知识与财富的链接
进一步论证了经验正交函数/主分量分析(EOF/PCA)在气象变量场诊断中的物理内涵,证明基于EOF/PCA的R型和Q型展开,可描述为气象变量场主要振荡型分解和主要空间分布型分解两种方案.前者表明,气象变量场的准周期振荡可分解为各主分量的周期振荡,它们各自等价于不同网格点(或站点)以其载荷为权重的迭加周期振荡,因此,气象变量场准周期振荡可视为来自不同周期源(网格点或站点)的准周期振荡逐层叠加的结果;后者表明,气象变量场的水平空间分布可视为各种主要空间分布型的叠加,而Q型展开才是对各种主要空间分布型的正交分解.由此深化了EOF/PCA气象变量场诊断的物理内涵.
The physics meaning of Empirical Orthogonal Function/Principal Component Analysis (EOF/PCA)is further verified for diagnoses of meteorological variable fields. EOF/PCA may shows the orthogonal resolution of principal oscillation patterns or the orthogonal decomposition of primary horizontal space distribution pattern. The former (i.e. R pattern EOF expansion) shows that the quasi-oscillation of meteorological variable fields may resolve many different oscillations which possesses different oscillation periods by means of each principal component, and the latter (i.e. Q pattern EOF expansion) shows that the horizontal space distribution of meteorological variable fields may decompose different primary space patterns. Thus, EOFs describe the complex quasi-oscillation of meteorological variable fields which is a composition of different periodic oscillations from different origins of the periods corresponding to different loading over the meteorological variable field by using each principal component.
