SAR-Based Flood and Waterlogging Extent Mapping Using Sentinel – 1 Time Series Backscatter Analysis: A Case Study of the FCT

Abstract

Flooding is still one of the most important environmental dangers for rapidly urbanising communities, especially when the prompt flood monitoring is hindered by the continuous cloud cover and shortage of hydrological measurements. This study established a multi-temporal flood mapping methodology using Sentinel-1 Synthetic Aperture Radar (SAR) images and Google Earth Engine (GEE) to examine flood dynamics in the Federal Capital Territory (FCT), Nigeria, between 2019 and 2024. The workflow comprised the seasonal dry and wet season compositing, adaptive Otsu thresholding, permanent water masking and terrain filtering to yield annual flood maps, flood frequency products and persistent flood prone locations. Processing the Sentinel-1 SAR scenes produced six annual flood maps and corresponding flood statistics. The yearly flood extent varied between 39.98km2 (2021) and 40.66km2 (2022), with an average of 40.35km2 and a total fluctuation of 0.69km2 for the research period. Results showed that flooding continued to be spatially concentrated in key river corridors and floodplain ecosystems, demonstrating a strong temporal persistence of flood-prone locations. The adaptive thresholding method produced annual thresholds ranging from 2.87 to 3.37, and showed stable classification performance under various SAR backscatter conditions. The created procedure was effectively able to identify recurrent flood hotspots and provide a replicable cloud-based framework for operational flood monitoring. Our results show that the use of Sentinel-1 SAR imagery combined with Google Earth Engine offers an efficient and scalable approach for long-term flood assessment and supports evidence-based flood risk management, sustainable urban planning and climate adaptation in fast-growing metropolitan areas.