| This article takes Zi Gui Country as the research area to analyze the types,characteristics,and distribution patterns of geological disasters in Zi Gui Country.Using fractal theory,quantitatively analyze and evaluate the impact of factors on the distribution of geological hazards,and screen out the main factors that affect geological hazards based on the correlation between fractal deformation results and various factors.Based on Arc GIS and SPSS software,multiple models were used for susceptibility assessment,and results were compared and accuracy validation was conducted.The conclusions are as follows:(1)There are 1193 geological disasters developed in the study area,of which 913 are landslides,269 are collapses,10 are ground collapses,and only 1 is debris flow.Landslides and collapses are the main disasters in the area;The disaster points along the Yangtze River and its tributaries are relatively concentrated;The landform is mainly composed of low mountain and hilly areas,and medium to low mountain areas;The townships with over 100 geological disasters include Guizhou Town,Guojiaba Town,Jiuwanxi Town,Shazhenxi Town,and Shuitianba Township;From May to September,there is a high incidence of geological disasters.(2)Analyzing the spatial distribution relationship between each factor classification and geological hazards,the conditions prone to geological hazards are: <1200m elevation section;slope 10°-30° is more concentrated;west,southeast and north slope direction;less than 3km to faults;muddy sandstone,sandstone type and carbonate type;less than 1500 m to water system;955-1020 mm rainfall section;vegetation NDVI value is 0.7-0.9;less than 400 m from the road;land use type is woodland,shrubland and grassland.(3)Based on the fractal theory,the fractal characteristics of the classification of each factor and the spatial distribution density of disaster points were analyzed.,and quantitatively analyze the contribution of each factor to the distribution of disaster points.According to the variable dimension fractal results,11 factors such as elevation show variable dimension fractal characteristics,indicating that the distribution of disaster points is affected by these factors.Based on the order of cumulative sum transformation and the absolute value of fractal dimension D,the contribution degree relationship is obtained: fault distance > slope > road distance > engineering rock group > vegetation NDVI > annual rainfall > elevation > water system distance> land use type> slope type> slope aspect.(4)Based on the results of variable dimensional fractal analysis and combined with the validation of the independence of various factor conditions,vegetation elevation,NDVI,and slope aspect factors were selected to be excluded.Finally,a total of 8 factors,including slope,slope type,fault distance,engineering rock group,water system distance,annual rainfall,road distance,and land use type,were used to participate in the construction of the model.(5)Based on the training set samples of geological hazard points,the susceptibility assessment was conducted using the information model(IM),deterministic coefficient model(CF),information logistic regression model(IM-LR),and deterministic coefficient logistic regression model(CF-LR).(6)Based on the validation set of geological hazard points,frequency ratio and ROC curve are used to verify the results and accuracy of susceptibility zoning.The AUC values of all four models are greater than 0.8,indicating high accuracy.The CF-LR model has the highest AUC value of 0.881,indicating the highest evaluation accuracy.Comparison of the accuracy of four models: CF-LR > IM-LR > CF > IM.Based on the optimal model CF-LR,it is divided into low susceptibility areas,medium susceptibility areas,high susceptibility areas,and extremely high susceptibility areas using the natural breakpoint method. |
PDF Full Text Request