...
首页> 外文期刊>ournal of the Meteorological Society of Japan >New Distribution Functions for Hourly and Daily Precipitation Intensities during the Snowless Season in Japan
【24h】

New Distribution Functions for Hourly and Daily Precipitation Intensities during the Snowless Season in Japan

机译:日本下雪季节每小时和每天降水强度的新分布函数

获取原文

摘要

In this study, various probability distribution functions for the entire intensity range of short-term precipitation, such as the hourly and daily precipitation intensities during the snowless season in Japan, were examined. The traditional distribution functions, i.e., the exponential, Weibull, Gamma, generalized Gamma, log-normal, and Johnson's SB distributions, were insufficient to express the precipitation intensities within the rainy periods. Three types of distribution functions were newly proposed in association with the extension of the Weibull distribution. The new distributions were constructed to vary asymptotically from the exponential distribution on the weak intensity side to the Weibull distribution on the strong intensity side. The number of parameters of the distribution functions was four except for the parameter of the rainy period ratio. One or two parameters were fixed as unity parameters in all observation points in the parameter estimation in order to overcome a multi-solution problem caused by the strong non-linearity of the distribution functions. The unity parameters were determined to be the values at which the all-points average of mean square errors for logarithms of exceedance probabilities had the minimum value. Other parameters were estimated by the maximum likelihood estimation method. The new distribution functions were more suitable to express the short-term precipitation, including both weak and strong intensities, than the traditional distribution functions were. For the hourly precipitation, the boundary intensity between the exponential and Weibull distribution properties was considered to be associated with that between the main domination ranges of stratiform and convective precipitations. A two-parameter (except for the rainy period and the unity parameters) function with Type-I and three-parameter functions with Type-I and -III were proposed as the best functions for hourly precipitation intensities. A two-parameter function with Type-III was also proposed for daily precipitation.
机译:在这项研究中,研究了短期降水整个强度范围内的各种概率分布函数,例如日本无雪季节每小时和每天的降水强度。传统的分布函数,即指数分布,Weibull分布,Gamma分布,广义Gamma分布,对数正态分布和Johnson SB分布,不足以表达雨季的降水强度。随着威布尔分布的扩展,新提出了三种分布函数。新的分布构造为从弱强度侧的指数分布到强强度侧的威布尔分布渐近变化。除雨季比率的参数外,分布函数的参数数目为四个。为了克服由分布函数的强非线性引起的多解问题,在参数估计中的所有观察点中将一个或两个参数固定为统一参数。确定单位参数为超出概率对数的均方误差的所有点平均的最小值。其他参数通过最大似然估计法进行估计。与传统的分布函数相比,新的分布函数更适合表达短期降水,包括强度的强弱。对于每小时的降水,指数和威布尔分布特性之间的边界强度被认为与层状和对流降水的主要控制范围之间的边界强度有关。建议将具有I型的两参数(雨季和单位参数除外)函数以及具有I型和-III的三参数函数作为小时降水强度的最佳函数。还提出了具有III型的两参数函数用于日降水量。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号