Flood Inundation Mapping Over Kerala Region During 2018 Flood—Based on Cloud Computing and Automatic Threshold Detection Approach

Flood inundation maps provide valuable information towards flood risk preparedness, management, communication, response, and mitigation at the time of disaster, and can be developed by harnessing the power of satellite imagery. The present study focuses primarily on the rapid mapping of flood inundated area using freely available high resolution Synthetic Aperture Radar (SAR) data during 2018 floods and its impact assessment in the Kerala state. Multi-temporal, dual-polarized (VH & VV) Sentinel-1 (SAR) data have been used for mapping the flood inundated area. SAR data is very useful and most preferred for detecting flood inundated area in all weather conditions and during day/night. Methodological advances included open-source APIs like Google Earth Engine (GEE) for monitoring flood inundated areas. In this study, automatic threshold detection technique was applied to separate satellite image pixel values into flooded and non-flooded group. Threshold for backscattering coefficient (σ0) between -2 to -24 dB have been applied for both the polarised band (VH & VV) to extract the maximum flood pixels. Further, a comprehensive analysis has been carried out by Global Precipitation Mission (GPM) data to investigate the intensity and frequency of precipitation over the basin. This study observed that flood inundation was in its peak during August 15 to 21, 2018 and impacted a large part of agricultural and urban patches during this time. The zonal statistics are calculated to find the inundated area over study region. The finding reveals that 995.49 square kilometres (sq. km) of land have been affected in the Kerala state. The results revealed that crop land was mostly inundated by flooding with low elevation and low slope. The proposed approach showed the effectiveness of combining SAR data and Google Earth Engine for flood inundation mapping. The flood inundation is associated with climatic variables such as high frequency and intensity rainfall, high runoff from high elevation to low elevation, etc. By leveraging these technologies, accurate and up-to-date information about flood inundation can be provided to decision-makers and emergency responders, ultimately contributing to more effective disaster management, and reducing the impact of flooding on vulnerable communities.