Empirically informed AHP-FR modeling for multi-hazard susceptibility mapping in a mountainous transport corridors

Abstract

Mountainous transport corridors are increasingly exposed to interacting natural hazards that intensify infrastructure, socio-economic, and land-use risks. However, most susceptibility assessments for high-mountain corridors still focus on single hazards or rely heavily on expert-based weighting, which can introduce subjectivity into spatial decision-making. This study develops an empirically informed Analytical Hierarchy Process–Frequency Ratio (AHP-FR) framework for integrated landslide, debris-flow, and flood susceptibility mapping along the Barakhan–Gilgit–Juglot section of the northern China–Pakistan Economic Corridor (CPEC). The main methodological contribution is the use of frequency-ratio-derived empirical evidence that was integrated with AHP-derived factor weights to calculate final subclass weights and thereby reducing reliance on purely subjective expert judgment. Field surveys, remote sensing datasets, historical hazard inventories, and 12 conditioning factors were integrated in a GIS environment to generate single-hazard susceptibility maps and an eight-class multi-hazard susceptibility map. Independent ROC-AUC validation showed acceptable to very good predictive performance, with AUC values of 0.803 for landslides, 0.762 for debris flows, and 0.872 for floods. The results indicate that 63% of the corridor is exposed to one or more hazards, including 59% exposed to two or more overlapping hazards; 41% are co-susceptible to landslides and debris flows, while 37% falls within the no/very low multi-hazard susceptibility class. The spatial outputs identify major hazard-prone sections, including Attabad, Chalt, and Passu, providing practical guidance for road-maintenance prioritization, slope-stabilization planning, drainage improvement, and emergency preparedness. The added value of this study lies in combining empirical FR-based weighting with AHP and translating separate hazard maps into explicit multi-hazard combination classes, offering a reproducible decision-support framework for resilient planning in mountainous transport corridors.