沙漠绿洲戈壁区域同化预报系统降水预报检验
Performance verification of Desert Oasis Gobi Regional Assimilation andForecast System for precipitation forecast
-
摘要: 基于沙漠绿洲戈壁区域同化预报系统(Desert Oasis Gobi Regional Assimilation and Forecast System,简称“DOGRAFS”),对2012 年11 月1 日—2013 年10 月31 日逐日四次预报的6 h 累积降水量和实况资料,利用TS、ETS 和BS 评分统计量,对该系统的降水预报能力进行客观检验评估,结果表明:总体而言,模式预报效果存在一定的季节差异,夏秋两季预报评分高于冬春两季;对于同一降水阈值,系统不同起报时间对降水的预报结果差别不大,但总体上12 UTC 略优于其它三个时次;模式对降水的预报能力随着降水阈值的增大而逐渐降低,其中,对0.1 mm·(6 h)-1阈值的降水预报性能最稳定,对3.1 mm·(6 h)-1和6.1 mm·(6 h)-1两阈值次之但预报范围与实况比较吻合,对12.1 mm·(6 h)-1以上的大阈值降水过程,整体把握能力还有待提高。另外,针对2013 年5 月26—29 日南疆一次极端强降水天气个例进行分析,不同起报时间对强降水时段的落区和量级的预报能力参差不齐,其中26 日12 UTC 的预报结果最优;对其增量场分析表明,同化的资料对预报初始场中、低层的风场、温度场以及湿度场都有明显的调整作用,对预报结果有较好的正效应。Abstract: Based on the 6 h rainfall forecasts of Desert Oasis Gobi Regional Assimilation and Forecast System (DOGRAFS)and observationsat four initialized times daily during 1 November 2012 to 31 October 2013,the capabilities to precipitation forecast is evaluated preliminarilyin terms of Threat Score,Equitable Threat Score and Bias Score. The results indicate that the performance of DOGRAFS changes closelycorresponding to the seasonal variation as a whole,which tends to be better in summer and autumn than in winter and spring. With the samethreshold,the difference in performance among the results of different initialized times is small,although the result at 12 UTC is slightly betterthan the others. The forecast skill decreases with the threshold increasing,among which the 0.1 mm·(6 h)-1 threshold behaves the most stableand the 3.1 mm·(6 h)-1 and 6.1 mm·(6 h)-1 are at the second place. For larger rainfall thresholds such as 12.1 mm·(6 h)-1 and 24.1 mm·(6 h)-1,the performance is relatively poor. In addition,the strong precipitation event occurred in the southwest of Xinjiang during 26th —29th May2013 is analyzed as an example. In this case,the forecast at different initialized times has an uneven performance for the rainfall area andgrade by comparing the observed and forecast precipitation; among which the forecast at 12 UTC 26th behaves the best. Furthermore,the assimilateddata have improved wind,temperature and humidity fields remarkably and have a positive effect on the forecast results by means ofanalysis increments.