Assessment of Risk to Terrestrial Receptors Using Uncertain Analysis—A Case Study

Abstract

Former industrial landfills on Crab Orchard National Wildlife Refuge were found to contain polychlorinated biphenyls (Aroclor-1254) at levels posing risks to wildlife. This case study reports on a risk assessment that used both deterministic and probabilistic (Monte Carlo) methods to provide risk managers with point estimates of risk, as well as multiple descriptors of uncertainty and variability around the risk estimate. Problem formulation identified mink (Mustela vison) as the higher trophic level assessment endpoint. Exposure analysis estimated Aroclor 1254 doses to mink using a multi-species food web model with probabilistic exposure parameters. Toxicity reference value point estimates and distributions were obtained from the literature. Risk characterization used quotient and probabilistic methodologies. Deterministic quotient analysis corroborated remedial goals originally established in the Record of Decision. Probabilistic risk analysis estimated projected post-remediation soil conditions as having a 7% probability of inducing significant reproductive impairment in mink. A probabilistic approach, while giving greater insight into the consequences of uncertainty and variability inherent in data and risk analyses, can present risk managers with a more ambiguous basis for proceeding with a remedial action. This suggests that more sophisticated risk assessment techniques will not necessarily provide an easier path to remedial decisions.