Comparison of PRZM and GLEAMS computer model predictions with field data for fluometuron and norflurazon behavior in soil

The ability of the pesticide root zone model (PRZM) and the groundwater-loa ding effects of agricultural management systems (GLEAMS) model to predict m ovement of two herbicides in soil was evaluated using site-specific environ mental data from sites in three states. Predictions of herbicide movement w ith site-specific data were compared to predictions using more generalized database soil and pesticide data within each model. Field experiments exami ned fluometuron and norflurazon movement in three soils representative of t he cotton-growing regions of the southeastern United States. In comparing t he use of site-specific vs. database values, the small increase in accuracy using site-specific inputs would not justify the large cost to obtain the data. The databases for each model gave predictions similar to those using the site-specific numbers. Both the PRZM and the GLEAMS model had similar a ccuracy levels in predicting the presence of fluometuron or norflurazon pre sent in three surface soils, although each model tended to overpredict move ment and total herbicide concentration, especially at lower herbicide conce ntrations. At higher herbicide concentrations, prediction accuracy was less than that probably needed to predict agronomically relevant herbicide conc entrations in surface soils.