闽江流域竹岐站年平均流量时间序列的混沌特征分析

Chaos Characteristics of Annual Flow Time Series at Zhuqi Station of the Minjiang River Basin

以闽江流域控制水文站竹岐站1936—2004年年平均流量的时间序列重构相空间,计算饱和关联维D2与Kolmogorov熵的稳定估计值.结果表明竹岐站年平均流量变化具有明显的混沌特性;相空间吸引子的关联维为5.14,至少需要设立6个独立变量建立竹岐站年平均流量的系统模型;当D2趋于稳定时饱和嵌入维数等于7,说明重构相空间所需要的最小嵌入维数为7;Kolmogorov熵的稳定估计为0.155,即系统的平均可预报时间约为6～7 a.

Based on time series of the annual mean flow at Zhuqi station of the Minjiang River Basin in the period of 1936—2004,its phase space reconstruction was conducted,and at the same time the steady estimate of correlation dimension and Kolmogorov entropy were calculated.The results show that there was an intrinsic dynamic mechanism behind the variation of annual mean flow at Zhuqi station,the correlation dimension of attractor of phase space was 5.14 and the saturation inserted dimension was 7.It also shows that to establish a mathematical model of the annual mean flow system at Zhuqi station,at least 6 independent variables would be required,and the required saturation inserted dimension for rebuilding phase space must be 7 or more.The steady estimate of Kolmogorov entropy was 0.155,indicating that the average predictable time of the annual mean flow variation at Zhuqi can be as long as about 7 years.