Application of decision analysis to evaluate recovery actions for threatened Snake River fall chinook salmon (Oncorhynchus tshawytscha)

We used models and decision analysis to incorporate uncertainties into evaluations of two recovery actions for Snake River fall chinook salmon (Oncorhynchus tshawytscha): (i) maximize transportation and (ii) natural river drawdown of four lower Snake River dams. In the retrospective analysis, we compared alternative stock–recruit models and selected one that was consistent with historical spawner–recruit data and allowed us to implement alternative hypotheses about hydrosystem, hatchery, harvest, and environmental effects. In the prospective analysis, we used this model and posterior distributions of its parameters in a decision analysis framework to compare projected escapements for the two actions over a range of alternative hypotheses. We found that drawdown was most risk averse, producing larger long-term escapements than maximizing transportation under most hypotheses and model assumptions. Maximizing transportation and drawdown produced similar escapements only if we assumed high or increasing estuary and ocean survival rates of transported fish coupled with either severe reductions in harvest rates or insensitivity of upstream survival rates to dam construction and removal. Although there was relatively little information available for Snake River fall chinook (particularly about estuary and ocean survival rates of transported smolts), decision analysis was a useful technique for organizing data, assessing actions over a range of uncertainties, and identifying research priorities.