湖北分县MOS预报系统建立与评分

A MOS Prediction System of Precipitation,Surface Air Temperature and Wind in Hubei Province

在T213L31数值模式产品释用基础上,设计计算并选用模式大气中分别与降水、地面气温和风场有天气学意义的线性化预报因子,采用Kalman滤波和相似-Kalman滤波,在湖北省及邻近范围,建立实时(逐日更新样本)多站点、多时次、多要素的统计MOS预报方程,做定点(分县)、定时(6小时分辨率)和定量降水与概率、地面气温和风场客观预报(MOS预报系统)。"系统"经2005、2006年6—8月业务试验表明,各个MOS预报都具稳定性和具有一定的评分水平。从而表明,相似-Kalman滤波在一定程度上消去单纯Kalman滤波"预报滞后"效应,并实现用Kalman滤波或相似-Kalman滤波做定性、定量降水MOS预报,和做地面气温和风场MOS预报,且用了实时历史样本雨/晴预报准确率,当作有/无降水发生概率。

Based on the explanation and application of T213L31 numerical forecast model products, an operational MOS prediction system and a set of real-time statistical prediction equations are established for stations of Hubei province. The system firstly adopted Kalman Filter and Analogue-Kalman Filter, and the equations selected several linear predictors that are relative to the model outputs, such as precipitation, surface air temperature and wind fields. This MOS system can forecast above meteorological elements quantitatively in the form of fixed point （every county）, fixed time （every six hours）. The experiments show that the MOS system has good stability and excellent forecast scores in the testing period in June-August from 2005 to 2006.