https://orcid.org/0000-0002-4735-9403
Figures
Abstract
Along the northeast Pacific coast, the salmon-eating southern resident killer whale population (SRKW, Orcinus orca) have been at very low levels since the 1970s. Previous research have suggested that reduction in food availability, especially of Chinook salmon (Oncorhynchus tshawytscha), could be the main limiting factor for the SRKW population. Using the ecosystem modelling platform Ecopath with Ecosim (EwE), this study evaluated if the decline of the Pacific salmon populations between 1979 and 2020 may have been impacted by a combination of factors, including marine mammal predation, fishing activities, and climatic patterns. We found that the total mortality of most Chinook salmon populations has been relatively stable for all mature returning fish despite strong reduction in fishing mortality since the 1990s. This mortality pattern was mainly driven by pinnipeds, with increases in predation between 1979 and 2020 mortality ranging by factors of 1.8 to 8.5 across the different Chinook salmon population groups. The predation mortality on fall-run Chinook salmon smolts originating from the Salish Sea increased 4.6 times from 1979 to 2020, whereas the predation mortality on coho salmon (Oncorhynchus kisutch) smolts increased by a factor of 7.3. The model also revealed that the north Pacific gyre oscillation (NPGO) was the most important large-scale climatic index affecting the stock productivity of Chinook salmon populations from California to northern British Columbia. Overall, the model provided evidence that multiple factors may have affected Chinook salmon populations between 1979 and 2020, and suggested that predation mortality by marine mammals could be an important driver of salmon population declines during that time.
Citation: Couture F, Christensen V, Walters C (2024) The combined effects of predation, fishing, and ocean productivity on salmon species targeted by marine mammals in the northeast Pacific. PLoS ONE 19(3): e0296358. https://doi.org/10.1371/journal.pone.0296358
Editor: Andrea Belgrano, Swedish University of Agricultural Sciences and Swedish Institute for the Marine Environment, University of Gothenburg, SWEDEN
Received: June 16, 2023; Accepted: December 11, 2023; Published: March 14, 2024
Copyright: © 2024 Couture et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data can be found within the manuscript (Table 4) and its Supporting Information files (Table A2-A7). Some data underlying the results presented in this study are third party data, and available from the Chinook Technical Committee at the Pacific Salmon Commission (https://www.psc.org/about-us/structure/committees/technical/chinook/). To access the specific dataset related to this research, please contact the Chinook Technical Committee at the Pacific Salmon Commission at info@psc.org. Some data were also directly provided by Fisheries and Oceans Canada. We hereby confirm that the authors of this study did not have any special privileges to access those data.
Funding: We acknowledge that FC received support through Mitacs Accelerate Award (https://www.mitacs.ca/en/programs/accelerate) IT13063 in cooperation with Ocean Wise (https://ocean.org/), and NSERC Discovery Grant RGPIN-2019-04901. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. there was no additional external funding received for this study.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The Southern resident Killer Whale (SRKW, Orcinus orca) population inhabits the coastal waters between northern British Columbia and central California [1]. SRKW are highly specialized predators, feeding primarily on Pacific salmon species [2–4]. Chinook salmon (Oncorhynchus tshawytscha) is considered a prey of choice for SRKW during the summer months in the Salish Sea, where they can constitute up to 90% of the diet [3]. The SRKW population has been at a very low level since its identification as a distinct ecotype in the early 1970s, and was listed as Endangered by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) in 2001, and under the Endangered Species Act (ESA) in the US in 2005 [1, 5]. As of July 2022, the population only counted 73 individuals, and the abundance, quality, and availability of prey have been designated as major drivers of the SRKW population trend over the last 50 years [6–8]. The question of the potential causes behind the low levels of the SRKW population is of critical importance, as another salmon-eating resident killer whale population (i.e. the Northern resident Killer Whale, NRKW, Orcinus orca) geographically overlaps with the SRKW [1]. Although largely feeding on Chinook salmon as well, the NRKW population has more than doubled since 1975, and scientists are still baffled on what could explain the different demographic trends of those two populations [1].
Most Chinook salmon populations along the northeast Pacific have decreased in abundance over recent decades, and an ongoing debate exists regarding the causes of these declines. While some studies [9–11] argue that the growing pinniped populations now have a substantial predation impact on those salmon populations, others suggest that fishing [12], climatic patterns [13–16], and deterioration of freshwater habitats [17, 18] could have significantly impacted the productivity of Chinook salmon populations over the years.
Several pinniped species occur along the northeast Pacific coast, including harbor seals (Phoca vitulina) and Steller (Eumetopias jubatus) and California (Zalophus californianus) sea lions [19–22]. All three species had been intensely depleted by commercial harvests and predator control programs in Canada and the US prior to their protection under the Fisheries Act in Canada in 1970 and the Marine Mammal Protection Act in the U.S. 1972 [23]. All populations have grown in size since then, with the harbor seal population stabilizing in the early 2000s throughout its range [20, 21, 24]. In the Strait of Georgia only, the harbor seal population has increased by about 20-fold between 1970 and the early 2000s [20]. Further, the population sizes of Steller and California sea lion have increased about six and four times since the early 1970s, respectively [22, 25]. Although constituting a limited portion of their diet, all three species feed upon Pacific salmonids, including Chinook salmon [11, 26–28].
The potential impact of increased predation pressure on depleted Pacific salmon populations by harbor seal and sea lion populations has received considerable attention in recent years. Chasco et al. (2017) estimated that the annual biomass of Chinook salmon consumed by pinnipeds along the northeast Pacific coast increased by almost ten-fold between 1970 and 2015 [10]. In 2015, that consumption would have been equivalent to that of all resident killer whales, and two times higher than all fisheries combined [10]. In the Columbia River only, Sorel et al. (2020) established a negative correlation between Chinook salmon survival and pinniped abundance, with overall mortality increasing by about 21.1% between years of low and high California sea lion abundance [29]. Despite those findings, there is still a real need to understand whether this growing predation pressure on Pacific salmon species over the last 40 years could have significantly impacted the overall availability of prey available to SRKW, ultimately limiting the growth of the population.
Climatic patterns have also been advocated as being an important driver for salmon population declines along the northeast Pacific coast. Crozier et al. (2008, 2021) suggested that changes in streamflow and temperatures associated with climate change might pose a direct threat to Chinook salmon survival during their freshwater stages [16, 30]. For instance, high temperature might reduce stream flow, hence potentially decreasing potential habitat size and further increasing food competition and vulnerability to predation [30]. Several large scale seasonal and annual environmental indicators patterns, such as the Pacific Decadal Oscillation (PDO), North Pacific Gyre Oscillation (NPGO), or El-Niño Southern Oscillation (ENSO), have also been of particular interests for scientists and fisheries managers in recent years. For instance, the NPGO has been shown to correlate with variations in Pacific salmon population productivity [14] and survival [31], while the PDO has been found to be associated with changes in returning spawner abundance [32]. In addition, Kilduff et al. (2015) suggested that the NPGO was the most important driver associated with changes in Chinook and coho salmon survival along the north American west coast since the 1980s [31]. Other studies [33, 34] found similar co-variations between NPGO and other organisms along the northeast Pacific Ocean, including invertebrates, demersal fish, and phytoplankton.
Fishing has been a major cause of Pacific salmon populations’ decline until the late twentieth century. Prior to European contact, Argue et al. (1990) estimated that about 18,000 metric tons of Pacific salmon were consumed annually by aboriginal communities along the coast, which would be equivalent to the fisheries levels observed in the early 2000s [35, 36]. In the early 1800s, salmon were also harvested by fur traders for subsistence and trade [36]. By the mid-1860s, Pacific salmon were considered the cornerstone of the fishing industry of British Columbia [35, 37]. At this time, new fishing techniques, such as trolling, allowed for more intensive and extensive fishing activities [36]. The fishing industry continued to expand over the next decades, with over 7,000 vessels and almost 100 salmon canneries operating in British Columbia by the 1920s [36]. Between 1970 and 1990, more than 1,000,000 Chinook salmon were caught annually along this coast [36]. Intense fishing activity was found to have significant impacts on the abundance, age, and size of returning Chinook salmon individuals [38]. More recently, Hard et al. (2008) showed that high fishing pressure could even induce evolutionary changes (i.e. morphology, migration timing, and maturation timing) on all Pacific salmon species [39]. However, in the face of the dramatic decline of most Chinook salmon stocks, drastic harvesting restrictions (e.g., seasonal closures, gear, licenses) were implemented in the mid-1990s along the northeast Pacific coast [40]. Altogether, recent evidence [9, 41] supports the argument that the current low levels of most Pacific Chinook salmon populations cannot not be solely attributed to fishing pressure. Ward et al. (2013) argued that a fishing reduction could have mainly had a beneficial impact in the early 1980s [12]. Following this claim, O’Farrell and Satterthwaite (2015) showed that the highest impact of fisheries on Chinook salmon in California would also have occurred in the mid-1980s [41]. Besides, despite the reduction in fishing activities over the past 40 years, the overall mortality rates of Chinook salmon along the northeast Pacific have kept increasing, likely driven mainly by predation by marine mammals [9]. In comparison to the findings presented by Chasco et al. (2017), this research will further investigate the impact of those declining Chinook salmon populations for the SRKW. Our study will also provide an additional level of understanding regarding the potential role of primary productivity fluctuations in those salmonid populations.
In this study we use an ecosystem modelling approach to evaluate if the decline of Pacific salmon populations could be due to a combination of factors, rather than to any single factor as evaluated by most previous studies. Choosing an ecosystem-based approach might also allow for developing more realistic and sustainable conservation and management plans for both Pacific salmon and SRKW. This is of significant importance, especially given that both salmon and killer whales are iconic and highly commercially and ecologically valuable species of great conservation concern.
Analytical methods became an important tool for scientists and ecologists to understand animal population dynamics in the 1930s [42]. Ecosystem modelling is increasingly recognized as a convincing method to provide an understanding of how specific perturbations could affect marine ecosystems while predicting future population trends [42, 43]. Such approach provides a global representation of an ecosystem’s mechanisms, as it incorporates species life-history characteristics, harvest data, as well as spatial and temporal ecological variations over time [44, 45].
The ecosystem modelling platform Ecopath with Ecosim (EwE) was ranked the most applied tool for modelling marine and aquatic ecosystems worldwide [46, 47]. The software has been used to address a variety of questions, including fisheries impacts [44, 48, 49], efficiency of marine protected areas [46, 50, 51], habitat loss [52, 53], interactions between marine species [54–58], environmental impact assessments [59–62], pollution [63] as well as climate change and associated environmental variations [64–66]. This modelling approach has been widely applied in the northeast Pacific, with about 37 EwE models created to date for different areas extending from the Aleutian Islands to Mexico [67]. Among those models, some [68, 69] aimed at understanding the bottom-up and top-down processes regulating the ecosystem, while others [70, 71] examined the adequacy of fisheries management plans or intended to provide a general overview of a static system [72, 73]. In this study, the use of such modelling platform would allow to gain valuable insight on the concurrent effects of the three most important factors impacting Pacific salmon species: marine mammal predation, fishing activity, and primary production.
In this paper, we aimed to evaluate to which extent marine mammal predation, fishing activities, and climatic patterns may have affected the population trends of prey available to SRKW between 1979 and 2020. The primary objectives of this paper were to 1) estimate the predicted variations in predation pressure on salmon populations from 1979 to 2020, 2) examine if competition for food might have occurred between the SRKW and other marine mammals during that time, and 3) assess whether environmental variables may have impacted the population trends of prey available to SRKW. The use of the EwE modelling approach will allow to use historical fishing data to drive the fishing pressure on the different Pacific salmon species over the years, while marine mammal diet studies and population demographics will be used as a proxy for marine mammal predation pressure. Ultimately, this research aimed at understanding the ecological linkages existing in the ecosystem while providing a detailed overview of the method and parameters used in an EwE model. This study will fit into the important research effort aiming at conserving economically and culturally valuable Pacific salmon species while ensuring the resilience of an already endangered killer whale population.
Materials and methods
The ecological modeling platform Ecopath with Ecosim (EwE, http://www.ecopath.org) was used to describe the population dynamics related to the interplay between salmon species, killer whales, pinnipeds, fisheries, and environmental conditions between 1979 and 2020 [74]. EwE combines a static analysis (Ecopath) of an ecosystem trophic mass balance (i.e biomass status and trophic flow) with a temporal (Ecosim) dynamic model for representing variations over time [74].
Study area
The study area of the model corresponded to the historical geographical range of SRKW, and expends from the northern tip of the Haida Gwaii Archipelago on the western coast of British Columbia (Canada) to the northern border of California (USA) (Fig 1). It included the continental shelves of British Columbia (~150,000 km2), Washington and Oregon State (~ 62,486 km2), and the northern portion of California (~30,137 km2). The total model area ([133°26’25W, 121°55’42E, 54°22’08N, 36°37’44S]) was estimated at 242,623 km2 [69, 75].
The study area of the model is the blue square, which corresponds to the historical geographical range of SRKW, and extends from the northern tip of the Haida Gwaii Archipelago (1) (central British Columbia, Canada) to Monterey Bay (4) in California (USA). The model area includes the Salish Sea (3) and the Strait of Georgia, which is open water part of the Salish Sea between Vancouver Island (2) and the mainland of British Columbia. This map was created using the R package rnaturalearth.
Ecopath model
The Ecopath model was built to represent the base year 1979, and represented the start for the time dynamic Ecosim analysis. The Ecopath analysis provided a static mass-balanced assessment of stocks and flows for the ecosystem (biomass estimates, mortality rates, and trophic flows) [74]. This analysis relies on two master equations used to describe the production term (Eq 1) and the energy flow between each group of an ecosystem (Eq 2) [74]. In addition to diet compositions for all consumer groups (biomass pools), four parameters were provided or estimated for each group prior to running the Ecopath analysis: the biomass (B, t/km2), the production/biomass ratio (P/B, (year-1)), the consumption/biomass ratio (Q/B, (year-1)), and the fisheries landings (Y, t/km2/year)). When a population is considered at equilibrium (i.e. stable population level), the production (P/B) represents the instantaneous total mortality rate of a functional group (usually referred to as Z in fisheries science) [76]. In the case of a growing population, the logarithm of the finite rate of increase of the population (usually referred to as λ) needs to be added to the calculation of the P/B ratio, so that P/B = Z + ln (λ). In the case of a continuous population growth rate, ln (λ) is also equal to the intrinsic rate of increase of the population, which can be defined as the maximum instantaneous growth rate that a species can achieve under ideal survival and fecundity conditions. The consumption/biomass ratio Q/B corresponds to the total consumption rate of a predator over its biomass.
In Eq 1, DCij * Bi * ∑j = 1(Q/B)i represents the sum of predation mortalities from predators j on a given prey i depending on the relative proportion of the diet predator j obtained from prey i (DCij). Yi represents the fisheries catch of functional group i, Ei the net migration of group i, BAi the bioaccumulation of group i, and EEi the Ecotrophic Efficiency (1- mortalities, i.e. proportion of the production used in the model) of group i. In Eq 2, Ri corresponds to the respiration rate of group i, and UNi to the unassimilated food for group i.
All the parameters sources used to build the Ecopath model for the year 1979 are presented in details in the Section 1 of S1 Appendix.
Functional groups.
Altogether, 74 functional groups were included in the model. A functional group can be defined as a species or a group of ecologically or taxonomically related species, or a life history stage or stanza for one species [77]. In total, species (i.e. Chinook salmon, coho salmon, chum salmon, and herring) were split into ‘multi-stanzas’, allowing to capture ontogenetic diet shifts and different age-specific predation and exploitation patterns [77, 78]. ‘Pedigrees’ indicating how well input data were rooted in local information were associated with most parameters entered in Ecopath, allowing to define confidence intervals around data based on their origin [74, 79]. Those pedigree estimates can be used by the Monte-Carlo simulations of the Ecosim routine to calculate confidence intervals in the different Ecopath ecological indicators [78]. All the parameters and literature sources used to build the Ecopath model for the year 1979 are summed up in this paper and presented in Table 1 (marine mammals), Table 2 (salmon species), and Table 3 (all other functional groups). The references listed in those tables are not intended to serve as reference citations. However, all associated references can be found in the S1 Appendix. All parameter estimates are listed in Table 4, whereas detailed descriptions of how the parameters were calculated for the different functional groups can be found in the Section 1 of S1 Appendix.
Those values were entered based on estimates found in previous literature and were not generated by the model.
Mass-balance.
Mass balancing was the next major step to ensure that ecological and thermodynamics rules were followed [74]. Following the equation that describe the production term (Eq 1), parameters were adjusted when the model-derived ecotrophic efficiency (EE) estimates (i.e. proportion of the production used in the model) were too high (>1) [74, 78]. The EE can also be written as EE = 1−other mortalities. In the case of an EE higher than 1, the ‘other mortalities’ estimate would thus be negative, and several scenarios were investigated, including underestimated biomass or production/biomass ratio, or overestimated predation mortality.
The initial biomass estimates were increased for all Chinook salmon Returning spawners stanzas, which seemed reasonable given that the salmon abundance estimates provided by the Pacific Salmon Commission do not include all salmon stocks present in the model area at a given time [80]. The catch was adjusted for all groups in order to maintain a similar initial fishing mortality F, based on the fact that F = catch/biomass [81].
Following a similar method, the initial biomass of the Adult herring stanza was increased. The biomass estimates initially entered in Ecopath solely involved the spawning biomass and are based on fishery landings of mature individuals (i.e. Age 4+ fish) provided by different governmental organizations [82]. Yet, virtual population analyses and age-structured population dynamics model fitting (i.e. age structure reconstruction) of herring populations suggest that the immature/mature biomass ratio has typically been around 1.33, suggesting that the initial biomass could be multiplied by at least 2.3. Finally, the diet proportions of several functional groups (i.e. pinniped species, salmon species) were adjusted in order to balance the model. All modifications were minor, and deemed reasonable given that most diet studies are carried out in small geographic areas during limited periods of time, hence potentially not accurately representing the annual diet of a species throughout its range.
Ecosim
Ecosim analysis provide dynamic simulation expressing biomass dynamics through a system of differential equations derived from the Ecopath master equations [77, 83]. In this simulation, temporal variations in biomasses (dBi) of a functional group i are estimated by multiplying its consumption (Qji) with the food conversion efficiency gi, adding immigration (Ii) and subtracting mortality terms, i.e. consumption Qij of all predators j, and the product of the group biomass Bi and the summed non-predation mortality rate M0i, emigration rate(ei), fishing mortality rate (Fi) and immigration rate (Ii), expressed as follows:
(Eq 3)
The consumption rates are calculated based on the ‘foraging arena’ concept, according to which prey move from ‘refuges’ to ‘foraging arenas’ where they are vulnerable to predators [81, 84, 85]. This concept is illustrated in Ecosim using ‘vulnerability multipliers’, which allow for prediction of the magnitude at which an increase in predator’s abundance will influence the predation mortality of a given prey [77]. The vulnerability multipliers express how much the predation mortality of a prey may increase if the predators were to increase to the carrying capacity. A vulnerability of 1 thus imply that the predator is at carrying capacity and hence cannot increase the predation mortality it is causing on its prey–at carrying capacity the predator is limited by available prey. Conversely, a high vulnerability multiplier imply that a predator is far from its carrying capacity and may be able to increase prey predation mortality as it grows towards its carrying capacity. Ecosim simulations permit to fit predicted temporal changes in biomass and mortality to time-series of reference data. Except if stated otherwise, all biomass time-series were entered as relative to the base Ecopath biomass estimates of the different functional groups in 1979 [77]. Detailed descriptions of how each Ecosim time-series was calculated and used can be found in the Section 3 of S1 Appendix. For multi-stanza populations, the differential equations are replaced by a discrete-time (monthly) age structured population model that uses rates calculated from the differential equation model rates for fish of all ages that are in each stanza each month [86].Finally, time-series of fishing mortality were entered for each functional group of the model between 1979 and 2020.
After the time-series were loaded and applied, we generated a statistical measure of goodness of fit (i.e. a weighted sum of sum of squared deviations of log biomasses from log predicted biomasses, often referred to as SS residuals) for each Ecosim run [77]. The SS residuals are sensitive to the vulnerability multipliers for each functional group and their prey, as those trophic flow between predators and their prey [77, 81]. Those vulnerability estimates were generated by Ecosim to improve the ‘rightness’ of fit (i.e. lower SS) based upon the influence of other driving factors, including primary productivity and fishing activity. In addition, we ran a search for primary production anomalies between 1979 and 2020, and applied the resultant as a forcing function for the ‘producers’ functional group (i.e. phytoplankton). We then compared this forcing function to three different environmental indices: NPGO, DPO, and ENSO. Anomaly indices for the NPGO were downloaded directly from the Georgia Institute of Technology website (http://www.o3d.org/npgo/) [87]. Further, both DPO (https://www.ncei.noaa.gov/access/monitoring/pdo/, [88]) and ENSO (https://www.ncei.noaa.gov/access/monitoring/enso/) anomaly indices were extracted from the National Centers for Environmental Information of the National Oceanic and Atmospheric Administration. For each climatic index, the primary productivity anomaly function extracted from Ecosim was rescaled for comparison. The NPGO, PDO, and ENSO environmental indexes were not initially entered in the model, as it was deemed important to understand whether the model itself would accurately predict changes in primary productivity between 1979 and 2020.
Results
The main objective of this model was to examine whether competition for food between marine mammals could be a limiting factor for the SRKW population, and to understand how the predation mortality associated with growing populations of marine mammals has changed over the northeast Pacific coast between 1979 and 2020. Resident killer whales are primarily feeding on mature fish on their way to spawning grounds, while harbor seals primarily target young smolts entering the marine environment. Steller and California sea lions target Pacific salmonids through different life stages but are mostly found foraging in coastal waters. To illustrate changes in predation mortality, we thus chose to present the results of six marine mammal groups (i.e. TKW, SRKW, NRKW, HS, SSL, CSL) and 16 salmon groups (i.e. Smolts and Returning Spawners stanzas for all Chinook and coho salmon). EwE simulation output with changes in biomass of all other functional groups can be found in the Figs B1-B7 of S2 Appendix. The biomass trajectories predicted by the EwE simulations and presented here are those for which we obtained the ‘best fit’ (i.e. the lowest sums of squared differences between the predicted and observed biomass estimates based upon optimal vulnerability parameters). The biomass predictions are presented for marine mammals and Returning spawners stanzas only. Detailed information about the different vulnerability settings used in the EwE simulations can be found in the Tables A5-A7 of S1 Appendix.
Population biomass
The overall sums of squared differences before model fitting was 86.8. The fit was improved to 55.60 when using search-derived vulnerability multipliers (Section 4, Tables A5-A7 of S1 Appendix). Finally, the primary productivity anomaly estimated by Ecosim was applied to all time-series and further improved the overall sums of squared differences to 49.7. After applying the primary productivity anomaly, the fit was improved for all Chinook salmon Returning spawners stanzas.
The model predicted that the population trend of SRKW has been fluctuating slightly since 1979 (Fig 2). This result is in line with the population censuses reported by the Center for Whale Research (2022), but with less amplitude [89]. However, the bioaccumulation for this population had to be lowered (to 0.5%) in order to achieve this fit. This modification is not surprising, as the growth rate of the population that we initially used (i.e. 1.3%) was an annual average between 1974 and 1987 [90]. Besides, the SRKW population has undergone periods of negative population growth, including in the early 1980s [91] (Fig 2). The model also predicted that the NRKW population would have increased by 2.1-fold between 1979 and 2020, which is in line with previous estimates [1, 92] (Fig 2).
Comparison of changes in SRKW (top) and NRKW (bottom) relative predicted biomass (line) versus biomass estimates from population censuses (points) between 1979 and 2020.
The model predicted that the three populations of pinnipeds consistently increased from 1979 on (Fig 3). The EwE simulation predicted that the harbor seal population rose by about 3.2 times between 1979 and the early 2000s, before plateauing in the early 2000’s. The population of Steller and California sea lions increased by about six and three times since 1979, respectively. The growth of the CSL population was predicted to slow down around 2015 (Fig 3).
Comparison of changes in HS (top), SSL (middle), and CSL (bottom) relative predicted biomass (line) versus biomass estimates from population censuses (points) between 1979 and 2020.
The predicted variations in population trends of Returning spawners of Chinook salmon were similar to the trends extracted from CWT recoveries for all three runs originating from the Salish Sea (Fraser River/Georgia Strait/Puget Sound) (Fig 4). A peak in abundance was predicted for all three groups in the early 2000s (Fig 4). Overall, the abundance of FRGSPS spring Chinook salmon group was predicted to only slightly decrease (~ -8%) between 1979 and 2020. The FRGSPS summer Chinook salmon group increased by about 43% between 1979 and 2020, whereas the FRGSPS fall Chinook group was predicted to decline by about 90% throughout the time series (Fig 4). Finally, the model predicted that the overall abundance of the Returning spawners Chinook salmon originating from the west coast of Vancouver Island decreased by about 50% between 1979 and 2020 (Fig 4).
Comparison of changes in relative predicted biomass (line) of spring (top left), summer (top right), and fall (bottom left) Returning Spawners Chinook salmon originating from the Salish Sea (Fraser River/ Georgia Strait/Puget Sound) and the west coast of Vancouver Island (bottom right) versus biomass estimates from CWT recoveries (points) between 1979 and 2020.
The predicted variations in population trends of Returning spawners Chinook salmon were similar to the trends extracted from CWT recoveries for all three runs originating from the Columbia River/Washington/Oregon/California coasts (Fig 5). Similarly to the FRGSPS groups, a peak in abundance was predicted for all three groups between 2000 and 2007. A second peak was predicted for the Returning spawners summer run between 2010 and 2018. The model predicted that this functional group would have increased by 61% during that time. Finally, an overall decline in abundance was predicted between 1979 and 2020 for the spring and fall run, which decreased by 50% and 20%, respectively (Fig 5).
Comparison of changes in relative predicted biomass (line) of Returning spawners Chinook salmon originating from the Columbia River/Oregon/Washington/California coast (CRWORC) in the spring (top), summer (middle), and fall (bottom) versus biomass estimates from CWT recoveries (points) between 1979 and 2020.
The model predicted that the coho salmon population would drastically decrease between 1979 and 2020 (Fig 6), which, however, was a slightly less severe decline than in previous population estimates [93, 94]. Overall, the model predicted a 66% biomass decrease between 1979 and 2020 (Fig 6).
Mortality
The model predicted that total mortality of all Chinook salmon Returning spawners groups originating from the Salish Sea (i.e. FRGSPS) and the west coast of Vancouver Island has been relatively constant between 1979 and 2020 (Fig 7). This is the case despite a consistent decrease in fishing mortality during this period, including for FRGSPS SP (-73.6%), FRGSPS SU (-83.8%), FRGSPS FA (-59.9%), and WCVI (-36.9%) (Fig 7). In contrast, the predation mortality associated with marine mammals has increased for all groups between 1979 to 2020 (Fig 7). The model predicted that the predation mortality on FRGSPS SP Returning spawners was multiplied by 3.9 between 1979 and 2020, and by 3.6, 3.1, and 1.8 for FRGSPS SU, FRGSPS FA, and WCVI, respectively (Fig 7). This increase in predation mortality is driven by different marine mammals depending on the Chinook salmon functional group (Fig 7). In 2020, California sea lion accounted for almost 40% of the predation mortality predicted for FRGSPS SU while Steller sea lion accounted for 54.2% of the predation mortality predicted for FRGSPS SP Chinook salmon Returning spawners. Finally, NRKW and California sea lion together accounted for 66.7% of the predation mortality predicted for WCVI Returning spawners in 2020.
Relative marine mammal predation mortality (left column) and total mortality (including fishing and predation mortality) (right column) on Returning spawners Chinook salmon originating from the Salish Sea (i.e. FRGSPS) and the west coast of Vancouver Island from 1979 to 2020. The graphs are presented for FRGSPS SP (first row), FRGSPS SU (second row), FRGSPS FA (third row), and the WCVI (fourth row).
The model predicted similar trends of predation and total mortality for Chinook salmon Returning spawners groups originating from the southern part of the model (i.e. CRWORC). The total mortality slightly decreased for CRWORC SU (-37%) and CRWORC FA (-26%) between 1979 and 2020, while their respective fishing mortality during that time was reduced by 72% and 45% (Fig 8). During this period, the predation mortality on CRWORC SU and CRWORC FA increased by factors of 8.5 and 2.6, respectively. The total predicted mortality for CRWORC SP Chinook salmon Returning spawners slightly increased (+18.6%) from 1979 to 2020 (Fig 8). During this time, the fishing mortality decreased by 49% while the predation mortality doubled (Fig 8). For this group, the combined predation of Steller and California sea lions accounted for 63.3% of the total predation mortality in 2020. During this year, Steller sea lion accounted for about 64% of the predation mortality predicted for CRWORC SU Chinook salmon Returning spawners. Finally, NRKW accounted for 25% of the predation mortality predicted for CRWORC FA (Fig 8).
Relative marine mammal predation mortality (left column) and total mortality (including fishing and predation mortality) (right column) on Returning spawners Chinook salmon originating from the Columbia River and the Washington, Oregon, and California coast (i.e. CRWORC) from 1979 to 2020. The graphs are presented for CRWORC SP (top row), CRWORC SU (middle row), and CRWORC FA (bottom row).
The predation mortality of FRGSP SP, FRGSPS SU, CRWORC SP, CRWORC SU, and CRWORC FA Smolts was not predicted to vary significantly throughout the time series, as very little predation was applied on those groups.
The predation mortality associated with harbor seals on FRGSPS FA Smolts was predicted to increase from 0.6 to 2.8 between 1979 and 2020 (Fig 9). This increase follows the same variation patterns as predicted by the model for total mortality. The total mortality was predicted to have doubled between 1979 and 2020, which is in line with observations and data from CWT recoveries [80] (Fig 9).
The first graph (left) shows the relative predation mortality of harbor seals and sea birds on FRGSPS FA Chinook salmon Smolts predicted by the model between 1979 and 2020. The second graph (right) is a comparison of the predicted total mortality Z (thick purple line) versus total mortality estimates from CWT recoveries (points) from 1979 to 2020 (top).
Despite a drastic reduction in fishing mortality (- 90%) on Returning spawners coho salmon between 1979 and 2020, the model predicted that the total mortality of this group would have stayed relatively stable during that time. This is because the predation mortality on this group was predicted to almost double between 1979 and 2020 (Fig 10). California sea lions accounted for more than 50% of this predation mortality throughout the 42 years of the model.
The first graph (left) shows the relative predation mortality of marine mammals on coho salmon Returning spawners predicted by the model between 1979 and 2020. The second graph (right) shows the predicted total mortality Z (thick purple line, includes fishing and predation mortality) for this group from 1979 to 2020.
The predation mortality due to harbor seals on coho salmon Smolts was predicted to increase from 0.32 to 2.2 between 1979 and 2020 (Fig 11). This increase follows the same variation pattern as predicted by the model for total mortality (Fig 11). The total mortality was predicted to have doubled between 1979 and 2020, which is in line with observations and data extracted from CWT recoveries (Fig 11).
The first graph (left) shows the relative predation mortality of harbor seals and sea birds on coho salmon Smolts predicted by the model between 1979 and 2020. The second graph (right) is a comparison of the predicted total mortality Z (thick purple line) versus total mortality estimates from CWT recoveries (points) from 1979 to 2020.
The primary productivity anomalies were estimated as 42 yearly values. These primary productivity anomalies (PPA) estimated by Ecosim showed an alternation of positive and negative fluctuations throughout the time series. These ‘oscillations’ lasted from one to six years (Fig 12). This primary productivity function extracted from Ecosim showed oscillation patterns that were very similar to the NPGO anomalies indices, with 25 years of the time-series being either positive or negative for both functions (Fig 12). The primary productivity anomaly function and the climatic patterns ENSO and PDO simultaneously showed negative or positive values for only 21 years and 19 years, respectively (Fig 12). Clear periods of overall negative (i.e. 1993–1997, 2015–2020) and positive (i.e. 1998–2002) anomalies were detected by Ecosim for primary productivity, all of which matched the timing and range of oscillations observed for NPGO (Fig 12).
Discussion
Our predictions of changes in biomass of the different marine mammal populations are in line with previous literature and recent population surveys [1, 21, 22, 24, 89, 95, 96]. The Chinook and coho salmon biomass estimates predicted by the model are also aligned with the mostly declining population trends suggested by CWT recovery data [80]. The main objective of this research was to quantify how predation mortality associated with growing marine mammal populations may have limited those salmonid populations so as to potentially reduce food availability for SRKW since 1979. To answer this question, it is important to understand which other sources of mortality, such as fishing, could have limited Chinook and coho salmon populations.
Since 1979, the fishing mortality on Returning spawners Chinook salmon was drastically reduced for all groups, with decreases ranging from -37% (WCVI) to -84% (FRGSPS FA). These declines in fishing mortality resulted from numerous conservation management plans and fishing restrictions implemented over the years, such as the Pacific Salmon Treaty (1985, [80]). Severe restrictions in gear, spatial and temporal closures, and fish quotas were gradually applied to both recreational and commercial fisheries over the years. For instance, troll fisheries in British Columbia can now only operate a few weeks per year in specific areas, and some years there are hardly any openings, even for bread-and-butter fisheries such as notably for sockeye salmon [40]. Similarly, recreational fishing is banned from some regions of the Fraser River in the spring and fall, and strict gear, fish size, and number restrictions apply during the rest of the year [40, 97]. Similar extensive regulations have been applied to coho salmon over the last 40 years, with fishing restrictions occurring in Washington, Oregon, and California from the early 1990s and in British Columbia from between 1996 and 1998 [98]. Despite those severe fishing regulations on both Chinook and coho salmon, the model suggests that the total mortality of those groups has not significantly decreased between 1979 and 2020.
During that time, estimated predation mortality increased for all Chinook and coho salmon Returning spawners and Smolts stanzas included in the model. Large pinnipeds (i.e. California and Steller sea lions) and NRKW appear to be the most important predators of Returning spawners Chinook salmon in recent years. The model suggests that the predation mortality of Steller and California sea lions on those Chinook salmon groups in 2020 ranged from 0.05 to 0.3, which would represent an annual consumption of about 0.7 and 0.8 million fish, respectively. Those estimates are in line with the results presented by Chasco et al. (2017) for the year 2015 [9]. Further, the model predicted that the overall predation mortality associated with NRKW on Returning spawners Chinook salmon doubled between 1979 and 2020. The intensifying predation by those marine mammals on mature fish can affect the average body size, maturation timing, and life-history characteristics of the Chinook salmon populations through the selective removal of larger fish [99]. In addition, leaving smaller Chinook salmon implies a reduction in caloric value, which could further affect other marine mammals feeding on similar preys, including SRKW [100, 101]. Although one could argue that SRKW might have access to larger fish the same way NRKW do, recent evidence suggests that SRKW might exhibit different foraging strategies than NRWK [102]. When compared to NRKW, SRKW females were found to be less efficient at capturing prey [102]. In this context, the size-selective predation by growing populations of more efficient hunters could directly affect the availability of high-quality prey for SRKW. In addition, no published study yet exists on the potential behavioral avoidance of SRKW when in presence of other marine mammals. Saulitis et al. (2000) observed aggressive non-predatory interactions between Steller sea lions and resident killer whales [103]. During their study, Steller sea lions were reported to charge and nip at both transient and resident killer whales.
The model suggested that the growing population of harbor seals could have been a primary cause for the increased mortality of FRGSPS FA Chinook salmon and coho salmon Smolts between 1979 and 2020. Estimated predation mortality by harbor seals on coho salmon Smolts increased by a factor of 6.4 between 1979 and 2020, against 4.6 for FRGPS FA Smolts. This result is in line with numerous recent studies [9–11, 104] suggesting considerable overall consumption of Pacific salmon smolts by harbor seals in recent years. Most fall-run Chinook salmon exhibit an ocean-type juvenile life-history and spend their first year rearing in the marine environment, potentially making them highly vulnerable to predation during this residency time [105]. In addition, fall-run Chinook smolts are known to grow at a faster rate than their spring- and summer counterparts [106]. In this context, it is possible that harbor seals exhibit a positive selection rather than an opportunistic behavior by targeting more profitable prey with higher calorie value. In addition, the hatchery contribution to the total production of FRGSPS FA smolts in the Salish Sea between 1979 and 2020 ranged from about 33% to 61%. Most hatchery rearing environments do not allow conditioning of young salmon to respond to predator stimuli, therefore potentially increasing their vulnerability to predation after release [107, 108].
In 2020, the model predicted that harbor seals could have consumed about 27 million Chinook salmon smolts in the Salish Sea alone, while Chasco et al. (2017) estimated that consumption at about 23.2 million in 2015 [9]. Further, our study suggested that harbor seal could have consumed about 47 million coho smolts in 2020. Thomas et al. (2017) indicated that harbor seals could consume about 5.7 million coho smolts in one month only in the Strait of Georgia, and that coho smolts were part of the harbor seal diet for at least 8 months of the year [11]. In both cases, the consumption of Chinook and coho smolts by harbor seals predicted by the model is uncertain, as the seasonal and spatial overlap between harbor seals and their prey is not considered in these computations. In order to more precisely quantify the consumption of Pacific salmons by marine mammals, it is essential to acknowledge the highly complex life cycle and seasonal movement patterns of Pacific salmonid species. Efforts should thus be made in future modeling to integrate information on the seasonal and spatial overlap of the main predators and their prey.
The high level of juvenile salmon mortality [109, 110] and declining smolt-to adult survival [111, 112] for Chinook and coho salmon have been well documented. Along with other reports [11, 113], our study suggests that the predation mortality by harbor seals could strongly contribute to the poor survival of those fish in the early marine phase of life. In asking the question of competition between marine mammals, one could argue that harbor seals and SRKW do not target Chinook salmon of similar age. However, a dramatic reduction of survival in the early life could directly result in a lower number of adult individuals available to fisheries and larger predators [111, 114, 115]. In this context, harbor seal could indirectly compete with SRKW by lowering the prey biomass available to the whales at a later stage.
Despite our findings, knowing with certainty whether harbor seal predation on juvenile salmon is additive or non-additive remains challenging and is an important limitation of the model. Non-additive mortality occurs if a prey becomes more vulnerable to predation because of external stress, physiological, or behavioral factors, yet would die even if not eaten by a predator [116]. Additive mortality would occur if each mortality sources were independent from each other [116, 117]. In the case of Chinook salmon, Mesa et al. (1998) suggested that juvenile fish infected with disease agents were significantly more vulnerable to predation than healthy individuals [118]. Other studies conducted on Chinook [119], coho [120], and steelhead [121] salmon suggested that pathogens could be related to lower growth rates, fitness indices, potential chances of successful migration, and overall survival. Besides, Walters and Christensen (2019) suggested that Chinook smolts could still be exhibiting high stress-dependent mortality rates even if seal populations were reduced [116]. Taking into consideration the geographical overlap between the different marine mammal populations may also be of critical importance when examining how marine mammal competition might be limiting for the SRKW population. This is especially needed given that the geographical range of marine mammal populations may well be changing over the coming decades. For instance, Edgell and Demarchi (2012) suggested that in recent years sea lions have been occurring more frequently and in higher numbers along the coast of Washington State and British Columbia [122]. Spatial integration should be recognized as a key aspect of future model examining the evolution of inter-specific competition between highly mobile predator and prey populations.
The model did not show a clear association between variations in primary productivity and changes in coho and chum salmon population trends. On the other hand, the model suggested that primary productivity anomalies linked to the North Pacific Gyre Oscillation could influence the productivity of Pacific Chinook salmon populations from British Columbia to California. This finding is of critical importance, and highlights the potential correlation between large-scale climatic indexes and the productivity of separate Chinook salmon stocks. Di Lorenzo et al. (2008) showed that the NPGO was correlated to significant changes in sea surface temperature, salinity, as well as in nutrient and chlorophyll-a concentrations [87]. Increased sea surface temperatures have been shown to influence salmon juvenile growth rates [123], prevalence to diseases [124], mortality [125], and productivity [126, 127] of different Pacific salmon species. In addition, studies in the north Pacific showed that sea-surface salinity could negatively affect the growth [128] of chum salmon. In the model, peaks in abundance of all Chinook salmon groups were observed in years where anomalies in primary productivity and NPGO were positive (i.e. early 2000s, and after 2010). These results are in line with several recent studies [31, 129, 130], which have concluded that a significant correlation exist between the NPGO and the productivity of several Pacific salmon populations. For instance, Debertin et al. (2017) observed that a positive NPGO was associated with an increased primary productivity and growth in chum salmon [129]. Kilduff et al. (2015) further suggested that the NPGO could be the most important climatic driver of coho and Chinook salmon survival along the northeast Pacific coast, which is also supported by the results of the model [31]. Our study emphasizes the need to consider large-scale changes in climatic conditions and primary productivity patterns when aiming to understand the causes for changes in Pacific salmon population trends. Finally, it is important to note that while the main research question focused on the impact of marine mammal predation on Pacific salmon populations, the use of an ecosystem model allowed for investigating both top-down and bottom-up driving forces in the ecosystem. Important future research should aim at understanding how different management measures could impact the state and evolution of this ecosystem in the coming decades.
Acknowledgments
We thank the Joint Chinook Technical Committee of the Pacific Commission for facilitating the access to Chinook salmon escapement and catch data.
References
- 1.
Fisheries and Oceans Canada. Recovery Strategy for the Northern and Southern Resident Killer Whales (Orcinus orca) in Canada. Species at Risk Act Recovery Strategy Series, Fisheries and Oceans Canada, Ottawa, 84 pp.
- 2. Ford J, Ellis G. Selective foraging by fish-eating killer whales Orcinus orca in British Columbia. Mar Ecol Prog Ser 2006; 316: 185–199.
- 3. Hanson MB, Baird RW, Ford JKB, Hempelmann-Halos J, Van Doornik DM, Candy JR, et al. Species and stock identification of prey consumed by endangered Southern resident killer whales in their summer range. Endanger Species Res 2010; 11: 69–82.
- 4. Hanson MB, Emmons CK, Ford MJ, Everett M, Parsons K, Park LK, et al. Endangered predators and endangered prey: Seasonal diet of Southern Resident killer whales. PLOS ONE 2021; 16: e0247031. pmid:33657188
- 5. US Environmental Protection Agency. Southern Resident Killer Whales. US EPA, https://www.epa.gov/salish-sea/southern-resident-killer-whales (2023, accessed 22 January 2020).
- 6. Ward EJ, Holmes EE, Balcomb KC. Quantifying the effects of prey abundance on killer whale reproduction. J Appl Ecol 2009; 46: 632–640.
- 7. Ford JKB, Ellis GM, Olesiuk PF, Balcomb KC. Linking killer whale survival and prey abundance: food limitation in the oceans’ apex predator? Biol Lett 2010; 6: 139–142. pmid:19755531
- 8. Wasser SK, Lundin JI, Ayres K, Seely E, Giles D, Balcomb KC, et al. Population growth is limited by nutritional impacts on pregnancy success in endangered Southern Resident killer whales (Orcinus orca). PLOS ONE 2017; 12: e0179824.
- 9. Chasco BE, Kaplan IC, Thomas AC, et al. Competing tradeoffs between increasing marine mammal predation and fisheries harvest of Chinook salmon. Sci Rep 2017; 7: 15439. pmid:29158502
- 10. Chasco B, Kaplan IC, Thomas A, Acevedo-Gutiérrez A, Noren DP, Ford M, et al. Estimates of Chinook salmon consumption in Washington State inland waters by four marine mammal predators from 1970 to 2015. Can J Fish Aquat Sci 2017; 74: 1173–1194.
- 11. Thomas AC, Nelson BW, Lance MM, Deagle BE, Trites AW. Harbour seals target juvenile salmon of conservation concern. Can J Fish Aquat Sci 2017; 74: 907–921.
- 12. Ward E, Ford M, Kope R, Ford J, Velez-Espino A, Parken C, et al. NOAA Technical Memorandum NMFS-NWFSC-123. Estimating the Impacts of Chinook Salmon Abundance and Prey Removal by Ocean Fishing on Southern Resident Killer Whale Population Dynamics. Natl Ocean Atmospheric Adm NOAA Tech Memo 2013; NMFS-NWFSC-1: 85.
- 13. Levin PS. Regional differences in responses of chinook salmon populations to large-scale climatic patterns: Climate responses of Pacific salmon. J Biogeogr 2003; 30: 711–717.
- 14. Dorner B, Catalano MJ, Peterman RM. Spatial and temporal patterns of covariation in productivity of Chinook salmon populations of the northeastern Pacific Ocean. Can J Fish Aquat Sci 2018; 75: 1082–1095.
- 15. Grant SCH, Department of Fisheries and Oceans. State of the Canadian Pacific salmon: response to changing climate and habitats, http://publications.gc.ca/collections/collection_2019/mpo-dfo/Fs97-6-3332-eng.pdf (2019, accessed 29 January 2020).
- 16. Crozier LG, Burke BJ, Chasco BE, Widener DL, Zabel RW. Climate change threatens Chinook salmon throughout their life cycle. Commun Biol 2021; 4: 222. pmid:33603119
- 17. Levin PS, Tolimieri N. Differences in the impacts of dams on the dynamics of salmon populations. Anim Conserv 2001; 4: 291–299.
- 18. Finn RJR, Chalifour L, Gergel SE, Hinch SG, Scott DC, Martin TG, et al. Quantifying lost and inaccessible habitat for Pacific salmon in Canada’s Lower Fraser River. Ecosphere; 12. Epub ahead of print July 2021.
- 19. Harvey JT, Brown RF, Mate BR. Abundance and Distribution of Harbor Seals (Phoca vitulina) in Oregon, 1975–1983. Northwest Nat 1990; 71: 65.
- 20.
Olesiuk PF. An assessment of the status of harbour seals (Phoca vitulina) in British Columbia. Fish Oceans Can 1999; Canadian Stock Assessment Secretariat Research Document: 71.
- 21. Jeffries S, Huber H, Calambokidis J, Laake J. Trends and Status of Harbor Seals in Washington State: 1978–1999. J Wildl Manag 2003; 67: 207.
- 22. Laake JL, Lowry MS, DeLong RL, Melin SR, Carretta JV. Population growth and status of California sea lions: Status of California Sea Lions. J Wildl Manag 2018; 82: 583–595.
- 23. NOAA’s National Marine Fisheries Service. The Marine Mammal Protection Act of 1972 As Amended, https://media.fisheries.noaa.gov/2021-04/mmpa_2018_revised_march_2019-508_%282%29.pdf?null (2018).
- 24. Brown RF, Wright BE, Riemer SD, Laake J. Trends in abundance and current status of harbor seals in Oregon 1977–2003. Mar Mammal Sci 2005; 21: 657–670.
- 25. Fisheries and Oceans Canada. Trends in Abundance and Distribution of Steller Sea Lions (Eumetopias Jubatus) in Canada. Can Sci Advis Secr Sci Advis Rep 2021; 035: 8.
- 26. Weise M, Harvey J. Temporal variability in ocean climate and California sea lion diet and biomass consumption: implications for fisheries management. Mar Ecol Prog Ser 2008; 373: 157–172.
- 27. Lance M, Chang W, Jeffries S, Pearson SF, Acevedo-Gutiérrez A. Harbor seal diet in northern Puget Sound: implications for the recovery of depressed fish stocks. Mar Ecol Prog Ser 2012; 464: 257–271.
- 28. Bromaghin JF, Lance MM, Elliott EW, Jeffries SJ, Acevedo-Gutiérrez A, Kennish JM. New insights into the diets of harbor seals (Phoca vitulina) in the Salish Sea revealed by analysis of fatty acid signatures. Fish Bull; 111. Epub ahead of print 2013.
- 29. Sorel M, Zabel RW, Johnson D, Wargo Rub AM, Converse SJ. Estimating population‐specific predation effects on Chinook salmon via data integration. J Appl Ecol 2020; 58: 372–381.
- 30. Crozier LG, Zabel RW, Hamlet AF. Predicting differential effects of climate change at the population level with life-cycle models of spring Chinook salmon: population-specific effects of climate change. Glob Change Biol 2008; 14: 236–249.
- 31. Kilduff DP, Di Lorenzo E, Botsford LW, Teo SLH. Changing central Pacific El Niños reduce stability of North American salmon survival rates. Proc Natl Acad Sci 2015; 112: 10962–10966.
- 32. Peterson WT, Morgan CA, Casillas E, Peterson JO, Fisher JL, Ferguson JW. Ocean Ecosystem Indicators of Salmon Marine Survival in the Northern California Current. Newport, Oregon: Fish Ecology Division, Northwest Fisheries Science Center, 2010.
- 33. Menge B, Chan F, Nielsen J, Di Lorenzo E, Lubchenco J. Climatic variation alters supply-side ecology:impact of climate patterns on phytoplankton and mussel recruitment. Ecol Monogr 2009; 79: 379–395.
- 34. Cloern J, Hieb K, Jacobson T, Sansó B, Di Lorenzo E, Stacey MT, et al. Biological communities in San Francisco Bay track large‐scale climate forcing over the North Pacific. Geophys Res Lett; 37. Epub ahead of print 2010.
- 35. Argue AW, Campbell D, Gee PA, Shepard MP. Records of Annual Salmon Harvest by British Columbia Indians Prior to 1951. Can Data Rep Fish Aquat Sci; 782.
- 36. Argue AW, Shepard MP. Historical Commercial Catch Statistics for Pacific Salmon(Oncorhynchus spp.) in British Columbia, 1828 to 1950. Fish Oceans Can; 2601.
- 37.
Nelson S, Turris B. The evolution of commercial salmon fisheries in British Columbia. Report to the Pacific Fisheries Resource Conservation Council, Vancouver, B.C.: Pacific Fisheries Resource Conservation Council, 2004.
- 38.
Ricker WE. Causes of the decrease in age and size of chinook salmon (Oncorhynchus tshawytscha). Fish Ocean Can Pac Biol Stn Nanaimo 1980; Canadian Technical Report of Fisheries and Aquatic Sciences: 32.
- 39. Hard JJ, Gross MR, Heino M, Hilborn R, Kope RG, Law R, et al. Evolutionary consequences of fishing and their implications for salmon. Evol Appl 2008; 1: 388–408. pmid:25567639
- 40.
Fisheries and Oceans Canada. Recreational fishing limits, openings and closures in British Columbia | Pacific Region | Fisheries and Oceans Canada, https://www.pac.dfo-mpo.gc.ca/fm-gp/rec/bc-zones-cb-eng.html (2021, accessed 12 June 2023).
- 41. O’Farrell MR, Satterthwaite WH. Inferred historical fishing mortality rates for an endangered population of Chinook salmon (Oncorhynchus tshawytscha). Fish Bull 2015; 113: 341–351.
- 42.
Haddon M. Modelling and quantitative methods in fisheries. 2nd ed. Boca Raton: CRC Press, 2011.
- 43.
Hilborn R, Mangel M. The Ecological Perspective. Princeton University Press, 2013.
- 44. Guenette S, Christensen V. Food web models and data for studying fisheries and environmental impacts on Eastern Pacific ecosystems.
- 45. Stock A, Murray CC, Gregr EJ, Steenbeek J, Woodburn E, Micheli F, et al. Exploring multiple stressor effects with Ecopath, Ecosim, and Ecospace: Research designs, modeling techniques, and future directions. Sci Total Environ 2023; 869: 161719. pmid:36693571
- 46. Colléter M, Gascuel D, Albouy C, Francour P, Tito de Morais L, Valls A, et al. Fishing inside or outside? A case studies analysis of potential spillover effect from marine protected areas, using food web models. J Mar Syst 2014; 139: 383–395.
- 47. Coll M, Akoglu E, Arreguín-Sánchez F, Fulton EA, Gascuel D,. Modelling dynamic ecosystems: venturing beyond boundaries with the Ecopath approach. Rev Fish Biol Fish 2015; 25: 413–424.
- 48. Couce Montero L, Christensen V, Castro Hernández JJ. Simulating trophic impacts of fishing scenarios on two oceanic islands using Ecopath with Ecosim. Mar Environ Res 2021; 169: 105341.
- 49. Rehren J, Coll M, Jiddawi N, Kluger LC, Omar O, Christensen V, et al. Evaluating ecosystem impacts of gear regulations in a data-limited fishery—comparing approaches to estimate predator–prey interactions in Ecopath with Ecosim. ICES J Mar Sci 2022; 79: 1624–1636.
- 50. Walters C, Pauly D, Christensen V. Ecospace: Prediction of Mesoscale Spatial Patterns in Trophic Relationships of Exploited Ecosystems, with Emphasis on the Impacts of Marine Protected Areas. Ecosystems 1999; 2: 539–554.
- 51. Pitcher T, Buchary E, Hutton T. Forecasting the benefits of no-take human-made reefs using spatial ecosystem simulation. ICES J Mar Sci 2002; 59: S17–S26.
- 52. Plummer ML, Harvey CJ, Anderson LE, Guerry AD, Ruckelshaus MH. The Role of Eelgrass in Marine Community Interactions and Ecosystem Services: Results from Ecosystem-Scale Food Web Models. Ecosystems 2013; 16: 237–251.
- 53. Lewis KA, de Mutsert K, Steenbeek J, et al. Employing ecosystem models and geographic information systems (GIS) to investigate the response of changing marsh edge on historical biomass of estuarine nekton in Barataria Bay, Louisiana, USA. Ecol Model 2016; 331: 129–141.
- 54. Blanchard JL, Pinnegar JK, Mackinson S. Exploring marine mammal-fishery interactions using Ecopath with Ecosim: modelling the Barents Sea Ecosystem. Sci Ser Tech Rep 2002; 117: 53.
- 55. Morissette L, Hammill MO, Savenkoff C. The trophic role of marine mammals in the Northern Gulf of St. Lawrence. Mar Mammal Sci 2006; 22: 74–103.
- 56. Surma S, Pitcher TJ. Predicting the effects of whale population recovery on Northeast Pacific food webs and fisheries: an ecosystem modelling approach. Fish Oceanogr 2015; 24: 291–305.
- 57. Lamb PD, Hunter E, Pinnegar JK, et al. Inclusion of jellyfish in 30+ years of Ecopath with Ecosim models. ICES J Mar Sci 2019; 76: 1941–1950.
- 58. Woodruff P, van Poorten B, Walters C, et al. Potential effects of invasive Dreissenid mussels on a pelagic freshwater ecosystem: using an ecosystem model to simulate mussel invasion in a sockeye lake. Aquat Invasions 2021; 16: 129–146.
- 59. Port of Vancouver. Roberts Bank Terminal 2. Appendix 10.B. Roberts Bank Ecopath with Ecosim and Ecospace Model Parameters Estimates.
- 60. Fretzer S. Using the Ecopath approach for environmental impact assessment—A case study analysis. Ecol Model 2016; 331: 160–172.
- 61. Lekshmi M, Sreekanth GB, Singh NP, Vennila A, Ratheesh Kumar R, Pandey PK. Ecopath modelling approach for the impact assessment of a small-scale coastal aquaculture system in Goa, India. Indian J Fish; 67. Epub ahead of print 30 September 2020.
- 62. Steenbeek J, Romagnoni G, Bentley JW, Heymans JJ, Serpetti N, Gonçalves M. Combining ecosystem modeling with serious gaming in support of transboundary maritime spatial planning. Ecol Soc 2020; 25: art21.
- 63. Serpetti N, Benjamins S, Brain S, Collu M, Harvey BJ, Heymans JJ. Modeling Small Scale Impacts of Multi-Purpose Platforms: An Ecosystem Approach. Front Mar Sci 2021; 8: 694013.
- 64. Ainsworth CH, Samhouri JF, Busch DS, Cheung WW, Dunne J, Okey TA. Potential impacts of climate change on Northeast Pacific marine foodwebs and fisheries. ICES J Mar Sci 2011; 68: 1217–1229.
- 65. Alva-Basurto JC, Arias-González JE. Modelling the effects of climate change on a Caribbean coral reef food web. Ecol Model 2014; 289: 1–14.
- 66. Cornwall C, Eddy T. Effects of near‐future ocean acidification fishing and marine protection on a temperate coastal ecosystem. Conserv Biol 2014; 29: 207–2015. pmid:25354555
- 67.
EcoBase. EcoBase: a database of Ecopath with Ecosim models, http://ecobase.ecopath.org/#discoverytools (2023).
- 68.
Martell S. Variation in pink shrimp populations off the West coast of Vancouver Island: oceanographic and trophic interactions. Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy, University of British Columbia, 2002.
- 69.
Preikshot D. The influence of geographic scale, climate and trophic dynamics upon North Pacific oceanic ecosystem models. Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy, University of British Columbia, https://doi.org/10.14288/1.0074902 (2007).
- 70.
Field JC. Application of Ecosystem-Based Fishery Management Approaches in the Northern California Current. A Dissertation Submitted in Partial Satisfaction of the Requirements for the Degree Doctor of Philisophy, University of Washington, 2004.
- 71. Walters C, Martell SJD, Mahmoudi B. An Ecosim model for exploring ecosystem management options for the Gulf of Mexico: implications of including multistanza life history models for policy predictions. Bull Mar Sci 2008; 6: 23.
- 72. Pauly D. Back to the Future: Reconstructing the Strait of Georgia Ecosystem. Fish Cent Res Rep; 6.
- 73. Ainsworth CH, Heymans JJ, Pitcher T, Vasconcellos M. Ecosystem Models of Northern British Columbia for the Time Periods 2000, 1950, 1900, and 1750. Fish Cent Res Rep; 10.
- 74. Christensen V, Walters CJ. Ecopath with Ecosim: methods, capabilities and limitations. Ecol Model 2004; 172: 109–139.
- 75.
Google Technology Company. Google Earth. Google Earth, https://earth.google.com/web/ (2022, accessed 12 June 2023).
- 76. Allen KR. Relation Between Production and Biomass. J Fish Res Board Can 1971; 28: 1573–1581.
- 77. Christensen V, Walters CJ, Pauly D, Forrest RE. Ecopath with Ecosim: User Guide version 6. Lenfest Ocean Futures Project 2008. 2008.
- 78. Heymans JJ, Coll M, Link JS, Mackinson S, Steenbeek J, Walters C, et al. Best practice in Ecopath with Ecosim food-web models for ecosystem-based management. Ecol Model 2016; 331: 173–184.
- 79. Ecopath Pauly D., Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. ICES J Mar Sci 2000; 57: 697–706.
- 80. Pacific Salmon Commission, Joint Chinook Technical Committee Report. Annual Report of Catch and Escapement for 2017. 2018; 18: 235.
- 81. Walters CJ, Martell SJD. Fisheries ecology and management. Princeton university Press, 2004.
- 82. Thompson SA, Sydeman WJ, Thayer JA, Weinstein A, Krieger KL, Hay D. Trends in the Pacific Herring (Clupea pallasi) metapopulation in the California current ecosystem. 2017; 58: 18.
- 83. Walters C, Pauly D, Christensen V, Kitchell JF. Representing Density Dependent Consequences of Life History Strategies in Aquatic Ecosystems: EcoSim II. Ecosystems 2000; 3: 70–83.
- 84. Walters CJ, Christensen V, Pauly D. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Rev Fish Biol Fish 1997; 7: 139–172.
- 85. Ahrens RNM, Walters CJ, Christensen V. Foraging arena theory. Fish Fish 2012; 13: 41–59.
- 86. Walters CJ, Christensen V, Walters W, Rose K. Representation of multistanza life histories in ecospace models for spatial organization of ecosystem trophic interaction patterns. Bull Mar Sci 2010; 86: 439–459.
- 87. Di Lorenzo E, Schneider N, Cobb KM, Franks PJS, Chhak K, Miller AJ, et al. North Pacific Gyre Oscillation links ocean climate and ecosystem change. Geophys Res Lett 2008; 35: L08607.
- 88. Mantua N. The Pacific Decadal Oscillation and Pacific Salmon Production. 1998; 5.
- 89. Center for Whale Research. The Southern Resident Killer Whale Population. Orcas // Killer Whales | United States | Center For Whale Research, https://www.whaleresearch.com/orca-population (2022, accessed 30 January 2020).
- 90. Olesiuk PF, Bigg MA, Ellis GM. Life history and population dynamics of Resident Killer Whales (Orcinus orca) in the coastal waters of British Columbia and Washington State. Rep Int Whal Comm.
- 91. Krahn MM, Wade PR, Kalinowski ST, Dahlheim ME, Taylor BL, Hanson MB, et al. Status Review of Southern Resident Killer Whales (Orcinus orca) under the Endangered Species Act. NOAA Tech Memo NMFS-NWFSC 2002; 54: 159.
- 92.
Towers JR. Photo-identification catalogue and status of the northern resident killer whale population in 2014. Pacific Biological Station, Nanaimo: Fisheries and Oceans Canada, http://www.deslibris.ca/ID/10050075 (2015, accessed 22 March 2019).
- 93. Fisheries and Oceans Canada. Coho Salmon in Georgia Basin. Stock Status Rep 2002; D6-07: 6.
- 94. Pacific Fishery Management Council. Salmon Fishery Management Plan Impacts to Southern Resident Killer Whales Risk Assessment. 2020; SRKW Workgroup Report 1: 165.
- 95. Olesiuk PF. Une évaluation des tendances démographiques et de l’abondance du phoque commun (Phoca vitulina) en Colombie-Britannique. 2010; 163.
- 96.
Olesiuk PF. Recent trends in Abundance of Steller Sea Lions (Eumetopias jubatus) in British Columbia. Fish Ocean Can Pac Biol Stn Nanaimo 2018; Canadian Science Advisory Secretariat Proceedings Series DFO: 72.
- 97.
Government of Canada. British Columbia Sport Fishing Regulations, 1996, http://laws-lois.justice.gc.ca (2023).
- 98. Shaul L, Weitkamp L, Simpson K, Sawada J. Trends in Abundance and Size of Coho Salmon in the Pacific Rim. North Pac Anadromous Fish Comm 2007; 4: 93–104.
- 99. Ohlberger J, Schindler D, Ward E, Walsworth TE, Essington TE. Resurgence of an apex marine predator and the decline in prey body size. Proc Natl Acad Sci 2019; 116: 26682–26689. pmid:31843884
- 100. Couture F, Oldford G, Christensen V, Barrett-Lennard L, Walters CJ. Requirements and availability of prey for northeastern pacific southern resident killer whales. PLOS ONE 2022; 17: e0270523. pmid:35759490
- 101. Lerner JE, Hunt BPV. Seasonal variation in the lipid content of Fraser River Chinook Salmon (Oncorhynchus tshawytscha) and its implications for Southern Resident Killer Whale (Orcinus orca) prey quality. Sci Rep 2023; 13: 2675.
- 102. Tennessen JB, Holt MM, Wright BM, Hanson MB, Emmons CK, Giles DH, et al. Divergent foraging strategies between populations of sympatric matrilineal killer whales. Behav Ecol 2023; 1–14. pmid:37192928
- 103. Saulitis E, Matkin CO, Barrett-Lennard LG, Heise K, Ellis GM. Foraging strategies of sympatric killer whale (Orcinus orca) populations in Prince William Sounds, Alaska.pdf. Mar Mammal Sci 2000; 16: 94–109.
- 104. Thomas AC, Deagle B, Nordstrom C, Majewski S, Nelson BW, Acevedo-Gutiérrez A, et al. Data on the diets of Salish Sea harbour seals from DNA metabarcoding. Sci Data 2022; 9: 68. pmid:35236843
- 105.
Healey 1991. Life history of Chinook salmon (Oncorhynchus tshawytscha). In: Pacific Salmon Life Histories. University of British Columbia Press, Vancouver B.C, 1991.
- 106. Taylor EB. Phenotypic Correlates of Life-History Variation in Juvenile Chinook Salmon, Oncorhynchus tshawytscha. J Anim Ecol 1990; 59: 455.
- 107. Berejikian BA, Tezak EP, Larae AL. Innate and enhanced predator recognition in hatchery-reared chinook salmon. Environ Biol Fishes 2003; 67: 241–251.
- 108. Olla BL, Davis MW. The role of learning and stress in predator avoidance of hatchery-reared coho salmon (Oncorhynchus kisutch) juveniles. Aquaculture 1989; 76: 209–214.
- 109. Beamish RJ, Riddell BE, Neville C-EM, Thomson BL, Zhang Z. Declines in chinook salmon catches in the Strait of Georgia in relation to shifts in the marine environment. Fish Oceanogr 1995; 4: 243–256.
- 110. Neville CM, Beamish RJ, Chittenden CM. Poor Survival of Acoustically-Tagged Juvenile Chinook Salmon in the Strait of Georgia, British Columbia, Canada. Trans Am Fish Soc 2015; 144: 25–33.
- 111. Welch DW, Porter AD, Rechisky EL. The coast-wide collapse in marine survival of west coast Chinook and steelhead: slow-moving catastrophe or deeper failure? BioRxiv; 476408. Epub ahead of print 2018.
- 112. Welch D, Porter A, Rechisky E. A synthesis of the coast‐wide decline in survival of West Coast Chinook Salmon (Oncorhynchus tshawytscha, Salmonidae). Fish Fish 2021; 22: 194–211.
- 113. Riddell B, Pearsall I, Beamish RJ, Devlin B, Farrell AP, McFarlane S, et al. Strait of Georgia Chinook and coho proposal. Pac Salmon Found; Strait of Georgia Proposal.
- 114. Riddell BE, Brodeur RD, Bugaev AV, Moran P, Murphy JM, Orsi JA, et al. Ocean Ecology of Chinook Salmon. first. American Fisheries Society, 2018.
- 115. Beamish RJ, Neville C, Sweeting R, Lange K. The Synchronous Failure of Juvenile Pacific Salmon and Herring Production in the Strait of Georgia in 2007 and the Poor Return of Sockeye Salmon to the Fraser River in 2009. Mar Coast Fish 2012; 4: 403–414.
- 116. Walters C, Christensen V. Effect of non-additivity in mortality rates on predictions of potential yield of forage fishes. Ecol Model 2019; 410: 108776.
- 117. Pearsall I, Schmidt M, Kemp I, Riddell B. Factors limiting survival of juvenile Chinook salmon, coho salmon and steelhead in the Salish Sea: Synthesis of the findings of the Salish Sea Marine Survival Project. 2021; 142.
- 118. Mesa MG, Poe TP, Maule AG, Schreck CB. Vulnerability to predation and physiological stress responses in juvenile chinook salmon (Oncorhynchus tshawytscha) experimentally infected with Renibacterium salmoninarum. Can J Fish Aquat Sci 1998; 55: 1599–1606.
- 119. Jeffries KM, Hinch SG, Gale MK, Clark TD, Lotto AG, Casselman MT, et al. Immune response genes and pathogen presence predict migration survival in wild salmon smolts. Mol Ecol 2014; 23: 5803–5815. pmid:25354752
- 120. Ferguson JA, Romer J, Sifneos JC, Madsen L, Schreck CB, Glynn M, et al. Impacts of multispecies parasitism on juvenile coho salmon (Oncorhynchus kisutch) in Oregon. Aquaculture 2012; 362–363: 184–192.
- 121.
Hostetter NJ. Susceptibility of juvenile salmonids to avian predation: Are Caspian terns and double-crested cormorants only taking the sick and injured? Thesis submitted in partial fulfilment of the requirements for the degree of master of science, Oregon State University, https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/3197xr26d (2009).
- 122. Edgell T, Demarchi M. California and Steller sea lion use of a major winter haulout in the Salish Sea over 45 years. Mar Ecol Prog Ser 2012; 467: 253–262.
- 123. Beauchamp DA. Bioenergetic Ontogeny: Linking Climate and Mass-Specific Feeding to Life-Cycle Growth and Survival of Salmon. Am Fish Soc Symp; 70.
- 124. Dietrich JP, Van Gaest AL, Strickland SA, Arkoosh MR. The impact of temperature stress and pesticide exposure on mortality and disease susceptibility of endangered Pacific salmon. Chemosphere 2014; 108: 353–359. pmid:24559935
- 125. Connor WP, Burge HL, Yearsley JR, Bjornn TC. Influence of Flow and Temperature on Survival of Wild Subyearling Fall Chinook Salmon in the Snake River. North Am J Fish Manag 2003; 23: 362–375.
- 126. Litzow MA, Ciannelli L, Cunningham CJ, Johnson B, Puerta P. Nonstationary effects of ocean temperature on Pacific salmon productivity. Can J Fish Aquat Sci 2019; 76: 1923–1928.
- 127. Jones L, Schoen E, Shaftel R, Cunningham CJ, Mauger S, Rinella DJ, et al. Watershed-scale climate influences productivity of Chinook salmon populations across.
- 128. Morita S, Morita K, Sakano H. Growth of chum salmon (Oncorhynchus keta) correlated with sea-surface salinity in the North Pacific. ICES J Mar Sci 2001; 58: 1335–1339.
- 129. Debertin AJ, Irvine JR, Holt CA, Oka G, Trudel M. Marine growth patterns of southern British Columbia chum salmon explained by interactions between density-dependent competition and changing climate. Can J Fish Aquat Sci 2017; 74: 1077–1087.
- 130. Malick MJ, Cox SP, Mueter FJ, Dorner B, Peterman M. Effects of the North Pacific Current on the productivity of 163 Pacific salmon stocks. Fish Oceanogr 2017; 26: 268–281.