Modeling the Influence of Environmental Factors on Spawning Migration Mortality for Sockeye Salmon Fisheries Management in the Fraser River, British Columbia

The impact of freshwater environmental factors on spawning migration mortality was modeled to provide a predictive tool for fisheries management of four run timing groups of Fraser River sockeye salmon Oncorhynchus nerka: early Stuart (Stuart Lake), early summer, summer, and late. We tested the significance of different measures of water temperature, discharge, fish abundance, and entry timing for forecasting discrepancies between lower-river and upriver escapement estimates using multiple regressions of principal component scores. Descriptive discrepancy models (i.e., management adjustment'' models) identified using Akaike's information criterion were consistent with the known biology of each group. For example, temperature and discharge thresholds were selected for early Stuart run discrepancy models, reflecting the extremes in both variables experienced by these early migrants. Predictive discrepancy models were also generated for each run timing group by using the limited number of environmental variables that are available in-season to fisheries managers. Even predictive discrepancy models using simple environmental metrics of average river temperature, flow, and river entry timing provide a valuable tool for forecasting relative indices of spawning migration mortality. This study provides an example of how environmentally based predictive tools can be used to inform fisheries management decisions and improve the probability of achieving spawning escapement targets.