基于经验似然比函数模型的降水型滑坡灾害概率风险分析与预测  

Analysis and Prediction of Probabilistic Risk of Rainfall-induced Landslides Based on Empirical Likelihood Function Model

作　　者：胡爱军[1,2] 李宁[1] 吴吉东[1] 黄炜敏 温玉婷[1] 黄小玉 

机构地区：[1]民政部/教育部减灾与应急管理研究院地表过程与资源生态国家重点实验室,北京师范大学环境演变与自然灾害教育部重点实验室,北京100875 [2]湖南省气象局,湖南长沙410007 [3]湖南省地质环境监测总站,湖南长沙410007

出　　处：《灾害学》2009年第3期1-6,共6页Journal of Catastrophology

基　　金：国家科技支撑计划课题(2008BAK50B08,2006BAD20B01);国家自然科学基金项目(40771008)

摘　　要：以湖南郴州地区为例,建立了降水型滑坡历史灾害空间分布、地质和地形等滑坡影响因子空间数据库;基于50×50 m的网格,选取地貌、地层岩性、土壤类型、坡度、坡向、高程等滑坡影响因子,利用经验似然比函数计算出每个网格的滑坡危险度,并进行分级,进而制作出郴州地区滑坡危险度区划图。在此基础上,利用另一组滑坡灾害空间分布数据采取交叉验证法估算每一类危险级别发生滑坡的经验概率。结果表明,应用经验似然比模型进行滑坡灾害概率风险分析评估效果较好。Taking Chenzhou in Hunan as an example, a spatial distribution of historical rainfall-induced landslides and a spatial database of landslide causative factors as geology and topography etc. are established . Landslide causative factors, such as landform, formation lithology, soil type, slope degree, slope direction, elevation etc. are selected in 50 × 50 m grids, and landslide hazard degree of each grid is calculated and graded by using empirical likelihood function. Then, the landslide hazard degree zoning map for Chenzhou is worked out. Using the spatial distribution data of another group of landslides, landslide empirical probability of each risk grade is estimated by using cross validation method. The results show that using empirical likelihood ratio for landslide hazard probabilistie＇ analysis and assessment is more effective.

关 键 词：滑坡 降水 经验似然比 风险评价 交叉验证 湖南郴州 

分 类 号：P642.22[天文地球—工程地质学]