【摘要】 以黄河花园口断面1953~2002年年平均径流量为时间序列资料,在G-P算法的基础上,用最小二乘法分别计算了关联维和Kolmogorov熵的稳定估计值。结果表明:①花园口年平均径流量变化存在着内在动力学机制,是由周期性和非周期性影响因子共同作用的结果,具有明显的混沌特性;②相空间吸引子的关联维为5.09,饱和嵌入维数为14,这说明要建立花园口年平均径流系统的数学模型,至少需要6个独立变量,重构相空间所需要的饱和嵌入维数为14;③Kolmogorov熵的稳定估计为0.14,说明花园口年平均径流量变化的平均可预报时间大约为7年;④用非线性确定性的混沌模型比完全的随机模型更适于描述黄河花园口年平均径流量的变化。更多还原

【Abstract】 Taking the annual mean runoff of Huayuankou section of the Yellow River in the period of 1953~2002 as time series data and based on G-P algorithm,the paper calculates the steady estimate of correlation dimension and Kolmogorov entropy by using the least-squares method.The results show that a) the variation of annual mean runoff of Huayuankou section exist an intrinsic dynamic mechanism,which is the results of joint action of affecting factor of periodicity and non-periodicity,having obvious chaos characteristic;b) the correlation dimension of attractor of phase space is 5.09 and saturation inserted dimension is 14.It shows that to establish a mathematical model of the annual mean runoff system of Huayuankou at least requires 6 independent variables and the required saturation inserted dimension for rebuilding phase space is 14;c) the steady estimate of Kolmogorov entropy is 0.14,showing the average predictable time of the annual mean runoff variation of Huayuankou is about 7 years and;d) using nonlinear deterministic chaos model is more suitable to describe the variation of annual mean runoff of Huayuankou.更多还原