China has been greatly affected by natural disasters, so that it is of great importance to analyze the impact of natural disasters on national economy. Usually, the frequency of disasters or absolute loss inflicted by disasters is the first priority to be considered, while the capability of regions to overcome disasters is ignored. The concept of vulnerability is used to measure the capability to overcome disasters in different regions with distinctive economies. Traditional methods for vulnerability analysis calculate sub-indices based on disaster frequency, loss, the economic impact and the population of each region, and then add the sub-indices to get a composite index for regional vulnerability. But those methods are sensitive to the weights selected for sub-indices when multi-indexes are added up to get an index of total vulnerability. The analytic results are less convincing because of the subjectivity of different weighting methods. A data envelopment analysis (DEA)-based model for analysis of regional vulnerability to natural disasters is presented here to improve upon the traditional method. This paper systematically describes the DEA method to evaluate the relative severity of disasters in each region. A model for regional vulnerability analysis is developed, based on the annual governmental statistics from 1989 to 2000. The regional vulnerabilities in China's mainland are illustrated as a case study, and a new method for the classification of regional vulnerability to natural disasters in China is proposed. (C) 2004 Elsevier Inc. All rights reserved.
