地基微波辐射计资料的质量控制及同化应用研究

Research on quality control and assimilation application of ground-based microwave radiometer data

针对地基微波辐射计反演资料的质量问题,利用北京区域7个站点的RPG-HATPRO型微波辐射计反演的温湿数据,在传统质量控制方法的基础上,通过增加双权重检查和偏差订正方法,设计了新的质量控制方案,而后针对质量控制后的温湿廓线开展了同化应用研究,结果表明:只有个别站点的反演温度误差较大,反演温度误差大的站点在质量控制前最大均方根误差为8℃,但在质量控制后均方根误差下降到4℃以内,误差下降50%以上,而对于反演温度误差小的站点,质量控制只对高层数据有所改善;进一步分析表明所有站点的反演相对湿度误差都比较大,在质量控制前反演的相对湿度均方根误差最大甚至达到50%,但在质量控制后所有站点反演的相对湿度均方根误差都降到20%以内,误差显著降低,同时经过质量控制后的反演温湿数据误差平均偏差降为0左右,误差满足高斯分布,符合同化的要求。同化试验表明:与控制试验相比,同化质量控制后的地基微波辐射计反演温湿廓线资料能有效改善模式对所有量级的6 h累积降水量预报,无论是降水中心位置预报还是降水量级预报,都使得模式降水预报更加趋近实况,同化试验还表明同化对温湿场有较大影响,使得模式温湿场更加合理,能有效提高模式的预报准确率。

In this paper,aiming at the quality problem of the retrieved data of ground-based microwave radiometer,on the basis of traditional quality control methods,a new quality control scheme was designed by adding double weight check and deviation correction methods by using the temperature and humidity data retrieved by RPG-HATPRO microwave radiometer at 7 stations in Beijing area.Then,the assimilation and application of temperature and humidity profile after quality control were studied.Results show that:only individual stations have large inversion temperature error,and the maximum Root Mean Square Error(RMSE) of stations with large inversion temperature error is 8 ℃ before quality control,but after quality control,the RMSE drops to less than 4 ℃,and the error drops by more than 50%.However,for sites with small inversion temperature error,quality control only improves the high-level data.Further analysis shows that the relative humidity inversion error of all stations is relatively large,and the RMSE of relative humidity inversion before quality control is the largest or even reaches 50%,but the RMSE of relative humidity inversion of all stations after quality control is reduced to less than 20%,and the error is significantly reduced.After quality control,the Average Deviation(AD) of inversion temperature and humidity data is reduced to about zero,and the error distribution meets Gaussian distribution,which can meet the requirements before assimilation.This quality control scheme provides technical support for the effective use of temperature and humidity profile data of ground-based microwave radiometer in numerical model.Assimilation test shows that the inversion of temperature and humidity profile data by ground-based microwave radiometer after assimilation quality control can effectively improve the model's 6-hour cumulative precipitation forecast for all orders of magnitude,which makes the model's precipitation forecast closer to the reality.Assimilation test also shows that assimilation has great influence on the temperature and humidity field,which makes the model's temperature and humidity field more reasonable and can effectively improve the model's forecast accuracy.