Sustainable Marine Structures

Monitoring flood inundation in data-poor areas is still a significant challenge for Disaster Risk Management, especially in dry areas where few in situ measurements and unfavorable weather conditions may limit traditional optical remote sensing. In this study, an approach to rapidly detect floods using Sentinel-1 synthetic aperture radar (SAR) images and utilizing the cloud computing platform provided by Google Earth Engine (GEE) is proposed. The workflow combines pre- and post-event SAR change detection, permanent water masking, climate information and field observations, for the purpose of outlining flood-affected areas and newly inundated ephemeral water bodies. The framework is tested on the extreme flood event of September 2024 in the Merzouga dune field, Morocco. The results indicate that the flood water was mainly focused around the drainage pathways and interdunal depressions where the local topography and geomorphological configuration influenced the water accumulation and temporary storage. The analysis also detected new depressions that are inundated after the exceptional rainfall event, which represent ephemeral water bodies. By incorporating permanent-water data, there was less potential for an event water surface to be confused with existing water surfaces. The results illustrate the capability of Sentinel-1 and GEE to provide timely flood assessments in arid areas. Adaptability to the coastal and near-shore environment is expected but not proven and would need to be demonstrated in representative coastal environments in the presence of tides, waves, intertidal surfaces, and complex coastlines.