On the assessment of health risks at Superfund sites using Monte Carlo simulations

In this paper, state-of-the-art risk assessment models, and the most current and applicable site specific exposure data are employed in the reassessment of baseline risks at three Superfund Sites for which a Record of Decision has been issued since 1990. The results of this reassessment are compared to health risk assessments based on EPA's conservative or default parameters in estimating exposure and risk. These state-of-the-art models are based on Monte Carlo simulation that accounts for uncertainty and variability in the input parameters, and uses parameter distributions based on the current literature. In general, the Monte Carlo analysis produced risk values at the mean and 90(th) percentile that were much lower than those obtained by the EPA using default point estimate values. As a result, the risks calculated at some sites appear to be at levels acceptable to the EPA without additional remediation and, therefore, ''no-action'' alternatives might be more appropriate.