A systematic assessment of regional landslide risk under typhoon rainfall: a case study of Taishun, Zhejiang, China in September 2016

To clarify the occurrence patterns and risk characteristics of typhoon-induced rainfall landslides, this study assessed landslide susceptibility in Taishun County, Zhejiang Province, and surrounding areas affected by Typhoon Meranti in September 2016. Based on a detailed inventory of 4,102 landslides, a susceptibility framework incorporating key topographic and rainfall-related factors was established using an automated machine learning approach. The results show that landslides are predominantly distributed at elevations of 400 ~ 800 m and slope gradients of 25°~40°, with frequent occurrences associated with cumulative rainfall of 160 ~ 180 mm. The optimal ensemble model achieved high predictive performance and demonstrated strong spatial agreement between predicted high-risk zones and observed landslide distributions. Factor contribution analysis indicates that accumulated rainfall, elevation, and slope are the primary controlling factors, with their interactions playing a critical role in landslide susceptibility. Overall, this study provides an effective and interpretable framework for understanding typhoon-triggered landslide mechanisms and supporting regional early warning and disaster mitigation efforts.