Answers are grounded in this paper and cite their sources.
Accurate mapping of winter catch crops is critical for monitoring sustainable agricultural practices and supporting climate mitigation policies. However, large-scale information on their spatial extent and crop type distribution remains limited. This study evaluates winter catch crop classification using multi-temporal Sentinel-1 (S-1) SAR and Sentinel-2 (S-2) optical data from 2021–2024 across four German regions: Lower Saxony (LS), Brandenburg (BB), North Rhine-Westphalia (NRW), and Saxony (SN). We assessed (i) the relative contribution of SAR, optical, and combined datasets for accurate classification, (ii) temporal transferability to an independent 2024 dataset, and (iii) spatial transferability to an independent region (SN). Results show that the combined SAR–optical model achieved a balanced overall accuracy (OA) of 90% for distinguishing catch crops from non-catch crops and demonstrated strong temporal transferability, with an OA of 86% in 2024. When transferred spatially to SN (2023), the model achieved an OA of 77%. Similarly, for individual catch crop type classification, the combined SAR–optical model achieved the highest overall accuracy of 84%, with notable improvements for minority classes such as niger, vetch, and phacelia. Temporal validation on 2024 data resulted in OA of 76%, while spatial transfer to SN achieved 63%. Persistent confusion among radish, mustard, and mixture was observed, largely due to overlapping phenological traits and shared structural characteristics. These findings confirms the complementary value of SAR and optical observations and provide a transferable framework for large-scale monitoring of winter catch crops in Germany.
No takes yet. Share an insight, caveat, or question.
Selvaraj et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: