| Geological disaster is a geological event that threatens the life and property of human beings and living environment seriously,and restrict the sustainable development of mankind.Yanchi county has severe geological disasters in Ningxia Autonomous Region.Taking geologic-al disaster in Yanchi County for research objects,relying on the Project of geological disaster detailed investigation in Yanchi County,based on RS and GIS technology the paper revealed the characteristics of geological disaster development,distribution law and geological condition-ns,analyzed the changes of vegetation and landscape pattern characteristic from 1989 to 2014(1989,1999,2006,2014),discussed the spatial-temporal variations of ecological environment quality and the cause during past 25 years employing remote sensing ecological index(RSEI)constructed by PCA method,analyzed the geological disaster susceptibility,hazard and risk assessment respectively with various models on three different levels quantitatively and comple-ted the assessment and zoning,explored the geological disaster risk management of study area preliminarily expecting to provide theoretical methods and technical support of geological disaster prevention and control in study area.Through these studies,the major progress and conclusions are obtained in this thesis as follows:(1)231 geological disasters which involved 125 landslides,84 collapses,19 debris-flows and 3 surface collapses were interpreted by researching RS images characteristics and building interpretation indication of geological disasters.According to regional geography and geological environmental characteristics,the basic feature of geological disasters points was summarized.The results show that geological disasters points mainly distributed in the town of Mahuang-shan,Dashuikeng and Huianbao,which show the characteristics of linear distribution along northwest-trending channel such as Kushuihegou,Lijiadawangou,Penjiagou,Shijiagedagou,Zhangnangou and Zengjiaqubeigou.Spatial distribution of geological disaster points showed the characteristics of regional aggregation,which point density is high in south and low in north.(2)The vegetation coverage was higher in the east and lower in the west and presented the characteristics of increase-decrease-increase.Correspondingly,the vegetation appeared the repetitive process of restoration-degeneration-restoration with the whole tendency of recovery.Landscape heterogeneity increased and landscape fragmentation degree was reduced.During 1989—2014,gravity center of vegetation patch moved mainly on the direction of NW-SE,and vegetation patches expanded in the southwest and northeast of the study area,but degraded in the northwest.The change of vegetation landscape pattern was the consequence of vegetation restoration measures and precipitation.(3)The mean value of RSEI is 0.41,0.54,0.51,0.56 in 4 years respectively,and the ecological environmental quality of study area showed the characteristics of rise-decline-rise,the overall ecological environment quality has improved apparently.The area of significantly improvement of ecological environment quality accounted for 20.55% of total area;howerver,the area of significantly deterioration of ecological environment quality just accounted for 1.43% of total area during 25 years.The precipitation has significant impact on change of ecological environmental quality of study area(4)Information Value(IV)+Logistic Regression(LR)model and CF+ Logistic Regression(LR)model were employed to evaluate the geological disaster susceptibility of Yanchi County.The results showed that the area under the curve for success rates are 0.868 and 0.829 respect-ively,and asymptotic Sig.b less than 0.05 for two models,which means two combined models both produced reasonable accuracy and IV+LR model has higher accuracy than CF + LR model.The area percentage of IV+LR model for high class accounted for 4.43%,which high class mainly distributed in Mahuangshan area and Huianbao.(5)Four evaluation models such as Analytics Hierarchy Process(AHP),Ordinary Least Square(OLS),Geographically Weighted Regression(GWR)and Support Vector Machines(SVM)were used to assess the hazard assessment respectively.The results showed that hazard density of 4 models all taken the characteristics of a continuing decrease in the density values from the very high to very low classes,and the area under curve(AUC)value of ROC of 4 models were 0.89,0.868,0.923 and 0.903 respectively,which means 4 models all produced reasonable accuracy and GWR model has highest accuracy.The area percentage of GWR model for high and very high class accounted for 18.24% and 6.78% respectively,which mainly distributed in the area of Mahuangshan with steep slope and two sides of rivers and roads.(6)The results of geological disaster risk shown that the area percentage of high class accounted for 17.64% and geological disaster density was 0.081,which mainly distributed in the area of southern Mahuangshan Dashuikeng,two sides of roads and area with relatively concentration of population.The area percentage of very high class accounted for 3.32% and geological disaster density was 0.454,which mainly distributed in the area of Mahuangshan with steep slope and in the area with relatively developed economy and relative concentration of population such as Mafang,Songbaozi,Liutiaojing,Xinquanjing,Baiyanjing and Hexinzhuang-Tangpingzhuang-Songjiashui. |
PDF Full Text Request