Eyes from Above: SAR-Based Remote Sensing for Flood and Landslide Risk – The Case That Shocked South Brazil in 2024

The hydrometeorological disaster that struck Rio Grande do Sul, Brazil, in 2024 resulted in floods and landslides of unprecedented magnitude, surpassing historical events recorded in 1941 and 2023. Among the most severely affected regions, the Baixo-Taquari-Antas Valley tood out due to its high population density, critical infrastructure concentration, and extensive socio-economic damage. This study developed an operational, multi-sensor workflow to rapidly delineate impacts in nine municipalities by fusing Sentinel-1 SAR and Sentinel-2 optical data within a DEM-constrained framework. Pre-/post-event median composites (pre: Dec-2020–Mar-2024; post: Apr–11 May 2024) were differenced to map backscatter changes, applying fixed thresholds (Δσ⁰ ≤ −5 dB for flood; Δσ⁰ ≥ +3 dB for landslides). Flood extent was further refined with NDWI from Sentinel-2, while landslide candidates were restricted to slopes > 15°. Object-based image analysis with SVM on optical scenes produced building footprints that were intersected with hazard layers to quantify exposure. Field surveys at 30 control points confirmed mapping accuracies of 92% (floods) and 85% (landslides). Results show ~ 174 km² of inundation and ~ 84.8 km² of landslides across the nine cities, with ~ 34,603 buildings affected by flooding and ~ 1,114 by slope failures. The integrated SAR–optical approach proved robust under cloud cover and heterogeneous urban fabrics, delivering actionable situational awareness for emergency response and early recovery. The results provided critical support for emergency response and post-disaster mitigation planning, demonstrating the operational value of integrated SAR–optical remote sensing for rapid disaster risk assessment in highly vulnerable regions.