Abstract
The geological and geographical conditions have made the Himalayas one of the regions most vulnerable to disasters. Northern Pakistan is made up of high-relief Himalayan mountains where landslides are the most frequently occurring catastrophic events. Floods, earthquakes, and landslides occur most often in this region, causing deaths and severe infrastructure damages. The studied area is experiencing significant landslide events triggered by destructive seismic activity, extreme seasonal rainfalls, and human activities. In this study, landslide susceptibility mapping is proposed using an integrated model of information value method by optimizing its capabilities and overcoming its limitations to enhance both precision and accuracy of the results. This approach quantifies and estimates the probability of landslide occurrence using statistical principles and techniques. Seven distinct landslide causative factors, including slope angle, distance to stream, drainage density, distance to road, land cover, normalized difference vegetation index, lithology, and distance to faults were evaluated to determine the areas at risk for future slope movements. A reliable landside inventory from a spatial database combined with ground-truthing field surveys was prepared for evaluating the relationship between the landslides and their causative factors. The accuracy and efficiency of the developed model are evaluated using sensitivity analysis of the receiver operator curve (ROC). The ROC-validated results showed an accuracy of 0.845 for model prediction rate and an accuracy of 0.837 for model success rate, indicating that the model is reliable to produce landslide susceptibility map. The results showed the reasonable efficiency of the applied method for assessing landslide susceptibility in the study area and applicable to other regions with similar geomorphological conditions.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have