C-band SAR interferometry based flood inundation mapping for gorakhpur and adjoining areas

Abstract

SAR or Microwave remote sensing today has proved itself to be a paradigm shift in the field of remote sensing. With its all-weather availability, it has found variety of applications in several fields today. Till now, and mainly, floods have been mapped using optical remote sensing data since a long time. Floods have been a common and yearly catastrophe which the people of areas situated in lesser elevations and on the foothills of mountains in the Terai area must face. Proper mapping of floods can help to delineate the areas with higher loss of life and property thus aiding the relief measures. This study aims to analyze and exploit the potential of SAR Interferometry (InSAR) for flood inundation mapping using a pair of C-Band dual PolSAR datasets from Sentinel-1. The datasets were preprocessed for speckle removal and radiometric corrections. Thereafter, subset of study area was taken, and the datasets were coregistered followed by interferometric processing. The results showed low phase shift in and around the river and to some extent in the urban areas which generally does not happen unless there is an increased single bounce scatter. The topographic phase removed interferogram showed high phase variations in the core of urban areas while coherence map showed an acute loss of coherence from permanent features as well which showed high degree of flood water accumulation. This flood was mapped by masking out the low values of high coherence areas from the image using a proper threshold value. High degree of flooding was observed in Urban areas and much of agricultural land was seen to have been submerged under water once the resulting inundation map was overlaid on Google earth