A Robust AHP–GIS Based Flood Susceptibility Mapping Using Stochastic Hill-Climbing Algorithm

Abstract

Flooding is one of the most frequent and devastating natural hazards affecting urban areas globally, leading to considerable social, economic, and environmental impacts. In the present study, a novel framework for flood susceptibility mapping is proposed using the stochastic Hill-Climbing (SHC) method in the Analytical Hierarchy Process (AHP) integrated with Geographic Information System (GIS) technique. For illustrating the proposed methodology, a set of ten geo-environmental and human-induced factors are considered, which are elevation, land cover, slope, lithology, drainage density, rainfall intensity, proximity to rivers, topographic wetness index (TWI), soil classification, and groundwater level for one of the fast-growing districts, Raipur in India. These factors are weighted according to their relative importance in influencing flood occurrence, and the Flood Hazard Index (FHI) is evaluated across the study area using the rainfall data of various time periods. The results indicate that proximity to rivers, elevation, and slope are the major contributing factors to the flood susceptibility level in the study area. Depending upon the calculated FHI values, the study grouped the area into five levels of flood susceptibility: very high, high, moderate, low, and very low. The model's performance was checked with the Receiver Operating Characteristic (ROC) curve, where the predicted data was plotted between the data obtained from the District Disaster Management Authority, Raipur, and it reached a score of 0.828, showing good prediction ability. The final flood susceptibility map gives useful information for district authorities, planners, and disaster management teams to support better decisions by identifying high-risk zones which strengthen Raipur district's resilience to floods.