A GIS-based method to define geographical determinants of environmental exposure to agricultural pesticides in France

Introduction Exposure to pesticides has been suggested as a risk factor for several diseases or adverse health outcomes. The use of geographic information systems (GIS) in epidemiological studies to y to assess environmental exposure to agricultural pesticides (EEAP) is growing. Residential proximity to agricultural crops is frequently used as a surrogate of EEAP. Previous studies have suggested that application of pesticides near residences results in contamination of indoor surfaces. US studies have recently reported a positive correlation between the proximity or crop acreage and the contamination of indoor dust by agricultural pesticides in nearby homes. There is however no validated and applicable model in France: the US model is based on pesticide use registries and adapted on large crops and local practices. However, most of these studies did not consider spatial information on environmental factors such as wind direction and physical barriers (structural or natural) to pesticide drift from the field to receptors. In France, the availability of public data on these aspects allows to describe the relationship between the proximity of agricultural areas and residential exposure to pesticides. We aimed to assess geographical determinants of the indoor dust contamination by agricultural pesticides in different crop areas (SIGEXPO project). Methods A GIS was developed for 239 households, from three different agricultural sectors (69 near orchards and cereals, 66 near cereals, 68 near vineyards) and an urban area ( n  = 36). For each residence, the GIS considers the agricultural land use, the proximity of household to agricultural zones (fields). For five different buffer sizes (250 m–1250 m), we used GIS to determine targeted crop acreage, prevailing winds, and vegetative, topographic and structural barriers to pesticides drift from crops. Two additional dimensions have been considered in the development of the GIS: prevailing winds direction that influence pesticide drift and obstacles that mitigate on pesticide drift from agricultural applications, i.e. the presence of topographical, vegetal (hedges, forests) or structural barriers (built areas). The effect of obstacles is reflected in the presence of topographical, vegetal (hedges, forests) or structural barriers (built areas). In parallel, both recent ( 6 month) indoor dust samples were collected from households without professional pesticide users and analyzed. Pesticide applicators were excluded from the study. Information household characteristics and domestic pesticide users were collected by questionnaire. Redundancy analyzes were conducted to define the geographical determinants that best explain the indoor dust concentrations of the 27 agricultural pesticides detected in more than 10% of study households. Results Overall, main determinants of agricultural pesticides contamination in indoor dust were crop acreage within 500 m (orchards) or 1000 m buffer size (vineyards and cereals), as well as prevailing winds, and the presence of vegetative barriers. Our different models explained up to 18.3% of the variability of the pesticide contamination in indoor dust. Conclusion While the explained variability remains modes overall, it was above or comparable to previous studies. The determinants identified should be taken into account in future GIS-based approaches aiming to assess EEAP in the French context. The approach developed to assess the impact of wind and barriers need to be validated on additional datasets. Future studies should include the analyses of the impact of seasonal variations into the geographical determinants selection.