Groundwater Protection

In surveying state officials across the country, GAO discovered that many states possess at least some of the data needed to assess the vulnerability of groundwater to pesticide contamination. Significant gaps in the data exist, however. Officials in only 15 states reported that data have been mapped in their state for all eight vulnerability factors necessary for doing valid assessments. Even when data are available, they often cover only part of the state and are not detailed enough to use in preparing valid assessments. In fact, no state has been completely mapped at a sufficient resolution for every factor. States need to collect more data if they are to conduct valid groundwater assessments. GAO found that the performance of vulnerability assessment models has been inconsistent. At best, existing models have been shown to predict groundwater vulnerability adequately only in some cases--that is, for some pesticides in some soils. Moreover, the model tests generally have not been done on the subcounty scale necessary to show whether the Environmental Protection Agency's differential strategy is either scientifically sound or economically viable. To be useful for regulatory purposes, both the models and their testing will have to be improved.

GAO found that: (1) there were significant gaps in the data needed to conduct valid vulnerability assessments; (2) half of the states did not have data on the vadose and confining zones; (3) existing information was generally not of a sufficient degree of geographic resolution to be useful in vulnerability assessments; (4) there was great variability across states in the availability and sufficiency of the data; (5) the most extensive tests of depth, recharge, aquifer, soil, topography, impact, conductivity (DRASTIC) have found no positive relationship between DRASTIC scores and pesticide contamination; (6) a discriminant analysis approach to predict groundwater contamination was validated once and for only one pesticide; (7) none of the mathematical models were adequately validated to justify their use for developing state management plans; and (8) states generally used unvalidated methods for their assessments and, in most cases, did not verify model predictions with monitoring data.

Recommendations

Our recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.

Director: Team: Phone: