Landslide Susceptibility Mapping and Comparison Using Frequency Ratio and Analytical Hierarchy Process in Part of NH-58, Uttarakhand, India

Abstract

In Himalayas, National Highway 58 (NH-58) is one of the important and the major lifelines, which is very badly affected by frequent landslide occurrences particularly during the monsoons and cause recurring problems to pilgrims and local people. In the present research, the 52 km stretch ghat road in part of NH-58 was chosen for the landslide susceptibility zonation (LSZ) mapping using frequency ratio (FR) and analytical hierarchy process (AHP) models and through integrated remote sensing and geographical information systems (GIS) techniques. The landslide inventory details were mapped out using high resolution satellite image, the data collected from secondary sources and field investigation. There are 11 landslide influencing parameters were considered for the analyses. LSZ maps were generated by calculating relationship between the landslide influencing factors with training landslide data in the case of FR model but for AHP model pair-wise comparison were made to derive the weights and final score. The LSZ map prepared using FR and AHP models and classified into five different susceptibility zones. The LSZ maps were compared and validated with validation landslide dataset using Area Under Curve (AUC) method. The AUC value of FR model is 0.8157 showing better prediction accuracy than the AHP model (AUC value is 0.6780).