Testing two simple pesticide runoff models in Northern Australian agriculture

The Paddock to Reef Integrated Monitoring, Modelling and Reporting program (P2R) (Department of Premier and Cabinet, 2009) is investigating ways to improve agricultural management practices and the quality of water entering the Great Barrier Reef. One of the major risks to GBR water quality is pesticide runoff from agricultural industries. Direct measurement of pesticide fate and behavior is both cost and time intensive. Modelling therefore offers a logical way forward in assessment of improved management scenario to reduce pesticide runoff to the GBR. Modelling also confers the advantage that it can extrapolate from a relatively small experimental database, to a wider range of conditions and years. The ‘Simplified Formula for Indirect Loading caused by runoff’ (SFIL) from The Organization for Economic Cooperation and Development (OECD) and the one-dimensional simulation model HowLeaky? are both paddock scale models used for the prediction of pesticide runoff. We tested the applicability of these models to predict paddock scale (edge-of-field) pesticide runoff against experimental data. The experimental data was from two North-East Australian sites with three kinds of cotton management, ‘conventional’, cotton grown in wheat stubble and Polyacrylamide (PAM) added to irrigation water. The simulations were carried out for the eight most frequently detected pesticides. The SFIL-OECD modelled total pesticide loads in runoff were 1.4 to 2.8 times lower than measured loads for various pesticides. The HowLeaky? modelled total pesticide load was 5-23% higher than the measured load using parameter values from the literature. HowLeaky? was able to model similar trends in pesticide concentration compared to measured data following small adjustments to the most sensitive parameters. A sensitivity analysis indicated that the most significant model parameter for the SFIL model was pesticide soil half-life, while for the HowLeaky? model soil half-life, runoff curve number and application rate or efficiency were most significant. The HowLeaky? model was then used to compare pesticide runoff estimates for ‘old’ (pre-genetically modified cotton) and ‘new’ pesticides. The rationale for this was that pesticide usage patterns changed significantly with the introduction of genetically modified cotton, particularly with replacement of soil residual herbicides by knockdown herbicides and insecticides with BT cotton. Therefore the pesticide runoff profile has changed accordingly. Comparing the old and new pesticide runoff profiles suggests that use of new pesticides will result in less pesticide runoff in the GBR. This work demonstrates the value of models for predicting the environmental fate of pesticides and thus accessing the amount of pesticide entering streams. Future work will investigate the ecotoxicological implications of a shift in application patterns towards the ‘new’ pesticides.