Alternative Recruitment Models of Adams River Sockeye Salmon, Oncorhynchus nerka

Abstract

Despite evidence of depensatory interactions among year-classes of Adams River sockeye salmon (Oncorhynchus nerka), the best management policy is one of equal escapement for all year-classes. We fit alternative models (Ricker model and Larkin model) to 32 yr of stock–recruitment data and checked, using simulation tests, that the significant interaction terms in the Larkin model are not caused by biases in estimating the parameters. We identified a parameter set (Rationalizer model) for which the status quo cyclic escapement policy is optimal, but this set fits the observed data very poorly. Thus it is quite unlikely that the Rationalizer model is correct or that the status quo escapement policy is optimal. Using the fitted stock–recruitment parameters, we simulated the sockeye population under several management policies. The escapement policy optimal under the Ricker model is best overall because of the high yields if it should be correct. If the three stock–recruitment models are equally likely to be correct, the simulations predict that adopting a constant-escapement policy would increase long-term yield 30% over the current policy and that an additional 15% increase in yield could be obtained if the policy were actively adaptive.