Font Size: a A A

Research On Precipitation-induced Regional Landslides Spatio-temporal Analysis And Comprehensive Assessment

Posted on:2016-01-08Degree:MasterType:Thesis
Country:ChinaCandidate:L ShengFull Text:PDF
GTID:2310330512475332Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
Landslide is one of the main types of geologic hazards in Fujian Province,which features large areas of damage,simultaneity,random,and complexity.Therefore,it is necessary to assess the landslide risk,to protect people and their properties,and to reduce the economic casualties.Based on the rich practical data,the paper makes the landslide susceptibility assessment at first.Then rainfall is imported as triggering factor,the paper describes an inversion model to determine the threshold expression of rainfall type regional landslide and the landslide probability of occurrence in order to make a strategic decision for landslide hazard warning.The main approaches and conclusions of the research are as follows:(1)The paper proposes correlation analysis to solve the problem of factor selection in the study area(Fujian province),and makes statistical analysis of the relationship between landslide hazard and factors;By combining the Analytic Hierarchy Process(AHP)and T-S Fuzzy Neural Network model(T-S FNN),the landslide susceptibility assessment of Fujian province is evaluated.The method not only solves the data layer of T-S Fuzzy Neural Network model initial weight but also improves the accuracy and efficiency of the model.Research shows that impact factors include slope,aspect,elevation,distance to fault,distance to river,distance to road and NDVI.Based on the T-S FNN model,the landslide susceptibility in Fujian Province is divided into 4 grades,namely high susceptibility,middle susceptibility,low susceptibility and none susceptibility.Inspection is carried out on the Mean Absolute Error(MAE)and Root Mean Square Error(RMSE).The result indicates that the evaluation result has high precision.(2)Based on the previous landslide susceptibility mapping,this dissertation further investigates the landslide spatio-temporal modeling by importing rainfall as the triggering factor.In this paper,Nanping City,the city of Fujian Province is chosen as the research target.The time-space relationship between the landslide and rainfall in this area is studied in this area.The expression of landslide threshold in the rainfall-induced region and the probability of occurrence of landslide are inversed based on logistic regression model.So that people can analyze the scale of landslide hazard in the study area in such rainfall situation,and make a strategic decision for landslide hazard warning by combining with the landslide-prone evaluation results.Research shows that the occurrence of landslide in the study area is relevant to four-day rainfall,which is the rainfall on that the same day and in the previous three days.Moreover,rainfall which happened on landslide day plays the most important role in the occurrence of rainfall type landslide hazards.The expression of landslide threshold in the rainfall type region landslide of study area is as follows:R0=-0.5052RL3 +95.227Considering four-day rainfall,the expression of the landslide probability of occurrence is as follows:(?)Based on the precipitation-induced landslides occurring in the history in Nanping City,exactness is 78.73%.
Keywords/Search Tags:Precipitation-induced Regional Landslide, Susceptibility Assessment, T-S Fuzzy Neural Network, Logic Regression Model, Rainfall Threshold
PDF Full Text Request
Related items