GIS BASED LANDSLIDE SUSCEPTIBILITY MAPPING ALONG THE NATIONAL HIGHWAY (N 15), PAKISTAN

Abstract

The N-15 highway is a 240 km long road connecting the cities of Mansehra and Chilas. The study area lies between latitude 34˚20' to 35˚24’ N and longitude 73˚11' to 74˚08' E. Geologically, the road passes through Himalayas with thrusts like MMT, MCT and MBT passing through the area making it unstable. The geological and environmental conditions of the area make the road section vulnerable to landslides. To assess the landslide susceptibility along this road, two statistical models were used that are Frequency Ratio (FR) and Information Value Method (IVM). Eleven triggering factors were used for susceptibility mapping which include Elevation, Slope Angle, Aspect, Curvature, Lithology, Peak Ground Acceleration (PGA), Distance to Faults, Distance to Roads, Normalized Differential Vegetation Index (NDVI), Rainfall and Topographic Wetness Index (TWI). 203 landslides were marked along the road using google earth imageries for landslide inventory. Then, the landslide inventory was divided into training data set and validation data set with the ratio of 70 and 30 percent. The conditional factor layers were overlapped with the mapped 142 training landslides to get the final landslide susceptibility map. The FR and IVM success and prediction curves were drawn using training and validation landslide data respectively. The success rates for FR and IVM were 58 % and 70 % respectively. Similarly, prediction rates for FR and IVM were 57 % and 69 %. Overall, the results for landslide susceptibility model using frequency ratio model were satisfactory however the results derived through Information Value Method were promising showing good success and predictive curves of 70 and 69% respectively. The results for the used statistical models could be increased using better quality DEM which will eventually enhance the research control over the area.