Predicting the probabilities of groundwater contamination by pesticides under variable recharge

Abstract

A major deficiency in many models used to assess the groundwater contamination potential of organic pesticides used in agricultural and horticultural systems has been the assumption of a constant average groundwater recharge rate. This paper describes an enhanced version of a simple model designed to screen out or identify those pesticides that have a high probability for causing groundwater contamination in a region (e.g. a catchment, farm, or cropping area), which enables the temporal variability of recharge to be taken into account and illustrates the influence of a seasonal pattern of leaching and variations in predicted mobilities and persistence from those associated with averaging the recharge rate. The model is available as a user-friendly software package (PESTSCRN 3).

The model was formulated based on the following assumptions: (1) linear, equilibrium, and reversible sorption; (2) first-order breakdown or degradation; (3) pesticide leaching by steady convective flow; and (4) recharge rate varying with time. Random values of the input parameters required by the calculations are generated from probability distributions as specified by the means (m) and standard deviations (s), assuming normal distributions. Outputs provide a statistical analysis of travel time and fraction of pesticide remaining at different soil depths. Most distributions for travel time are near normal with slight positive skewness and can be approximated as normal distributions. The frequency distributions of residue fractions for all the pesticides examined to date show significant positive skewness (Beta distributions), with the peak frequency towards the lower bound of 0. Simulations using metalaxyl as an example demonstrate that, depending on recharge conditions, the use of daily data instead of mean recharge data can result in important differences in the predicted values of both travel times and residue percentages.