GTS同化对DOGRAFS预报新疆区域一次特大暴雪的影响

Impact of GTS Assimilation upon DOGRAFS in Prediction of a Streeme Snowstorm in Xinjiang

受欧洲脊发展和乌拉尔低槽东移影响,2015年12月9日20时到12月13日08时,新疆大部出现了大暴雪、寒潮和大风天气过程,其中乌鲁木齐、米东区、小渠子、白杨沟等站日降雪量均突破当日历史记录值。为定量评估新疆数值天气预报业务系统DOGRAFS（DesertOasis-Gobi Rapid Analysis＆Forecast System）对此次暴雪过程的预报准确率,以NCEP的GFS（Global Forecast System）预报场作为背景场,基于WRFv3.5.1和WRFDAv3.5.1同化WMO的常规观测资料GTS（Global Telecommunications System）进行积分预报。结果表明,DOGRAFS预报的逐24 h累计降水在北疆的落区和量级均与实况接近,GTS同化对暴雪极值中心乌鲁木齐站的逐1 h降水预报为负效果。对DOGRAFS预报区域内的所有站点而言,2 m温度预报的平均偏差在-3~4.2℃,绝对偏差在-1.6~2.5℃。高空温度场预报的绝对误差小于3.5℃。加入GTS后,乌鲁木齐站温度预报偏差大幅度增大,预报效果明显变差,不同化GTS预报的乌鲁木齐单站逐小时2 m温度及其变化趋势与实况非常接近。

Influenced by the development of the European ridge and the eastward movement of Ural low trough, an extremely heavy snowstorm accompanied by strong cold wave and strong gale occurred in majority region of Xinjiang from December 9th till 12nd, 2015. To quantitatively evaluate the forecast capacity of the Desert-Oasis-Gobi Rapid Analysis ＆ Forecast System （DOGRAFS）, a series of experiments were conducted by this research, integrating by WRFV3.5.1 using the prediction field from Global Forecast System （GFS） as the background, while assimilating the traditional observations from Global Telecommunications System （GTS） by WRFDAv3.5.1 to forecast the process. It is concluded that, the forecasted distribution and volume of 24h accumulation precipitation is thoroughly consist with the actual situation. However, in Urumqi station, the forecast performance of hourly snowfall is worse with GTS assimilation. For all the stations in the simulation domain, the mean bias of 2 m temperature is between -3℃-4.2℃, while the absolute error is -1.6 ℃-2.5 ℃. The maximum prediction RMSE of the spatial temperature is 3.5 ℃. Similarly, there is a negative impact on 2 m temperature forecast with GTS assimulation in Urumqi station.