An indicator of pesticide leaching risk to groundwater

Abstract

Since the 90s an increasing number of assessment methods using operational tools like indicators have been proposed for environmental issues linked to pesticides, among them, groundwater contamination by pesticide transfer. To our knowledge none of these indicators address preferential flow, an important process determining pesticide leaching. The objective of this study is twofold: (i) to develop a new groundwater sub indicator for an existing indicator, I-Phy (former Ipest), that explicitly take preferential flow into account, and (ii) to test the possibility of developing an indicator by means of data-mining methods using simulations of a mechanistic model. The groundwater sub indicator developed is in the form of decision trees based on fuzzy inference systems. It was derived through neuro-adaptive learning on data sets from simulations running the process-based MACRO model. Unlike the previous version, the new indicator considers preferential flow, climatic differences and differences in soil texture with depth. Other benefits are less dependency on expert knowledge and the possibility to integrate a broad range of conditions.