Rainfall-Triggered Landslide Hazard Assessment in the Lower Mekong River Basin Using Model and Satellite-Based Estimates

Landslides can result in devastating loss of life and property damage and are a growing concern from a local to global scale. In recent years, increased rainfall-triggered landslide events in the Lower Mekong River Basin (LMRB) have put communities at increased risk of damage from these disasters. This dissertation focuses on how remote sensing can be used to address landslide hazard at global scale and specifically in the Lower Mekong region. In the first part of this research, two satellite-based rainfall products, CHIRPS and TMPA, were compared with daily rain gauge observations from 2000 to 2014 in the Lower Mekong River Basin. Both products showed higher correlation with in-situ data during the wet season (June–September) as compared to the dry season (November–January). Our validation test showed TMPA to correctly detect precipitation or no-precipitation 64.9% of all days and CHIRPS 66.8% of all days, compared to daily in-situ rainfall measurements, indicating CHIRPS may be better at representing precipitation in this region. The second part of this research, the influence of land use and land cover (LULC) change and other causative factors of landslide susceptibility are evaluated in the LMRB using Frequency Ratio analysis and Logistic Regression models. Results indicate LULC change from agricultural land to forest have a positive correlation with landslide occurrence. However, the most statistically significant factors in the models are found to be slope and distance to roads. The third part of this research evaluates global patterns in landslide reporting from events in the Global Landslide Catalog. The most notable landslide hotspots are in the Pacific Northwest of North America, High Mountain Asia, and the Philippines. More landslides are reported in areas with high population density compared to remote locations. A bias towards English-speaking countries was also discovered in the catalog reports. Finally, the last part of this research assesses how remotely sensed hydrological products can be used to improve rainfall-triggered landslide monitoring and prediction in data-sparse regions like the Lower Mekong. High-resolution soil moisture is better able to capture the soil moisture profile than coarser resolutions. By incorporating high resolution soil moisture we can better investigate landslide hazard and prediction and evaluate landslide prone areas. The results of this dissertation can be used in local decision making and disaster preparation and mitigation in the Lower Mekong River Basin