Probabilistic landslide susceptibility mapping along Tipri to Ghuttu highway corridor, Garhwal Himalaya (India)

- Semi-automatic methods for the landslide mapping using satellite image.
- Application of Frequency ratio method for landslide susceptibility assessment.
- A conceptual analysis of reactivation of landslides.
- Landslide characteristics along the highway corridors.

Landslides are very crucial natural hazard that pose serious threats to lives and livelihood in the Garhwal Himalaya, particularly during the monsoon season. The aim of this study was analysis of distribution probability of landslides along the highway corridors and susceptibility mapping. Hence, a comprehensive landslide inventory has been prepared using Landsat 8 satellite images for 2009, IRS LISS-IV MX satellite image for 2011; and Google earth images for 2014. In order to validate the landslide locations, repeated field investigations were carried out along the road corridors in the area. It has been analyzed that approximately 76% of the landslides have occurred along the highway corridors and streams in the study area during the monsoon season (July to September). The factors which cause landslides; such as relative relief, slope, slope curvature, aspect, geology, soil, surface dissection, compound topographic wetness index (CTI), normalized difference vegetation index (NDVI), land use, distance to stream and distance to road were created using ArcGIS 10.1 software. These parameters were analyzed to prepare landslide susceptibility zones, which is depicting the probability of occurrence of landslides. This study deduces that present scenario will lead to the future; weightages of each parameter were calculated on the basis of landslide inventory (2014), as a Frequency Ratio (FR). Higher is the landslide frequency, higher will be weightages. The sum of weightage of FR of the variables was used as landslide susceptibility index (LSI). Five landslide susceptibility zones were classified based on the LSI and were validated using 50 active random landslide locations and grid pixels. The AUC value was calculated 0.812 shows that the 81.2% of landslides have occurred in the high and very high landslide susceptible zones. Thus FR method is very significant for the probabilistic analysis of landslides along the highway corridors in the Himalayas for policy implications.

456