摘要
本文用非线性动力系统理论探讨了现代太阳周（１８５０年１月─１９９２年５月）黑子相对数月平均变化过程的可预报性。用时间延迟方法重构吸引子，计算它的最大Ｌｙａｐｕｎｏｖ指数（λ_１＝０．０２３±０．００４ｂｉｔｓ／月），估算了用这些黑子数进行确定性预报的理论时限（ｔ＝３．６±０．６年）．结果表明，动力系统的可预报性与它的最大Ｌｙａｐｕｎｏｖ指数有直接关系，黑子数月平均变化过程的演化不是周期的，也不是拟周期的，而是混沌的。即使今后找到了描述该过程的确定性方程，它的长期行为也不可能准确地预报，只能作短期预报，这是黑子数本身的混沌特性决定的。用于黑子数预报的纯粹数值统计方法仅对短期预报才有效。
he predictability of the monthly mean variations of the relative sunspot number for the period of time from January 1850 to May 1992 is examined based upon the theory of nonlinear dynamical system. The attractor of the variation process is reconstructed using the method of time delays. The attractor is used to compute its largest Lyapunov exponent (λ1= 0.023±0.004 bits/month). The upper limit (t=3.6±0.6 years) of the theoretical time-scale on which the monthly mean sunspot number can be used to make deterministic prediction is estimated. The results indicate that the predictability of a dynamical system is directly related to the largest Lyapunov exponent of the system, and the evolution for the montlily mean variation process of the relative sunspot number is neither periodic nor quasiperiodic, but is chaotic. Its long term deterministic behaviour is unpredictable, even if we have known a deterministic set of equations describing the process in the future, the deterministic behaviour of the process allows for short term predictions only, because of the variation process of the relative sunspot number is an inherent chaotic system.The purely numerical statistical approach used to predict long term behaviour of the sunspot number seems to be useless.
基金
国家自然科学基金
云南省应用基础研究基金
