基于WACCM+DART的临近空间SABER和MLS温度观测资料同化试验

基于WACCM+DART（Whole Atmosphere Community Climate Model，Data Assimilation Research Test-Bed）临近空间资料同化预报系统，以2016年2月的一次平流层爆发性增温（SSW）事件为例，开展了临近空间SABER（Sounding of the Atmosphere using Broadband Emission Radiometry）和MLS（Microwave Limb Sounder）温度观测资料集合滤波同化试验.结果表明:同化SABER和MLS温度观测资料可显著降低WACCM模式在中间层和平流层中上部（0.001～10hPa）大气温度场的预报误差，改善CR试验在SSW发生时中间层变冷现象偏强、纬向风场首次发生反转的层次偏低以及增温恢复阶段0.1～10hPa的东风层提前消退、纬向风速偏大、平流层顶位置偏高等现象.基于ERA5（The Fifth Generation of ECMWF Reanalyses）再分析资料的检验表明:同化SABER和MLS温度资料明显有利于减小北半球高纬度地区（60°-90°N）平流层中上层和下中间层（0.1～14hPa）纬向风场以及平流层和中间层中下层（0.01～100hPa）温度场的分析误差；同化低层大气观测也有利于减小0.1～14hPa纬向风场和0.01～100hPa温度场的分析误差，但是不如同化SABER和MLS温度资料对临近空间纬向风场和温度场分析误差的改善效果显著. 

Assimilation of Near Space Temperature Data from SABER and MLS Observations into the Whole Atmosphere Community Climate Model and Data Assimilation Research Test-bed

This study performs SABER(Sounding of the Atmosphere using Broadband Emission Radiometry) and MLS(Microwave Limb Sounder) temperature data assimilation experiments to simulate a SSW(Stratospheric Sudden Warming) process occurred in February 2016,based on WACCM+DART(Whole Atmosphere Community Climate Model,Data Assimilation Research Testbed).The following main conclusions are obtained.First,assimilating SABER and MLS temperature observations significantly reduces WACCM’s forecast error of temperature fields in mesosphere and middle-upper stratosphere(0.001~10 hPa),and effectively improves control experiment’s several discrepancies with observations and reanalysis,such as colder mesosphere during SSW,lower layer height that zonal wind direction firstly changes when SSW occurs,east zonal wind layers in0.1~10 hPa prematurely vanishing,stronger zonal wind and higher stratopause height during SSW recovery phase.The verification based on ERA5 reanalysis suggests that assimilating SABER and MLS temperature observations is in favor of reducing analysis error of zonal wind in low mesosphere and middle-upper stratosphere(0.1~14 hPa) and temperature in stratosphere and middle-lower mesosphere(0.01~100 hPa) above high-latitude areas(60°-90°N)in the northern hemisphere.In addition,assimilating low atmospheric observations is also beneficial for reducing analysis error of zonal wind in 0.1~14 hPa and temperature in 0.01~100 hPa,but this reduction effect is not as significant as that of assimilating SABER and MLS temperature observations.