Using Time clustering to anlyze how precipitation cause landslides occurrence

Landslides are closely related to precipitation events. However, the mechanism of how different types of precipitation events clustering actuate the landslides remains poorly understood. We collected and analyzed the precipitation and landslides data in the upper reach of Min River, Sichuan, China, which is a typical area composed of four counties with frequent landslides and seasonal variations in precipitation. Based on the time cluster algorithm, we identified six models of different landslides with precipitation events clustering drivers. 70% of landslide events are associated with short-term precipitation within 7 days before the occurrences. Nearly all the massive landslide events are preceded by the short-term precipitation and temporal clustering of moderate to extreme precipitation events. We conclude a causal relationship between precipitation event clustering and landslides. Our results represent a clear example of the potential for precipitation events to landslides or other hazards.