论文部分内容阅读
基于国家级地面气象站基本气象要素日值数据集的均一化降水序列计算了19612014年青藏高原中东部71个站夏季极端降水指数,选取二参数和三参数的Weibull分布、广义极值分布、皮尔逊Ⅲ型分布(Gamma分布)、对数逻辑斯特分布拟合其极值,分析了青藏高原中东部夏季极端降水的极值分布特征.结果表明:青藏高原夏季极端降水由东南向西北递减,大值中心位于四川东部地区,西藏东南部有一较大值中心,小值中心位于青海西北部.通过极值分布函数对极端降水指数的拟合,发现不同的指数适用的函数不同,需采用多种概率分布模式进行对比,并结合实际物理意义加以选择最适合的拟合分布函数.利用Gumbel分布计算夏季极端降水指数的多年一遇水平、50年一遇和100年一遇水平,均呈现出完全一致的空间分布特征:东南降水多,西北降水少;根据滑动t检验,强降水量和极强降水量均于2006年突变,由Gumbel分布估计突变后青藏高原中东部的极端降水有所增加.“,”Based on the homogenous historical precipitation dataset in the eastern and central Qinghai-Tibetan Plateau ( QTP) , the two-parameter and three-parameter Weibull distribution, generalized extreme value distribu-tion, Pearson type-III distribution and log-logistic distribution selected to fit the summer extreme precipitation in-dices to study the characteristics of the probability distribution of summer extreme precipitation indices over the QTP during 19612014. The results are showed as follows:The distribution of extreme precipitation decreased gradually from southeast to northwest of the eastern and central QTP, with a maximum center located in the southeast of Sichuan province, and the Rikeze station at southeast of Tibet Autonomous Region was the other max value center, while the minimum center was seated at the northwest of Qinghai province. Through the fitting of extreme precipitation indices by the extremum probability function, we found that different probability func-tions are fitted for different indices, and which probability function to choose is depends on the comparison of a variety probability function models combined with the actual physical sense. Then, Gumbel distribution was used to calculate different return levels of summer extreme precipitation indices, and it had the same spatial distribution of the 50 a and 100 a return levels that the precipitation was more at southeast and less at northwest of the region. After testing by moving t-test, R95 and R99 average value abruptly changed at the year 2006, and the extreme precipitation increased after 2006 according to the return level values of Gumbel probability distribution for sum-mer extreme precipitation indices.