| With increasing evidence of agricultural pesticide pollution of ground water in the nation, it is pertinent to have a reliable method of determining where the problems are likely to occur. A number of comprehensive computer simulation models are available for site-specific evaluations of pesticide behavior in the root zone. Such models are usally data intensive and require knowledge of a number of specific soil, environment, crop and pesticide parameters. In a majority of cases, such parameters are neither available nor likely to be available in the near future. As an alternative, various simple screening approaches for assessing the relative ground water pollution potential of various pesticides have been devised; these can be classified as empirical index methodologies.;No extensive documentation of the reliability of these methodologies is available, and there is a need for accuracy tests against field data. The reliabilities of three index methodologies, namely, DRASTIC, LEACH, and Rao's Methodology were determined by applying the methodologies to numerous selected areas where pesticide contamination has been reported. Statistical validity testing was performed to assess their respective correlations to actual measured pollution. The testings indicated that the three methodologies cannot be reliably correlated with actual field data on pesticide pollution. However, one of the methodologies, LEACH, seemed to have promise if its results are rank ordered and then used.;The development of an empirical index via statistical models, using readily available existing information, brought to light interesting results. Four models developed via multiple regression analysis indicated correlation with field data; however, the correlations were too low to be considered desirable. Further investigation of statistical models, via discriminant analysis utilizing 17 parameters variously representing the pesticide, soil, crop and hydrologic characteristics, produced significant results. The model or classification criterion yielded a remarkable percentage of correct classification of pollution areas, especially for identification of areas with no evidence of pollution. |
