Rec-EOF质量控制方法在地面观测2m比湿中的应用

针对地面观测站资料,从资料同化的角度出发,通过引入递归的概念,提出了递归经验正交质量控制(Rec-EOF QC)方法,并将该方案应用于不同时段的地面观测资料和不同种类的背景场资料中。结果表明:Rec-EOF QC方法一般在递归4次之后,质量控制过程达到稳定状态。通过递归方法,该质量控制方案能够通过资料本身调整质量控制的临界值,对资料具有普适性。而且与传统EOF QC方法相比,Rec-EOF QC方案能识别出更多的离群值,并且可以使观测偏差也更接近于高斯分布。另外还研究了背景场差异对质量控制的影响,分别利用FNL、ERA-Interim和JRA-25三组资料作为背景场进行地面质量质量控制试验,结果表明JRA-25剔除的资料最多,ERA-Interim次之,FNL最小。无论选用何种资料作为背景场,冬季的资料剔除率始终大于夏季的。

Application of recursive EOF quality control to 2 m specific humidity from ground-based observations

n order to improve the surface observation data assimilation from ground-based stations, through introducing the recursive theory, a recursive EOF quality control(Rec-EOF QC) method is put forward and is applied to the surface observation data with different periods and the background field data with different varieties. Results show that the quality control process of Rec-EOF QC method is stable as a result of four times of recursion. The Rec-EOF QC method can adjust the threshold of the quality control by the data itself and is universally applicable to different data. Compared with the traditional EOF QC method, the Rec-EOF QC scheme can identify more outliers and can make the observation error more close to Gaussian distribution. In addition, FNL, ERA-Interim and JRA-25 datasets are selected as background fields to conduct surface quality control test, whose results show that outliers identified by JRA-25 as background data are the maximum, ERA-Interim takes the second and FNL is the minimum. Moreover, outliers are identified more in winter than those in summer regardless of what kinds of data were chosen as background.