Landslide Mitigation in Sri Lanka Using a Remote Sensing Based Multisensory Framework with LIDAR, SAR, and LULC Data

Landslides are one of the major natural disasters of Sri Lanka, and they can potentially impact human life, infrastructure, and the environment, particularly in hilly and mountainous regions. This research integrates multi-source remote sensing and terrain analysis for enhanced characterization and prediction of landslide susceptibility in Passara and Lunugala DSDs. High-resolution Airborne LiDAR data from known landslide sites capture fine-scale slope variations and are compared with coarser resolution SRTM DEM data to quantify geometric changes in the terrain. The results show significant slope alteration due to displacement and accumulation of material. Backscatter intensity and GLCM-based texture changes before and after the landslide events were analyzed using Sentinel-1 Synthetic Aperture Radar (SAR) imagery. The overall results indicate that urban features have relatively consistent backscattering, while rural features show strong variation, hence showing significant alterations in surface roughness and scattering properties after the events. Additionally, LULC classification from Sentinel-2 imagery revealed that bare land and low vegetation areas on steep slopes are highly prone to landslides, while built-up areas though generally located on flatter surfaces show high variability in slope conditions and are therefore also susceptible to landslide hazards. Interferometric SAR (InSAR) analysis effectively captures spatial deformation patterns, supporting the identification of slope instability in both rural and urban landslide zones. In the proposed methodology, slope conditions derived from the locations of actual landslides are used to produce a landslide susceptibility map. A kernel-based terrain analysis is presented that incorporates a landslide specific slope kernel aligned with the local slope direction to better detect the areas with slope patterns similar to past landslides, increasing detection accuracy. This study integrates LiDAR, SAR, InSAR displacement analysis, and optical LULC data to improve landslide hazard mapping and support effective mitigation planning.