Mapping farming practices in Belgian intensive cropping systems from Sentinel-1 SAR times-series

The environmental impact of the so-called conventional farming system calls for new farming practices reducing negative externalities. Emerging farming practices such as no-till and new inter-cropping management are promising tracks. The development of methods to characterize crop management across an entire region and to understand their spatial dimension offers opportunities to accompany the transition towards a more sustainable agriculture. This research takes advantage of the unmatched polarimetric and temporal resolutions of Sentinel-1 SAR Cband to develop a method to identify farming practices at the parcel level. To this end, the detection of changes in backscattering due to surface roughness modification (tillage, inter-crop cover destruction ...) is used to detect the farming management. The final results are compared to a reference dataset collected through an intensive field campaign. Finally, the performances are discussed in the perspective of practices monitoring of cropping systems through remote sensing.