【作者】 张旭；

【导师】 曾春华；

【作者基本信息】 昆明理工大学 ， 系统理论， 2023， 硕士

【摘要】 近年来,由于全球的气候变化和极端天气的增多,使得研究人员对复杂气候系统的关注程度日益提升。在复杂气候系统的研究过程中,物理学、数学和计算机科学等学科的相关理论和方法都是必不可少的,并且这些学科的一些理论方法已经被应用在气象和空气质量的研究中并取得了很好的结果。但是,在现有的一些关于复杂气候系统的研究中,对于一种比较新颖的时间关联的研究还不是很多,这种时间关联是系统中过去的动力学涨落与未来的动力学变量的数值之间的关联。因此,本文应用这种关联对复杂气候系统进行相关的研究。在本论文的开始部分,首先对相关的研究背景和研究现状进行了介绍并给出了本文的研究意义,然后阐述了研究中所需要的理论方法,最后给出了数据来源和相关的数据处理方法。在研究过程中,本文以复杂气候系统中一些指标为研究对象,分别讨论上述关联在稳态和非稳态两种前提条件下的结果。当假设复杂气候系统处于稳态时,一些指标的涨落具有较强的自关联性,但是涨落与变量之间的关联却很弱甚至不存在;然而,如果在研究中考虑了非稳态动力学效应的影响,涨落与变量之间则表现出了一定的时间关联性。在得到这一结果后,我们还对计算方法的科学性和数据缺失处理方法在应用方面的合理性进行了验证。本文同时也对结果进行了更深入的分析:首先对产生上述关联的主要因素进行了讨论,发现“较小的涨落”是涨落–变量关联的主要构成因素,我们还发现了这种关联在时间尺度上具有一定的稳定性;然后,借助于“保持概率”这一工具对一些气候指标在非局域时间尺度内的性质进行了讨论,发现这些指标的涨落在非局域时间尺度内确实具有一定的长程关联性;最后,我们对方法进行了一些改进,从而得到了更加合理且符合实际要求的结果。完成了上述的研究工作后,我们对本文工作进行总结并对未来可能需要研究的工作进行展望。本文或许可以为某些气候指标的预测工作提供一些理论支持,为气候变化背后的机制提供更多的见解,同时也为在更多的复杂系统中计算受到非稳态影响的时间关联函数提供了经验。更多还原

【Abstract】 In recent years,the global climate change and the increase of extreme weather have led to an increasing interest in the complex climate system.Theories and methods from physics,mathematics,and computer science are essential in the study of the complex climate system,some of these disciplines have been applied to meteorological and air quality studies with good results.However,among the existing studies on the complex climate system,there are not many studies on a relatively novel temporal correlation,which is the correlation between the past dynamical fluctuations and the values of future dynamical variables in complex systems.Therefore,this paper applies this correlation to the study of the complex climate system in a relevant way.In the beginning of this paper,the relevant research background and research status are firstly introduced,the significance of this paper is also given;then the theoretical methods needed in the study are described,and finally the data sources and relevant data processing methods are given.In the research process,some indicators in the complex climate system are taken as research objects to discuss the results of the above-mentioned correlation under both stationary states and non-stationary states preconditions.When the complex climate system is assumed to be in the stationary state,the dynamic fluctuations of some indicators have strong autocorrelation,but the correlations between the dynamic fluctuations and the dynamic variables of some indicators are weak or even nonexistent;however,if the dynamic effects of the non-stationary states are considered in the study,there is a certain temporal correlation between the dynamic fluctuations and the dynamic variables.After obtaining this result,we also verify the scientific validity of the computational method and the rationality of the missing data treatment method in terms of application.In this paper,we also analyze the results in more depth: firstly,the main factors of the correlation are discussed and it is found that “the small fluctuation”is the main component of the fluctuation-variation correlation,we also find that this correlation is stable over time scales;then,we discuss the nature of some climate indicators at non-local time scales with the help of the “the persistence probability” and find that the fluctuations of these indicators do have some long-range correlations at non-local time scales;finally,we make some improvements to the method,so as to obtain more reasonable and realistic results.After completing the above-mentioned research,the work of this paper is summarized and an outlook on possible future research is given.The research in this paper may be useful for the analysis and prediction of climate.This paper may provide some theoretical support for the prediction of some climate indicators,provide more insights into the mechanism behind climate change,and provide experience for calculating the temporal correlation functions affected by non-stationary states in more complex systems.更多还原