The Relative Risk Model (RRM) is a tool used to calculate and assess the likelihood of effects to endpoints when multiple stressors occur in complex ecological systems. In this study, a Bayesian network was used to calculate relative risk and estimate uncertainty (BN-RRM) in the Puyallup River Watershed. First, we calculated the risk of prespawn mortality of coho salmon. Second, we evaluated the effect of low impact development (LID) as a means to reduce risk. Prespawner mortality in coho salmon within the Puyallup watershed was the endpoint selected for this study. A conceptual model showing causal pathways between stressors and endpoints was created to show where linkages exist. A relative risk gradient was found throughout the watershed. The lowest risk was found in risk regions with the least urban development, and the greatest risk of prespawner mortality was found in the highly urbanized risk regions with the largest amounts of impervious surface. LID did reduce risk but only when implemented at high intensities within the urban watersheds. The structure of the BN-RRM also provides a framework for water quality- and quantity-related endpoints within this and other watersheds. The framework is also useful for evaluating different strategies for remediation or restoration activities. The adaptability of using BNs for a relative risk assessment provides opportunities for the model to be adapted for other watersheds in the Puget Sound and Salish Sea region. Integr Environ Assess Manag 2014;10:269-278. (c) 2013 SETAC
