Mapping Flood Prone Areas Using Sentinel-1 Satellite Imagery In Northern Kelantan, Malaysia

Abstract

Natural hazards including landslides, flash floods, earthquakes and hurricanes have been constituted a significant problem in many countries. The most considerable natural hazard in Malaysia is flood, which mostly triggered by heavy rainfall. The recent flood in November 26, 2017 has induced many problems in northern part of Kelantan. This paper aims to map flooded areas in the region by using RADAR imagery of SENTINEL-1 without using any automated classifier. A total number of two satellite imageries of SENTINEL-1 for the dates 18/11/2017 (Before event) and 30/11/2017 (After event), were collected. However, using SENTINEL application platform (SNAP) software the images were corrected, stacked and combined in order to map flooded areas. Results indicate that this kind of flood mapping is precise, fast and easy to use. Furthermore, using confusion matrix and a hundred ground control points (GCPs) the final map was validated. The overall accuracy and Kappa coefficient was 94.5023% and 0.9189, respectively. This paper can be helpful to policy and decision makers in order to map flooded areas immediately, especially when we need quick action for settling camps for evacuees.