• 责任者：Jiao Fangqian (College of Geology and Mining Engineering ; Xinjiang University ; Urumqi 830047 ; China) ; Zhao Xinsheng
• 刊名：Arid Land Geography
• 出版年：2013
• 卷期：36(6)
• 关键词：泥石流 ; debris flows ; 证据权重法 ; evidence weight methods
• 页码：p.1111-1124
• 图表：17 illus., 10 tables, 17 refs.
• 分类号：1800
• issnNo：ISSN1000-6060
• isbnNo：CN65-1103/X

     Xinjiang is China＇ s debris flow prone and heavy disaster area, debris flow＇ s harmfulness and destruc tiveness is very serious. The nature to use spatial statistical method for debris flow quantitative susceptibility evalua tion is to measure the spatial relationship between impacting factors （altitude, rainfall） and the response factors （debris flow）, finally giving the results of all influencing factors＇ combined effects and getting the results of debris flow susceptibility evaluation. It contains two processes, the first step is to carry on the single factor analysis, which is to calculate the spatial correlation degrees between each impact factor and response factor, respectively; Based on the results of the first process, the second step is to choose the impact factors that have close spatial relationships with response factors to make comprehensive evaluation. At this time, the multiple synthesis of impact factors is often re quired to get the final evaluation results. Synthesis or complex often adopts some data integration model. This paper uses Weighted Evidences Model to evaluate the debris flow disaster susceptibility of the Tianshan Mountains and their foothills, selects altitude, slope, aspect, ground undulation, lithology, rainfall and water system distance as 7 im pact factors of debris flow. The terrain data is mainly composed by DEM data, DEM data obtaines slope, aspect, ground undulation by spatial analysis tools in the ArcGIS platform. Lithology data is collected from Bureau of Xinji ang Geology and Mineral Resources Development. Rainfall data is collected from the Xinjiang meteorological observation points. Water system distance data is the basic geographic data. The Weighted Evidences Model obtains the weighted values of evaluation indices by overlay analysing the evidence factor layers and debris flow distribution lay ers, objectively reflects the size of the influence of the evaluation index on the debris flow＇ s dangerous degrees, weighting method is relatively clear and intuitive. It makes the conclusions that the main influencing factors of the study area＇ s debris flow are altitude, slope, ground undulation and water system distance, the secondary influence factors is lithology, rainfall is not involved in the evaluation of the secondary influence factor, non-influential factor is aspect. With the aid of ArcGIS platform, this paper classifies the susceptible degrees of study area＇ s debris flow using natural discontinuity method and gets the study area＇ s debris flow prone probability graph, obtaines higher confidence degrees. The method is simple and easy with high reliability. In the final interpretation, an important as pect is to recognize the existence of uncertainty and to quantify the uncertainty. The reasons causing uncertainty mainly come from two aspects, variance in the weight calculation existing and the data in the predictive factors graph covering incompletely. The check of the uncertainty can be carried out through the variance calculation respective or combined. Using Weighted Evidences Model to evaluate debris flow proneness can also be generalized, we can eval uate the debris flow susceptibility of Xinjiang or China, and even the global area. It also can be generalized and ap plied in the susceptibility evaluation of other natural disasters such as landslide. This is the first application of Weighted Evidences Model and spatial analysis in Xinjiang＇ s natural disaster prone. Weighted Evidences Model is one of classical statistical methods, other statistical models can also be used to evaluate the susceptibility of natural disasters, such as the simple Weighted Superposition Model, Fuzzy Logic Model, Regression and Neural Network Model.