Document Type : Original Research Paper
Authors
Department of Civil Engineering, Chandigarh University, Mohali, India
Abstract
Landslides pose significant risks to human life, infrastructure, and the environment, particularly in geologically unstable regions like the Himalayas. This study aims to develop and validate landslide susceptibility maps using Frequency Ratio (FR) and Information Value (IV) models within a GIS framework. Employing high-resolution geospatial data, including geomorphological, topographical, and hydrological factors derived from high-resolution digital elevation models (DEMs) and other geospatial datasets. The susceptibility maps were classified into five categories: Low, Moderate, High, Very High, and Extremely High. The models were trained and validated using a landslide inventory of 1313 landslide events, with a 70:30 split for training and testing datasets. The predictive performance of the models was evaluated using the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve, yielding AUC values of 84.1 for the FR model and 83.9 for the IV model. The Landslide Density Index (LDI) further confirmed the models' reliability, indicating higher landslide densities in the predicted high-susceptibility zones. The study demonstrates that both FR and IV models are effective tools for landslide susceptibility mapping and its validation. The findings highlight the FR model's superior predictive accuracy in this specific area. Future research should leverage advanced machine learning techniques, such as XGBoost, Random Forest (RF), Naive Bayes (NB), and K-Nearest Neighbors (KNN), to enhance the reliability and precision of landslide susceptibility models.
Keywords
- Landslide Susceptibility Mapping (LSM)
- Frequency Ratio (FR)
- Information Value (IV)
- Geospatial Analysis
- Landslide Density Index (LDI)
Main Subjects
)