MODELING THE PROBABILITIES OF GROUNDWATER CONTAMINATION BY PESTICIDES

Field soils show significant spatial variations in properties, such as organic matter content, bulk density, and moisture content, that can affect the mobility and persistence and thus fate of organic pesticide s in the soil environment. A simple model incorporating the variations in soil and pesticide parameters has been developed to assess the gro undwater contamination potential of pesticides. The model is based on linear, equilibrium, and reversible sorption, first-order degradation, and steady piston flow, and allows the unsaturated soil zone to be di vided into a number of layers of different thickness and properties. F or each input parameter, 500 random data were generated from normal di stributions that characterize the variability of the parameters. The f ate of 29 pesticides were assessed using soil and environmental condit ions of the Swan Coastal Plains of Western Australia and pesticide pro perties reported in the literature. The predicted pesticide residue fr actions remaining at 150-, 300-, and 500-cm depths were described by b eta distributions, and the corresponding travel times by normal distri butions. Fourteen out of the 29 pesticides are predicted to reach 150 and 300 cm and 13 are predicted to reach the 500-cm depth with mean re sidue fractions greater than or equal to 0.01%. The predicted mean tra vel times for pesticides that may reach 300 cm vary from about 2 mo to about 18 yr. Significant standard deviations are associated with thes e mean residue fractions and travel times. Six pesticides, fenamiphos, simazine, metribuzin, linuron, fenarimol, and metalaxyl, have been id entified as having high cumulative probabilities of 0.82 to 1.00 for r esidue concentrations to be >0.01% at the 300-cm depth.