Superfund Research Program
A Holistochastic Approach to Human Exposure Assessment
Project Leader: George Christakos
Grant Number: P42ES005948
Funding Period: 1995 - 2006
This project is concerned with the quantitative risk assessment of environmental processes in heterogeneous environments. The goal involves two different components: modeling of physical processes by means of mathematical models, and estimation of exposure and health impact at various scales. Meaningful estimates of human and ecological exposure to anthropogenic contaminants that account for uncertainty, and an improved understanding of subsurface processes based on stochastic analysis are to be provided.
The variability of the subsurface environment is an important factor in assessing the large-scale spatial and temporal distributions that emerge in environmental processes, such as in the transport of pollutants in the groundwater. Prediction of these patterns is very important for evaluating the potential for human and ecological risk. The modeling of processes in which spatiotemporal variabilities play a major role must account for the incomplete characterization, that is for the limited number of available measurements due to socio-economic and technological considerations. Stochastic methods are capable of estimating the level of uncertainty, and thus they provide a very useful tool for the modeling and the estimation of processes in space and time.
The research efforts of this group during the last year have focused on four major directions. First, researchers have investigated the solution of flow equations by means of space transforms. Progress during the last year involves the development and testing of numerical codes for space transform solutions of flow and transport equations with emphasis on optimizing the efficiency of the codes, and the development of methods for a more accurate representation of the boundary effects. Second, diagrammatic analysis of effective parameters that represent the behavior of the porous media at large scales have been conducted. In this direction, the calculation of effective parameters that incorporate the size of the domain have been emphasized. Thus, estimates of effective permeability have been obtained that show an explicit dependence on the domain size and the correlation range. Such expressions are useful for the upscaling of effective parameters. Third, with respect to stochastic analysis of multiphase flow, a new method for solving the problem of two-phase flow within a porous medium has been developed which is based on tracking the streamlines of the two fluids. This method not only is intuitively appealing, but it also has an advantage over existing methods, because it can handle flow under conditions that do not justify neglecting the capillary pressure. Fourth, in the area of space/time estimation methods, emphasis has been on developing numerical codes for the estimation of exposures and their impact on human health that vary both in time and space. Space/time estimation is based on the theory of spatiotemporal random fields, which can account for exposures that have definite trends as opposed to stationary and homogeneous processes that have constant means. Random fields provide a more accurate representation of the physical reality than homogeneous and stationary fields, since the former allow for a spatial or temporal dependence of the means. The results of these projects will help to provide more meaningful estimates of human and ecological exposure to contaminants that take into account the uncertainties and variabilities of groundwater flow and reactive contaminant transport.