A simulation-enabled slope digital twin for real-time assessment of rain-induced landslides

Abstract

Digital twin technology has emerged as a promising tool for effective geological hazards management in the digital era. By creating a digital replica that mirrors the real world, it enables real-scene restoration, timely disaster analysis, and coordinated decision-making. Presently, the maturity of slope digital twins remains low. A digital twin platform that can achieve prompt full-process landslide simulation (including initiation and mobility) is highly needed for real-time assessment of regional landslide hazards. This study develops a simulation-enabled slope digital twin, which includes a digital model of slope entities and mitigation structures, a real-time rainfall data transmission network, and a virtual simulator for assessing landslide consequences. In the virtual simulator, a probabilistic slope unit-based landslide initiation model and a rapid position-based landslide mobility simulation model are proposed for real-time assessment within a digital twin environment. Critical landslide metrics (e.g., number of landslides, flow depth, impact force and factor of safety of rigid barriers against sliding and overturning failures) that are closely tied to loss estimates are visualized in real-time for emergency assessment. The position-based landslide mobility simulation model is calibrated and verified through historical events and shows much higher efficiency than traditional numerical simulation methods. The simulation-enabled slope digital twin is demonstrated through a case study of the Yu Tung Road landslides in Hong Kong. With its ability to provide timely and realistic feedback regarding the evolution of landslide processes and intensity, the slope digital twin can serve as an effective tool for landslide hazard assessment.