Multi-Hazard Susceptibility Analysis using Multi-Criteria Decision analysis, an application of Analytical Hierarchy process: A Case Study of Ghizer District, Gilgit Baltistan

Abstract

Natural hazards can be of various types such as Earthquakes, Landslide, Avalanche, Flash flooding and etc, but all in all these hazards bring long lasting destruction whenever and where ever they occur. The northern regions of Pakistan particularly the Himalayan mountain range are prone to these natural hazards. The main objective of this research is to perform susceptibility analysis on three particular natural hazards in Ghizer District namely: Landslides, Flash floods and snow avalanches and find suitable mitigation measures for each natural hazard in said region. Data of the study area was collected through satellite imagery, while the data of the triggering factors (factors responsible for natural hazards) such as lithology, precipitation and drainage density was acquired through manipulation of said satellite image and through the use of various GIS techniques. The susceptibility analysis of the mentioned natural hazards was calculated through the use of multi-criteria decision analysis techniques primarily Analytical Hierarchy Process (AHP). The AHP model to be used was operated through the extension AHP2.0 for Arc-Map. The results indicated that the potential susceptibly zones were found near fault lines and high slope zones, high drainage density and high precipitation zones, and near regions of greater snow cover respectively for landslide, Flash-Flood and Snow Avalanche respectively. The accuracy of prediction of the AHP model was determined through the use of the AUC and ROC curve method where the AUC values for Landslide, Flash-Flood and Snow Avalanche Susceptibility were 71%, 72% and 89% respectively. The accuracy of AHP model can be improved through increasing the number of true positives/ historic data. The study also suggested various mitigation factors for said natural hazards while considering the economic conditions of the area.