Development of methods to evaluate management options for achieving the recovery of endangered salmon stocks

I developed a stochastic population model in a Bayesian decision analysis framework to evaluate management options for the depleted Cultus Lake, British Columbia, sockeye salmon stock. I sought state-dependent harvest rules that met three management objectives reflecting the probability of recovery within a specified period, the probability of abundance remaining above a conservation threshold, and the economic value of the harvest. This method produced quantitative information about tradeoffs between competing objectives. I found that recovery is feasible for the Cultus Lake sockeye stock under a number of harvest rules that allow harvesting in most years. Results were highly sensitive to pre-spawning mortality rate, indicating the need for a better understanding of that factor. Allowing the Cultus stock to recover may permit other late-run stocks to rebuild, thus partially offsetting the economic losses associated with reduced catches during recovery of the Cultus stock.