三种基于均生函数的统计模型在厄尔尼诺预测中的应用

为了探究厄尔尼诺的预测方法,提高其预报准确度与预报时效,本文基于均生函数的基本原理,在普通均生函数模型中分别加入气候序列的本征模函数和预报因子变量,构建了基于EEMD(Ensemble EmpiricalModeDecomposition)的均生函数模型和多变量均生函数模型;并应用三种方案对NINO3区海温指数进行了预报试验。结果表明,两种改进模型对厄尔尼诺的预报效果好于普通均生函数模型,是提高预报准确度的有效手段;同时,利用统计的预报模型,可以在一定程度上有效延长厄尔尼诺的预报时效,具有一定实践意义。

The application of three statistical models based on MGF to El Niño prediction

To explore the prediction method of El Niño and improve its forecast accuracy and forecast period, in this study, on the basis of the fundamental principles of the mean generating function, the intrinsic mode function and predictor variable of a climosequence are added to the common mean generating function model to build the EEMD-based and multivariable mean generating function models. The forecasting experiment on sea temperature indices in Zone NINO3 is conducted through three forecasting schemes. The results show that the forecasting effects for El Niño by the two improved models are better than that by the common mean generating function model, which effectively improves the forecast accuracy. Simultaneously, with the statistical forecasting model, the forecast period of El Niño can be improved effectively to some extent, with some practical significance.