Crop-protection compounds are useful tools that enhance the quality of the food we enjoy. However, crop-protection products can enter aquatic systems either by direct or by indirect application. To better understand the possible frequency and magnitude of exposure to water resources, the regulatory community has developed a set of relatively straightforward models for estimating exposure to these water systems. The focus of this research was to compare how well the estimates of exposure to drinking water based on model calculations relate to actual monitoring data. Physical/chemical property data were entered in the EPA's exposure model FIRST and into PRZM/EXAMS. The predictions from FIRST and PRZM/EXAMS were then compared to actual monitoring data from a USGS/EPA cooperative program, which monitored for pesticides in vulnerable surface drinking water supplies during 1999 and 2000. Results from this examination indicate the exposure from the models can overpredict concentrations found in water by several orders of magnitude. An overprediction factor is presented that corrects model predictions to more closely approximate concentrations found in reservoirs (p = 0.05).