Using a convolution integral model for assessing pesticide dissipation time at the end of catchments in the Great Barrier Reef Australia

Abstract

A model based on convolution integrals derived from a pesticide application function and the firstorder kinetic decay function (Cook et al. 2011a) was applied to pesticide monitoring data collected from end-of-system (EOS) sites, i.e. on the catchment scale above the tidal zone. The convolution model, at the catchment scale, is based on the summation of pulse inputs over an application period within a catchment area, followed by a period of concentration decay. This model has previously been successfully applied at the block, multi-block and multi-farm scales to predict the concentrations of pesticide lost from runoff. Here we investigate the applicability of the model to determine pesticide concentrations at the catchment scale. The model was applied to atrazine and diuron concentration data collected over the 2010-2011 wet season from three EOS sites (Barratta Creek, Pioneer River and Sandy Creek) that discharge into the Great Barrier Reef lagoon. Temporal trends observed in atrazine and diuron concentrations, fitted the convolution integral model for these catchments. Multiple 'decay events' were observed at each catchment, indicating periods of reapplication throughout the wet season. For each 'decay event', the dissipation half-life (d 1/2) was estimated as well as a global d 1/2 representative of the whole wet season. The results indicated that the dissipation half-life of atrazine and diuron at the catchment scale was much shorter than what has previously been observed at the paddock and farm scale. The results presented here will be of value to pesticide runoff models that use an up-scaling approach from paddock scale point models to catchment scale models.