Flood mapping and monitoring are common emergency response applications using Synthetic Aperture Radar (SAR) data, particularly Sentinel-1, which provides regular observations at a global scale. Nevertheless, local-scale flood mapping and damage proxy maps are typically performed using high-resolution satellite data, and most of the studies and the present operational services, which use medium-resolution Sentinel-1 data, primarily focus on regional-scale floods using SAR intensity data. This study uses Sentinel-1 data and proposes a methodology for improved mapping of inundation and a coherence-based damage proxy map associated with local-scale floods using both incoherent and coherent change detection techniques. We analyse incoherent changes using Ground Range Detected (GRD) and coherent changes using interferometric coherence images estimated from Single Look Complex (SLC) Sentinel-1 data. The method is applied to detect regular flood events between 2018 and 2024 and assess the flood maps and coherence-based damage proxy map in Afghanistan. We examined six local flood events and validated the results in five of them against high-resolution PlanetScope imagery. Incoherent change detection identified floods in irrigated agriculture and bare lands with F1 scores ranging from 79% to 83%. In contrast, coherent change detection reveals flood extent and coherence-based damage proxy map in built-up areas with F1 scores ranging from 69% to 73%. In comparison with global flood products from the Global Flood Monitoring (GFM) system, our results identified approximately 5 times more inundation areas in some of our study areas, which highlights the effectiveness of the proposed methodology for near-real-time local-scale flood mapping, supporting timely disaster management and recovery planning.
Hotaki et al. (Sun,) studied this question.