作　　者： ; ; ;

机构地区： 暨南大学经济学院

出　　处： 《数理统计与管理》 2014年第1期31-41,共11页

摘　　要： 该研究致力于建立台风路径模拟模型,目的在于提高风险评估的稳健性和实用性。首先,采用中国台风网提供的CMA-STI热带气旋最佳路径数据集,以广东省为研究区域,通过引入空间统计模型,结合马尔可夫—蒙特卡罗方法模拟了西北太平洋的热带气旋路径。进一步,参数和非参数假设检验结果表明模拟结果是合理的,可以作为风险评估的数据来源。在此基础上,本研究模拟出620年的台风路径,给出了以地级市为基础的重现期估计和风险图绘制方法。研究发现,模拟数据克服了历史数据的时空局限性,与历史数据相比较,基于模拟数据估计的重现水平更具准确性和稳定性;其次,采用模拟方法还可以绘制出历史数据无法实现的大比例风险图,很大程度上提高了风险评估的实用性.研究结果可以作为保险行业开展巨灾保险业务,政府部门制定防灾减灾规划的科学依据. This study aims to set up the simulation model of tropical cyclone tracks. The purpose is to improve the robustness of the risk assessment. First, the CMA-STI tropical cyclone best track data set is used in this paper. Then, the Spatial Statistics Model and Markov Chain Monte Carlo Methods are introduced to simulate the tropical cyclone tracks of Pacific Northwest. The results of hypothesis testing show that simulation results quite consistent with the historical data. Subsequently, large amount of data is generated by means of an implementation of this method. This data overcomes the historical limitations of time and space. It, thus, creates a sounder basis for risk assessment. On this basis, this study takes Guangdong Province for an example and gives the city-level risk assessment. Meanwhile, the simulation data can also be used to draw large-scale risk maps. By doing this, it greatly improved the usefulness of risk assessment. The conclusion of this study can provide scientific reference for the government department and insurance business.

关 键 词： 空间点随机过程 马尔可夫蒙特卡罗模拟 风险评估 重现期 风险图

领　　域： [理学] [理学]