基于GIS和多变量决策树的地震滑坡道路中断风险应急评估模型  

An Emergency Evaluation Approach to Road Disruption Risk Influenced by Earthquake-induced Landslide Using GIS and Multivariate Decision Tree

作　　者：白仙富 杨志全[2] 罗伟东 王杰 田鹏 戴雨芡 BAI Xianfu;YANG Zhiquan;LUO Weidong;WANG Jie;TIAN Peng;DAI Yuqian(Kunming Institute of Earthquake Prediction,China Earthquake Administration,Kunming 650224,Yunnan,China;Faculty of Public Safety and Emergency Management,Kunming University of Science and Technology,Kunming 650093,Yunnan,China;Earthquake Administration of Xishan District,Kunming 650118,Yunnan,China)

机构地区：[1]中国地震局昆明地震预报研究所,云南昆明650224 [2]昆明理工大学应急管理学院,云南昆明650500 [3]昆明市西山区防震减灾局,云南昆明650118

出　　处：《地震研究》2023年第3期343-353,共11页Journal of Seismological Research

基　　金：中国地震局地震科技星火计划攻关项目(XH222509C);云南省重点研发计划(202203AC100003);中国地震局地质研究所中央级公益性科研院所基本科研业务专项(IGCEA2106).

摘　　要：在地震应急阶段,如何定量评估地震滑坡道路中断风险是一项亟待完善的关键技术。为解决这一难题,以2008年汶川M_(S)8.0地震、2014年鲁甸M_(S)6.5地震及2012年彝良M_(S)5.7、5.6地震为案例开展地震滑坡道路中断风险应急评估模型的构建和检验。汶川研究区用来建立地震滑坡道路中断风险多变量决策树的应急评估模型,并对模型作有效性评价,鲁甸和彝良研究区用来对所建模型开展相似区域外延适用性的评价。通过P值检验模型统计学的显著性,使用Kappa值评价模型推断结果与实际情况的一致性。汶川研究区的P值为2.52×10^(-203),Kappa系数为0.91。说明使用模型计算出的道路中断风险是地震滑坡道路是否中断的良好指标。鲁甸和彝良研究区的P值为9.7×10^(-107),Kappa系数为0.81。这表明在允许一定误差的情况下,本研究建立的地震滑坡道路中断风险多变量决策树应急评估模型可以推广应用到其它类似地区。In mountainous areas,roads are often damaged by earthquake-induced landslides.The degree of road damage and the existing functional state will have a very important impact on the whole earthquake emergency response.To solve this problem,in this paper we propose a new modle.To test our new model,we select four historical earthquake events for case study:the 2008 Wenchuan,Sichuan M_(S)8.0 earthquake,the 2014 Ludian,Yunnan M_(S)6.5 earthquake,and the 2012 Yiliang,Yunnan M_(S)5.6&5.7 double earthquakes.Here,the Wenchuan earthquake case is for testing the effectiveness of our new model.The other three cases serve for testing the practicability of our model for the future earthquake disasters emergency assessment in other places.P-value and Kappa coefficient are used to examine the model s statistical portability.The computed P-value for the Sichuan area was less than 0.001(just 2.52×10^(-203))and the Kappa coefficient is 0.91.This suggests that the computed RIR values are good indicators of blocked-road occurrences.In the Yunnan study areas,the computed P-value is less than 0.001 too(just 9.7×10^(-107))and the Kappa coefficient is 0.81.The results from the Yunnan study areas suggest that the Multivariate Decision Tree can be introduced to another study areas and our approach can be used in other earthquake events.

关 键 词：道路中断风险 地震滑坡 多变量决策树 应急评估 

分 类 号：P315.94[天文地球—地震学]