Flood susceptibility assessment and mapping using GIS-based analytical hierarchy process and frequency ratio models

Abstract

In this study, flood susceptibility maps were produced for district Nowshera by using Geographic Information System (GIS)-based Multi-Criteria Decision Analysis (MCDA) Analytical Hierarchy Process (AHP) and Bivariate Statistical Frequency Ratio (FR) models. This study analyses twelve Hydro-geomorphological flood conditioning factors selected based on the scale and characteristics of the study area. These factors include elevation, slope, distance from the river, rainfall, drainage density, land use land cover (LULC), topographic wetness index (TWI), height above nearest drainage (HAND), normalized difference vegetation index (NDVI), distance from the road, curvature, and soil type. The weighted factors were then integrated using the weighted overlay technique in ArcGIS 10.8 to produce flood susceptibility maps, which were classified into five zones: very high, high, moderate, low, and very low susceptibility. The AHP model classified 145.1 km² (8.34 %) and 252.2 km² (14.49 %) of the area as very high and high flood susceptibility zones, respectively. While the FR model classified 94.6 km² (5.4 %) as very high and 230.5 km² (13.2 %) as high susceptibility zones. The models' results were validated using Receiver Operation Characteristics (ROC) Area Under Curve (AUC) and collected flood inventory. The calculated AUC values were 0.921 for the AHP model and 0.924 for the FR model, corresponding to accuracy of 92.1 % and 92.4 %, respectively. Through this study, we gained a deeper understanding of how hydro-geomorphological factors interact to influence flood susceptibility, emphasizing the importance of integrating multi-criteria and statistical approaches for accurate flood risk mapping. Our methodology provides a transferable framework that can be applied in diverse geographical contexts, aiding in flood mitigation planning for both data-rich and data-scarce regions worldwide.