A rapid, spatially explicit approach to describe cropping systems dynamics at the regional scale

Land managers need spatially explicit information about agricultural practices to address issues that arise from the use of natural resources in agriculture. One main characteristic of agriculture is its great variability in space and time. However, describing the spatial distribution of “cropping systems”, i.e. crop sequences and crop management systems at the regional scale, remains a major scientific challenge. This study presents a new, simple and rapid approach to model the spatial distribution of irrigation management practices. It was developed in two large watersheds in southwestern France (about 1500 and 3000 km2). Based on a previous study consisting of 27 farmer interviews in a study area about one-sixth the size of these watersheds, we interviewed 12 key informants who had an integrative vision of the study area and spent only one-fourth as much time collecting and processing the relevant data. One major innovation was to combine knowledge from generic databases and ad-hoc intermediate objects, such as diagrams, tables and maps, to interact with the key informants. These objects helped them focus on specific local information that we had missed and facilitated data processing. Interview results were used to spatially allocate cropping systems formalized as dynamic IF-THEN decision rules. We evaluated our approach by using a cropping system model to simulate irrigation withdrawals over a ten-year period. Its predictions reproduced well annual amounts and inter-annual dynamics of irrigation water withdrawals recorded by the regional Water Agency. This approach, combining diagrams with IF-THEN rules, appears easy to adapt to study other areas and agricultural practices besides irrigation, as well as to manage annual and perennial crops.