Flood inundation mapping in the Kelantan River Basin, Malaysia, using Sentinel-1 SAR and Google Earth Engine

Abstract

One of the most severe floods in Peninsular Malaysia occurred during 2021-2022, displacing over 20,000 people and resulting in two deaths in Kelantan. Accurate flood extent data during such events is crucial for effective flood management, however, gathering this information is challenging due to limited access to affected area. Google Earth Engine (GEE) offers rapid satellite image processing for flood inundation mapping, making it an effective tool for this purpose. In this study, GEE was utilized to generate flood inundation maps for the Kelantan River Basin (KRB) using Sentinel-1 SAR data. Site inspections and Sentinel-2 Multispectral Instrument (MSI) satellite images of the actual flood regions were then used to validate the flood inundation maps. Additionally, this study evaluated the effects of three distance thresholds (3-, 4- and 5-pixel) to differentiate inundated area from preliminary water surfaces. The findings showed that the flood inundation maps achieved an accuracy of 57 – 60%, with the highest accuracy observed under the 5-pixel threshold. The 2021-2022 flood, with an inundated area of 8.92 km2, was one of the worst experiences in Kota Bharu. These findings provide valuable insights to support local authorities in designing better flood mitigation strategies for the future.