动力延伸预报产品释用方法的研究

RESEARCH ON REEXPLANATION AND REANALYSIS METHOD OF DYNAMICAL EXTENDED RANGE FORECAST PRODUCTS

目前以数值形势预报为基础的动力产品释用方法主要有模式输出统计法（ＭＯＳ）和完全预报方法（ＰＰＭ）。由这两种方法建立的预报方程缺乏明确的物理意义，所以文中提出了一种动力与统计相结合的方法。该方法从大尺度大气动力学方程组出发，推导出月降水距平百分率与月环流场的关系，从而建立了月降水距平百分率预报方程，随后利用月动力延伸预报的５００ｈＰａ高度场和实际降水场资料反演出月降水距平百分率预报方程的系数。该方程经过独立样本检验，表明这种方法对利用动力延伸集合预报的环流形势作月降水距平预报具有一定的能力。

Currently, there are two dominant methods, Model Output Statitics (MOS) and Perfect Prediction Method (PPM), in reexplanation and reanalysis of dynamic products based on potential height of model outputs. The forecast equations which are estabalished by using these two methods have statistical meanings except physical senses. In this paper, we present a blending method of dynamical and statistical approaches, which has established a relationship between monthly precipitation anomaly and monthly circulation. The relationship is monthly precipitation forecast equation and its coefficients are determined by using 500 hPa potential height anomaly of ensemble dynamical extended range forecast(DERF) and observed precipitation. The results show that the method is useful for forecasting monthly rainfall anomaly using potential height of ensemble DERF through independent sample tests.