The pesticide sub-model of CREAMS was tested to describe pesticides transfer into runoff generated by a rainfall simulation. Adsorption coefficient, Kd, was found to be a sensitive parameter of the model. So, we developed an experimental method in order to measure Kd in a dynamic way. We studied four pesticides (atrazine, simazine, lindane, methidathion) and three soils (loam sand, sand loam, calcareous silt loam). The values of Kd obtained were compared to the values measured on the soils by the standardized method of batches and to the values of the literature. Mean Kd measured in the runoff of a rainfall simulation were considerably greater, especially for low suspended particles loads. We ran the model with these different data for Kd and we found that predictions were more accurate with Kd obtained by rainfall simulation particularly for lindane and simazine. We improved the modelisation for atrazine and methidathion. Nevertheless, the relative error between observed and predicted concentrations was not the same according to the pesticide, which made prediction of risk dependent on pesticides classes of behaviour.

You do not currently have access to this content.