Abstract:

The fine forecast user support environment(FUSE )forecast products issued by the Central Meteorological Observatory were applied in Changsha, it forms an useful display and extended platform for forecasters, and forecasters were convenient at any time to check the temperature and precipitation forecasts, understand intuitively forecast performance of FUSE products on the platform. Forecast time can be shortened by using FUSE forecast products, and prediction efficiency can be improved. Firstly, the five FUSE temperature forecast products at Changsha, Liuyang and Ningxiang stations in 2013 were scored, then the errors of maximum temperature forecast were analyzed in this paper. If it was a single forecast period, we calculated forecast accuracy, and for multiple forecast period, we calculated the total score with a total score formula. Results are as follows:(1) FUSE temperature forecast products scores of each time in 2013 indicated that 72 h maximum temperature forecast and 24, 48, 72 h minimum temperature forecast initiated at 08:00 BST were close to the average level of forecasters, and forecast effect of daily maximum and minimum temperature which initiated at 08:00 BST was better than that initiated at 20:00 BST. The forecast accuracy generally decreased with forecast period. (2)The monthly test results show that maximum and minimum temperature forecast accuracy was lowest in January and highest in July, forecast accuracy of the second half year was higher than that in the first half year which initiated at 08:00 BST. The monthly test results show that minimum and maximum temperature forecast that initiated at 08:00 BST and 20:00 BST presented two wave peaks in March and July, three wave valleys in January, April and September. (3)Through principal component analysis of maximum temperature forecast at Changsha and Liuyang stations, it was found that forecast results had a great (small) deviation with observations before (after)June, and forecast results were lower than observations before June at Changsha station, while the deviation at Liuyang station was larger than that in Changsha , which indicated that forecast results were more accurate in Changsha than that of Liuyang station.

Key words:  FUSE forecast products, temperature forecast, forecast accuracy, error analysis, Changsha, principal component analysis

摘要：

将精细化气象要素预报支撑环境（FUSE）产品经过本地化应用开发，形成对预报员有用的显示和扩展平台，方便预报员随时查看温度、降水预报，了解FUSE产品预报性能，缩短使用FUSE产品的时间，提高预报效率。在此基础上，对长沙、浏阳、宁乡3站2013年5个时效的FUSE温度预报产品进行评分，并对长沙、浏阳2站高温预报误差的主成分进行分析。结果表明：（1）08时起报的长沙、浏阳、宁乡3站72 h高温预报和72 h时段以内低温预报准确率已经接近预报员平均水平，08时起报的高低温预报效果均要优于20时，随着预报时效延长，08时和20时起报的预报准确率均表现出降低趋势；（2）长沙、浏阳、宁乡3站高低温预报准确率都是1月最低，7月最高，20时起报的低温预报准确率下半年总体比上半年高，08时和20时起报的高低温预报逐月检验结果均呈现出2个波峰、3个波谷的特征，波峰出现在3月和7月，波谷出现在1月、4月和9月；（3）长沙、浏阳2站高温预报6月之前预报值偏离实况大，6月之后预报值偏离实况小，其中长沙站6月之前预报值较实况偏低，6月之后预报值较实况偏高、偏低的频次差别不大；浏阳站预报值偏离实况比长沙站大。说明FUSE系统对长沙站的预报比浏阳站更准确。

关键词: FUSE预报产品, 温度预报, 预报准确率, 误差, 长沙, 主成分分析