| Hydropower development in the lower reaches of the Yarlung Zangbo River has been explicitly proposed in the 14 th Five-Year Plan for national Economic and social Development and the 2035 vision.The lower reaches of the Yarlung Zangbo River contain abundant water energy due to its huge difference in land potential and rapid current,so there is a large space for resource development.However,the lower reaches of Yarlung Zangbo River are high in the east and low in the west,with abundant hydrological and climatic conditions,abundant rainfall and active fault zones.The natural geographical environment is conducive to breeding debris flow disasters,and rainfall debris flow is the frequent natural geological disaster in this region.In order to avoid debris flow disaster to the lower reaches of the Yarlung Zangbo River in the future damage to construction facilities,reasonable removal of safety risks,it is necessary to systematically analyze the vulnerability of debris flow in the lower reaches.In this paper,the area within 2km width from both sides of the lower reaches of the Brahmaputra River is selected as the research object.By collecting 84 debris flow points,selecting 11 disaster factors and analyzing satellite elevation data,the susceptibility of debris flow in the study area was evaluated.The information of disaster factors was extracted from Arc GIS platform and analyzed.The certainty factor model(CF)was used to calculate the deterministic coefficients of different levels of 11 disaster factors,including fault distance,elevation,slope aspect,vegetation coverage rate,average annual rainfall,profile curvature,relief degree,silt content,TWI,SPI and land use rate.The models of analytic Hierarchy Process(AHP),binary logistic regression(LR),random forest(RF)and CF coupling were respectively used to analyze and predict the occurrence of debris flow.Confusion matrix and ROC curve were used to verify the accuracy of the three models,and the coupling model with the highest accuracy was selected to calculate the evaluation map of debris flow susceptibility in the study area.Rainfall is an important factor to induce the eruption of debris flow.Different levels of annual rainfall are used to find the regularity of rainfall and debris flow occurrence.The main conclusions are as follows:(1)Through the CF model,it can be concluded that: if the study area is 12000-14000 m away from the fault;The altitude is 2900 to 3100m;NDVI is 0.4-0.6;Profile curvature is 0-5;Fluctuation is 0-10;The silty sand content is 30%-40%;TWI at 10-12;In the geological and geomorphologic conditions where SPI is-4-0 and slope direction is located in the northwest,debris flow is most likely to occur.(2)It is found that the index of the CF-LR coupling model is better than the other two models under the verification of the confusion matrix.According to the AUC value of ROC curve chart,the AUC value of CF-LR model is 0.884,which is the largest among the three coupling models,indicating the highest accuracy.Therefore,the coupling model of deterministic coefficient and logistic regression is the most suitable for this study.(3)Under the GIS platform and based on the CF-LR model,the low-prone area accounted for 54.57% of the study area.Said the study area was relatively safe.The proportion of high-prone areas was 33.76%,and the proportion of mildly prone and medium-prone areas was not large.The danger areas are mainly located in the Grand Canyon and the basins before it,and are concentrated in the areas with large turning tendency of the river and the areas with multiple turns of the river.(4)The CF values of four levels of rainfall were introduced into the model respectively,and the annual rainfall threshold of debris flow in the study area was found to be in the range of 600-700 mm.When the annual rainfall is 700-800 mm,the debris flow is concentrated.After the annual rainfall of 800 mm,the proportion of low-prone areas gradually increased,while the proportion of high-prone areas gradually decreased. |
PDF Full Text Request