Agron. Sustain. Dev. 25 (2005) 465-472
DOI: 10.1051/agro:2005046

Biases in the spatial estimation of pesticide loss to groundwater

Yves Coquet^a, Dalila Hadjar^a, Jean-Marc Gilliot^a, Marie-Paule Charnay^a, Julien Moeys^a, Alexandre Dufour^a and Nicolas Beaudoin<sup>b

a</sup>  UMR, INRA/INAPG, Environment and Arable Crops, BP 01, 78850 Thiverval-Grignon, France
^b  UR, INRA Agronomy, rue Fernand Christ, 02007 Laon Cedex, France

(Accepted 12 May 2005)

Abstract - Prediction of pesticide fate in soils is highly sensitive to parameters describing sorption and degradation processes, namely the K_oc partioning coefficient between the soil solution and organic carbon fraction and the half-life DT₅₀ for degradation. This paper explores the impact of getting K_oc and DT₅₀ values either from databases or from site-specific measurements on the predicted fate of atrazine, isoproturon and metamitron on the catchment scale. Pesticide fate on the scale of the Bruyères-et-Montbérault catchment, France, was predicted using the SEAMS software that couples a one-dimensional local-scale model of pesticide fate to a geographic information system. The results show that the use of database average values for K_oc and DT₅₀ underestimates the average risk of pesticide leaching calculated from site-specific K_oc and DT₅₀ values, whereas maximised risk scenarios based on extreme K_oc and DT₅₀ values may be overestimated when using database values. Whenever available, site-specific data should be preferred to limit bias in pesticide leaching risk assessments on the catchment scale.

Key words: pesticide fate / herbicide / spatial variability / soil / environmental risk assessment

Corresponding author: Yves Coquet coquet@grignon.inra.fr

© INRA, EDP Sciences 2005