Marine Climate Variation and Population Forecasting in Pacific Salmon Recovery Planning

Population forecasting plays two distinct roles in species conservation and recovery planning. Long-term forecasts (e.g., population viability analysis) are used for risk assessments and strategic recovery planning, while near-term forecasts are used for year-to-year tactical management decisions. Applications of long-term forecasts for salmon conservation have been reviewed recently, so we focus on the near-term forecasting problem. Traditionally, management forecasts have been based on stock-recruit analysis, but simple stock-recruit models fail badly when conditions affecting stock capacity or productivity change. Including single or a few environmental indicators into statistical forecasts have led to short-term improvements, but often failed within a few years of adoption. There are three problems that make forecasting difficult: (1) There is a wealth of environmental indicators linked to plausible hypotheses of population control; (2) There is a dearth of process-oriented research that can directly evaluate these hypothesized processes; (3) Salmon are embedded in a complex, adaptive ecosystem where the importance of causal processes can change in response to shifting climate or ecosystem phases. We review a number of approaches for incorporating climate variation into salmon forecasts and suggest new directions for incorporating multiple factors within an ecosystem context for salmon management and recovery.