利用二维极值分布模拟我国几个代表站的强降水概率特征

Probability characteristics of heavy precipitation in China based on bivariate Gumbel-Logistic model

基于中国6个代表站5—9月的逐日降水资料,利用二维Gumbel-Logistic分布,研究了中国不同区域的过程降水量和日最大强降水雨量的联合概率特征。结果表明,各代表性台站的过程雨量和强降水雨量的联合分布均符合二维Gumbel分布。强降水雨量与过程降雨量联合分布所描述的极端事件是更小的小概率事件。相同强降水雨量条件下,过程雨量越大,重现期越长;当强降水雨量增大时,同一过程雨量的重现期也延长。在同级强降水雨量出现的条件下,各地过程降雨量往往是愈往南方其条件概率愈大,而其出现的过程雨量也随之增大。这为研究强降水极端状况的全方位特征做出了新的试验,也更加客观地揭示了极端气候事件的多方面概率特征。

Based on the daily precipitation data of six representative stations in China from May to Sep tember, the joint probability characteristics of the heavy precipitation and the corresponding course pre cipitation are investigated by the bivariate GumbelLogistic model. Results show that the bivariate Gum belLogistic model is suitable for the joint distribution of the heavy precipitation and the corresponding course precipitation at the selected stations. The probability of the joint distribution is smaller than the probability of either. On the certain heavy precipitation condition, the more the course precipitation is, the longer the return period is. The return period of the corresponding course precipitation can be pro longed by the increased amount of heavy precipitation. The probability of course precipitation in south is larger than that in north and the same as the amount of course precipitation under the same heavy pre cipitation conditioan. The study is a new attempt of research on extreme precipitation in all domains and reveals the multisided probability characteristics of extreme climate events.