You are here:
USE OF WEIBULL FUNCTION FOR NON-LINEAR ANALYSIS OF EFFECTS OF LOW LEVELS OF SIMULATED HERBICIDE DRIFT ON PLANTS
Citation:
OLSZYK, D. M., E. LEE, AND T. G. PFLEEGER. USE OF WEIBULL FUNCTION FOR NON-LINEAR ANALYSIS OF EFFECTS OF LOW LEVELS OF SIMULATED HERBICIDE DRIFT ON PLANTS. Presented at Society of Environmental Toxicology and Chemistry, Milwaukee, WI, November 11 - 15, 2007.
Impact/Purpose:
Presentation
Description:
We compared two regression models, which are based on the Weibull and probit functions, for the analysis of pesticide toxicity data from laboratory studies on Illinois crop and native plant species. Both mathematical models are continuous, differentiable, strictly positive, and monotonically decreasing to zero with increasing rates of pesticide application on a log scale. The Weibull model was reparametrized so that one of its parameters was equal to EC25 (effective concentration producing a 25% reduction in dry weight response). Similarly, the probit model was reparametrized so that its inflection point corresponds to log(EC25). Unlike the Weibull model, the probit model is derived from the cumulative normal distribution and assumes that the absolute difference in plant response is the same for a unit increase or decrease in the pesticide application rate about log(EC25). We compared EC25 values for dry weight for a variety of plant species treated with herbicides at below field application rates (f.a.r.) using both the Weibull and probit models. In cases where there were significant reductions in dry weight at the lowest level of herbicide used (0.001 x f.a.r.), the probit model would either not fit the data or would result in a poorer fit. The probit model did not fit the data as well as the Weibull model because its constraint of symmetry about the inflection point was not a feature of the data. Thus, the Weibull model provides a better alternative function to evaluate the risks associated with new pesticides prior to their release into the environment.