The cropping systems mosaic: How does the hidden heterogeneity of agricultural landscapes drive arthropod populations?

► Heterogeneity of crop and farming practice mosaic is rarely considered in landscape ecology. ► Spatio-temporal patterns of cropping systems emerge from farmers’ decisions. ► The cropping system mosaic drives arthropods population dynamics and persistence. ► Different approaches can be used to map this hidden functional heterogeneity. ► Biodiversity could be managed by designing the cropping system mosaic. Landscape ecology has emphasized the key role of spatial heterogeneity as a driver of ecological processes, which has led to a shift from a patch-matrix to a landscape mosaic representation of landscapes. However, in agricultural landscapes, the role of heterogeneity for biodiversity is often studied by considering (semi-) natural elements regardless of the functional heterogeneity of the cultivated mosaic. This mosaic, which emerges from the organization of cropping systems by farmers, is not just heterogeneous in space but also strongly dynamic (i.e. temporally heterogeneous) due to crop phenology, management and sequence. In this paper, based on agronomic and ecological literature we highlight the heterogeneous properties of the cropping system mosaics in space and time, and the implications of this “hidden” heterogeneity for controlling of arthropod population dynamics and persistence. An illustration is then given to propose different methodological approaches to map the hidden heterogeneity of agricultural landscapes. This conceptual and methodological framework concerns various species in agro-ecosystems, including crop and multi-habitat species, as well as species that only disperse through the crop mosaic. This paper suggests that the design of the cropping system mosaics could be complementary to the design of natural areas to manage populations, notably for species of economic interest (pests, natural enemies or pollinators). More empirical work on this topic is crucial, which will require new methodological approaches.