Modeling pesticide dynamics of four different sites using the model system SIMULAT

This study aimed to assess the accuracy of SIMULAT, a computer model primarily designed for predicting the fate of pesticides in soil. The evaluation was carried out by comparing simulated results on herbicide degradation with results obtained from ®eld and lysimeters experiments. For model validation four different data sets were available. The data sets included ®eld, lysimeter, and laboratory experiments from Germany (Weiherbach), The Netherlands (Vredepeel), Great Britain (Brimstone), and Italy (Tor Mancina). The applied herbicides and the determined soil and water parameters varied substantially among the four empirical data sets used for model evaluation. In a ®rst step simulations were run with the model still being uncalibrated. Afterwards a calibration of hydraulic parameters was performed using measured water and bromide contents in soil (Weiherbach, Vredepeel and Brimstone) or leachate (Tor Mancina). In contrast to the hydraulic parameters, the sorption and degradation parameters were not calibrated. Simulations were run with the calibrated model and the results compared with those obtained from ®eld measurements. The wide range of implemented boundary conditions such as lysimeter, free drainage or ̄uctuating groundwater table enabled applying SIMULAT to all four data sets. However, usage of parameters obtained in laboratory experiments gave no satis®able simulation of the degradation of the herbicides. In contrast to the macroporous loam soil at Tor Mancina, water and bromide transport were accurately simulated in the loess (Weiherbach) and in the sandy soil (Vredepeel). Severe problems occur while simulating the ̄uctuating groundwater and drain ̄ow in the clay soil at Brimstone. # 2000 Elsevier Science B.V. All rights reserved.