河川径流演变的非趋势波动分析

Detrended fluctuation analysis of river runoff evolvement

运用非趋势波动分析法,对黄河上游河川径流近百年月平均流量序列进行了分析,揭示了河川径流的长程相关性及其内在规律性。研究结果表明:黄河上游近百年河川径流序列波动的标度指数a=0.5862,存在内在的正长程相关性,长程幂律关系至少可持续54月或4.5年;大于54月尺度区间的标度发生改变,可能是不同时空尺度下信号突变或序列长度受限所致,这也从某种程度上体现了河川径流系统演变过程中复杂的物理机制。

By using the detrended fluctuation analysis, the approximately 100 years time series of the average monthly river runoff in the upstream of the Yellow River is analyzed to depict its inherenced law and the long-range correlation. The preliminary results show as follows. Firstly the fluctuation of the runoff time series is characterized by a scaling index which is 0.5862. It indicates that the time series are of long range power-law correlation which has a duration of at least 54 months or 4.5 years. The different from that of less than 54 months, which can be scaling index and larger scale of more than 54 months is regarded as the change of signal in the state of different spare-time scales or the limit of the length of the series. This also embodies the complex physical mechanism during the process of the river runoff evolvement in some degrees.