U.S. Geological Survey Toxic Substances Hydrology Program--Proceedings of the Technical Meeting, Colorado Springs, Colorado, September 20-24, 1993, Water-Resources Investigations Report 94-4015

A Review of Methods for Assessing the Potential for Regional Ground-Water Contamination

Abstract

The potential for ground-water contamination, or the sensitivity of an area to ground-water contamination, can be estimated by a quantitative, deterministic model that predicts fate and transport of water and contaminants locally, or by a broad, regional qualitative model. Although regional models do not predict ground-water contamination at specific sites as accurately as do the local models, the former are useful tools for regional planning and research. All regional models depict relative contamination potential. Some models focus on contamination potential of ground water in aquifers, either at the land surface or confined beneath low-permeability sediments, whereas other models focus on ground water at the unconfined water table. For a given area, the focus of ground-water protection efforts will determine which models are most appropriate. Some regional models use dimensionless numbers and weighting factors to derive a relative measure of contamination potential, whereas others assemble source data into a composite map and rank the resulting map units for contamination potential according to a set of rules. Both approaches have certain advantages. However, because most contamination potential maps in the area considered (Central United States) show roughly similar patterns of contamination potential, some models may be more practical than others. For example, if simple models give results similar to models requiring cumbersome numeric manipulation of data, there is little justification for the effort needed to execute the more complex model. Availability of information required by a model also should determine its utility. Regional, stochastic models that provide actual, not relative, predictions of contamination potential are needed. Those models should consider the uncertainty in source information in order to estimate contamination potential with specified levels of uncertainty.