基于线性分解时序方法的径流序列长度影响研究

应用线性分解时间序列分析方法揭示径流序列特征的成果较多,但很少涉及序列长度对径流线性分解结果的影响问题.利用闽江流域竹岐站年径流和松花江流域白山水库入库年径流近70 a的径流序列对此进行了探讨.将约70 a的径流资料分成各相差5 a的6种不同径流序列进行分解,结果表明,序列长度对径流序列线性分解的结果有十分明显的影响,调整序列长度导致径流的趋势、跳跃和周期等特征产生变化,且没有明显的规律性.因此,径流时间序列分析要选取合理的序列长度,且合理的序列长度可能与时间尺度有关.

Effects of Time Series Length on Runoff Characteri stics by Using Linear Decomposition Method

Linear decomposition time series analysis has become one of useful tools to investigate runoff characteristics, but how the sequence length affects the results of linear decomposition has not yet been studied in detail. In this paper,the runoff time series at Zhuqi station of the Minjiang River and the inflow time series at Baishan reservoir of Songhuajiang River are selected to investigate the effects of time series length on rurr o ff characteristics. Six sequence lengths are selected according to the available runoff time series over 70 years, and the linear decomposition for different length is conducted. Results show that the characteristics of runoff,such as shift trend,jump change, and periods have been affected obviously by sequence length. It is suggested that the reasonable sequence length necessary to conduct linear decomposition and reasonable sequence length may be correlated with the time scale of time series.