A multi-scale method to assess pesticide contamination risks in agricultural watersheds

• We implemented a spatial cognitive model to assess surface water contamination risk.
• The model is built with a GIS at pixel level of remote sensing.
• Variables are slope, soil type, distance to the stream and agricultural pressure.
• A map is obtained for each variable which is considered as an indicator.
• Pixels used as Reference Spatial Object can be aggregated to many managing scales.

The protection of water is now a major priority for environmental managers, especially around drinking pumping stations. In view of the new challenges facing water agencies, we developed a method designed to support their public policy decision-making, at a variety of different spatial scales. In this paper, we present this new spatial method, using remote sensing and a GIS, designed to determine the contamination risk due to agricultural inputs, such as pesticides. The originality of this method lies in the use of a very detailed spatial object, the RSO (Reference Spatial Object), which can be aggregated to many working and managing scales. This has been achieved thanks to the pixel size of the remote sensing, with a grid resolution of 30 m × 30 m in our application.The method – called PHYTOPIXAL – is based on a combination of indicators relating to the environmental vulnerability of the surface water environment (slope, soil type and distance to the stream) and the agricultural pressure (land use and practices of the farmers). The combination of these indicators for each pixel provides the contamination risk. The scoring of variables was implemented according knowledge in literature and of experts.This method is used to target specific agricultural input transfer risks. The risk values are first calculated for each pixel. After this initial calculation, the data are then aggregated for decision makers, according to the most suitable levels of organisation. These data are based on an average value for the watershed areas.In this paper we detail an application of the method to an area in the hills of Southwest France. We show the pesticide contamination risk by in areas with different sized watersheds, ranging from 2 km2 to 7000 km2, in which stream water is collected for consumption by humans and animals. The results were recently used by the regional water agency to determine the protection zoning for a large pumping station. Measures were then proposed to farmers with a view to improving their practices.The method can be extrapolated to different other areas to preserve or restore the surface water.

Figure optionsDownload as PowerPoint slide