Residual landslide susceptibility analysis based on integrated machine learning framework Shou Hao,Chiang, Tung Cheng,Lu
First Author^s Affiliation: Professor, Center for Space and Remote Sensing Research, National Central University, Taiwan
Second Author^s Affiliation: Research Assistant, Center for Space and Remote Sensing Research, National Central University, Taiwan
Abstract
In this study, residual landslides refer to slopes characterized by substantial accumulations of landslide-derived failure materials, forming unstable colluvium that may serve as primary source zones for debris flows, especially under conditions of intense rainfall, thereby posing significant hazards to downstream areas. In Taiwan, such residual landslides present a critical challenge for watershed management and sediment-related disaster mitigation. The inaccessibility of mountainous terrain, combined with the increasing frequency and intensity of extreme precipitation events driven by climate change, has exacerbated the risks and unpredictability associated with landslides and debris flows. To address these challenges, this study proposes an advanced machine learning framework for assessing residual landslide hazards. The approach involves the development of a predictive model to evaluate the potential activity of residual landslides through the integration of temporal remote sensing data, including time-series satellite observations from Sentinel-1 and Sentinel-2. This methodology enables systematic inventory mapping and activity assessment across mountainous regions, thereby enhancing early warning capabilities and informing more effective sediment disaster prevention and management strategies.
