BCC-GODAS

BCC-GODAS, as one part of climate dynamic model operational system, has been employed to offer routinely the ocean initial fields since October of 2001. This system mainly contains 4 parts, such as pre-processing data, calculating realtime global wind stress, variational analysis and interpolating and ocean dynamic model.

Handling data includes collecting data, decoding data and data quality control. The observation data used in this system contains both GTSPP data (from 1981) and Argo data (from 1998). The real-time GTSPP data are obtained from the data bank of Meteorological Information Center of CMA (China Meteorological administration). Argo data are downloaded from French Research Institute for Exploitation of the Sea(IFREMER). It should be pointed out that we have used the observation data from NOAA and MEDS to retrieve history observation data.

Having been calculated in the observation points, the wind stress data are processed by means of optimal interpolation. The wind data, we accessed, for years from 1982 to 1990 in the tropic Pacific are a bit lack. we will manage to retrieve these data.

The scheme of 3-dimensional variation is adopted in BCC-GODAS. Meanwhile, a time-window, about 4-weeks, is opened in order to obtain as much observation information as possible. In BCC-GODAS, both sea temperature and salinity observation data, if available, are input. The background errors covariance involves a vertical weak correlation and a loose correlation between temperature and salinity.

The dynamic model used in this system is L30T63 OGCM Version 1.0, which is established and developed by LASG/IAP. This model has the same horizontal resolution as T63 atmosphere model, and there are 30 layers in the vertical direction, in which the first 10 layers are from 0 to 250m, the second 10 layers span from 250m to 1000m, and the last 10 layers are located deeper till 5600m. There are two vertical mixing parameterization schemes in this model. One is based on Richardson number applied in the tropic region from 30S to 30N. Another is Isopycnal mixing scheme. While using this model, we have made some improvement on the vertical mixing parameterization scheme: a transition zone to connect two areas mentioned above is designed, and the criterion of gradient of density in Isopycnal mixing scheme is redefined as a spatial function rather than a constant.

Acknowledgment: we wish to thank IFREMER, CDC/NOAA, NODC/NOAA, MEDS(Canada) and China Argo Realtime Data Center for their observation data.

References

Yimin Liu, Jiangxing Zhou and Qiang Ma: Study on The Pacific and The Indian Oceanic Data Assimilation System, documents for national key project-studies on short-term climate prediction system in China (1996-2000), Chinese Meteorological Press, 2000. 401-407 pp. In Chinese.
Yimin Liu ????Optimal estimation of the model covariance matrix of Ocean data assimilation system by neural network method, a poster in "Workshop on Advances in Marine Climatology" held in Brussels on 3/11/2003
Yimin Liu, Renhe Zhang, Yonghong Yin and Tao Niu: The application of Argo data to optimal estimation of the model errors covariance matrix of ocean data assimilation system, a poster in the first Argo Science workshop held in Tokyo on 11/11/2003.
Yimin Liu, Weijing Li, Peiqun Zhang: "NCC 4-Dimensional Ocean Data Assimilation System and the Studies on Its Results in the Middle Pacific", to be published.
Yimin Liu, Renhe Zhang, Yonghong Yin and Tao Niu: "The Application of ARGO Data to NCC 4-Dimensional Ocean Data Assimilation System", to be published.