摘要
为研究Copula函数在黄河上游随机径流模拟中的应用及不同抽样方法对模拟结果的影响,采用黄河上游唐乃亥水文站1956至2012年的实测径流资料,利用4种常用分布函数(P-Ⅲ分布、Gamma分布、Logn分布、Gev分布)对各月径流进行边缘分布拟合并优选后,用3种Copula函数(Clayton Copula、Frank Copula、Gumbel Copula)对相邻月间径流进行联合分布拟合并优选,再分别用直接抽样方法与Gibbs抽样方法进行月径流随机模拟,并与实测数据进行对比验证。结果表明:直接抽样方法与Gibbs抽样方法的模拟结果均能保留实测值的统计特征和月间径流的相关性,Gibbs抽样方法模拟结果的丰平枯划分更接近实测值划分;采用不同边缘分布的径流模拟结果优于单分布的模拟结果,避免了单分布个别月份不适用的情况。本研究结果对长序列径流模拟具有一定的参考价值。
The measured runoff data of Tangnaihai hydrological station in the upper Yellow River from 1956 to 2012 are taken as examples to investigate the application of Copula function in the random simulation of runoff in the region and the effects of different sampling methods on the simulation results.Four common distribution functions(P-Ⅲdistribution,Gamma distribution,Logn distribution and Gev distribution)are used to combine and optimize the marginal distribution of monthly runoff.Three copula functions(Clayton Copula,Frank Copula and Gumbel Copula)are used to conduct joint distribution and optimization of adjacent monthly runoff.The methods of direct sampling and Gibbs sampling are used to conduct random simulation of monthly runoff,and the results are compared with the measured data.The results show that both the direct sampling method and Gibbs sampling method can retain the statistical characteristics of the measured values and the correlation between monthly runoff.The simulation results of Gibbs sampling method are closer to the measured value division in terms of wet season,flat season and dry season.The runoff simulation results with different marginal distributions are better than that of single distribution,which avoids the inapplicability of single distribution in some months.The results of this study have certain reference values for long-term runoff simulation.
作者
王诗瑞
李芳芳
裘钧
WANG Shirui;LI Fangfang;QIU Junn(School of Civil Engineering and Water Resources,Qinghai University,Xining 810016,China;State Key Laboratory of Plateau Ecology and Agriculture,Qinghai University,Xining 810016,China;State Key Laboratory of Hydroscience and Engineering,Tsinghua University,Beijing 100084,China)
出处
《青海大学学报》
2023年第5期1-10,共10页
Journal of Qinghai University
基金
青海省科学技术厅项目(2021-ZJ-934Q)。
