Item

Abstract

Flood is amongst the hydro metrological hazards most frequently occurring and responsible for massive destruction of lives and properties in developing countries like Pakistan. A most devastating flood episode occurred during 2022 in lower Indus basin where roughly more than 33 million people were affected. The aim of this study was to evaluate previous major floods episodes temporally and extraction of susceptible sites based upon multiple causal parameters, over the lower Indus plain, during past 22 years. To achieve this, we utilized the relevant resolution planet imageries for the study region. The susceptible sites were extracted using multiple causal parameters. A suit of geodetic observations was monitored to enhance the precise extraction of susceptibility. The deep learning (deep boost) and machine learning (naive bayes tree) algorithms were utilized and spatial coverage of vulnerability for each algorithm was calculated and mapped. Both algorithms were compared for the accuracy assessment. Our approach will greatly be beneficial for planners in managing and protecting this river basin from this natural hazard.