Research On Assessment And Pre-evaluation For Typhoon Disasters

Typhoon is one of a main natural disasters, it has caused large economic losses in China’s coastal area every year. Although a lot of research work about the typhoon disaster assessment and risk assessment has been carried out, in business applications, there is yet no mature method to assess typhoon disasters and the accuracy of assessment also needs to be further improved. Therefore, the establishment of effective typhoon disaster assessment and forecast method is still a very difficult task. So some researches on assessment and pre-evaluation of typhoon disasters have been carried out in this paper.(1) Factor analysis for disaster assessment of the typhoon. In order to overcome the shortage of human subjectivity in the selection of assessment indices, selection method of assessment indices has been studied. From disaster-causing factors, hazard-bearing factors and factors for disaster prevention and mitigation capabilities, taking Zhejiang Province as a example,25 factors were selected to analyze their impact on the number of deaths, the number of collapsed houses, affected areas of the crops and direct economic losses. The key factors which impact on disaster assessment were determined by correlation analysis method from 25 factors. These indices were validated by regression analysis.(2) Comparison of methods for assessment of typhoon disaster. There are two types of methods for assessment of typhoon disaster. One is indirect assessment which is carried by disaster-causing factors, hazard-bearing factors and factors for disaster prevention and mitigation capabilities; the other is direct assessment which is carried by disaster factor directly. As a comparison, taking Zhejiang province as an example, indirect assessment was carried by, direct assessment was carried by grey relation analysis method and fuzzy math method, then their results obtained by indirect assessment and direct assessment were compared.(3) Pre-evaluation model of the typhoon disaster based on the meteorological factors and the index of disaster prevention and mitigation was established. Firstly, we use grey relational analysis method to compute relational degree of comprehensive disaster according to the number of deaths, the number of collapsed houses, affected areas of the crops and direct economic losses, the correlation between meteorological factors and relational degree of comprehensive disaster was verified by correlation analysis. Secondly, grey relational analysis method was used to obtain the index of disaster prevention and mitigation according to statistical data which is closely related to disaster prevention and mitigation capabilities. Lastly, a pre-evaluation model of the typhoon disaster was established by the BP neural network., and an experimental was conducted to predict the level of comprehensive disaster with independent samples. The result indicates that the proposed method can be devoted to the pre-evaluation of typhoon disasters.(4) Pre-evaluation model of the typhoon disaster for the direct economic loss of Zhejiang Province was established. After analysis of the correlation between direct economic loss and disaster-causing factors, hazard-bearing factors and factors for disaster prevention and mitigation capabilities, there are 25 factors were found to be related to direct economic loss of the typhoon disaster. Considering the coupling between the various factors and nonlinear, as well as few sample data, a pre-evaluation model of direct economic loss was established by using the Generalized Regression Neural Network. An experimental was conducted to predict direct economic loss and the predictive values were compared with the real values with two independent samples. The result of comparison indicates that the proposed method can be devoted to the pre-evaluation of direct economic loss.