Review Article

Remote Sensing Applications for Flood Control and Stormwater Management: A Systematic Review

Abstract

Flooding is a major and persistent risk to cities and riverine communities worldwide. We conducted a systematic review to assess how remote sensing is being applied in flood control and stormwater management in operational contexts. Following PRISMA 2020 guidelines, we screened studies published between 2000 and 2024 and identified 20 empirically rigorous articles that reported quantitative validation metrics using satellite, airborne, or UAV-based data. Across these studies, synthetic aperture radar (SAR), particularly Sentinel-1, emerged as the most reliable platform for all-weather flood extent mapping, with typical accuracies above 85%. Optical sensors such as Landsat and Sentinel-2 performed well under cloud-free conditions and were especially useful for surface water detection, vegetation dynamics, and monitoring green stormwater infrastructure through indices such as NDVI. LiDAR provided the most precise topographic information, generating centimeter-level digital elevation models that significantly improved urban flood modeling. UAV platforms offered ultra-high spatial resolution (< 10 cm), supporting localized damage assessment, infrastructure inspection, and post-event validation. We found that machine learning approaches, including convolutional neural networks and vision transformers, consistently outperformed traditional threshold-based methods, often achieving F1-scores above 0.90. However, challenges remain, particularly limited validation data, weak uncertainty quantification, scale mismatches, and difficulties with operational integration. Overall, our review shows that remote sensing has become a practical and indispensable tool for flood and stormwater management when sensor platforms are strategically combined.