During the 9th Five-Year National Development Plan period, the Institute of Atmospheric Physics (IAP) of the Chinese Academy of Science, NCC, and the State Oceanic Administration collaboratively developed a global oceanic circulation model (T63L30 OGCM_1.0), a high-resolution Indian Ocean-Pacific basin model and a thermodynamic-dynamical sea-ice model.
The OGCM model was developed on the basis of the IAP’s 20-layer oceanic model (with 4 x 5 degrees resolution). Its horizontal resolution is 1.875 x 1.875 degrees and its vertical resolution is 30 layers, including l0 layers within top 250 meters, another 10 layers between 250 to 1000 meters, and the rest 10 layers between 10005600 meters. This results in the resolution of thermocline increased greatly. No “rigid” approximation is applied in the model. Therefore, the sea level is a predictand.
The control equations are based on a set of primitive baroclinic equations. In convention to describe the free surface and to deal with the complicated topography at oceanic bottom, ηcoordinate system is adopted vertically in the model. Due to the introduction of free sea surface, the model’ s primitive equations could include a rapid external gravity wave. The state equation of sea water is computed based on the UNESCO formula of 3rd order polynomial, in which only density disturbance related to sea temperature and salinity is calculated, that is, the reference stratification at a specific depth is deducted (Bryan and Cox, 1972, UNESCO 1981). In the process of model integration, a methodology of separation-couple of barotropic and baroclinic equations is utilized for saving computation hours. Within the baroclinic mode, the thermo-salinity process is further separated from the momentum process. The frog-leap scheme is adopted in the integrations of barotropic, baroclinic and thermo-salinity process, each taking a different time length, i.e. 2 minutes, 4 hours and 8 hours respectively.
Newton relaxation boundary condition is applied on the oceanic surface, and its surface heat flux is calculated from Haney (1971) formula. The sea ice state depends on the principles of Parkinson and Washington (1979) and it is calculated by a thermodynamic sea-ice model. When model predicted SST reaches freezing point, sea ice is generated. Since “lead” is taken into consideration, the sea temperature within each grid covered by sea-ice is determined by the air-sea heat exchanges through “lead”. Hence, this model not only responses to the changes of sea ice thickness, but also to the status of sea ice domain and lead coverage. Concerning the sub-grid-scale physical process parameterization, a mixed scheme along isopycnal surface developed by Gent and McWilliams (1990) is adopted in the model so as to improve the simulation of main oceanic thermocline. In the upper layer of the tropical ocean between 30oN and 30oS, the vertical mixed scheme (Pacanowski and Philander 1981) is adopted to improve the simulation of the equatorial thermocline. In addition, the parameterization scheme of solar short-wave transmission (Rosati, 1988) is also adopted, but relevant parameters are adjusted according to model requirements.
The control experiment of oceanic circulation model (L30T63) has maintained a stable integration for over 3000 years. The oceanic model has been coupled with several atmospheric circulation models and a series of model experiments have been done. The basic assessment results indicate that the performance of L30T63 has reached, at least, the mean level of all the models n the world.
The Indian and Pacific Oceanic circulation model is developed on the basis of GFDL MOM2, covering the Indian and Pacific basin to the north of 35°S (25°E～70°W，35°S～60°N). Its horizontal resolution is 1/3°～1°×1.5° and it has 31 vertical layers. The upper 12 layers are distributed equidistantly with an interval of 10m, and there are 22 layers above 400m. In the model’s southern boundary, the Stevens active open boundary condition is used to nest the model, in one-way, with the global oceanic model. Compared with the global oceanic model, this model could simulate more specifically and accurately the climatological and seasonal variation in both Indian and Pacific Oceans, particularly the corresponding variations in thermocline layer and the wave propagation process in tropical oceans. The experiment of forcing model with observational sea-surface wind stress indicates that the model can simulate, in a better way, the EI Nino events in the tropical Pacific Ocean, the dipoles in the tropical Indian Ocean, and the changes of the throughout current around Indonesia.
The thermodynamic – dynamic ice model has been coupled with the global oceanic model. Compared with similar foreign models, the result from our simulating experiment indicates that simulation of sea-ice coverage is at an equal level, while that of sea-ice drifts lagged behind. The proposed new scheme of ice-sea thermodynamical coupling has been applied to regional ice-sea coupling experiments in Bohai Sea and Baltic Sea.