Résumé
One of the main aims of sustainable agriculture is to reduce the risk of disease in cropping systems and thereby contribute to preserving public health and conserving the environment. In particular, if we are to reduce our recourse to agrochemicals, we need to acquire a deeper understanding of ecological processes, for instance biological pest control by predatory arthropods such as Carabids. We analyzed the population dynamics of a carabid (
Pterostichus melanarius
) in various cropping systems (conventional, integrated, and organic) to improve our limited knowledge of its potential impact on insect populations. To do this, we built a model composed of three coupled modules, each referring to different assumptions: a phenology module of the carabid species was based on a periodic Leslie-matrix model and a 1:1 sex ratio, the population was spatialized in Brittany (Western France) using a spatially implicit landscape module that simulated the capacity of grass margins to provide shelter for this carabid beetle, and an agronomic module was used to assess the effects of different crop management systems on the beetle population. Our results showed that the spread of
P. melanarius
larvae by farming practices was key to species presence in crop fields. This was greatly dependent upon the farming calendar, the intensity of farming practices, and the presence of shelter zones. Maintaining and managing uncultivated habitats such as grass margins is thus as crucial for this beneficial predator as managing cultivated fields (asymptotic growth of the population was found to be 25% higher in integrated and organic management systems when 20% grass margins were present). At the field level, integrated management and direct drilling (with minimum disruption of the soil surface) are the best crop management systems for
P. melanarius
survival. Organic management is less effective (asymptotic growth of the population is only half that of integrated and direct-drilling management systems), for it requires multiple soil tilling to compensate for the absence of herbicides. Winter oilseed rape is the most detrimental crop (asymptotic growth of the population is half that seen with other crops), as it gives rise to summer molluscicides and soil tillage. This is the first study developing a numerical model of the long-term effects of landscape composition and cropping system on this beneficial predator species. This approach provides a powerful tool to test hypotheses and quantify their possible effects in agroecology.