Estimating the Effects of Basin Characteristics on Stream Pesticide Concentrations Using Coupled Watershed and Regression Models

A process-based watershed model (TOPMODEL) was used with a regression model to estimate stream chemistry in a network of 65 agricultural basins located throughout the conterminous United States. First, a simple version of TOPMODEL was applied to each of the basins to predict the relative contributions of saturation overland flow, infiltration-excess overland flow, and subsurface flow to streamflow. The TOPMODEL estimates of flow-path sources then were used along with other basin characteristics and measured stream chemistry in a regression model to predict concentrations for various pesticides. The modeling results indicated that the intensity of pesticide use explained most of the variability (r-squared value of about 0.6) in pesticide concentrations among the basins. In addition, infiltration-excess overland flow was positively correlated with pesticide concentrations, whereas saturation overland flow was negatively correlated with pesticide concentrations. (Typically, the flow-path related variables together increased the r-squared value to about 0.8.) This analysis suggests that infiltration-excess overland flow increases the transport efficiency of land-surface applied pesticides, but saturation overland flow has a diluting effect.