AMDAR温度资料的偏差订正及对GRAPES系统的影响  

Bias Correction of AMDAR Temperature Data and Its Impact on GRAPES System

作　　者：王瑞文[1] 龚建东[1] 韩威 张林 WANG Ruiwen GONG Jiandong HAN Wei ZHANG Lin(National Meteorological Center, China Meteorological Administration, Beijing 100081, China)

机构地区：[1]中国气象局国家气象中心,北京100081

出　　处：《高原气象》2017年第5期1346-1356,共11页Plateau Meteorology

基　　金：GRAPES全球数值预报系统专项(GRAPES-FZZX-2016-05;GRAPES-FZZX-2016-14);国家科技支撑计划项目(2012BAC22B02)

摘　　要：通过对AMDAR(Aircraft Meteorological Data Relay)温度资料进行了偏差订正来改善数值预报同化预报效果。方法是以NECP(National Centers for Environmental Prediction)的GFS(Global Forecast System)的6 h预报场作为参考场来统计AM DAR观测温度减去参考场的偏差,根据飞机上升率和下降率,应用最小二乘法拟合温度资料偏差的线性回归方程,然后用回归方程结合探空资料对温度观测进行偏差订正。AMDAR资料来源是国家气象信息中心数据库,偏差统计时间为冬季(2014年12月到2015年1—2月)和夏季(2014年6—8月)。另外,应用GRAPES(Global Regional Assimilation and Prediction System)系统对订正后的AM DAR温度资料进行了1个月(2013年12月)的同化预报影响试验,结果表明,同NCEP的FNL(Final Operational Global Analysis data)资料相比,加入订正后的AMDAR温度资料可以减小高度分析场的偏差和均方根误差,尤其在平飞阶段(300~150 hPa),其偏差、均方根误差减小了2 gpm,对温度分析场的影响也有正效果,偏差和均方根误差最多减小0.15℃。高度场和温度场的预报距平相关系数在250 hPa以上也比温度资料订正前有所提高,温度场的预报时效提高了0.5天。Bias correction of AMDAR（ Aircraft Meteorological Data Relay） temperature data is used to improve numerical weather forecast effect in this paper. 6 hour forecast of global forecast system in national center for environmental prediction is as reference field,and bias statistic of AMDAR temperature data minus reference filed has been done. The equation of linear regression of bias of AMDAR temperature is fitted by the method of least square method according to the rate of rise and descend,and bias of AMDAR temperature can be corrected by combing the radiosonde data. The AMDAR data is from real-time database in National Meteorological Information Centre（ NMIC）,and the two period for the data is from June to August 2014 and from December 2014 to February 2015. One month numerical experiment has been done by the use of corrected AMDAR temperature data in GRAPES（ Global Regional Assimilation and Prediction System）. The results showthat Bias and RMSE（ root mean square error） of height analysis are decreased by assimilating corrected AMDAR temperature data than uncorrected data. Bias and RMSE of height decrease 2 gpm on 300 - 150 hPa. The maximum reduction of bias and RMSE of temperature is close to 0. 15 ℃. Forecast anomaly correlation of height and temperature is increased above 250 hPa than before correction. The period of validity of temperature improve 0. 5 days.

关 键 词：偏差订正 GRAPES AMDAR资料 回归方程 

分 类 号：P456.7[天文地球—大气科学及气象学]