Comparative Performance of Sentinel-1 SAR Polarization and Orbit Configurations for Flood Detection in Aceh Tamiang, Indonesia

Abstract

Sentinel-1 SAR imagery has potential for rapid flood mapping in tropical regions because it can see through clouds and operates day and night. Nevertheless, knowledge gaps remain in using different polarizations and SAR data orbits for flood monitoring. This study analyzes the outcomes of Sentinel-1 SAR polarizations and orbits for flood detection in Aceh Tamiang, Indonesia, after the Senyar Cyclone of November 2025. The six scenarios considered used single and dual polarizations (VV+VH) with both ascending and descending orbits. After pre-processing the SAR data in Google Earth Engine with speckle filtering and a classification threshold of 1.1 to define flooded areas, the classification's validity was assessed against reference data derived from optical imagery. Validation results showed that SAR imagery in VH polarization with an ascending orbit achieved the highest accuracy (92.4%). This study demonstrates the importance of selecting appropriate configurable SAR imagery parameters for flood mapping in tropical regions and provides a case study in disaster monitoring. Future work should examine different polarization combinations and other thresholding procedures and/or values to improve the modeling system.