FLOOD MAPPING AND IMPACT ASSESSMENT

This project focusing on Flood Risk Mapping around the Lake Baringo region and its environs characterized a significant flooding event and assessed its critical impact on the surrounding environment using integrated geospatial technology. This study leveraged Synthetic Aperture Radar (SAR) data from Sentinel-1, specifically utilizing the Vertical Transmit, Vertical Receive (VV) polarization in a change detection approach, alongside advanced machine learning algorithms like Random Forest (RF), all processed efficiently within the cloud-based Google Earth Engine (GEE) platform. By comparing pre-flood and post-flood SAR imagery and integrating Sentinel-2 optical data for Land Use and Land Cover (LULC) classification—which achieved an impressive Overall Accuracy of 0.964 and a Kappa Coefficient of 0.953—detailed damage assessment was accomplished. The analysis successfully delineated the flood extent, revealing an estimated total inundation of 2,441 hectares. Crucial impact metrics were provided regarding the vulnerability of the region, including an estimated 657 people exposed, alongside 566 hectares of cropland and 3 hectares of built-up areas that were affected. This robust methodology has demonstrated the efficient application of SAR data and machine learning for rapid flood inundation mapping, monitoring, and environmental impact analysis, providing valuable results to enhance efficiency in flood management strategies in the Lake ecosystem.