摘要
基于历史灾害数据的灾害风险分析与评价,采用信息扩散技术弥补历史灾情数据样本小、资料不充分的缺陷,运用投影寻踪聚类模型对评价指标进行客观赋值、综合评价。以贵州省近12年雷电灾害事故为样本、不同等级雷灾风险为评价指标,对全省雷电灾害风险进行评价。结果表明:毕节发生特重大、重大、较大雷灾事故的风险均高于其他市州,一般雷灾事故高发区出现在贵阳,≥45次的风险为0.0506,约为20a一遇;进一步进行综合灾害风险评价,应用自然间距断点法进行风险等级划分,毕节、黔南、铜仁属雷电灾害极高风险区,贵阳、六盘水、黔西南属高风险区,安顺、黔东南属中风险区,遵义属低风险区。
Disaster risk analysis and evaluation based on historical data not only compensates for such defects as limited samples and insufficient information of historical disaster data with the information diffusion technology, but also deals with the evaluation index in an objective and comprehensive way with the projection pursuit classification model. This article takes 12 years' lightning disasters in Guizhou province as samples and lightning disasters of different grades as evaluation index to evaluate the lightning disaster risk of the whole province. The result indicates that extraordinarily serious, serious and relatively serious lightning accident risk in Bijie is higher than that in other cities or prefectures. The high incidence area of general lightning accident is in Guiyang, where the risk of greater than or equal to 45 times is 0. 0506, about once in 20 years. A further comprehensive disaster risk assessment with the natural spacing breakpoint method for grading finds that Bijie, Qiannan, Tongren are extremely high lightning risk areas, Guiyang, Liupanshui, Qianxinan are high risk areas, Anshun and Qiandongnan are medium risk areas and Zunyi is a low risk area.
出处
《防灾科技学院学报》
2015年第4期26-31,共6页
Journal of Institute of Disaster Prevention
关键词
投影寻踪聚类
信息扩散理论
灾害重现率
风险评价
projection pursuit classification
information diffusion theory
disaster return rate
risk evaluation