Debris Flow Risk Prediction And Simulation In The Alpine Valley Of Western Yunnan

Debris flow as a geological disaster with strong destructive power has always threatened the safety of people's lives and property.The alpine valley area of Western Yunnan Province is one of the most serious debris flow hazard area in China.Due to the inconvenience of monitoring,difficulties in field surveying,etc.,debris flow hazard prediction work is difficult to carry out.The government and scientific research institutions are actively taking various protective measures to mitigate the damage caused by disasters.In order to improve the accuracy of the prediction and early warning work of debris flow disasters in the alpine valley area of western Yunnan Province.This paper firstly analyzes the features of disaster debris flow in Nujiang Lisu Autonomous Prefecture by selecting a more complete pregnancy risk factor.And then this paper uses the combined information method and analytic hierarchy process,logistic regression method and artificial neural network method to comply the risk prediction of debris flow in study area.Finally,based on the secondary development technology of ArcGIS,the debris flow prediction simulation platform of Nujiang Lisu Autonomous Prefecture was constructed to provide users with regional visualization maps,online prediction of debris flow hazard and other services.The main contributions of this paper are as follows:(1)Due to the debris flow pregnant factors are extremely complexity in the alpine valley area of western Yunnan Province.This paper choose five aspects: geographical environment,geology,climate hydrology,topography,and human activities.That include 12 prediction factors such as landform,soil,vegetation,NDVI,lithology,fault,river distribution,precipitation,slope,aspect,land cover and road distribution.Based on the above-mentioned 12 factors,the features of the debris flow in the study area are comprehensively mastered,which lays a foundation for constructing the debris flow risk prediction system.(2)This paper establishes a debris flow risk prediction system by combining information method and analytic hierarchy process,logistic regression method and artificial neural network method.Firstly,the system was constructed by randomly selecting 80% of debris flow disaster and non-disaster points data as training samples,and the remaining 20% of the data was verified as test samples.Secondly,the ROC curve is used to verify the effectiveness of the three methods in the prediction of debris flow hazard in the alpine valley area.(3)This paper based on the above three debris flow hazard prediction systems,the basic process of software development and ArcGIS secondary development technology,a debris flow prediction simulation platform for Nujiang Lisu Autonomous Prefecture was constructed.The simulation platform completed the basic operation of the map,the management of the debris flow data and the pregnancy factors data,the input of the weights of the prediction factors,the prediction and early warning functions,etc..This satisfies a number of business needs,such as predictive queries by users of mudslide disasters in the area.