Salish Sea Ecosystem Conference

Refined estimates of premature coho salmon spawner mortality in Puget Sound urban basins in relation to landscape condition and precipitation patterns

Room 608-609

We provide an overview of a landscape ecotoxicology project with the goal of identifying the causes of coho salmon (Oncorhynchus kisutch) spawner mortality that occurs widely in urbanized streams throughout the Puget Sound lowlands. Urban stream surveys since the late 1990s have shown that returning coho salmon spawners are dying at high rates (e.g., 40-90% across a total run) before they can spawn. The weight of evidence to date suggests that toxic urban stormwater runoff is the likely cause of the recurring fish kills. Spawner die-offs at these high rates pose a potential threat to wild coho populations, particularly in basins undergoing land use conversion in response to regional human population growth. We previously used survey data from a limited number of Seattle-area urban streams to identify a close correlation between the severity of coho mortality and the proportions of impervious surfaces, roads, and commercial zoned property types within a basin. These relationships were strengthened when rainfall patterns were qualitatively incorporated into the analysis. To improve the predictive strength of our analyses, we added spawner survey data from 50 additional monitoring sites. We also explicitly incorporated precipitation patterns into the statistical modeling structure; used dynamic representations of human population growth and associated changes in land cover and land use; and, incorporated uncertainty into the model predictions. The incorporation of additional sites improved model output confidence and also captured a larger geographic area within Puget Sound. The time series for regional growth and changing landscape patterns yielded a more precise coupling between observed coho mortality and these covariates. Overall, our new analyses identified the potential interplay between anthropogenic (e.g., development) and climatic (precipitation) drivers of coho mortality. Consistent with other lines of evidence, our current results suggest that motor vehicles are the likely source of an as-yet unidentified chemical (or chemical mixture) that washes off the urban landscape into coho spawning habitats. We used our refined analyses to predict likely hotspots for coho spawner die-offs in unmonitored basins across a gradient of urbanization in central Puget Sound.