DEFINING ACCEPTABLE LEVELS OF HERBICIDE DEPOSIT VARIATION FROM AERIAL SPRAYING

The distance between flight lines (lane separation) selected for aerial application of herbicides influences both spray deposit variation (expressed as the coefficient of variation or CV) and efficacy of weed control. Increasing lane separation increases productivity, thereby reducing costs, but at the same time it generally causes an increase in CV. A database of 33 field trial treatments was used to model the relationship between CV and the biological response of weeds to variable doses. Although previous field studies indicated that a lognormal distribution provided a reasonable approximation of deposit variation, the more extensive database analyzed in this study showed that this distribution is only appropriate when the variation is high. It was found that deposit distributions could be successfully transformed to normality using a power transformation where the power is a linear function of the CV. These results also indicated the importance of knowing the distribution of spray deposits when calculating biological consequences of deposit variation. Deposit distributions simulated using SpraySafe Manager (SSM), an aerial application decision support system that links models of herbicide deposition to herbicide/weed dose-response models, were compared with the field data. Results showed that normal procedures for defining lane separation to match a target CV using SSM may often underestimate actual CV because of random variation in factors such as release height, wind speed, and deviation from flight line. Although the assumption that there will be little effect on levels of weed control if CVs of deposition are maintained at levels of 30% or less appeared to be true at recommended application rates, CVs as high as 80% produced good levels of weed control. Especially for low CVs there appears to be scope for reducing recommended herbicide rates. Consequences of reducing herbicide rates are discussed and recommendations are made for additional research.