Modelling pesticides transfer to surface water at the catchment scale: a multi-criteria analysis

The demand for operational tools at a catchment scale is growing to assess both the sustainability of agricultural practices and the efficiency of mitigation measures on pesticide transfer to surface water. Here a literature review of 286 investigations highlights the large number of indicators and hydrochemical models developed at the catchment scale. Given this large number of indicators and models, the choice is difficult for potential users. Therefore, this article proposes a multi-criteria analysis applied to ten existing tools including physically based and conceptual models, indicators and multi-agent systems. We found the following major points: (1) Indicators and conceptual models are the most popular approaches to assess the transfer of pesticides to surface water at the catchment scale due to a trade-off between environmental relevance and adaptation to user's needs. (2) The latest indicators developed are inferred from the results of conceptual or physically based models to combine the strengths of each approach. (3) Only a handful of physically based models have addressed both flow and pesticide transport at the catchment as affected by the internal heterogeneity of the system. However, it is only physically based models that can simulate the impact of changes to the catchment. Physically based models integrate feedbacks between hydrological and chemical processes not possible from conceptual models or indicators alone. (4) The ability of models to assess the pesticide loads both in the dissolved and particulate phases is a key issue not properly addressed by many indicators or models. A key way forward is the integration of erosion processes with the fate of pesticide adsorbed to these particles. (5) At the catchment, the hydrological connectivity is perhaps the primary hydrological variable required to correctly assess rapid flow processes as surface runoff and associated pesticide transfer. This in turn implies using tools that explicitly represent the connectedness of surface and/or sub-surface water pathways including mitigation measures to correctly assess the risk of pesticide transfer.