A fisheries risk-assessment framework to evaluate trade-offs among management options in the presence of time-varying productivity

Empirically based simulation models can help fisheries managers make difficult decisions involving trade-offs between harvests and maintaining spawner abundance, especially when data contain uncertainties. We developed such a general risk-assessment framework and applied it to chum salmon (Oncorhynchus keta) stocks in the Arctic-Yukon-Kuskokwim region of Alaska, USA. These stocks experienced low abundance in the 1990s, which led to declarations of economic disaster and calls for changes in harvest strategies. Our stochastic model provides decision makers with quantitative information about trade-offs among commercial harvest, subsistence harvest, and spawner abundance. The model included outcome uncertainty (the difference between target and realized spawner abundances) in the subsistence and commercial catch modules. We also used closed-loop simulations to investigate the utility of time-varying management policies in which target spawner abundance changed in response to changes in the Ricker productivity parameter (a), as estimated with a Kalman filter. Time-varying policies resulted in higher escapements and catches and reduced risk across a range of harvest rates. The resulting generic risk-assessment framework can be used to evaluate harvest guidelines for most salmon stocks.