Evaluating the Efficacy of Source Control Strategies for Managing Public Health Risks of Benzene. David L. MacIntosh, Halűk Özkaynak and Jianping Xue, Harvard University, School of Public Health, Boston, MA
A probabilistic model of human exposures and absorbed doses of benzene to residents of U.S. EPA Region V was used to evaluate the efficacy of alternative source control strategies for managing public health risks of benzene. Monte Carlo simulation was used to propagate distributions of inter-individual variability and parameter uncertainty through to estimated lifetime exposures and doses. Subsequently, lifetime exposures, doses, and cancer risks were estimated for various fractiles of the population distribution under four situations: (1) a base case that represents current exposures, (2) reduced tailpipe emissions of benzene commensurate with the 1990 Clean Air Act Amendments, (3) removal of benzene from gasoline, and (4) removal of benzene from building materials and other consumer products. Under the base case, the population distribution of risk was estimated to be lognormally distributed with geometric mean (GM) 6.5 x 10-5and geometric standard deviation (GSD) 2.0; with population risk 8.1 x 10-5. The results for cases (2)-(4) have yet to be completed, although it can be predicted in advance that cases (2) and (3) will have substantial effects on the population distribution of benzene exposures and case (4) will have little effect. The examples presented here illustrate the utility of probabilistic models in evaluating the efficacy of alternative pollution control strategies to protect public health.