Student Research: Robert C. Lee

, , 1993
Faculty Advisor: John C. Kissel

Prediction of Exposures to Arsenic Contaminated Residential Soil Using Monte Carlo Simulation: A Comparison with Biomonitoring Data


The derivation and promulgation of cleanup standards for soil contaminants are impeded by the complex and uncertain science of assessing exposures to these contaminants. Monte Carlo simulation (MCS) is proposed as a more statistically and scientifically valid method of asessing soil contaminant exposures than deterministic methods currently used by regulatory agencies.

The nature and predictive value of MCS are evaluated in an exposure scenario involving arsenic contaminated residential soil near the formmer ASARCO smelter in Tacoma, Washington. The population of interest is children, aged 2-6, living within one-half mile of the smelter site. Models are derived which predict urinary arsenic levels based upon unintentional soil ingestion and inhalation exposure pathways. Distributions for variables are derived based upon site-specific data and previous exposure studies.

Simulated distributions of urinary arsenic levels are compared to two urinary biomonitoring studies performed during the late 1980s. Simulated and biomonitored urinary arsenic distributions are significantly different. Inadequate data for important variables, as well as inaccurate biomonitoring data, are likely contributors to the discrepancies.

This study demonstrates that MCS can be a valuable tool in determining the statistical nature of exposures to contaminated soil. Evaluation of uncertainty and varibaility in exposure estimates can be valuable in teRMS of determining exposure pathways and areas which require futher research, even in simple exposure scenarios. Evaluation of uncertainty and varibaility may result in more cost-effective public health and hazardous waste remediation decisions.

Unintentional soil and dust ingestion appears to be a major exposure pathway of concern for children in the study area. However, soil ingestion is not well quantified, and may overlap with other exposure pathways such as dermal absorption and contaminated food intake. Further research is required to determine distributions relating to contaminated soil intake and absorption in order to maximize the utility of MCS.

A possible strategy of assessment of exposures and risks, or for the derivation of soil standards for contaminants, is to develop and routinely use stochastic predictive modeling, supplemented with adequate biomonitoring assays for the contaminants of interest when possible. This combination is more informative than either method alone, and provides a system for validation.