摘要
利用现行的Copula函数来描述模型参数之间的相关性,提出基于Copula-Glue的水文模型参数不确定性估计方法。选取采样次数为10 000、50 000和100 000,阈值为0.5、0.6和0.7的6种组合,进行了模型参数的不确定性模拟,并与Glue方法的模拟结果进行了对比。结果表明基于Copula-Glue的水文模型参数不确定性估计方法用联合分布代替单独的参数的分布,能够很好地考虑参数之间的相关性,在同样的采样次数下能够生成更多有效的参数组合,从而更加全面地估计参数的不确定性。
In this paper, copula was employed to describe the correlation of parameters and an uncertainty estimation method of hydrologial model based on Copula and Glue was proposed. Six scenarios were selected by taking the sampling number as 10 000, 50 000, and 100 000, and seting threshold as 0. 5, 0. 6, and 0. 7, respectively. Using the six scenarios, the proposed and the Glue methods were compared to study the parameter uncertainty in hydrological modelling. The results showed that the proposed method can reflect the correlation of the parameters by using a joint probability distribution of the parameters to replace two separate one-variable probability distributions, and it can obtain more valid parameter sets, so as to estimate the parameter uncertainty completely.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第3期109-115,共7页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
国家自然科学基金青年科学基金资助项目(50809078)
国家自然科学基金重点资助项目(50839005)
国家科技支撑计划基金资助项目(2006BAB14B07)
中山大学青年教师科研启动基金资助项目
水资源与水电工程科学国家重点实验室开放基金资助项目(2008B043)
关键词
COPULA
Glue方法
水文模型
不确定性
copula
generalised likelihood uncertainty estimation method
hydrological model
uncertainty
