水利学报2018，Vol.49Issue(3)：332-342,11.DOI:10.13243/j.cnki.slxb.20170675
基于Copula函数的多变量水文不确定性处理器
Multivariate hydrologic uncertainty processor based on Copula function
摘要
Abstract
The traditional hydrologic uncertainty processor (HUP) belongs to the univariate structure type,which only independently provides a marginal Bayesian posterior probability density function of observed discharge for each lead time and does not consider and characterize the inherent dependence among these variables.In this paper,the analytical expression of Bayesian posterior transition density was derived by using Copula function,and therefore the Copula-based BTF (CBTF) method and Copula-based multivariate HUP (CMHUP) was proposed.Subsequently,the Copula-based BEF (CBEF) was developed.Application results of Three Gorges Reservoir (TGR) indicate that the proposed methods are practical and effective,of which the CBTF method and CMHUP not only can quantitatively evaluate the uncertainty of transition forecast for inflows of the TGR,but also reveal the evolution characteristic with time of uncertainty in hydrological forecasting.Moreover,the uncertain information about the maximum inflow forecast within specified lead time is provided by the CBEF method.The proposed methods relax the linear-normal assumption and capture the nonlinear and non-Gaussian characteristics of discharge process adequately,which lead to more extensive application scope and support the flood control and disaster mitigation,and reservoir operation better.关键词
水文预报/贝叶斯理论/水文不确定性处理器/转移概率预报/极值概率预报/Copula函数Key words
hydrological forecasting/Bayesian theory/hydrologic uncertainty processor/probabilistic transition forecast/probabilistic extremum forecast/Copula function分类
建筑与水利引用本文复制引用
刘章君,郭生练,何绍坤,巴欢欢,尹家波..基于Copula函数的多变量水文不确定性处理器[J].水利学报,2018,49(3):332-342,11.基金项目
国家自然科学基金重点项目(51539009) （51539009）
国家重点研发计划项目(2016YFC0402206) （2016YFC0402206）
