摘要
三江并流区是一个多民族聚居区,也是我国著名的旅游景区,同时又是典型的生态脆弱区和滑坡泥石流灾害多发区。本文采用聚类分析方法评价了三江并流区泥石流的危险性,这种非监督分类的方法适用于没有足够训练数据的情况,可以快速对一个区域的泥石流危险性进行评估。泥石流危险性评价指标选择了能够反映三江并流区泥石流灾害主控因素的8个影响因子:地形坡度、坡向、植被、地层岩性、断裂带密度、河流切割密度、降雨量和道路网密度。聚类分析算法采用k-means聚类方法,将整个研究区域划分为四个子类,分别对应极低、低、中等和高危险区,其评价结果与研究区泥石流灾害分布的点密度值相吻合。从危险性评价图可以发现,即使在极低危险区域附近,可以看到有高危险区域的存在,因此,这一分区图可以用来寻找危险区中的防灾避灾安置点(即小块的低危险区);从危险性分区图中还可以发现一点,在高危险区域和中等危险区,以及高危险区与极低危险区分界线附近,泥石流灾害点分布比较密集。
The Three Parallel Rivers area is not only the concentrated area inhabited by minority people and a famous tourist hotspots, but also an ecological fragile area and disaster-prone （landslide and debris flow） area in our country. Especially debris flow disaster is threatening the social and economic development in the Three Parallel Rivers area. Therefore, hazard assessment of debris flow disaster has the important meaning in revealing the ecological environment problems, establishing regional environment construction policy and development strategy in the Three Parallel Rivers area. Cluster analysis and maximum likelihood classification （MLC） are exploited to map the hazard assessment of debris flow in the Three Parallel Rivers area. The methodology is applicable even if there is short of training data. Debris flow hazard can be viewed as a kind of spatial probability of hazard occurrence at a certain terrain unit and usually calculated as a function of effective factors or predictor variables such as lithology, fault, slope gradient, land-use type etc. The hazard assessment factors of debris flow can be divided into three types： topography conditions, the sources of loose solid material and dynamic conditions. Based on the field investigation and indoor analysis, eight effective factors are chosen for mapping the hazard assessment, including slope gradient, aspect, vegetation, lithology, fault zone density, river cutting density, rainfall and road network density. Because each measurement unit of effective factors is different, and its value is also belonging to different extents, therefore, we must deal with the above eight factors in unified handling. Four clusters are grouped from sampling grid cells by k-means clustering approach using above eight effective factors. MLC classifies all the cells in the study area into the four clusters according to their statistical characteristics. Four hazard classes （extreme low, low, moderate, and high） are assigned to these clusters applying expert experience and hazard density. The final map gives a reasonable assessment of debris flow hazard in the Three Parallel Rivers area. The result using cluster and MLC classification has a better agreement with the dot density value of debris flows in the Three Parallel Rivers area. The hazard assessment map of the Three Parallel Rivers area can be used to identify safety spots within the high danger area, which are suitable for habitations and facilities. It is also found that more debris flows are densely concentrated at the boundary between high and moderate regions, and between high and extreme low regions.
出处
《资源科学》
CSSCI
CSCD
北大核心
2012年第7期1257-1265,共9页
Resources Science
基金
国家科技基础性工作专项项目:"澜沧江中下游与大香格里拉地区山地灾害考察"(编号:2008FY110300-06-1)
国家自然科学基金项目:"岷江上游山区聚落易损性及其灾变阈值"(编号:41101164)
四川省科技厅项目:"基于RS和DEM的震后泥石流沟危险度评估研究"(编号:2010SZ0308)
关键词
三江并流区
危险性评价
泥石流
聚类分析
Three Parallel Rivers area
Hazard assessment
Debris flow
Cluster analysis
作者简介
作者简介：丁明涛，男，山东日照人，博士，副教授，主要从事地质灾害风险评估及其预测预报研究。E-mail：mingtaoding@163．com
