Estimating pesticide concentrations in U.S. streams from watershed characteristics and pesticide properties

AGRO 140

Wesley W. Stone, wwstone@usgs.gov, Charles G. Crawford, cgcrawfo@usgs.gov, and Robert J. Gilliom, rgilliom@usgs.gov. U.S. Geological Survey, 5957 Lakeside Boulevard, Indianapolis, IN 46278
Empirical regression models have been developed by the U.S. Geological Survey (USGS) for estimating the concentration of atrazine and other pesticides in streams; the models are based on pesticide-use intensity and watershed characteristics. Separate models were developed to estimate time-weighted annual mean and selected percentile concentrations. The models were developed from USGS National Water-Quality Assessment Program monitoring data collected from more than 100 streams throughout the United States. Pesticide use in a watershed was the most-significant explanatory variable, but several hydrologic and soil parameters were useful in explaining the variability in observed pesticide concentrations. The potential influence of regional factors on model performance was evaluated through application of the models to about 62,000 Reach File 1 watersheds across the United States. Trends in pesticide occurrence in streams were investigated through use of the models with 1992, 1997, and 2002 data.