摘要
对短持续类模型中ＡＲ（１）、ＡＲ（２）及ＡＲＭＡ（ｌ，１）和长持续类模型中ＦＦＧＮ的赫斯特（Ｈｕｒｓｔ）系数的统计规律进行了研究，结果表明：短持续性模型不能在很长的时间水平上但可在较短的时间水平上保持某一指定的赫斯特系数；而长持续类模型可在很长或无限长的时间水平上保持某一指定的赫斯特系数值。因此在设计标准不高时，若强调模型对赫斯特系数的保持，可以采用适当的短持续模型，否则宜采用长持续类模型。
he persistence properties of several hydrological stochastic models expressed as AR (1),AR(2),ARMA(1,1)and FFGN are investigated. By means of Monte-carlo simulation,comparisons have been made of these models in preserving Hurst coefficient.It is shown that the expected Hurst coefficient can be preserved in a relatively short period by specially selecting parameters of a short term persistence model, which guarantees the application of an ARMA model in preserving persistence of a simulated series if the degined standard is not high;
出处
《河海大学学报（自然科学版）》
CAS
CSCD
1994年第6期48-53,共6页
Journal of Hohai University(Natural Sciences)
关键词
随机水文模型
赫斯特系数
统计试验
Hurst coefficient Long(short) term persistence model Statistic test
