Probability analysis of landslide based on typhoon track

Typhoon rainstorm is the main inducing factor of landslides in southeast hills and mountains in our country. Revealing the correlation between typhoon path and landslides has important theoretical and practical significance for monitoring and early warning of typhoon-rainstorm soil landslides in southeast hills and mountains. In this paper, based on the typhoon data directly passing through Fujian Province or indirectly affecting Fujian Province from 2015 to 2019, as well as the rainfall data and typhoon-rainstorm landslide data during these typhoon events, the Kriging interpolation method in ArcGIS software is used to couple the typhoon path, rainfall data and typhoon-rainstorm landslide data. Through the logistic regression method in SPSS software, the landslide probability model based on typhoon track in Quanzhou City is obtained. Using Kriging interpolation method and logistic regression method, we take the landslide occurred during typhoon Dujuan as an example to verify the model. The correct probability of landslide caused by rainfall data measured by actual rainfall station is 77%, and the correct probability of actual landslide is 100% with a satisfactory prediction effect. Based on the actual rainfall calculation data, the false alarm rate is 21%, but the actual landslide false alarm rate is 0, which is in line with the safety. Therefore, the model can be used to predict the occurrence probability of typhoon-rainstorm landslides under the influence of typhoon passing through Quanzhou City.