Geospatial assessment and integrated multi-model approach for landslide susceptibility mapping in Meghalaya, India

Landslides in Meghalaya, India, inflict severe property and life damage due to mountains, steep slopes, and heavy rains. Landslide Susceptibility (LS) maps are extremely useful for disaster management. The primary goal of this work is to generate an LS map for the state of Meghalaya. In the first phase, a Landslide Inventory (LI) map was created, which included 855 landslides that occurred from 2019 to 2023. The LI map was then split into 70 % (601) and 30 % (254) for training and testing, respectively. In the second phase, the study selected fourteen conditioning factors as thematic layers for LS mapping and performed multicollinearity and Pearson's correlation analysis; all the parameters were identified as optimal for the prediction model. Eight different scenario models (Frequency Ratio (FR), Evidence Belief Function (EBF), FR + EBF, FR*EBF, (FR*EBF)/2, (2*FR) + EBF, (2*EBF) + FR and (EBF + FR)/3) have been used to generate LS maps. The created maps were validated using Receiver Operating Characteristics (ROC) curve and the corresponding Area Under the Curve (AUC) value, statistical measures (recall, precision, F1 score, overall accuracy, and balanced accuracy) and on-site verification with recent landslides. Finally, the result of the best scenario was compared with the outcome of the Analytical Hierarchy Process (AHP) method. Results showed that scenario 4 (EBF*FR) has an overall accuracy of 82.3 %, whereas AHP has an overall accuracy of 77.6 %. It is indicated that scenario 4 achieved 4.7 % higher overall accuracy than that of the AHP method. Recent landslides selected for on-site verification occurred in an area classified as very highly susceptible by the scenario 4 model. These maps provide vital conceptions of landslide mechanisms, assisting land use planning and disaster management. This approach can be applied in similar areas by investigators, which indicates the originality of the study, and the result of this study will be beneficial for the Meghalaya region. (c) 2024 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.