ARIMA and SPSS statistics based assessment of landslide occurrence in western Himalayas

Abstract

The Jammu-Srinagar National Highway is the critical road connection between Kashmir valley and the rest of India. It passes through extremely steep slopes and high mountains prone to mass movements, particularly landslides and rockslides. Most mountainous roads are constructed on fragile and rocky slopes, and any natural (e.g., precipitation) or human-triggered disturbance (e.g., heavy traffic) can cause a fatal and devastating landslide under the influence of gravity. Many landslide-prone sites along the Highway are responsible for the continuous blockade almost throughout the year but peaking particularly during winters. As a result, it has a high toll on the state's economy and is responsible for many human casualties yearly. The present study aims to characterize various factors and their threshold values responsible for triggering a landslide. Through extensive field surveys, we evaluated different geotechnical parameters of soils at the most landslide-prone site along the Highway and augmented it with the satellite remote sensing datasets to determine the threshold values that trigger a landslide and assess the probability of occurrence of landslide events in the future using Autoregressive Moving Average (ARIMA) model and IBM SPSS Forecasting Model. This work shall help devise countermeasures for managing the landslides in the study area locally and shall serve as the guiding framework for using artificial intelligence and machine learning techniques for hazard management in the Himalayas.

Keywords

• Landslide
• Statistical analysis
• Remote sensing
• Disasters