Landslide susceptibility modelling applying user-defined weighting and data-driven statistical techniques in Cox’s Bazar Municipality, Bangladesh

Abstract

Landslides are common geophysical hazards in the highly urbanized hilly areas of Cox’s Bazar Municipality, Bangladesh. Every year, during the monsoon, landslides cause human casualties, property damage, and economic losses. Indiscriminate hill cutting, developing settlements in dangerous hill slopes, and torrential rainfall in short period of time are responsible for triggering landslide disasters. The aim of this paper is to produce landslide susceptibility maps (LSM) to help reduce the risks of landslides. Geographic information system and remote sensing-based techniques were used for LSM considering 12 relevant factor maps (i.e. slope, land cover, NDVI, geology, geomorphology, soil moisture, rainfall pattern, distance from road, drain, stream, structure, and faults–lineaments). For the modelling purpose, four techniques were implemented—artificial hierarchy process (AHP), weighted linear combination (WLC), logistic regression (LR), and multiple logistic regression (MLR). A landslide inventory map with 74 historical landslide locations was prepared by field surveying. The modelling results are validated using the area under the relative operating characteristics curves (AUC). AUC values of AHP, WLC, LR, and MLR methods are calculated as 88, 85.90, 74.90, and 90.40 %, respectively.