Using Bayesian State-Space Models to Evaluate the Effects of Density Dependence, Hatchery Conspecifics, and Environmental Conditions on the Productivity of Threatened Skagit River Steelhead

Successful recovery of threatened populations of Puget Sound steelhead (Oncorhynchus mykiss) will undoubtedly require knowledge of the predominant processes affecting population productivity. Here we fit a hierarchical Bayesian spawner recruit model to a 33 year data set of escapement, catch, and age composition estimates from a terminal fishery for adult wild steelhead in the Skagit River basin to examine the effect of density dependence and other life-stage specific covariates on population dynamics. Our analysis provides four important results regarding factors affecting productivity of wild Steelhead in the Skagit River basin: (1) productivity is strongly limited by the availability of  habitat, presumably in freshwater, due to density dependent processes; (2) productivity is negatively correlated with releases of hatchery reared steelhead smolts; (3) productivity is negatively correlated with large peak winter flow during the first freshwater winter; and (4) productivity is positively correlated with warm phases of the Pacific Decadal Oscillation (PDO) during the first year of ocean residency. The modeling framework employed in spawner recruit analyses not only allows for a better understanding of the point estimates of model parameters including intrinsic productivity, density dependence, and covariates, but also the uncertainty around them.