论文部分内容阅读
鉴于集合预报能在一定程度上减少单一数值预报的不确定性,以清河水库以上流域为例,根据TIGGE提供的降雨预报信息,选取ECMWF、NCEP、CMA三个降雨预报中心中的扰动预报模式,利用TS评分、Brier(BS)评分和Talagrand分布三个指标,对2007~2013年5~10月预见期为24、48、120、240h的降雨预报信息进行评估。结果表明,三个中心对低量级的降雨预报效果较好,整体表现为随着预见期的增长准确率逐渐降低,并存在一定的空报现象。并从实用性出发,进一步对ECMWF预见期为240h的降雨预报信息进行三率分析,验证了该预报信息可用于水库的实时防洪调度决策。
In view of the fact that ensemble forecasting can reduce the uncertainty of single numerical forecast to a certain extent, taking the watersheds above Qinghe Reservoir as an example, the disturbance forecasting models in three rainfall forecasting centers of ECMWF, NCEP and CMA are selected based on the rainfall forecast information provided by TIGGE, By using three indexes of TS score, Brier (BS) score and Talagrand distribution, the rainfall forecast information from May to October from 2007 to 2013 with a forecast period of 24, 48, 120 and 240 h was evaluated. The results show that the three centers are good for forecasting low-level rainfall, and the overall performance is that the accuracy decreases gradually with the increase of the forecast period, and there is a certain phenomenon of empty report. Based on the practicality, further three-rate analysis of the rainfall forecast information of ECMWF with a forecast period of 240h is conducted to verify that the forecast information can be used in the real-time flood control decision-making of reservoirs.