新疆果子沟区域是我国与中亚地区社会、经济、文化交流的重要通道以及我国石油、天然气等能源资源安全大通道,沿途滑坡灾害频发,威胁人类安全、影响社会、经济发展,但该区域针对滑坡灾害的研究程度较低,需借助滑坡灾害易发性分析与风险等级评估结果指导防灾减灾.本文基于GF-1号卫星影像数据进行滑坡解译,选取地层岩性、断层密度、坡度、坡向、地表高程、植被指数等6个评估因子,探讨运用GIS、RS技术及统计分析模型进行滑坡灾害易发性分析与建模.基于频率比法分析各因子敏感性,利用二元Logistic回归模型进行灾害易发性分析,将研究区滑坡灾害风险分为极低、低、中、高、极高五个等级.将模型计算结果与历史滑坡信息进行比较,并借助ROC Curve检验模型准确性,AUC为0.844,表明模型预测结果具有较高准确性,因此建立的分析模型可以满足新疆果子沟区域滑坡灾害分析与评估应用,研究成果可为研究区重大线型工程保护、边坡加固提供辅助决策支持.%As an entrance to link Central Asia for social,economic and cultural exchanges,Xinjiang Guozigou re-gion secure a route for transportation of oil,natural gas and other energy resources.Unfortunately,active landslide disasters along this route are endangering all aspects of local communities, including human security, local social and economic development.In the case that studies on landslide hazards relevant to this region are rarely done and have little to be achieved,it is necessary to motivate landslide susceptibility analysis and risk assessment over this area for the guidance of disaster prevention and mitigation.Based on satellite image interpretation method, this study collected 6 factors,namely rock,fault density,surface elevation,slope gradient,slope aspect,vegetation in-dex from GF-1 images,to integrate into GIS and RS application for construction of statistical analysis model for a scientific evaluation of landslide susceptibility.Based on frequency analysis,it was aimed to evaluate sensitivity of each factor,and then binary logistic regression model was employed to analyze geohazard potentials.According to the referred analysis,landslide risk was categorized into five levels as described as very low,low,medium,high, very high.This risk zoning result was further verified by historical landslide data, and AUC 0.844 was obtained with ROC Curve test model,justifying high accuracy of our prediction.In conclusion,our proposed analysis model can be used to guide landslide control in Guozigou region of Xinjiang and it would be contributed to engineering protection of major liner engineering at similar geo -environment circumstances.
展开▼
