基于环流分型的逐时气温预报技术探讨

论文利用2009年1-2月T213模式产品、30年来的常规观测报文以及场区逐时气温记录，综合运用环流分型和Kalman滤波方法，基于逐月环流型的气温增率建立了场区72h逐时气温释用预报模型，并针对场区2009年3-7月各天72h逐时气温进行了具体的预报。结果表明:Kalman滤波算法能够得到较为精确的温度最值预报，通过环流分型得到的逐时气温预报准确率能够达到73.2%，可以作为逐时气温预报一种较好的方法。

THE HOURLY SURFACE TEMPERATURE FORECASTING TECHNIQUES BY CIRCULATION TYPES

This paper introduced the Kalman filtering and circulation type classification, established the interpretation forecasting model of hourly temperature in 72 hours by temperature increase rate which is due to the different circulation types in terms of T213 prediction products from January to February at 2009, conventional observation data and hourly temperature records at site area more than 30 years. And use the site area for testing from March to July at 2009. The results show that: Combining with circulation types and Kalman filtering which is can get accurately results of the most values of temperature, the accurate rate can reach 73.2 percent of hourly temperature. The method in this paper can play well for hourly temperature forcasting .