USING RZWQM TO PREDICT HERBICIDE LEACHING LOSSES IN SUBSURFACE DRAINAGE WATER

Abstract

Improvements have been made in the pesticide component of the Root Zone Water Quality Model (RZWQM) sinceits release in 1999 for the Management System Evaluation Areas (MSEA) project. This study was designed to evaluate theherbicide leaching component of the model using data on subsurface drainage flow and herbicide leaching losses for a 6-year(1992 to 1997) period. A sensitivity analysis was conducted for the key parameters important in the pesticide calibration process.The model was calibrated using 1992 data and validated using 1993 to 1997 data collected from a tile-drained fieldwithin the Walnut Creek watershed in central Iowa. The model evaluation criterion was based on percent difference betweenthe predicted and measured data (%D), root mean square error (RMSE), and model efficiency (EF). Atrazine and metolachlorwere applied to corn in 1993, 1995, and 1997, and metribuzin was used during the soybean growing seasons in 1992, 1994,and 1996 at the standard application rates used in Iowa. The predicted subsurface drainage volumes were in close agreementwith the measured data showing %D = 1, RMSE = 8, and EF = 0.99, when averaged over the validation years. Herbicidehalf-life (t1/2) and soil organic based partitioning coefficient (Koc) were found to be the most sensitive parameters forsimulating herbicide leaching losses in subsurface drainage water. Both t1/2 and Koc affected the mass and temporaldistribution of the herbicide leaching losses in subsurface drainage flows. The predicted herbicide leaching losses insubsurface drainage water were the same order of magnitude as the measured data, when averaged across the validationyears. The study also revealed that herbicide leaching losses were significantly (P < 0.05) controlled by the drainage volume(R2 = 0.97). The model, however, underpredicted herbicide leaching losses after crop harvest and during early spring,possibly because of preferential flow paths developed during these periods. More improvements may be needed in the RZWQMto consider the dynamics of the preferential flow paths development in cultivated soils similar to that of the study area.