Pesticide spray drift monitoring in the evaluation of air dispersion models : a South African atrazine case study

Abstract

Air dispersion software models that evaluate pesticide spray drift during application have been developed. These models can potentially serve as a cheaper and more convenient alternative to field monitoring campaigns. Such models require validation against field monitoring data in order for them to be employed with confidence, especially when they are used to implement regulatory measures or to evaluate potential human exposure levels. In this project, a pesticide active ingredient, namely atrazine, was used as a tracer to monitor spray drift up to 400 m downwind for comparison and validation of the AGricultural DISPersal (AGDISP) model outputs. Airborne drift samples were collected using high volume air sampling onto polyurethane foam (PUF) at six downwind locations while ground deposition drift was captured with chromatography fallout paper samplers. Additional data, including meteorological information and some application parameters required to simulate spray drift with AGDISP, was collected. Airborne samples were extracted with a plunger method using a hexane:acetone solvent mixture and analysed by Gas Chromatography coupled to a Nitrogen Phosphorus Detector (GC-NPD) which performed well (94.5% recovery, 3.3% RSD and LOD 8.7 pg). Atrazine airborne concentrations ranged from 4.55 ng L-1 adjacent to the field to 186 pg L-1 at 400 m downwind. The experimental results correlated favourably with the modelled output, suggesting that the AGDISP model can be used to provide a good estimate for airborne drift in risk assessment studies or for regulatory purposes. A simple and rapid screening method using a Direct Sample Analyser coupled to a Time-of-Flight Mass Spectrometer (DSA-TOFMS) was employed for semi-quantitation of atrazine deposition. This method was shown to be quick (30 min extraction and 25 s analysis) and useful for the large sample set that was collected. The deposition sample extracts were also analysed by GC-NPD using a method similar to that used for airborne samples. Compared to the AGDISP-simulated deposition output, the model under-predicted the deposition by up to one order of magnitude compared to the GC-NPD results and even more compared to the DSA-TOFMS results. This suggested that the model should be used cautiously for predicting pesticide deposition. For the first time this project has shown the use of a pesticide active ingredient to validate the AGDISP ground application model under local South African conditions up to 400 m downwind of the application area.