Concern over global climate change is widespread, but quantifying relationships between temperature change and animal fitness has been a challenge for scientists. Our approach to this challenge was to study migratory Pacific salmon (Oncorhynchus spp.), fish whose lifetime fitness hinges on a once-in-a-lifetime river migration to natal spawning grounds. Here, we suggest that their thermal optimum for aerobic scope is adaptive for river migration at the population level. We base this suggestion on several lines of evidence. The theoretical line of evidence comes from a direct association between the temperature optimum for aerobic metabolic scope and the temperatures historically experienced by three Fraser River salmon populations during their river migration. This close association was then used to predict that the occurrence of a period of anomalously high river temperatures in 2004 led to a complete collapse of aerobic scope during river migration for a portion of one of the sockeye salmon (Oncorhynchus nerka) populations. This prediction was corroborated with empirical data from our biotelemetry studies, which tracked the migration of individual sockeye salmon in the Fraser River and revealed that the success of river migration for the same sockeye population was temperature dependent. Therefore, we suggest that collapse of aerobic scope was an important mechanism to explain the high salmon mortality observed during their migration. Consequently, models based on thermal optima for aerobic scope for ectothermic animals should improve predictions of population fitness under future climate scenarios.

Since 1995, several stocks of Fraser River sockeye salmon (Oncorhynchus nerka) have begun upriver spawning migrations significantly earlier than previously observed. In some years, the timing of peak migration has shifted more than 6 weeks. Coincident with this early migration are high levels of en route and pre‐spawning mortality, occasionally exceeding 90%. These phenomena pose risks to the perpetuation of these fisheries resources. At present, although there are many competing hypotheses (e.g., energetics, osmoregulatory dysfunction, oceanic conditions, parasites) that may account for early migration and high mortality, there are no definitive answers, nor any causal evidence that link these issues. With poor predictive ability in the face of uncertainty, fisheries managers have been unable to effectively allocate harvest quotas, while ensuring that sufficient fish are able to not only reach the spawning sites, but also successfully reproduce. If trends in mortality rates continue, several important sockeye salmon fisheries and stocks could collapse. Indeed, one sockeye stock has already been emergency listed as endangered under Canadian legislation.

Adult sockeye salmon Oncorhynchus nerka destined for the Fraser River, British Columbia are some of the most economically important populations but changes in the timing of their homeward migration have led to management challenges and conservation concerns. After a directed migration from the open ocean to the coast, this group historically would mill just off shore for 3-6 weeks prior to migrating up the Fraser River. This milling behaviour changed abruptly in 1995 and thereafter, decreasing to only a few days in some years (termed early migration), with dramatic consequences that have necessitated risk-averse management strategies. Early migrating fish consistently suffer extremely high mortality (exceeding 90% in some years) during freshwater migration and on spawning grounds prior to spawning. This synthesis examines multidisciplinary, collaborative research aimed at understanding what triggers early migration, why it results in high mortality, and how fisheries managers can utilize these scientific results. Tissue analyses from thousands of O. nerka captured along their migration trajectory from ocean to spawning grounds, including hundreds that were tracked with biotelemetry, have revealed that early migrants are more reproductively advanced and ill-prepared for osmoregulatory transition upon their entry into fresh water. Gene array profiles indicate that many early migrants are also immunocompromised and stressed, carrying a genomic profile consistent with a viral infection. The causes of these physiological changes are still under investigation. Early migration brings O. nerka into the river when it is 3-6° C warmer than historical norms, which for some late-run populations approaches or exceeds their critical maxima leading to the collapse of metabolic and cardiac scope, and mortality. As peak spawning dates have not changed, the surviving early migrants tend to mill in warm lakes near to spawning areas. These results in the accumulation of many more thermal units and longer exposures to freshwater diseases and parasites compared to fish that delay freshwater entry by milling in the cool ocean environment. Experiments have confirmed that thermally driven processes are a primary cause of mortality for early-entry migrants. The Fraser River late-run O. nerka early migration phenomenon illustrates the complex links that exist between salmonid physiology, behaviour and environment and the pivotal role that water temperature can have on population-specific migration survival.

Mean summer water temperatures in the Fraser River (British Columbia, Canada) have increased by $1.5 1C since the 1950s. In recent years, record high river temperatures during spawning migrations of Fraser River sockeye salmon (Oncorhynchus nerka) have been associated with high mortality events, raising concerns about long-term viability of the numerous natal stocks faced with climate warming. In this study, the effect of freshwater thermal experience on spawning migration survival was estimated by fitting capture-recapture models to telemetry data collected for 1474 adults (captured in either the ocean or river between 2002 and 2007) from four Fraser River sockeye salmon stockaggregates (Chilko, Quesnel, Stellako-Late Stuart and Adams). Survival of Adams sockeye salmon was the most impacted by warm temperatures encountered in the lower river, followed by that of Stellako-Late Stuart and Quesnel. In contrast, survival of Chilko fish was insensitive to the encountered river temperature. In all stocks, in-river survival of ocean-captured sockeye salmon was higher than that of river-captured fish and, generally, the difference was more pronounced under warm temperatures. The survival-temperature relationships for ocean-captured fish were used to predict historic and future (2010-2099) survival under simulated lower river thermal experiences for the Quesnel, Stellako-Late Stuart and Adams stocks. A decrease of 9-16% in survival of all these stocks was predicted by the end of the century if the Fraser River continues to warm as expected. However, the decrease in future survival of Adams sockeye salmon would occur only if fish continue to enter the river abnormally early, towards warmer periods of the summer, as they have done since 1995. The survival estimates and predictions presented here are likely optimistic and emphasize the need to consider stock-specific responses to temperature and climate warming into fisheries management and conservation strategies.

We used a multi-stock comparison to identify spatial and temporal characteristics of environmentally driven sources of variability across four decades in the productivity of 29 sockeye salmon (Oncorhynchus nerka) stocks from British Columbia (B.C.) and Alaska. We examined patterns of covariation among indices of survival rate (residuals from the best-fit stock-recruitment curve) and found positive covariation among Fraser River sockeye stocks (southern B.C.) and, to a greater extent, among Bristol Bay stocks (western Alaska) but no evidence of covariation between these two regions or with stocks of other regions in B.C. and Alaska. This indicates that important environmental processes affecting variation in sockeye survival rate from spawners to recruits operate at regional spatial scales, rather than at the larger, ocean-basin scale. The observed covariation in survival rates of Bristol Bay stocks appears to be due to a combination of both freshwater and, to a greater degree, marine processes. Bristol Bay sockeye stocks showed a dramatic and persistent increase in survival rates coinciding with the abrupt changes in the North Pacific environment in the mid-1970s; however, there was little evidence of a similar response for Fraser River stocks.