Study On Early Warning Of Rainfall-induced Landslide In Bazhou District

Landslide disasters have caused incalculable losses to the national economy,life and property because of their wide distribution,frequent occurrence,and their multiple,regional and serious characteristics.According to the statistics of the ministry of natural resources,in recent years,landslide disasters can account for more than 70% of the total number of geological disasters.In China,90% occurrence of landslide events are directly induced by rainfall or indirectly related to rainfall.The geographical location of Bazhou district,Bazhong city,has internal and external factors for landslide occurrence,and the climate belongs to the subtropical monsoon humid climate.The proportion of rainfall-induced landslides accounts for more than 70% of the total geological disasters.Therefore,this thesis conducts a study on landslide susceptibility and rainfall model in Bazhou district,so as to realize the rainfall-induced landslide warning in Bazhou District.This thesis constructs a machine learning algorithm to carry out the landslide susceptibility study in Bazhou district,establishes the rainfall intensity-rainfall duration model of Bazhou district to carry out the rainfall model research,and establishes the weather warning model of rainfall-induced landslide in Bazhou district based on the two research results.The main research work of this thesis is as follows:(1)Firstly,based on the geographic information system technology,combined with the historical landslide hazard data of Bazhou district,four machine learning algorithms are used to study the landslide susceptibility of Bazhou district.The results show that the overall accuracy of BP neural network classification is the highest,reaching 98.00 %,which is higher than other algorithms 2.00 ~ 6.00 %,the Kappa coefficient is 0.96,which is higher than other algorithms 0.04 ~ 0.12.The average accuracy of the three groups of test data is 95.33 %,which is higher than other algorithms 2.66 ~ 9.33 %,the average Kappa coefficient is 0.91,which is higher than other algorithms 0.06 ~ 0.19.(2)Secondly,based on the Tropical Rainfall Measuring Mission 3B42 rainfall product data,the rainfall intensity-rainfall duration model of Bazhou district is established,and the rainfall model parameters are obtained to predict the rainfall-induced landslide time in Bazhou district.The research results show that theprediction accuracy of rainfall-induced landslide is 81.82 %.Based on the established model,the landslide hazard prediction is carried out for potential landslides with an accuracy of 100.00 %.In summary,the landslide hazard prediction is carried out for rainfall-induced landslides and potential landslides,with a prediction accuracy of 90.91 %.Therefore,the rainfall intensity-rainfall duration model of Bazhou district established in this thesis has good reference value for rainfall-induced landslides and potential landslides prediction.(3)Finally,establish a probabilistic quantitative model for the disasters caused by geological disasters in Bazhou district,conducting meteorological early warning.Probabilistic quantification of geological factors uses experimental methods of landslide susceptibility research.And probabilistic quantification of meteorological factors uses experimental methods of rainfall model research,improving the quantitative method based on the relationship between the number of statistical landslides and the amount of rainfall.The research results show that the meteorological early warning model results are consistent with the actual inspection results.Therefore,the meteorological early warning model of this thesis can be used as a reference model for rainfall-induced landslide warning in Bazhou district.