PARENT SESSION

(06-01) Demonstration of a Substantial Reduction in Uncertainty after Application of Site-Specific Data - A Case Study with Wildlife Exposure to PCBs via Plants in the Diet.

Zwiernik, Matt^*,1, Blankenship, Alan^1,2, Jones, Paul¹, Kay, Denise², Bradley, Patrick¹, Strause, Karl¹, Pastva, Stephanie¹, Holem, Ryan², Giesy, John¹, Kannan, Kurunthachalam¹, ¹ Michigan State University, East Lansing, MI² ENTRIX, Inc., Okemos, MI

ABSTRACT- Wildlife ecological risk assessments are integral components of remedial investigations in the United States. The Aquatic Toxicology Laboratory of Michigan State University (MSU-ATL) has been conducting a comprehensive site-specific evaluation of the risk of harm to resident wildlife populations inhabiting the Kalamazoo River Superfund site which is contaminated with PCBs. The collection of information on actual concentrations of PCBs in tissues and field populations afforded an opportunity to compare observed and predicted exposures and effects to a range of organisms. Site-specific data collected for targeted receptors included PCB concentrations in receptor tissue, site-specific dietary composition, and PCB concentrations of dietary items. This data was then utilized in both a top-down (tissue-based) as well as a bottom-up (dietary-based) risk evaluation. Previously, a similar evaluation had been conducted based, in part, on the limited site-specific data available at the time, and to a greater extent on modeled data. Modeled and/or estimated data in the predictive assessment included receptor dietary composition, concentration of PCBs in dietary items, and foraging ranges. The results of the predictive evaluation concluded that on-site PCBs in floodplain soils and associated vegetation along the river may represent a risk of harm to resident ecological receptors. Substituting MSU-ATL site-specific, measured data into the predictive assessment substantially reduced the calculated risk of harm. Differences between modeled and measured data were found to be primarily due to the estimated exposures via consumption of on-site vegetation. The 95% upper confidence limit of the arithmetic mean for site-specific measured plant tissue PCB concentrations was 0.033 mg/kg ww, while the modeled estimate was approximately 650-fold greater at 22.0 mg/kg ww. In this presentation, we will compare and contrast two separate risk assessment approaches, effects of their underlying assumptions, and their relevance to decision making as part of the on-going remedial investigation and feasibility studies.

Key words: pcbs, wildlife, risk, uncertainty