society, 140, 221-234.  Perry, I. ,2008. Status of the Strait of Georgia. Symposium on Sustaining Wild Salmon: moving from words to actions. Available: http://www.pac.dfompo.gc.ca/science/oceans/detroit-Georgiastrait/documents/Perry%20Presentation%2016Aug2008%20afternoon.pdf.  Perry, R.I., Barange, M., Hofmann, E., Moloney, C., Ottersen, G., Sakurai, Y. 2010a. Introduction to the GLOBEC 3rd Open Science Meeting: From ecosystem function to ecosystem prediction. Progress in Oceanography, 87, 1-5.  Perry, R.I., Cury, P., Brander, K., Jennings, S., Möllmann, C., Planque, B. 2010b. Sensitivity of marine systems to climate and fishing: concepts, issues and management responses. Journal of Marine Systems, 79, 427-435.  Perry, R.I., Schweigert, J.F. 2008. Primary productivity and the carrying capacity for herring in NE Pacific marine ecosystems. Progress in Oceanography, 77, 241-251.  Perry, R.I., Ommer, R.E. ,2003. Scale issues in marine ecosystems and human interactions. Fisheries oceanography, 12, 513-522.  Peterson, W.T., Schwing, F.B. 2003. A new climate regime in northeast Pacific ecosystems. Geophysical Research Letters, 30, 1896.  131  Pinchuk, A.I., Hopcroft, R.R. 2006. Egg production and early development of Thysanoessa inermis and Euphausia pacifica (Crustacea: Euphausiacea) in the northern Gulf of Alaska. Journal of Experimental Marine Biology and Ecology, 332, 206-215.  Pitcher, T.J. 2001. Fisheries managed to rebuild ecosystems? Reconstructing the past to salvage the future. Ecological Applications, 11, 601-617.  Pitcher, T.J., Hart, P.J.B. 1995. The impact of species changes in African lakes. Chapman & Hall.  Pitcher, T.J., Kalikoski, D., Short, K., Varkey, D., Ganapathiraju, P. 2009. An evaluation of progress in implementing ecosystem-based management of fisheries in 33 Countries. Marine Policy, 33, 223-232.  Plaganyi, E.E., Butterworth, D.S. 2004. A critical look at the potential of Ecopath with Ecosim to assist in practical fisheries management. African Journal of Marine Science, 26, 261-287.  Plaganyi, E.P. 2007. Models for an ecosystem approach to fisheries. FAO Fisheries Technical Paper (FAO).  Polovina, J.J. 1984. Model of a coral reef ecosystem. Coral Reefs, 3, 1-11.  Pospelova, V., Esenkulova, S., Johannessen, S.C., O'Brien, M.C., Macdonald, R.W. 2010. Organic-walled dinoflagellate cyst production, composition and flux from 1996 to 1998 in the central Strait of Georgia (BC, Canada): A sediment trap study. Marine Micropaleontology, 75, 17-37. 132  Power, M.E., Tilman, D., Estes, J.A., Menge, B.A., Bond, W.J., Mills, L.S., Daily, G., Castilla, J.C., Lubchenco, J., Paine, R.T. 1996. Challenges in the quest for keystones. Bioscience, 46, 609-620.  Preikshot, D. ,2007. The influence of geographic scale, climate and trophic dynamics upon north Pacific oceanic ecosystem models. Ph.D thesis. University of British Columbia.  Raven, P.A., Devlin, R.H., Higgs, D.A. 2006. Influence of dietary digestible energy content on growth, protein and energy utilization and body composition of growth hormone transgenic and non-transgenic coho salmon (Oncorhynchus kisutch). Aquaculture, 254, 730-747.  Razak, S.A., Hwang, G., Rahman, M.A., Maclean, N. 1999. Growth performance and gonadal development of growth enhanced transgenic tilapia Oreochromis niloticus (L.) following heat-shock-induced triploidy. Marine Biotechnology, 1, 533-544.  Read, A.J. 2008. The looming crisis: Interactions between marine mammals and fisheries. Journal of Mammalogy, 89, 541-548.  Ross, R.M., Quetin, L.B. 1989. Energetic cost to develop to the first feeding stage of Euphausia superba Dana and the effect of delays in food availability. Journal of Experimental Marine Biology and Ecology, 133, 103-127.  Rumsey, S.M., Franks, P.J.S. 1999. Influence of variability in larval development on recruitment success in the euphausiid Euphausia pacifica: elasticity and sensitivity analyses. Marine Biology, 133, 283-291.  133  Schweigert, J.F., Thompson, M., Fort, C., Hay, D.E., Therriault, T.W., Brown, L.N. Accepted. Factors linking Pacific herring (clupea pallasi) productivity and the spring plankton bloom in the Starit of Georgia, British Columbia, Canada. Progress in oceanography.  Schwing, F.B., Mendelssohn, R., Bograd, S.J., Overland, J.E., Wang, M., Ito, S. 2010. Climate change, teleconnection patterns, and regional processes forcing marine populations in the Pacific. Journal of Marine Systems, 79, 245-257.  Schwing, F.B., Murphree, T., Green, P.M. 2002. The Northern Oscillation Index (NOI): a new climate index for the northeast Pacific. Progress in Oceanography, 53, 115-139.  Shurin, J.B., Borer, E.T., Seabloom, E.W., Anderson, K., Blanchette, C.A., Broitman, B., Cooper, S.D., Halpern, B.S. 2002. A cross‐ecosystem comparison of the strength of trophic cascades. Ecology Letters, 5, 785-791.  Siegel, V. 2000. Krill (Euphausiacea) life history and aspects of population dynamics. Canadian Journal of Fisheries and Aquatic Sciences, 57, 130-150.  Sinclair, A.R.E., Mduma, S., Brashares, J.S. 2003. Patterns of predation in a diverse predator–prey system. Nature, 425, 288-290.  Stachowicz, J.J., Bruno, J.F., Duffy, J.E. 2007. Understanding the effects of marine biodiversity on communities and ecosystems. Annual Review of Ecology, Evolution, and Systematics, 38, 739-766.  134  Stenseth, N.C., Mysterud, A. 2005. Weather packages: finding the right scale and composition of climate in ecology. Journal of Animal Ecology, 74, 1195-1198.  Stenseth, N.C., Mysterud, A., Ottersen, G., Hurrell, J.W., Chan, K.S., Lima, M. 2002. Ecological effects of climate fluctuations. Science, 297, 1292.  Stenseth, N.C., Ottersen, G., Hurrell, J.W., Mysterud, A., Lima, M., Chan, K.S., Yoccoz, N.G., Adlandsvik, B. 2003. Studying climate effects on ecology through the use of climate indices: the North Atlantic Oscillation, El Nino Southern Oscillation and beyond. Proceedings of the Royal Society of London, Series B: Biological Sciences, 270, 2087-2096.  Sun, S., Tao, Z., Li, C., Liu, H. 2011. Spatial distribution and population structure of Euphausia pacifica in the Yellow Sea (2006–2007). Journal of Plankton Research, 33, 873-889.  Sundström, L.F., Lõhmus, M., Devlin, R.H. 2010. Migration and growth potential of coho salmon smolts: implications for ecological impacts from growth-enhanced fish. Ecological Applications, 20, 1372-1383.  Sundström, L.F., Lohmus, M., Devlin, R.H., Johnsson, J.I., Biagi, C.A., Bohlin, T. 2004. Feeding on profitable and unprofitable prey: comparing behaviour of growth‐enhanced transgenic and normal coho Salmon (Oncorhynchus kisutch). Ethology, 110, 381-396.  135  Sundström, L.F., Devlin, R.H., Johnsson, J.I., Biagi, C.A. 2003. Vertical position reflects increased feeding motivation in growth hormone transgenic coho salmon (Oncorhynchus kisutch). Ethology, 109, 701-712.  Sundström, L.F., Lohmus, M., Devlin, R.H. 2005. Selection on increased intrinsic growth rates in coho salmon, Oncorhynchus kisutch. Evolution, 59, 1560-1569.  Sundström, L.F., Lohmus, M., Johnsson, J.I., Devlin, R.H. 2004. Growth hormone transgenic salmon pay for growth potential with increased predation mortality. Proceedings of the Royal Society of London, Series B: Biological Sciences, 271, 350-352.  Sundström, L.F., Lohmus, M., Tymchuk, W.E., Devlin, R.H. 2007. Gene-environment interactions influence ecological consequences of transgenic animals. Proceedings of the National Academy of Sciences, USA, 104, 3889-3894. Swartz, W., Pauly, D. ,2008. Who’s eating all the fish? The food security rationale for culling cetaceans. IWC 60.  Tanasichuk, R.W. ,1998. Interannual variations in the population biology and productivity of Euphausia pacifica in Barkley Sound, Canada, with special reference to the 1992 and 1993 warm ocean years. Marine Ecology Progress Series, 173, 163-180.  Taylor, F.J.R., Horner, R.A. 1994. Red tides and other problems with harmful algal blooms in Pacific Northwest Coastal Waters. Canadian technical report of fisheries and aquatic sciences, 1948, 175-186.  136  Trenberth, K.E., Hurrell, J.W. 1994. Decadal atmosphere-ocean variations in the Pacific. Climate Dynamics, 9, 303-319.  Trites, A.W., Christensen, V., Pauly, D. 1997. Competition between fisheries and marine mammals for prey and primary production in the Pacific Ocean. Journal of Northwest Atlantic Fishery Science, 22, 173-187.  Trites, A.W., Miller, A.J., Maschner, H.D., Alexander, M.A., Bograd, S.J., Calder, J.A., Capotondi, A., Coyle, K.O., Lorenzo, E.D., Finney, B.P. 2007. Bottom-up forcing and the decline of Steller sea lions (Eumetopias jubatas) in Alaska: assessing the ocean climate hypothesis. Fisheries Oceanography, 16, 46-67.  Trites, A.W., Donnelly, C.P. 2003. The decline of Steller sea lions Eumetopias jubatus in Alaska: a review of the nutritional stress hypothesis. Mammal Review, 33, 3-28.  Turvey, S.T., Pitman, R.L., Taylor, B.L., Barlow, J., Akamatsu, T., Barrett, L.A., Zhao, X., Reeves, R.R., Stewart, B.S., Wang, K., Wei, Z., Zhang, X., Pusser, L.T., Richlen, M., Brandon, J.R., Wang, D. 2007. First human-caused extinction of a cetacean species? Biology Letters, 3, 537-540.  Van Holmes, C. 2007. Mean spring and fall upwelling transition dates off the Oregon and Washington coasts. , Available: http://www.cbr.washington.edu/data/trans_data.html.  Wallace, J.M., Gutzler, D.S. 1981. Téléconnexions in the geopotential height field during the Northern Hemisphere winter. American Meteorological Society, 109, 784-812.  137  Walters, C., Korman, J. 1999. Linking recruitment to trophic factors: revisiting the BevertonHolt recruitment model from a life history and multispecies perspective. Reviews in Fish Biology and Fisheries, 9, 187-202.  Walters, C., Martell, S.J. 2004. Fisheries ecology and management. Princeton University Press. Princeton, NJ.  Walters, C.J., Juanes, F. 1993. Recruitment limitation as a consequence of natural selection for use of restricted feeding habitats and predation risk taking by juvenile fishes. Canadian Journal of Fisheries and Aquatic Sciences, 50, 2058-2070.  Walters, C., Christensen, V., Pauly, D. 1997. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries, 7, 139-172.  Walters, C., Kitchell, J.F. 2001. Cultivation/depensation effects on juvenile survival and recruitment: implications for the theory of fishing. Canadian Journal of Fisheries and Aquatic Sciences, 58, 39-50.  Ward, P., Myers, R.A. 2005. Shifts in open-ocean fish communities coinciding with the commencement of commercial fishing. Ecology, 86, 835-847.  Whitney, F.A., Welch, D.W. 2002. Impact of the 1997-1998 El Niño and 1999 La Niña on nutrient supply in the Gulf of Alaska. Progress in Oceanography, 54, 405-421.  Worm, B., Barbier, E.B., Beaumont, N., Duffy, J.E., Folke, C., Halpern, B.S., Jackson, J.B.C., Lotze, H.K., Micheli, F., Palumbi, S.R., Sala, E., Selkoe, K.A., Stachowicz, J.J., 138  Watson, R. 2006. Impacts of biodiversity loss on ocean ecosystem services. Science, 314, 787-790.  Yang, M.S., Nelson, M.W. ,1999. Food habits of the commercially important groundfishes in the Gulf of Alaska in 1990, 1993, and 1996. NOAA Tech. Memo. NMFS-AFSC 22:150p.  Yin, K., Harrison, P.J., Beamish, R.J. 1997. Effects of a fluctuation in Fraser River discharge on primary production in the central Strait of Georgia, British Columbia, Canada. Canadian Journal of Fisheries and Aquatic Sciences, 54, 1015-1024.  Yodzis, P. 2001. Must top predators be culled for the sake of fisheries? Trends in Ecology & Evolution, 16, 78-84.  Zhu, Z., He, L., Chen, S. 1985. Novel gene transfer into the fertilized eggs of gold fish (Carassius auratus L. 1758). Journal of Applied Ichthyology, 1, 31-34.  139  Appendices Appendix A Basic parameters of the Ecopath with Ecosim model Trophic levels are estimated from the diet matrix (Appendix B) and the vulnerabilities are scaled linearly with trophic level of the prey, and range from 1 to 15.  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39  Groups Res. Orca Trans. Orca Humpbacks Sea lions Seals Sea birds O. Sharks Dogfish Ratfish Halibut J. Pollock A. Pollock A. Hake J. Hake Cod Lingcod Rockfish Soles O. Demersals Trans. salmon Res. salmon Eulachon A. Herring J. Herring sand lance O. Pelagics Jellyfish Squid Crabs Clams Shrimps O. Benthos Krill Copepods O. Zooplankton Kelp/Sea Grass Macro algae Phytoplankton Detritus  Trophic Level 4.64 5.51 3.86 4.01 4.6 3.86 3.89 3.75 3.71 3.81 3.49 3.99 4.01 3.81 4.06 4.3 3.96 3.35 3.53 3.64 3.87 3.4 3.52 3.23 2.64 3.55 3 2.99 2.86 2.23 2.91 2.18 2.88 2.16 2.04 1 1 1 1  Vulnerabilies 12.3 15 10 10.3 12.3 10 9.8 9.6 9.3 9.5 8.7 10.2 10.4 9.8 10.4 11.3 10.2 8.1 8.7 9.4 9.9 8.7 9.8 8.4 6 9.1 7.2 7.1 6.6 4.7 7 4.5 6.9 5.3 4.1 1 1 1 1  Biomass (t·km-2) 0.015 0.015 0.01 0.037 0.287 0.02 0.096 5.71 2 0.05 0.0396 1.143 3 1.326 0.07 0.556 0.429 0.778 0.4 1 0.126 0.0063 6.6 9.997 3 0.4 12.5 0.7 8 9 1 4.69 20 15 30 20.3 8 40 38.7  Sources of Biomass Estimation based on (1) Estimation & tuning (2) (3; 4) (5) (3) Based on (2) McFarlane Pers. Comm. Fu Pers. Comm. Ecopath estimate Ecopath estimate McFarlane Pers. Comm. (6) Ecopath estimate Ecopath estimate (7) McFarlane Pers. Comm. (3) McFarlane Pers. Comm. Irvine Pers. Comm. Irvine Pers. Comm. (8) (9) Ecopath estimate (10) Estimation based on (11) (2) (2) (3) Estimation based on catch data Estimation based on (2) Ecopath estimate Estimation based on (2) Estimation based on (2) Estimation based on (2) (11) Estimation based on (2) (2) (3)  140  Appendix A (cont.)  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38  Groups Res. Orca Trans. Orca Humpbacks Sea lions Seals Sea birds O. Sharks Dogfish Ratfish Halibut J. Pollock A. Pollock A. Hake J. Hake Cod Lingcod Rockfish Soles O. Demersals Trans. salmon Res. salmon Eulachon A. Herring J. Herring sand lance O. Pelagics Jellyfish Squid Crabs Clams Shrimps O. Benthos Krill Copepods O. Zooplankton Kelp/Sea Grass Macro algae Phytoplankton  P/B (year-1) 0.04 0.04 0.02 0.16 0.14 0.14 0.1 0.19 0.06 0.4 0.8 0.4 0.55 1.8 0.66 0.5 0.19 0.55 1.58 2.48 2.2 1.8 0.7 1.1 1.6 2.3 9.6 3 1.5 0.9 11.48 4.5 6.1 27 30 4.43 9 130  Sources (1; 12) (12) (1) (11) (2) (16) (2) (2) (12) (12) (2) (2) (11) (11) & tuning (2) (2) & tuning (2) (2; 12) Ecopath (12) (11) (12) & tuning (2, 12) (2) (17) and P/Q (2) (2) (2) (2) (2) (12) (12) (12) (12) (2) & tuning (11) (2) (2)  Q/B (year-1) 13 7.4 9.1 15.08 13.5 80 1 2.7 1.4 1.1 6.72 2 5 10.98 1.8 2.4 2.6 2.3 5.26 8.33 7.3 8.4 4.4 7.58 5.24 7.67 20 12 3.5 3 45.9 14 24.82 90 80  Sources (1) (13) (1) (1) (14; 15) (16) (2) (2) (12) (12) Ecopath estimate (2) (11) Ecopath estimate (2) (2) (2) (2) (2) (12) (11) (12) (2) Ecopath estimate (10) Ecopath estimate (2) and P/Q (2) (2) Ecopath estimate (12) Ecopath estimate (12) (12) (2) Ecopath estimate Ecopath estimate  141  Appendix B Diet matrix of the Strait of Georgia 2005 Ecopath model Diet of whales is based on (1; 18), seals (14), other sharks (19), hake (20), dogfish (12, 21), pollock, jellyfish, clams and other benthos (2), sand lance (10), zooplankton (16) and other groups (12). Predator Res. Orca  Trans. Orca  Humpbacks  Sealions  Seals  Prey O. Sharks Dogfish Trans. salmon Res. salmon A. Herring J. Herring Sand lance Squid Import Sealions Seals Import A. Herring J. Herring Krill Copepods Dogfish Halibut J. Pollock A. Pollock A. Hake J. Hake Cod Lingcod Rockfish Soles O. Demersals Trans. salmon Res. salmon A. Herring J. Herring Sand lance O. Pelagics Squid Crabs Clams O. Benthos Import Dogfish J. Pollock A. Pollock  Diet 0.03 0.03 0.05 0.2 0.01 0.02 0.02 0.04 0.6 0.05 0.25 0.7 0.05 0.05 0.8 0.1 0.02 0.003 0.001 0.001 0.02 0.016 0.003 0.002 0.003 0.01 0.02 0.005 0.0005 0.01 0.1 0.0001 0.045 0.05 0.045 0.056 0.09 0.5 0.001 0.003 0.05  Predator  Sea birds  O. Sharks  Prey A. Hake J. Hake Cod Lingcod Rockfish Soles O. Demersals Trans. salmon Res. salmon Eulachon A. Herring J. Herring Sand lance O. Pelagics Squid O. Demersals Res. salmon Eulachon A. Herring J. Herring Sand lance O. Pelagics Jellyfish Squid Crabs Shrimps O. Benthos Krill Copepods Import Dogfish Ratfish Halibut J. Pollock A. Pollock A. Hake J. Hake Cod Lingcod Rockfish Soles  Diet 0.2 0.171 0.003 0.02 0.001 0.0005 0.016 0.057 0.026 0.0003 0.08 0.28 0.03 0.034 0.027 0.04 0.002 0.001 0.129 0.08 0.172 0.02 0.036 0.09 0.041 0.04 0.041 0.112 0.156 0.04 0.1 0.006 0.008 0.001 0.04 0.052 0.003 0.026 0.02 0.018 0.03  142  Appendix B (cont.) Predator  Dogfish  Ratfish  Halibut  Prey O. Demersals Sand lance Squid Crabs Clams Shrimps O. Benthos Krill Ratfish A. Hake J. Hake Cod Rockfish Soles O. Demersals Trans. salmon Res. salmon Eulachon A. Herring J. Herring Sand lance O. Pelagics Jellyfish Squid Crabs Clams O. Benthos Krill Copepods Detritus Eulachon A. Herring O. Pelagics Crabs Clams Shrimps O. Benthos Krill O. Sharks Cod Lingcod Rockfish Soles O. Demersals Eulachon A. Herring O. Pelagics  Diet 0.005 0.001 0.05 0.1 0.12 0.1 0.31 0.01 0.005 0.02 0.026 0.0001 0.0001 0.0005 0.005  Predator  J. Pollock  A. Pollock  0.015 0.002 0.0001 0.03 0.121 0.006 0.006 0.035 0.0001 0.083 0.019 0.03 0.4 0.116 0.08 0.0001 0.079 0.02 0.248 0.15 0.1 0.2 0.203 0.053 0.07 0.005 0.011 0.13 0.09 0.003 0.02 0.014  A. Hake  J. Hake Cod  Prey Squid Clams Shrimps O. Benthos Detritus O. Demersals O. Pelagics Jellyfish Squid Crabs O. Benthos Krill Copepods O. Zooplankton J. Pollock A. Hake J. Hake Cod Soles O. Demersals Trans. salmon Res. salmon O. Pelagics Crabs Shrimps O. Benthos Krill Copepods O. Zooplankton A. Pollock J. Hake O. Demersals Eulachon A. Herring J. Herring Sand lance O. Pelagics Squid Shrimps Krill Krill Copepods O. Sharks Dogfish Cod Lingcod O. Demersals  Diet 0.128 0.28 0.03 0.09 0.076 0.079 0.079 0.007 0.01 0.017 0.032 0.183 0.2 0.393 0.001 0.03 0.14 0.008 0.073 0.035 0.01 0.03 0.035 0.02 0.071 0.145 0.301 0.04 0.06 0.005 0.02 0.002 0.00001 0.02 0.31 0.123 0.01 0.02 0.22 0.27 0.9 0.1 0.005 0.005 0.02 0.014 0.135  143  Appendix B (cont.) Predator  Lingcod  Rockfish  Soles  O. Demersals  Prey Eulachon A. Herring J. Herring Sand lance O. Pelagics Shrimps Dogfish Halibut J. Pollock A. Pollock A. Hake J. Hake Cod Lingcod Rockfish Soles O. Demersals Eulachon A. Herring J. Herring O. Pelagics Shrimps O. Benthos O. Demersals Eulachon A. Herring J. Herring Sand lance O. Pelagics Crabs Shrimps O. Benthos Krill Lingcod Sand lance O. Pelagics Crabs Clams Shrimps O. Benthos Krill O. Demersals Eulachon Sand lance O. Pelagics Crabs  Diet 0.021 0.023 0.03 0.071 0.393 0.06 0.111 0.001 0.01 0.078 0.1 0.015 0.002 0.05 0.015 0.082 0.06 0.0001 0.1 0.099 0.02 0.1 0.1 0.045 0.001 0.02 0.33 0.001 0.06 0.123 0.1 0.17 0.15 0.003 0.001 0.001 0.08 0.07 0.141 0.7 0.004 0.01 0.0001 0.01 0.001 0.34  Predator  Trans. salmon  Res. salmon  Eulachon  A. Herring  J. Herring Sand lance  O. Pelagics  Jellyfish  Squid  Prey Clams Shrimps O. Benthos Krill Copepods Detritus Jellyfish Krill Copepods Import Eulachon A. Herring J. Herring Sand lance O. Pelagics Squid Crabs Krill Copepods Jellyfish O. Benthos Krill Copepods Krill Copepods Import Krill Copepods Copepods O. Zooplankton Kelp/Sea Grass Detritus Jellyfish Krill Copepods Detritus Jellyfish Krill Copepods O. Zooplankton Phytoplankton J. Herring Sand lance Jellyfish Squid O. Benthos  Diet 0.138 0.15 0.25 0.025 0.048 0.028 0.1 0.15 0.15 0.6 0.001 0.02 0.02 0.05 0.02 0.2 0.02 0.6 0.069 0.2 0.1 0.1 0.6 0.1 0.1 0.8 0.1 0.9 0.1 0.5 0.2 0.2 0.32 0.25 0.38 0.05 0.05 0.12 0.2 0.43 0.2 0.025 0.082 0.06 0.06 0.2  144  Appendix B (cont.) Predator  Prey Krill Copepods Kelp/Sea Grass Detritus Crabs Soles Crabs Clams Shrimps O. Benthos Macro algae Detritus Clams O. Zooplankton Phytoplankton Detritus Copepods O. Zooplankton Phytoplankton  Diet 0.127 0.073 0.1 0.273 0.002 0.05 0.15 0.153 0.245 0.1 0.3 0.222 0.444 0.333 0.15 0.85  Predator Shrimps  O. Benthos  Krill  O. Zooplankton  Prey O. Benthos Krill Copepods Macro algae Detritus O. Benthos O. Zooplankton Kelp/Sea Grass Macro algae Phytoplankton Detritus Copepods O. Zooplankton Phytoplankton Phytoplankton  Diet 0.1 0.3 0.2 0.1 0.3 0.05 0.115 0.1 0.1 0.3 0.335 0.4 0.4 0.2 1  145  Appendix C List of Abbreviations in Appendix A and B Res.  Resident  Trans.  Transient  O.  Other  J.  Juvenile  A.  Adult  146  Appendix D Appendix references  1. Trites, A., and Heise, K., 1996. Marine mammals. Page 51-55 in: Pauly, D., Christensen, V. and Haggan, N. (Eds.) Mass-Balance Models of North-Eastern Pacific Ecosystems. Fisheries Centre Research Reports. 4(1): 131 pp.  2. Preikshot, D., 2007. The influence of geographic scale, climate and trophic dynamics upon north pacific oceanic ecosystem models. Ph.D thesis. University of British Columbia, Vancouver, BC, Canada.  3. Beamish, R.J., McFarlane, G.A., Neville, C.M., and Pearsall, I., 2001. Changes in the Strait of Georgia ECOPATH model needed to balance the abrupt increases in productivity that occurred in 2000. PICES Sci. Rept., 17: 5-9.  4. DFO., 2008. Population assessment: Steller sea lion (Eumetopias jubatus). DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2008/047.  5. DFO., 2010. Population assessment Pacific harbour seal (Phoca vitulina richardsi). DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2009/011.  6. King, J.R., McFarlane, G.A., 2006. Shift in Size-at-age of the Strait of Georgia Population of Pacific hake (Merluccius productus). California Cooperative Oceanic Fisheries Investigations Report, 47, 111-118.  7. DFO., 2005. Strait of georgia lingcod (Ophiodon elongatus) assessment and advice for fishery management. DFO Can. Advis. Sec. Sci. Advis. Rep. 2005/042.  147  8. DFO., 2007. Fraser river eulachon (Thaleichthys pacificus): 2007 population assessment and harvest recommendations for 2008. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2007/048.  9. DFO., 2009. Stock assessment report on Pacific herring in British Columbia. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2009/059.  10. Mackinson, S., and Daskalov, G., 2008. An ecosystem model of the North Sea to support an ecosystem approach to fisheries management: description and parameterisation. Science Series Technical Report. 142.  11. Martell, S.J.D., Beattie, A.I., Walters, C.J., Nayar, T., and Briese, R., 2002. Simulating fisheries management strategies in the Strait of Georgia ecosystem using Ecopath and Ecosim. Fisheries Centre Research Reports. 10(2): 16-23.  12. Ainsworth, C., 2006. Strategic marine ecosystem restoration in northern British Columbia. Ph. D. Thesis. University of British Columbia, Vancouver, BC, Canada. 422 pp.  13. Winship, A., 1998. Pinnipeds and cetaceans in the Strait of Georgia. Back to the Future: Reconstructing the Strait of Georgia Ecosystem. Fisheries Centre Research Reports. 6(5): 53-57.  14. Olesiuk, P.F., Bigg, M.A., Ellis, G.M., Crockford, S.J., and Wigen, R.J., 1990. An assessment of the feeding habits of harbour seals (Phoca vitulina) in the Strait of  148  Georgia, British Columbia, based on scat analysis. Canadian Technical Report of Fisheries and Aquatic Sciences. 1730: 1-135.  15. Olesiuk, P.F., 1999. An assessment of the status of harbour seals (Phoca vitulina) in British Columbia. Canadian Stock Assessment Secretariat Research Document 1999/33 (33): 130.  16. Aydin, K.Y., McFarlane, G.A., King, J.R., and Megrey, B.A., 2003. PICES-GLOBEC International Program on Climate Change and Carrying Capacity. PICES Sci. Rept. 25: 93 p.  17. Robards, M.D., Willson, M.F., Armstrong, R.H., and Piatt, J.F., 1999. Sand lance: a review of biology and predator relations and annotated bibliography. Res. Pap. PNWRP-521. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station.  18. Ford, J.K.B., Ellis, G.M., Barrett-Lennard, L.G., Morton, A.B., Palm, R.S., and Balcomb III, K.C., 1998. Dietary specialization in two sympatric populations of killer whales (Orcinus orca) in coastal British Columbia and adjacent waters. Canadian Journal of Zoology 76(8): 1456-1471.  19. Cortes, E., 1999. Standardized diet compositions and trophic levels of sharks. ICES Journal of Marine Science 56(5): 707.  20. McFarlane, G.A., and Beamish, R.J., 1985. Biology and fishery of Pacific whiting, Merluccius productus, in the Strait of Georgia. Marine Fisheries Review 47(2): 23-34.  149  21. Tanasichuk, R.W., Ware, D.M., Shaw, W.,McFarlane, G.A., 1991. Variations in diet, daily ration, and feeding periodicity of pacific hake (Merluccius productus) and spiny dogfish (Squalus acanthias) off the lower west coast of Vancouver Island. Canadian Technical Report of Fisheries and Aquatic Sciences. 48: 2118-2128.  150  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073447/manifest

Comment

Related Items