摘要
描述时间连续日径流过程散粒噪声过程是非正态随机模型。本文对G.weiss提出的单散粒噪声模型中的涨水退水曲线的形式改用为Γ-分布的密度函数表示,推导出改进后的散粒噪声模型的一、二阶矩的计算公式和以日平均流量资料为依据的离散序列的一、二阶矩的递推公式。并用物理-统计相结合的方法估计有关模型参数,改进拟合程度。本文还对weiss提出的单散粒噪声综合数据的生成方法上做了修正。最后通过潮白河密云水库等站的日平均径流资料说明这模型的应用。
A shot noise process, which is used as basis for describing continuons time daily streamflow process, is an non-Gaussian stochastic model. In this paper we have a new try for the form of system transfer function of the simple shot noise models which was proposed by G. Weiss in 1970s. In order to improve the fitting, The shapes of the rise and recession curve in model are represented by probability density function of gamma distribution instead of exponential function. Some formulae some of the moments of improved shot noise process and some recurrent formulae of mo ments for discrete sequence of mean daily data are deduced. Some model parameters are estimated by associating physical method with statistical method. We also make an important modification for the method of generation of synthetic shot noisedata which wsa proposed by Weiss. The application of the models to some streamflow data of the Chao Bai river Mi Yun reservoir is illustrated.
出处
《水电能源科学》
北大核心
1989年第3期275-287,共13页
Water Resources and Power
基金
国家自然科学基金
关键词
日径流
散粒噪声过程
脉冲响应函数
Γ-分布的密度函数
daily streamflow, shot noise process, pulse response function, probability density function of gamma distribution.
