Box-Cox正态分布及其在降雨极值分析中的应用
发布时间：2018-12-11 14:39
【摘要】：Box-Cox变换和正态分布有机结合构建新的Box-Cox正态分布,可以用来研究降水极值分布拟合问题,采用极大似然估计方法估计Box-CoX正态分布的参数,并基于1951-2010年河北省21个气象站逐日降水观测资料,拟合逐年日最大降水量序列,借助K-S与A-D方法进行拟合优度的比较,结果表明BoX-Cox正态分布能适应不同站点的降雨极值分布的拟合,且大部分优于降雨极值分布拟合中常用的广义极值(GEV)分布、Weibull分布、Gamma分布,因而对掌握降雨极值分布规律,分析降水极值重现期、时空特征和变化趋势具有重要意义。
[Abstract]:Box-Cox transform and normal distribution are combined to construct new Box-Cox normal distribution, which can be used to study the fitting problem of precipitation extremum distribution. The maximum likelihood estimation method is used to estimate the parameters of Box-CoX normal distribution. Based on the daily precipitation observation data of 21 meteorological stations in Hebei Province from 1951 to 2010, the daily maximum precipitation series was fitted, and the goodness of fit was compared with K-S and A-D methods. The results show that the BoX-Cox normal distribution can fit the rainfall extreme value distribution at different stations, and most of them are better than the generalized extreme value (GEV) distribution, Weibull distribution and Gamma distribution, which are commonly used in the fitting of rainfall extreme value distribution. Therefore, it is of great significance to master the distribution law of rainfall extremum and analyze the recurrence period, space-time characteristics and variation trend of precipitation extremum.
【作者单位】： 上海财经大学统计与管理学院;首都师范大学资源环境与旅游学院;首都师范大学数学科学学院;
【基金】：国家自然科学基金面上项目(No.11231010,11471223) 北京市教委重点项目(No.KZ201410028030) BCMIIS项目的资助
【分类号】：O213
本文编号：2372730
[Abstract]:Box-Cox transform and normal distribution are combined to construct new Box-Cox normal distribution, which can be used to study the fitting problem of precipitation extremum distribution. The maximum likelihood estimation method is used to estimate the parameters of Box-CoX normal distribution. Based on the daily precipitation observation data of 21 meteorological stations in Hebei Province from 1951 to 2010, the daily maximum precipitation series was fitted, and the goodness of fit was compared with K-S and A-D methods. The results show that the BoX-Cox normal distribution can fit the rainfall extreme value distribution at different stations, and most of them are better than the generalized extreme value (GEV) distribution, Weibull distribution and Gamma distribution, which are commonly used in the fitting of rainfall extreme value distribution. Therefore, it is of great significance to master the distribution law of rainfall extremum and analyze the recurrence period, space-time characteristics and variation trend of precipitation extremum.
【作者单位】： 上海财经大学统计与管理学院;首都师范大学资源环境与旅游学院;首都师范大学数学科学学院;
【基金】：国家自然科学基金面上项目(No.11231010,11471223) 北京市教委重点项目(No.KZ201410028030) BCMIIS项目的资助
【分类号】：O213
【相似文献】
相关期刊论文 前4条
1 谭常春;张诚;周嘉章;;基于Box-Cox变换的城市火灾起数的模型研究[J];合肥工业大学学报(自然科学版);2013年01期
2 杨衡,李光泉,唐万生;最优Box-Cox变换模型的反向动态规划求解[J];陕西工学院学报;2003年03期
3 崔恒建;张娟;田玉斌;;火工品感度分布的Box-Cox正态分布拟合法[J];兵工学报;2007年10期
4 ;[J];;年期
，本文编号：2372730
本文链接：https://www.wllwen.com/kejilunwen/yysx/2372730.html
