Deconstructing Adsorption Variability: Investigating Controls on Pesticide Adsorption in Soil and Modelling Koc

Abstract

Due to potential environmental risks of pesticides, it is important that the fate of pesticides is known and that safer pesticides are developed in the future. This thesis focused on identifying controls on the Koc of pesticides in soil based on their structural parameters. This thesis also developed quantitative structure-activity relationship (QSAR) models to predict the environmental fate of new pesticides.

To understand the controls on Koc, a range of multivariate statistical techniques were used including; principal component analysis, and analysis of variance. Predictive models were created using logistic regression, and multiple linear regression.

The study found adsorption of pesticides in soil is controlled by a combination of size and solubility parameters. Logistic regression models were able to predict the adsorption potential of metabolites, relative to their parent based on metabolite structures. This study found that adsorption behaviour of pesticides was fairly specific to different chemical groups. A QSAR model for Koc was constructed for a group of early stage compounds and could predict Koc to just over an order of magnitude.

The results of this study have implications for the pesticide development process. If developed further to include a wider range of chemical groups then the models have the potential to reduce the dependence on laboratory tests in the early stages of the development process. However, this study also questions the use of Koc as a predictive parameter and offers alternative solutions to predicting environmental fate of pesticides.