基于Copula函数的径流随机模拟
Streamflow Simulation Based on Copula Function
作者：闫宝伟(武汉大学 水资源与水电工程科学国家重点实验室，湖北 武汉430072)；郭生练(武汉大学 水资源与水电工程科学国家重点实验室，湖北 武汉430072)；刘攀(武汉大学 水资源与水电工程科学国家重点实验室，湖北 武汉430072)；郭靖(武汉大学 水资源与水电工程科学国家重点实验室，湖北 武汉430072)；陈璐(武汉大学 水资源与水电工程科学国家重点实验室，湖北 武汉430072)
Author：Yan Baowei(State Key Lab. of Water Resources and Hydropower Eng. Sci., Wuhan Univ.，Wuhan 430072, China)；Guo Shenglian(State Key Lab. of Water Resources and Hydropower Eng. Sci., Wuhan Univ.，Wuhan 430072, China)；Liu Pan(State Key Lab. of Water Resources and Hydropower Eng. Sci., Wuhan Univ.，Wuhan 430072, China)；Guo Jing(State Key Lab. of Water Resources and Hydropower Eng. Sci., Wuhan Univ.，Wuhan 430072, China)；Chen Lu(State Key Lab. of Water Resources and Hydropower Eng. Sci., Wuhan Univ.，Wuhan 430072, China)
收稿日期：2008-11-03 年卷（期）页码：2010,42(1):5-9
期刊名称：工程科学与技术
Journal Name：Advanced Engineering Sciences
关键字：Copula函数；马尔柯夫；随机模拟；AR(1)模型
Key words：Copula function; Markov; stochastic simulation; AR(1) model
基金项目：水利部公益性行业科研专项资助项目（200701015）；国家“十一五”科技支撑计划课题（2006BAC14B06;2008BAB29B09）
中文摘要
随机模型的核心问题是构建联合分布或条件概率分布。建立了基于Copula函数的一阶非平稳时间序列模型，即季节性CAR(1)模型，并与季节性AR(1)模型进行比较。以宜昌站月径流模拟为例，研究了季节性CAR(1)模型的实用性。结果表明，所建模型能较好的模拟原序列的统计特性，尤其是偏态特性、非线性相关性和概率密度特征的保持上，为水文水资源随机模拟研究提供了一种新的途径。
英文摘要
The kernel of stochastic model is the construction of joint distribution or conditional probability distribution. A first-order non-stationary time series model, namely seasonal CAR(1), was established based on Copula function and compared with seasonal AR(1) model. Monthly streamflow data of the Yichang station was investigated and verified by using the proposed seasonal CAR(1) model. Results suggested that the proposed model can preserve the statistical properties of the recorded data series, especially the skew, non-linear dependence and probability density properties, and provides a new way for the study of hydrology and water resources simulation.
【关闭】
