两种自适应统计天气预报建模方法的比较

Comparison of Two Adaptive Statistical Weather Forecast Modeling

阐述了应用卡尔曼滤波法和BP网络法进行天气预报建立模型的基本原理,通过具体的预报试验说明两种自适应方法建模明显优于传统的逐步回归方法建模,为数值天气预报产品的统计解释应用提供了新思路。

In this paper, principles of weather forecast modeling based on Kalman filtering and BP neural networks are introduced. Forecast results indicated that weather forecast modeling based on Kalman filtering and BP neural networks was superior to that based on stepwise regression analysis, and this provide new way for statistical interpretation of mathematical forecast products.