BACKGROUND: Spray trials were conducted to determine the variation in primary spray drift deposition between trials under very similar conditions, in order to assess the feasibility of developing a computational tool to aid post-event investigations of pesticide spray incidents. Pesticide deposition was examined by analysis of filter paper and vegetation samples. RESULTS: Considerable variation in the drift profile was found. The overall estimate of the spray drift decay term was -1.13 (95% confidence interval -1.02 to -1.24), with statistically significant differences between plots. Variation in the drift profile between neighbouring essentially identical plots indicates the variation in deposition that might be expected over small distances. Vegetation samples were found to have considerably lower capture efficiency than filter papers. Importantly, degradation of pesticides was found to have little effect on the pesticide drift profile over a 14 day period. CONCLUSIONS: The levels of spatial variation in spray drift deposits between runs and plots observed in this study suggest serious limitations to the inferences that may be drawn from limited numbers of post-incident samples. In particular, they would limit inferences about the spray conditions that could be drawn from an estimate of the drift profile derived from limited post-incident samples.
BACKGROUND: Spray trials were conducted to determine the variation in primary spray drift deposition between trials under very similar conditions, in order to assess the feasibility of developing a computational tool to aid post-event investigations of pesticide spray incidents. Pesticide deposition was examined by analysis of filter paper and vegetation samples. RESULTS: Considerable variation in the drift profile was found. The overall estimate of the spray drift decay term was -1.13 (95% confidence interval -1.02 to -1.24), with statistically significant differences between plots. Variation in the drift profile between neighbouring essentially identical plots indicates the variation in deposition that might be expected over small distances. Vegetation samples were found to have considerably lower capture efficiency than filter papers. Importantly, degradation of pesticides was found to have little effect on the pesticide drift profile over a 14 day period. CONCLUSIONS: The levels of spatial variation in spray drift deposits between runs and plots observed in this study suggest serious limitations to the inferences that may be drawn from limited numbers of post-incident samples. In particular, they would limit inferences about the spray conditions that could be drawn from an estimate of the drift profile derived from limited post-incident samples.