Recovery tactics for sockeye blocked from anadromy evaluated through decision analysis and value of information

Choosing among recovery tactics for fish populations is challenged by high costs and uncertain outcomes. Depressed populations of salmonids are often supplemented using hatchery propagation; these strategies may involve releasing fish at different locations, densities, and life stages. We describe a unique decision context where supplementation is used to improve anadromous returns of a mixed anadromous-resident coastal sockeye salmon (Oncorhynchus nerka) population. Current anadromous returns in this system are declining, while resident abundance is stable. The recovery of anadromous sockeye salmon depends on depends in part on the heritability of anadromy, since anadromous sockeye are used as hatchery broodstock. Managers have long assumed that anadromy is a heritable trait, but the mounting evidence shows that both genetic and environmental processes shape the expression of migratory phenotypes. We use decision analysis to evaluate a series of decisions regarding whether to allow fish passage above the dam, whether to stock anadromous progeny above or below the dam, at what life stage, and by how much. We further conduct a value of perfect information analysis to evaluate whether outcomes can be improved with an experiment aimed at estimating anadromy. When model projections are integrated over uncertain heritability, the optimal solution is to stock into the reservoir at a high rate and allow passage of adults back into the reservoir. Value of perfect information suggests that this decision is insensitive to heritability, implying that restoration decisions should not wait for a time-intensive experiment. This work emphasizes careful consideration of objectives and the value of experimentation in driving decisions regarding population recovery and management.