摘要
以淮河干流蚌埠站64a(1950—2013年)的月径流资料为例,研究Archimedean Copula函数在月径流随机模拟中的应用.先利用4种一维分布函数对每个月径流进行单变量的分布拟合,再利用3种二维Archimedean Copula函数进行相邻月份的联合分布拟合,并对一维分布函数及二维联合分布函数的拟合优度进行判断,以确定拟合效果最优的边缘分布函数和二维Copula函数.基于选定的Copula函数,结合Gibbs采样实现月径流随机模拟,通过比较实测与模拟月径流对该方法进行验证.研究结果表明,Archimedean Copula函数结合Gibbs采样可有效建立相邻月份间的相关性结构,为水文水资源随机模拟提供一种新的途径.
The application of Archimedean Copula functions in probabilistic simulation of monthly streamflow is studied, and the case study is carried out with 64-year （1950--2013） observations of Bengbu hydrological gauge in Huaihe River. Four one-dimensional distribution functions are fitted to the marginal variates, three types of two-dimensional Archimedean Copula functions are fitted to the two adjacent months, and then a goodness-of-fit analysis is made for one and two dimensional distribution functions to determine the most appropriate marginal distribution and bivariate joint distribution function. By using the selected Copula function type and the Gibbs sampling technique, the monthly streamflow is simulated statistically, and the method is tested according to the compar- ison between the observations and the simulations. The result indicates that the method is effective for constructing the dependent structure for adjacent months, and it proposes a new method for hydrological probabilistic simulation.
出处
《扬州大学学报（自然科学版）》
CAS
北大核心
2016年第4期33-37,共5页
Journal of Yangzhou University：Natural Science Edition
基金
江苏省自然科学基金资助项目(BK20160470)
江苏省高校自然科学研究资助项目(15KJD170003)
