Modelling the future hydroclimatology of the lower Fraser River and its impacts on the spawning migration survival of sockeye salmon

Short episodic high temperature events can be lethal for migrating adult Pacific salmon (Oncorhynchus spp.). We downscaled temperatures for the Fraser River, British Columbia to evaluate the impact of climate warming on the frequency of exceeding thermal thresholds associated with salmon migratory success. Alarmingly, a modest 1.0 degrees C increase in average summer water temperature over 100 years (1981-2000 to 2081-2100) tripled the number of days per year exceeding critical salmonid thermal thresholds (i.e. 19.0 degrees C). Refined thresholds for two populations (Gates Creek and Weaver Creek) of sockeye salmon (Oncorhynchus nerka) were defined using physiological constraint models based on aerobic scope. While extreme temperatures leading to complete aerobic collapse remained unlikely under our warming scenario, both populations were increasingly forced to migrate upriver at reduced levels of aerobic performance (e.g. in 80% of future simulations, >= 90% of salmon encountered temperatures exceeding population-specific thermal optima for maximum aerobic scope; T-opt=16.3 degrees C for Gates Creek and T-opt=14.5 degrees C for Weaver Creek). Assuming recent changes to river entry timing persist, we also predicted dramatic increases in the probability of freshwater mortality for Weaver Creek salmon due to reductions in aerobic, and general physiological, performance (e.g. in 42% of future simulations >= 50% of Weaver Creek fish exceeded temperature thresholds associated with 0-60% of maximum aerobic scope). Potential for adaptation via directional selection on run-timing was more evident for the Weaver Creek population. Early entry Weaver Creek fish experienced 25% (range: 15-31%) more suboptimal temperatures than late entrants, compared with an 8% difference (range: 0-17%) between early and late Gates Creek fish. Our results emphasize the need to consider daily temperature variability in association with population-specific differences in behaviour and physiological constraints when forecasting impacts of climate change on migratory survival of aquatic species.