A REGIONAL MULTIPLE-STRESSOR RANK-BASED ECOLOGICAL RISK ASSESSMENT FOR THE FJORD OF PORT-VALDEZ, ALASKA

We conducted an ecological risk assessment of the marine environment o f Port Valdez, a fjord in south-central Alaska. Because the assessment was regional rather than site-specific and contained a large number o f different stressors in a variety of environments, we required a nont raditional method to estimate risks. We created a Relative Risk Model to rank and sum individual risks numerically within each subarea, from each source, and to each habitat. Application of this model involved division of Port Valdez into 11 subareas containing specific ecologica l and anthropogenic structures and activities. Within each subarea, th e stressor sources were analyzed to estimate exposure of receptors wit hin habitats leading to effects relevant to the chosen assessment endp oints. The subareas were analyzed and compared to form a Port-wide per spective of ecological risk. Available chemical concentrations from se diment and mussels collected from the Port were compared to various to xicological benchmarks as a partial confirmation of the risk analysis. An estimation of the risk of polycyclic aromatic hydrocarbons (PAHs) to marine invertebrates indicated low risk. The municipal boat harbor had the highest estimate, which reflected our relative risk rankings.T he Relative Risk Model approach appears robust and has potential for u se in situations where multiple stressors are of concern and for asses sments covering broad geographic areas. In the Port Valdez assessment the approach provided relative risk rankings for chemical and physical stressors from various sources. But data were available for confirmat ion of risk estimates only for the chemical stressors. The rankings ar e relative, and extrapolation beyond the scenario in which they were d eveloped is not warranted. Uncertainty is large, and the numerical sco res collapse a multidimensional space into a single value. Use of just the numerical score out of context is no more valid than with other i ndexes. The value of the approach lies in the relative rankings and th e accounting of the components of the relative risk score.