SAR SENTINEL-1 DATA, NDFI AND NDFVI FOR DETECTING AND MAPPING THE FLOOD HAZARD IN OUED SAKIA-EL HAMRA (LAAYOUNE, SOUTH MOROCCO)

Abstract

A new automated approach for flash extent delineation and mapping and risk assessment was applied in the context of the Oued Sakia-Al Hamra flood (Laayoune, southern Morocco) of October 2016. Normalized Difference Flood Index (NDFI) mapping was employed to distinguish between flood-prone and non-flood-prone areas across an extensive region, leveraging the Bragg backscattering properties of active radar pulses on flat water surfaces, which exhibit minimal signal reflection. The study utilized Sentinel-1 satellite SAR images in Wide Swath (IW) interferometric mode and Ground Range Detected (GRD) product type. Pre-processing and processing chain, which combines water classification, multi-temporal and contextual filtering, topographic correction of a total number of 2 images: before and after the flood event. Consequently, a high classification accuracy of 99.05% using Sentinel-1 C-band images for the flooded area of Oued Sakia-El Hamra was obtained after validation using both optical images of Landsat-8 and Google-Earth. The results demonstrate the effective utilization of SAR data for identifying flooded areas, assessing their extent, evaluating associated hazards, and subsequently recommending appropriate mitigation measures as needed.