[A-AS07] Extreme weathers and disasters in urban environments in East Asia and relationship to climate change

[AAS07-P01] Assessment of Landslide susceptibility along the mountainous road under Climate Change in the future in Central Taiwan

Chun-Hung Wu¹, *Kuan-Hsun Huang² (1.Associate Professor, Department of Water Resources Engineering and Conservation, Feng Chia University, Taiwan (R.O.C.), 2.Postgraduate, Department of Water Resources Engineering and Conservation, Feng Chia University, Taiwan (R.O.C.))

Keywords:landslide susceptibility, climate change, frequency ratio method, Taiwan

The weather in Taiwan, located in East Asia, is significantly affected by climate change. The rainfall characteristic in the future will gradually tend to be characterized by stronger short-latency rainfall intensity. The possibility of rainfall-induced landslide in the future will increase, especially along the mountainous roads. The study focuses on the evolution of landslide susceptibility along the mountainous roads under the climate change in the future in Renai Township, Nantou County, Central Taiwan. The mountainous roads in Renai Township, Nantou County, Central Taiwan, suffered severe sediment disasters during Typhoon Sinlaku in 2008 and Typhoon Khanun in 2023. The study uses the geomorphological data, rainfall data and landslide inventory after 2008 Typhoon Sinlaku as the basis for building the landslide susceptibility assessment model, and uses the landslide inventory after 2023 Typhoon Khanun in 2023 as a model verification. The study combines the landslide susceptibility assessment model and the daily rainfall in the future in Taiwan which was proposed in the Assessment Report Sixth by Intergovernmental Panel on Climate Change to predict the landslide susceptibility along the mountainous roads in Central Taiwan.
The frequency ratio method is the methodology for building the landslide susceptibility in the study, and 10 landslide-related factors, including elevation, slope, aspect, geology, land use, terrain roughness, terrain wetness, terrain curvature, landslide frequency, and daily rainfall, were selected as the main factors in the landslide susceptibility model. The daily rainfall during 2008 Typhoon Sinlaku and the landslide inventory after 2008 Typhoon Sinlaku were the basis for building the landslide susceptibility model. The daily rainfall during 2008 Typhoon Sinlaku ranged from 162.5 to 480 mm, and 12 landslide cases were induced in the study area. The landslide susceptibility model along the No. 14 Provincial highway uses 7 landslide-related factors, including geology, terrain roughness, aspect, terrain curvature, slope, landuse, and daily rainfall, and the AUC value of the landslide susceptibility model is 0.757. The correct ratio of the landslide susceptibility model by using the error matrix is 75.7%. The landslide susceptibility model along the Aowanda road uses 6 landslide-related factors, including landuse, geology, terrain roughness, daily rainfall, aspect, and slope, and the AUC value of the landslide susceptibility model is 0.861. The correct ratio of the landslide susceptibility model by using the error matrix is 75.5%. The daily rainfall during 2023 Typhoon Khanun and the landslide inventory after 2023 Typhoon Khanun were used to verify the model, and the AUC value and the correct ratio exceed 0.7.
The study will combine the daily rainfall from four climate change scenario models from 2024 to 2100 and the landslide susceptibility model to predict the landslide susceptibility in the future in Central Taiwan. The four climate change scenarios used in the study are the sustainability scenario (SSP 126), the middle of the road scenario (SSP 245), the regional rivalry scenario (SSP 370), and the fossil-fueled development scenario (SSP585). The daily rainfall with 50 year return periods based on the four climate change scenarios will estimated and used to predict the spatial and temporal distribution of landslide susceptibility along the mountainous roads in Central Taiwan.