Stochastic Simulation of Field‐Scale Pesticide Transport Using Opus and GLEAMS

Abstract

AbstractIncorporating variability in soil and chemical properties into root zone leaching models should provide a better representation of pollutant distribution in natural field conditions. Our objective was to determine if a more mechanistic rate‐based model (Opus) would predict soil water and pesticide mass in the soil profile more accurately than a capacity‐based model (GLEAMS) when spatial variability and uncertainly in parameters are considered. Predictions of spatial variations of soil water content and movement of aldicarb [2‐methyl‐2‐(methylthio)‐propionaldehyde O‐(methylcarbamoyl) oxime] and metolachlor [2‐chloro‐N‐(2‐methoxy‐1‐methylethyl) acetamide] in the root zone were compared using 3 yr of observed data from a 3.9‐ha agricultural field in southwest Georgia. Spatial variability of soil physical properties, pesticide properties, and pesticide application were described using probability distributions fitted to measured field data, after removing spatial trends that were physically meaningful. There were significant differences in mean soil water content predicted by the two models, although variations around the mean were comparable. Pesticide mass predictions were different on most post‐application dates in both mean and spatial variation. The less rigorous GLEAMS predicted mean depth‐averaged soil water content and pesticide mass in the 1.2‐m profile at least as good as the more mechanistic Opus, although it did not simulate depth distributions of water or pesticide mass as well as Opus. GLEAMS simulated spatial variations of depth‐averaged soil water content and pesticide mass in the field with reasonable accuracy while employing fewer parameters that exhibit lower spatial variability.