Research On Prediction Method Of The Typhoon Disaster Loss And Development Of Decision Support System

Natural disasters cause different levels of life and property loss to our country and people every year. Typhoon, as one of the biggest influential risk natural disasters becomes one kind of disaster which affects the global coastal areas of our country mostly and has serious loss. As an example, Zhejiang province is one of the coastal provinces and cities which are seriously affected by typhoon disaster. What previous scholars research on typhoon is mainly inclined to the typhoon’s own moving rule, path characteristics, disaster-causing danger assessment, situation of disaster assessment and rank division etc, however, they do less research on absolute loss of typhoon disaster, otherwise, researches about typhoon disaster-causing strength and the relationship between hazard-affected body vulnerability and loss of typhoon disaster are not comprehensive and deep. As a result, this article do some research on typhoon disaster of Zhejiang province as the following several aspects:First, from the perspective of system theory, various selected factors influencing the vulnerability of hazard-affected bodies in Zhejiang province such as population, agriculture, buildings, lifeline engineering and so on, scientific and reasonable vulnerability index system of hazard-affected body is established, and major impact typhoon data from1949to2007in Zhejiang province is collected. Using the AHP to weight the evaluation indexes, using fuzzy discernibility method to establish a comprehensive evaluation model of vulnerability of hazard-affected body, make quantitative evaluation of vulnerability of hazard-affected body for Zhejiang province. Through the model evaluation results, the reasons of the impact of the typhoon of different years in Zhejiang province are analyzed, which has practical guiding significance to the disaster prevention and mitigation work in Zhejiang province.Second, through the correlation analysis of typhoon disaster losses, the typhoon disaster-causing factor intensity and the vulnerability of hazard-affected body. determine the above has a great influence with typhoon disaster losses. Nonlinear regression model is used to quantitative analysis of the relationship among the typhoon disaster-causing factor intensity, the vulnerability of hazard-affected body, the direct economic loss rate, rate of farmland flooded area, and the rate of collapsed houses disaster loss to explore the causes of the formation of typhoon disasters in Zhejiang province.Third, through PCA method to make reduction dimension analysis for multiple aspects of vulnerability of hazard-affected body, and the resulted three principal components combined with the typhoon disaster-causing intensity as input of RBF neural network. A prediction model is established to predict for various typhoon disaster losses. The experimental results show that the PCA-RBF neural network prediction method can improve the prediction accuracy greatly, and achieve predictive result, which provides a reliable reference for the typhoon disaster loss prediction.Fourth, Typhoon of Zhejiang province decisive and supportive system based on WEBGIS is developed. The system consists of typhoon track, typhoon risk, situation of risk assessment, typhoon disaster assessment, disaster forecast, meteorological data collection management and personnel transfer, etc.Based on GIS spatial database, supported by the mathematical model of evaluation and prediction, and demonstrated by the method of data simulation, the effect of decision support of disaster prevention and mitigation is achieved, finally develop the system.