Improving preseason forecasts for U.S. coho salmon management units by accounting for spatially structured temporal variation in age-at-maturity

We propose to develop a new class of forecast models for Southern U.S. naturally spawning coho salmon Management Units (MUs) that accounts for possible environmental and demographic effects on the proportion of coho jacks within cohorts. The age composition of southern coho populations consists of jacks — precocious males that mature after ~ 6 months in the ocean — and older males and females that mature after ~18 months at sea. Integrating such information into the sibling relationship may improve forecasts by disentangling the extent to which the observed jack abundance in a given year reflects cohort size, marine survival, and future adult returns versus variation in the age composition. Moreover, previous studies have identified synchronous temporal variation in jack prevalence among southern coho populations at local and regional scales, providing an informative basis by which to coordinate the sharing of information between well-monitored management units and those for which demographic data are sparse.

Recent advances in statistical modelling have given rise to a number of methods for inference and forecasting which account for shared trends and spatially-structured co-variation among populations that have distinct potential for improving forecasts of Southern U.S. coho salmon. We will develop a suite of alternative forecast models using statistical frameworks that incorporate information on spatially structured environmental effects on age-at-maturity (i.e. jack prevalence within cohorts), and evaluate their forecast accuracy to identify the model or set of models with the best performance. Ultimately, this project will generate an alternative preseason forecast framework for coho salmon MUs that will be readily available for use by managers to promote the conservation and sustainable harvest of this culturally iconic and economically valuable species.

S19-SP11 Improving short-term recruitment forecasts for coho using a spatiotemporal integrated population model 2019 Report