综合灾情指数--一种自然灾害损失的定量化评价方法

摘要

Assessments of natural disaster losses are related to various aspects,such as population,crops, houses,and economics,which are typical multi-variant analysis issues.How to design an overall and reasonable assessing index is of crucial importance to disaster risk management,loss assessment and policy making.Current approaches are either sensitive to data of biased measurements or not applicable to simultaneously temporal-special comparisons.In this paper,we propose a new computational model of natural disaster index (NDI)based on geo-metric means.Compared with state-of-the-art,the proposed one can assess the losses of a region quantitatively a-long time and space at the same time,and it is robust to multi-variant data which has biased distributions of meas-urements.Simulations conduced on the real loss data of the 201 3 floods in China validate the effectiveness of the proposed NDI method.%自然灾害损失评价涉及人口、农业、房屋、经济等多项灾情指标，是典型的多元指标综合分析问题。如何设计合理的定量化评价指数，对灾害风险管理、损失评估和相关政策制定，具有重要意义。现有灾情评价方法和灾情指数，或对量纲差异较大的灾情指标敏感，或无法同时对不同区域的灾情时空分布进行比较，缺少综合性、兼顾时空分布的评价方法。针对上述问题，提出了一种新的基于几何平均的综合灾情指数计算方法。相较现有方法，该文提出的指数可同时在时间和空间两个维度对给定区域的灾害损失进行定量化的评价，且对指标分布偏倚的多元灾损数据不敏感。在2014年我国分省洪涝（含地质灾害）损失数据上的计算结果验证了新指数的有效性。