DEVELOPMENT AND FIELD VALIDATION OF THE PESTFATE MODEL IN SOUTHERN ONTARIO

Abstract

A new pesticide movement model called PESTFATE (PESTicide Fate and Transport in Environment) has beendeveloped by combining DRAINMOD, a well-known water table management model, and the pesticide submodel of PESTFADE.The pesticide sorption in the new model is based on two different techniques, namely, conventional mechanism anda new two-stage sorption method called Gamble kinetics. The model was validated by comparing the simulated midspan watertable depths and atrazine [2-chloro-4-ethylamino-6-isopr opylamino-1, 3, 5-triazine] concentrations against an independentlycollected dataset from a research site in southern Ontario. The experimental field consisted of 16 plots with twodifferent water table management and four different tillage practices, replicated twice. Only the plots with conventional tillageand controlled drainage were used in this study. The model performed well in predicting the daily water table depths.Although the intercept and slope of the regression between the observed and predicted water table depths were significantlydifferent from 0 and 1, respectively, the model efficiencies for 1992, 1993, and 1994 were 28%, 81%, and 64%, respectively,which shows a better accuracy for the last two years of the study. The normalized standard errors were within 15% to 20%,which indicates good model performance for water flow simulation. However, the correlation between the observed and simulatedpesticide concentrations with both the conventional and Gamble kinetics was poor. Moreover, the Gamble kinetics didnot significantly improve pesticide simulations (P < 0.05) as compared to the conventional method. The experiments wereconducted on Brookston clay loam soil, which is known for developing soil cracks and consequently preferential flow. Thiscould be one of the reasons for poorer model performance, especially with the Gamble kinetics. This was not the case in previousstudies, which reported better simulation results with the new sorption mechanism, so the model requires additional fieldtesting before any concrete conclusions can be drawn about its performance. There is also a need to test the model with otherpesticides.