大气科学进展(英文版)
大氣科學進展(英文版)
대기과학진전(영문판)
ADVANCES IN ATMOSPHERIC SCIENCES
2004年
2期
181-192
,共12页
data assimilation%background error%model output%covariance
The background error covariance plays an important role in modern data assimilation and analysis systems by determining the spatial spreading of information in the data. A novel method based on model output is proposed to estimate background error covariance for use in Optimum Interpolation. At every model level, anisotropic correlation scales are obtained that give a more detailed description of the spatial correlation structure. Furthermore, the impact of the background field itself is included in the background error covariance. The methodology of the estimation is presented and the structure of the covariance is examined. The results of 20-year assimilation experiments are compared with observations from TOGATAO (The Tropical Ocean-Global Atmosphere-Tropical Atmosphere Ocean) array and other analysis data.
