摘要
本文针对气象要素与其过去的数据自相关程度一般都较差,而与其自身的显著周期分量却存在较高的线性相关关系的特点,提出了一种改进的多层递阶预报模型。它将多层递阶方法与逐步回归双重分析相结合,用显著周期分量取代经典多层递阶预报模型中的自回归部分,使之能更好地反映气象要素自身的历史演变规律,从而稳定其预报效果。
In this paper, a improved forecasting scheme is presented. In this new forecasting model, the obvious periodic components of time series already replaced the autoregression component in classical recursive estimate model. This model not only eonsiders the important action of various predictors, but also fairly reflects its periodical variable rules of prediction; therefore, its forecast effect is more stable.
出处
《大气科学》
CSCD
北大核心
1991年第1期69-73,共5页
Chinese Journal of Atmospheric Sciences
