Spatial Evolution And Susceptibility Assessment Of Geological Disasters In Wenchuan County In The Past Decade After 5.12 Earthquake

Wenchuan County is located near the Longmenshan fault zone.This area is one of the most severely affected areas of the “5·12 Wenchuan Earthquake” in 2008.The geological structure in the area is complex and geological disasters occur frequently.After nearly ten years of natural recovery and government management,most of the geological disasters in the region have become stable,However,some geological disasters still pose a great threat to the safety of people's lives and property.This paper selects Wenchuan County,one of the most severely affected areas of the Wenchuan earthquake,which was finally announced by the Ministry of Land and Resources as a research area,analyzes the spatial variation of geological disaster distribution in the past ten years,and evaluates the regional geological hazard susceptibility to the existing disaster points.This research has important practical significance and can also provide a reliable reference for the prevention and control of geological disasters in the study area.The paper uses Landsat remote sensing image data,combined with post-earthquake disaster interpretation data,to extract information on geological disasters in the study area in 2008 and 2017.The spatial variation law of geological disasters in the study area in the past ten years is analyzed,and the main factors affecting the occurrence of geological disasters are obtained.Based on the deterministic coefficient method combined with the information model and the logistic regression model,the 2017 geological disaster susceptibility index of the study area is calculated,and the most suitable model is selected to evaluate the geological hazard susceptibility in the study area.The main workload of this paper is as follows:(1)Geological hazard information extraction.This paper selects Landsat data in2008 and 2017,combined with improved normalized differential water index,normalized vegetation index,improvement normalized building index,Impervious surface index and bare soil index formula to extract two phases of bare landinformation.The interpretation results of the geological disaster points of the high-resolution images after the earthquake are superimposed with the bare land information in 2008,and the point-to-face matching degree is as high as 99.31%.It can be concluded that the reliability of the geological hazard points interpreted in that year is relatively high and can be used for follow-up research,and the bare ground covered by the disaster points is removed to obtain the distribution range of post-earthquake geological disasters.Then the bare land information extracted from the remote sensing images in 2017 is superimposed with the previous geological hazard points,and the geological hazard points and their distribution range that still exist in the past ten years after the earthquake are screened out.(2)Analysis of spatial evolution of geological hazards.The paper classifies and compares the two geological hazard points by density.The results show that the proportion of low-density areas in geological disasters is increasing,and the distribution is relatively concentrated;the proportion of medium-and high-density areas is decreasing,showing a strip-like distribution.Comparing natural factors and human factors with geological disaster density and counting their relationships separately.Finally,the main factors affecting the distribution of geological disasters in the study area are obtained.(3)Spatial prediction of geological disaster susceptibility.By summarizing the temporal and spatial distribution law of geological hazards in the study area,we select suitable evaluation indicators for digitalization,use deterministic coefficient method to improve the index quantification system of information quantity model,and get the sensitivity value of each grade evaluation index.The susceptibility index of geological hazards in the study area is calculated by using information model and logistic regression model respectively.The digital results of the two models show that their spatial susceptibility trends are consistent as a whole.Among them,the high susceptibility is mainly distributed in the areas where human activities are concentrated and rainfall is large,while the non-favorable areas are mainly distributed in areas with higher altitude and better vegetation coverage.The results of the information model and the logistic regression model were verified by the ROC curve and the volatility partitioning method respectively.The verification results show that the logistic regression model is superior to the information model in the evaluation of geological hazard susceptibility.