Creating Landslide Susceptibility Map Using Pairwise Comparison Model in Soran City, Irbil Province, Iraq

Abstract

Landslide susceptibility maps are essential for planning and dealing with natural disasters in mountainous areas. Soran City, Iraq, was selected to use pairwise comparison models to produce the landslide area's susceptibility map. Three data sources were used to create a landslide inventory map: a satellite image, a field investigation, and previous studies. The satellite image utilized in this investigation is LANDSAT-8, downloaded from the USGS website, consisting of 11 bands with 16-days and 30 m temporal and spatial resolutions, respectively. Consequently, 72 landslides were identified and mapped, with 53 (70%) chosen randomly to create the landslide susceptibility model and the remaining 19 (30%) used to validate the models. Eight landslide conditioning factors were studied to build pairwise comparison models to produce a landslide susceptibility map. These factors encompass slope aspect, slope degree, land use, elevation, curvature, NDVI, road distance, and distance to stream. Consequently, the area under the curve (AUC) analysis showed that the measure of the level of disorder in the model with an AUC value of 0.7 has the maximum prediction accuracy at 73.3 %. The susceptibility maps generated by Pairwise comparison models produce an accurate prediction of the susceptibility of the Kamarbandi Soran-Papshtia road segment to landslides, which can be used as a tool for planning land use and reducing disasters.