A Landslide Predicting System Integrating Transient Hydrologic-stability Model And Precipitation Forecasts
Hui Chen(a*), Chih-Yuan Huang(b), Shou-Hao Chiang(b)

a)Master of Science Student in Department of Civil Engineering, National Central University
b)Associate Professor in Center for Space and Remote Sensing Research, National Central University

Abstract

Landslides pose a significant hazard in Taiwan, primarily driven by geological conditions, rainfall patterns, seismic activity, and human development. Among these factors, rainfall is one of the primary external triggers, as it directly alters slope hydrology. This study aims to develop a landslide forecasting system that integrates a transient hydrologic-slope stability model with rainfall forecasts to simulate slope responses under predicted rainfall scenarios. The stability model is based on SHALSTAB (Shallow Landslide Stability Model) and has been modified to incorporate a dual-layer soil structure. A fracture layer between the soil and bedrock is introduced to better represent subsurface hydrological processes. By coupling this framework with a time-series soil moisture simulation, a transient hydrologic-slope stability model is constructed to dynamically link hydrological changes with slope stability. The Yusuei Watershed in Kaohsiung City was selected as the study area, with the model evaluated using Typhoon Lupit (August 2021). Calibration was performed based on observed hourly rainfall, followed by validation with historical forecasts. Predicted landslide timing and locations were compared to inventory data for performance assessment. Future work will extend the model to varied geological settings and integrate it into a real-time system using rainfall forecasts and trigger mechanisms to enable timely simulation and risk assessment, supporting landslide forecasting and emergency response.

Keywords: Landslide- Fracture layer- Transient hydrologic-stability model- Landslide forecasting system- Yusuei watershed

Topic: Topic D: Geospatial Data Integration