Unexpected predominance of wine grower location over soil trafficability for vine management in southern France

International audience

إنجليزي. Abstract On a regional scale, many management concerns, such as agricultural planning and water contaminationmanagement, require information regarding the diversity and location of agricultural practices. One means of obtaining this information is to search for spatially explicit indicators that correspond to factors that drive agricultural practices. The objective of our study was to assess the role of soil trafficability on the distribution of soil surface management practices in a Mediterranean vine-growing region characterised by water contamination by herbicides. Soil surface management relies on different combinations of technologies such as shallow tillage, chemical weeding, and grass cover. We hypothesised that soils characterised by low trafficability were associated with the use of chemical weeding or grass cover in alleys that tractors use while spraying pesticides. Here, data on practices were collected by survey at the plot resolution. Soil trafficability was evaluated based on an expert classification of the soil units of a 1:25,000 soil map. Using classification trees, we tested the ability of the following three explanatory variables to determine the choice of practice: (1) the trafficability of the plot, (2) the percentage of plots with low trafficability within the vineyard of the farm holding, and (3) the wine grower residence place. Our results show that the trafficability classifies 59% of plots. The percentage of plots with low trafficability classifies 76% of plots. The wine grower residence place classifies 84% of the plots. Although the choice of practice correlated with soil trafficability, the residence place of the wine grower unexpectedly overdetermined the practice choice. As a consequence, our findings evidence a spatial variability of the role of soil in the distribution of soil surface management practices. In addition, soil trafficability cannot be used as major indicator of the practice spatial distribution.