Evaluation of Chum Salmon Fishery Performance Using Ricker and Beverton-Holt Stock Recruitment Approaches in a Bayesian Framework

摘要

To improve the ability of managers to make appropriate decisions, we analyzed chum salmon Oncorhynchus keta data from central British Columbia and evaluated alternative escapement goals and the potential impacts of environmental change on fishery performance. Fishery performance was expressed in terms of mean catch and the probability of annual catches exceeding 100,000 fish. We investigated Ricker and Beverton-Holt stock-recruitment relationships and tested if environmental variation as reflected bysea surface temperature anomalies (SSTA) improved these relationships. Uncertainties associated with stock-recruitment model parameters were evaluated using a Bayesian approach. Posterior distributions of the parameters were generated using Monte CarloMarkov Chain procedures. Our data set was relatively informative as reflected by the well defined probability distributions of the stock-recruitment parameters. Fishery performance was assessed through stochastic projections over three generations (12 years) under various scenarios of escapement goals and climate conditions. Sea surface temperature anomalies significantly improved both models and recruitments were higher for chum salmon that entered the sea during cold years than in warm years. Recruitments tended to be more correlated with SSTA from north of the point of fry sea entry than from where fry actually entered the sea. Using a Ricker model, an escapement goal of roughly 100,000-150,000 spawners (depending oif the climate scenario) yielded the highest probability of a projected yearly catch of at least 100,000 chum, while with the Beverton-Holt model, only -45,000 spawners were needed. Since the Ricker model was insensitive to the type of priors and is more precautionary than the Beverton-Holt, we conclude that it is the most appropriate.