Validation of the PESTLA model: Definitions, objectives and procedure

Abstract

The simulation model PESTLA was developed to produce estimates of accumulation and leaching of pesticides in soil to facilitate classification of pesticides in the Dutch registration procedure. Before PESTLA can be used for quantitative assessment of expected pesticide concentrations in groundwater, its accuracy to predict field-scale pesticide leaching needs to be assessed. The project "Validation of PESTLA" aims at evaluation of the applicability of PESTLA for the current assessment procedure for pesticide registration in the Netherlands. The present report is the first in a series of reports and provides a short overview of the history of PESTLA. The objectives of the project are outlined, and some background information on the process of model development is given. The specific tasks of the project are defined in this context. A list of definitions of the terms used during the project is included. Several methods are described with which model predictions can be compared with (field) measurements of pesticide leaching. This is preceded by a discussion on statistical hypotheses, which is needed to understand the Chapters that follow. The so-called factor-of-f approach is selected as the tool to be used in the evaluation of PESTLA performance. The factor-of-f approach acknowledges that both model predictions and field measurements usually cover a range of possible values instead of one fixed value. This implies that comparison of the two must be done by statistical methods. The factor-of-f approach allows uncertainty of a size f around the predicted value. It concludes the model to be 'acceptable' when a range of measured values fall within the range of model predictions as defined by the size of the factor f. Finally, the overall approach of the project is outlined in terms of several selected activities that intend to lead to validation of PESTLA.<br>