Using metapopulation models to estimate the effects of pesticides and environmental stressors to Spring Chinook salmon in the Yakima River Basin, WA

Population-level endpoints provide ecological relevance to Ecological Risk Assessments (ERAs), because this is the level at which environmental management decisions are made. However, many population-level risk assessments do not reflect the spatial and temporal heterogeneity of the populations they represent, and thus preclude an understanding of how population dynamics and viability are affected by toxicants on a regional scale. We have developed a probabilistic ERA (specifically, a Bayesian Network-Relative Risk Model (BN-RRM)) that integrates an Adverse Outcome Pathway (AOP) framework, to quantify the sub-lethal and lethal effects of toxicants and environmental stressors on the metapopulation dynamics of salmonids. As a case study for developing this model, we have examined the impacts of organophosphate (OP) insecticides, water temperature, and dissolved oxygen on the Spring Chinook (Oncorhynchus tshawytscha) salmon metapopulation in the Yakima River Basin (YRB), Washington. A stochastic Matrix Metapopulation Model was developed using demographic data for three Spring Chinook salmon populations and one supplemental hatchery population in the YRB. Site specific data on OP contaminated habitats utilized by various salmonid life stages were incorporated into the metapopulation model by incrementally reducing survival parameters based on levels of exposure. Exposure scenarios were simulated for 200 replications of 50-year population projections using RAMAS Metapop©, and the results were incorporated into the BN-RRM. The results of this modeling effort indicated that small, wild Spring Chinook populations in the YRB have a greater probability of altered population dynamics when exposed to stressors than larger, supplemented populations. Additionally, the results indicated a seasonal effect of the stressors, with summer conditions posing a greater risk to salmon populations than winter conditions. This probabilistic ERA framework shows promise for estimating the spatiotemporal impacts of stressors on ESA-listed species (i.e., Pacific salmon) at the metapopulation level, where population dynamics and spatial structure create complex risk dynamics