Integration of SAR and multi-spectral imagery in flood inundation mapping – a case study on Kerala floods 2018

Abstract

Torrential floods had placed a record in the history of Kerala, the state on the south-western coast of India, in 2018. This study focusses primarily on the assessment of flood inundation mapping using the Sentinel-1 SAR imagery applying the Otsu’s thresholding algorithm for the Alappuzha region. Though the pre-processing of SAR imagery is difficult, the availability of software and enhanced hardware facilitates efficacy in image processing. Normalized Difference Water Index (NDWI) was computed to aid in differentiating the permanent waterbody from the flooded region. The results implicate the spread of flood-inundated zones has considerably increased in August, thereby indicating the devastating effect of the flood in August. Based on the results obtained, 13.86% area of the region has been covered by flood in August, excluding the permanent water bodies, which is comparatively high compared to the previous months. It is necessary to note that, extreme rainfall accompanied by the full reservoir level of the dams in the region served to be the key to the disaster in August. The resultant flood inundation map could be utilized quite effectively for relief measures and damage assessment.