Background Error Covariance Is Estimated That The Technology Applied Research

The background error covariance of the true atmospheric is nearly isotropic. It is a key problem of variational data assimilation to describe the background error covariance. The background error covariance included error variance and characteristic scale-length is not the same in different areas of the global. Here we estimate and tune the background error covariance matrix which adapts to China region. The estimating and tuning combines NMC method which bases on forecast samples and IV method which bases on observation samples. During the course of estimating and tuning the isotropic covariance function is adopted. Three dimensional structure characters of background error covariance of regional numerical model in the China area is obtained from the statistic results of NMC method. And the variance of background error covariance obtained by NMC-method is tuned with the IV-method results. With used it in the limited area meso-scale model WRF 3DVAR, the POST-test and sound data assimilation validates the background error covariance is reasonable.On the other hand, the Doppler Radar Velocity is anisotropic. The application of the background error covariance function in the Radar data quality control is presented which aims at obtaining the VAD wind profile from Doppler radar aliased velocity data without any extra wind information. This is based on the ECMWF optimal interpolation quality control technique. As for the characters of the Doppler Radar radial velocity data, the single variable background error covariance is used. And the anisotropic covariance function is adopted in the horizontal plane. For the tested data on the circle and tilt from which is retrieved the VAD wind profile, the predigested multiple variables statistic interpolation technique, just the Cressman interpolation is used to provide a reference value. According to the reference velocity the Nyquist is decided. Base on the dealiasing the velocity, we could get the horizontal wind velocity field by the classical VAD technique at different acquired, and then obtain vertical wind profile. In order to verify the method, the comparison among the Tabary slope method, the Bayes function method, the Yamada continuity method and the background error covariance function method is presented. The results show that the background error covariance function method is the best one. The comparison between wind profile obtained by wind profiler and the one obtained by the background error covariance function method, and the comparison between the aliased PPI map and dealiased PPI map are also presented. The result shows that it has the ability of obtaining accurate wind profile from fold data.