Application of sentinel-1 SAR data for flood monitoring in the lower Ganges Basin: a time-series analysis of 2021 flood in Bihar
Mohammad Sajid, Haris Hasan Khan, Arina Khan, Rizwan Ahmad, Armugha Khan, Ghazala Siraj, Abdul Ahad Ansari
Abstract
Seasonal monsoon rains and upstream water inflows make the Lower Ganges Basin one of the most flood-prone regions in India. Bagmati, Gandak, Burhi Gandak, Adhvara group and Kosi, coupled with heavy rainfall, contribute significantly to the flooding, which havoc on people’s settlements and the naturally built environment. This comprehensive research was conducted to map the flood inundated area of the 2021 flood, using the time-series sentinel-1 SAR data from May to November in central part of Bihar. To accurately delineate flooded areas, an optimal threshold-based classification was applied to VH-polarized backscatter images, with threshold values ranging from − 19.5 dB to -22.3 dB. Fourteen Sentinel-1 images were processed in SNAP, followed by a detailed flood impact assessment on Land Use and Land Cover (LULC) using ArcMap. In this process, we carefully examine various LULC categories—such as cultivated land, Built-up, open scrub, trees, and bare ground—to quantify the extent of the flood’s impact. The key results show that the months of July and August marked the peak of the disaster, with inundated areas measuring an astounding 4,453 km² and 4,400 km², respectively. Overall, the floodwaters overtook 4.22%, 10.42%, 10.30%, 9.22%, 5.24%, and 3.22% of the land in June, July, August, September, October, and November 2021, respectively. Most alarmingly, our analysis underscored the severe impact on land use and land cover (LULC), particularly affecting agricultural lands, with approximately 4,159 km² of cultivated areas submerged during the peak flooding months. The data and methodologies we employed can be adapted for use in flood mapping across the globe, empowering communities to better prepare for and respond to future disasters.