Research On Early Warning System Of Highway Landslide In Nanchong City Induced By Heavy Rainfall

Highway landslides are very common in mountainous and hilly areas,which seriously affect the safety of people’s travel and local economic development.Taking the highway landslide disaster induced by heavy rainfall as the research object,this paper researched the highway landslide prone zoning and the highway landslide early warning model in Nanchong City,and developed the highway landslide disaster early warning and prediction system,which provides decision support for the relevant departments for the highway landslide disaster early warning and prediction.The main research contents are as follows:(1)Based on the statistical analysis of the historical landslide disaster in Nanchong City,from the four aspects of geological factors,human factors,induced factors and other factors,selected elevation,slope,precipitation from June to September,river influence distance,road influence distance,population density distribution,normalized difference vegetation index,geological lithology,soil type and land use type to complete the south of Nanchong City by combining the historical landslide disaster.According to the ROC curve,the result of the evaluation of the landslide susceptibility was tested.The AUC value was 0.724,which shows that the prediction effect was good.On the basis of the above,the main highway landslide hazard prone zoning in Nanchong city was completed by superposition analysis,and the results of the zoning were verified by using the hidden danger points of Lanhai highway landslide hazard.(2)Based on the historical landslide disaster data and precipitation data,the relationship between the frequency of landslide disaster and the precipitation intensity and the occurrence time of maximum precipitation was explored.The results showed that the frequency of landslide disaster decreased with the delay of the maximum precipitation.The greater the intensity of daily precipitation,the smaller the time lag of rainfall-induced landslide;The occurrence frequency of landslide disaster had the best correlation with the accumulated precipitation in the previous two days.(3)Based on 70%historical landslide disaster data,the probability prediction model of highway landslide disaster induced by precipitation was established by binary logistic regression equation,and the prediction model was verified by 30%landslide disaster point data.The accuracy of the model was 84.0%,and the prediction model was reliable.On this basis,combined with the result of probability prediction of landslide susceptibility,an early-warning model of geological and meteorological coupling was constructed based on the binary logistic regression equation,and the verification showed that the early-warning accuracy was 83.7%.The model was applied to the early warning of road landslide disaster in Nanchong city,and the accuracy was 84.5%.(4)Using GIS and spatial analysis technology,database technology,network technology and other key technologies to design and implement the Nanchong highway landslide disaster early warning system,the system included a landslide disaster grid early warning processing background service subsystem and landslide disaster early warning release service subsystem.Realized the regular release and webpage display of highway landslide warning and regional landslide warning,and put it into business operation.