Application of hierarchical Monte Carlo simulation to the estimation of human exposure to mercury via consumption of king mackerel (Scomberomorus cavalla)

Data from mercury poisoning episodes in Japan and Iraq have been used to derive threshold levels for health effects in humans, and the U.S. Environmental Protection Agency (EPA) has established a reference dose (RfD) for methylmercury of 0.30 μg/kg/day. It is desirable, therefore, to predict exposures for various populations, and examine these exposures in relation to the established RfD. Given data on mercury concentrations in fish, and rates of fish consumption among human populations, it is possible to calculate exposure in one of several ways. One of the simplest ways involves the use of average values to compute a single point estimate of exposure. Monte Carlo methods provide more information for characterizing the full range and variation of exposures, but fail to account for dependencies among input parameters, and often result in “impossible” parameter combinations.; A Hierarchical Monte Carlo Simulation (HMCS) method is presented which overcomes the limitations of the point estimate and conventional Monte Carlo approach by making parameter selections in a hierarchical manner, thereby preserving dependencies among input parameters and eliminating the chance of “impossible” parameter combinations. A HMCS model is developed using the S-distribution to approximate the respective input parameters, and is used to evaluate exposures under various scenarios.