Accurate flood mapping in arid regions using Sentinel-1 Synthetic Aperture Radar

Kurzfassung

Floods are among the most frequent and costliest natural disasters. Satellite remote sensing offers a cost-effective and widely adopted method for near real-time flood monitoring. Over the past decade, Sentinel-1 Synthetic Aperture Radar (SAR) imagery has emerged as a valuable tool in operational flood management, overcoming the challenges posed by optical sensors. SAR is an active imaging technique that provides cloud-free images day and night by utilizing specular reflection from smooth water surfaces. However, flood mapping using SAR is not universally effective. In areas with contrasting radar backscatter between water and non-water surfaces, SAR can distinguish floods, but this advantage is lost where the backscatter of landcover is similar to that of water, such as in dry sandy regions. Since dry sand appears similar to water, flood mapping in these areas becomes nearly impossible using amplitude data alone, a long-standing problem in the flood mapping community. The CEMS Global Flood Monitoring System (GFM) currently excludes arid regions from processing to avoid misclassification, highlighting a significant challenge. We conducted a detailed investigation using three different arid floods in three countries—Turkmenistan, Pakistan, and Iran—to understand how to best utilize SAR (Sentinel-1) information to address this challenge. Through multiple case studies, we employed the random forest method to train, test, and validate our model predictions against flood masks derived from cloud-free optical imagery. We designed several scenarios to investigate the contribution of different information layers in improving flood mapping accuracy in arid regions, along with feature importance analysis to understand the role of each feature in reducing model complexity. Our results demonstrate the effectiveness of fusing amplitude and coherence information in flood mapping compared to using coherence or amplitude alone. By utilizing key features derived through permutation feature importance, flood mapping accuracy improved by approximately 50% and reduced response time, which is crucial for effective emergency management. The findings highlight the potential of the proposed approach across different sensors and scenes, offering significant promise for global flood mapping in arid regions, particularly in countries with limited resources. As future missions and advancements in SAR systems evolve, flood detection capabilities will further improve, enhancing flood monitoring in arid areas.