混沌系统的局域特征与可预报性
THE LOCAL FEATURES AND PREDICTABILITY OF THE CHAOTIC SYSTEM
-
摘要: 讨论了混沌系统的时间和空间的局域特征。首先分析了研究时间和空间局域特征的必要性。接着引进了有限时间不稳定和局域时间不稳定的概念,并对有关的计算问题进行了研究。对Lorenz系统的具体计算表明,随着轨线在混沌吸引子上的演变,局域不稳定特征有很大的变化,相应误差增长也有很大的变化。相应于误差迅速增长的轨线部分局限于很有限的相空间范围内,而且同误差增长缓慢的轨线部分占据的相空间区域截然可分。每一个例的可预报性依赖于轨线在相空间中所处的区域。混沌系统的这种局域特征可以是导致个例业务预报技巧之间有很大差别的主要原因。Abstract: The local features in phase space and predictability are studied in this paper.It is illustrated that the local features in phace space have to be considered in determining predictability.The notions of local instability including the finit-time and local-time instability which determine the growth rate of the error are introduced and calculation methods are discussed in detail.The results from the calculation of the 3-component Lorenz model show that the local instability, correspondingly the growth rate of error, vary dramatically as trajetories evolve on the chaotic attractor.The region where the growth rate of error is small.is localized considerably and is separable from the region, where the growth rate of error is large.The local predictability depends on the region through which trajectories evolve.It is suggested that the local features are the main reasons why there is a great deal of case to case avariability of the predictive skill in operational forecasts.