中国日降水量的概率分布  

Probability distribution of daily precipitation in China

作　　者：顾学志 叶磊[1] 赵铜铁钢 欧阳文宇 张弛[1] GU Xuezhi;YE Lei;ZHAO Tongtiegang;OUYANG Wenyu;ZHANG Chi(College of Hydraulic Engineering,Dalian University of Technology,Dalian 116024,China;College of Civil Engineering,Sun Yat-sen University,Zhuhai 519082,China)

机构地区：[1]大连理工大学水利工程学院,辽宁大连116024 [2]中山大学土木工程学院,广东珠海519082

出　　处：《水利学报》2021年第10期1248-1262,共15页Journal of Hydraulic Engineering

基　　金：国家自然科学基金项目(51909010);中央高校基本科研业务费(DUT20RC(3)019)。

摘　　要：针对国内目前没有通用、普适性的日降水概率分布的现状,以全国820个气象站实测的日降水数据为研究对象,根据线性矩比值图选定7种参数个数不同的潜在概率分布进行拟合,选取均方根误差作为拟合优度的评价指标,初步判定KAP和P-Ⅲ分布在全国范围内最适用。在此基础上,权衡概率分布参数多少和拟合效果,进一步制定最优分布的选择策略,对KAP和P-Ⅲ分布拟合效果相近的站点,开展基于天气发生器的蒙特卡洛模拟,评价这两种分布模拟日降水序列不同分位点及降水量年际波动的能力,最后按照气候、地理和流域分区给出全国范围内最优概率分布空间分布图。结果表明:KAP分布的适用范围更广,模拟降水全频段的能力更强,参数更少的P-Ⅲ分布作为最优概率分布有3~4个明显的集中分布区域,而目前普遍接受的G2分布拟合效果并不好,不适宜广泛使用。本文的结果可为日尺度降水概率分布选择提供依据,结合研究目的和应用需求来确定最优概率分布,以最大程度减少因选择概率模型不当而造成的偏差。In view of the fact that there is no universal suitable daily precipitation probability distribution in China,the daily precipitation data measured by 820 meteorological stations in China were taken as the research object.According to the L-moment diagram,seven potential probability distributions were selected for fitting the daily precipitation data.Root mean square error was selected to evaluate the goodness of fit,and the distributions of KAP and P-Ⅲwere preliminarily determined to be the most suitable in the whole country.Then the selection strategy of optimal distribution was further formulated.For the stations with simi⁃lar fitting effect of KAP and P-Ⅲ,Monte Carlo simulation based on weather generator was carried out to evaluate the ability of these two distributions to simulate different quantiles and the interannual fluctuation of precipitation series.Then the spatial distribution maps of optimal probability distribution in climate,geog⁃raphy and watershed zones of China were drawn.The results show that:KAP has a wider area of applica⁃tion,and the ability to simulate the whole frequency range of precipitation is stronger;P-Ⅲhas three to four obvious concentrated areas as the optimal distribution;and G2 which is generally accepted is not suit⁃able for wide use because of poor fitting effect.Researchers can determine the optimal daily precipitation probability distribution according to the results of this paper and combining with the research purposes and application requirements,so as to minimize the deviation caused by improper selection of probability model.

关 键 词：日降水 最优概率分布 线性矩 天气发生器 蒙特卡洛模拟 KAP分布 P-Ⅲ分布 

分 类 号：TV125[水利工程—水文学及水资源]