Currently, regulatory practice in the United States is to estimate potential concentrations in drinking water from surface water by using an index reservoir scenario. This approach extrapolates results from the modeling of a single field with maximum application rates to the watershed scale, based on a percent crop area estimate. Since 1998, Bayer CropScience and its predecessor companies have conducted drinking water monitoring studies with a number of different compounds. The results from these studies show that the index reservoir scenario greatly overpredicts residues in surface water. The most important factor is the overestimation of use within a watershed. Other factors contributing to the overestimation of concentrations are the conservative procedures used to obtain the chemical fate related input parameters and the simplified hydrology. A new procedure based on the USGS WARP model, being developed by a group of scientists from the EPA, USGS, USDA, and industry, will provide more realistic estimates of concentrations of pesticides and their metabolites in drinking water.