Numerical simulations of Fraser River sockeye salmon homing migration routes in a dynamic marine environment

A spatially-explicit individual-based model was used to simulate adult sockeye salmon (Oncorhynchus nerka) return migration routes from the Gulf of Alaska to the Fraser River through temporally variable, spatially-explicit environmental fields. Examined was whether coastal migration route variability could be explained by the interactions between simple behavior rules and a dynamic ocean environment described by historical temperature observations and estimated surface currents. Assuming that sockeye were broadly distributed throughout the central Gulf of Alaska prior to homeward migration, and that during homeward migration they oriented on a fixed compass bearing, the following mechanisms were invoked to produce migration route variations among years: 1) the distribution prior to homing was constrained by thermal limits, 2) sockeye were advected by surface currents during open ocean migration, and 3) sockeye tended to avoid high water temperatures. The behavioural component of the model was numerically optimized to maximize the fit between simulated and observed coastal migration routes, while maintaining swimming speeds and migratory timing consistent with the literature. The optimized model suggested that southern thermal limits and current advection could not explain much of the observed coastal migration route variability. The tendency to avoid high temperatures explained aout 33% of the variation and suggested that coastal processes may be more important than offshore.