A global ocean general circulation model （L30T63） is employed to study the uptake and distribution of anthropogenic CO2 in the ocean. A subgrid-scale mixing scheme called GM90 is used in the model. There are two main GM90 parameters including isopycnal diffusivity and skew （thickness） diffusivity. Sensitivities of the ocean circulation and the redistribution of dissolved anthropogenic CO2 to these two parameters are examined. Two runs estimate the global oceanic anthropogenic CO2 uptake to be 1.64 and 1.73 Pg C yr^-1 for the 1990s, and that the global ocean contained 86.8 and 92.7 Pg C of anthropogenic CO2 at the end of 1994, respectively. Both the total inventory and uptake from our model are smaller than the data-based estimates. In this presentation, the vertical distributions of anthropogenic CO2 at three meridional sections are discussed and compared with the available data-based estimates. The inventory in the individual basins is also calculated. Use of large isopycnal diffusivity can generally improve the simulated results, including the exchange flux, the vertical distribution patterns, inventory, storage, etc. In terms of comparison of the vertical distributions and column inventory, we find that the total inventory in the Pacific Ocean obtained from our model is in good agreement with the data-based estimate, but a large difference exists in the Atlantic Ocean, particularly in the South Atlantic. The main reasons are weak vertical mixing and that our model generates small exchange fluxes of anthropogenic CO2 in the Southern Ocean. Improvement in the simulation of the vertical transport and sea ice in the Southern Ocean is important in future work.
A global ocean general circulation model, called LASG/IAP Climate system ocean model （LICOM）, is employed to study the influence of climate change on the uptake and storage of anthropogenic CO2 in the global ocean. Two simulations were made： the control run （RUN1） with the climatological daily mean forcing data, and the climate change run （RUN2） with the interannually varying daily mean forcing data from the NCEP （National Centers for Environmental Prediction） of the US. The results show that the simulated distributions and storages of anthropogenic dissolved inorganic carbon （anDIC） from both runs are consistent with the data-based results. Compared with the data-based results, the simulations generate higher anDIC concentrations in the upper layer and lower storage amount of anDIC between the subsurface and 1000-m depth, especially in RUN1. A comparison of the two runs shows that the interannually varying forcing can enhance the transport of main water masses, so the rate of interior transport of anDIC is increased. The higher transfer rate of anDIC in RUN2 decreases its high concentration in the upper layer and increases its storage amount below the subsurface, which leads to closer distributions of anDIC in RUN2 to the data-based results than in RUN1. The higher transfer rate in RUN2 also induces larger exchange flux than in RUN1. It is estimated that the global oceanic anthropogenic CO2 uptake was 1.83 and 2.16 Pg C yr-1 in the two runs in 1995, respectively, and as of 1994, the global ocean contained 99 Pg C in RUN1 and 107 Pg C in RUN2 of anDIC, indicating that the model under the interannually varying forcing could take up 8.1% more anthropogenic carbon than the model under the climatological forcing. These values are within the range of other estimates based on observation and model simulation, while the estimates in RUN1 are near the low bound of other works. It is estimated that the variability of root mean square of the global air-sea anthropogenic carbon flux from the simulated
The coupled models of both the Global Ocean-Atmosphere-Land System （GOALS） and the Atmosphere- Vegetation Interaction Model （GOALS-AVIM） are used to study the main characteristics of interannual variations. The simulated results are also used to investigate some significant interannual variability and correlation analysis of the atmospheric circulation and terrestrial ecosystem. By comparing the simulations of the climate model GOALS-AVIM and GOALS, it is known that the simulated results of the interannual variations of the spatial and temporal distributions of the surface air temperatures and precipitation are generally improved by using AVIM in GOALS-AVIM. The interannual variation displays some distinct characteristics of the geographical distribution. Both the Net Primary Production （NPP） and the Leap Area Index （LAI） have quasi 1-2-year cycles. Meanwhile, precipitation and the surface temperatures have 2-4- year cycles. Conditions when the spectrum density values of GOALS are less than those of GOALS-AVIM, tell us that the model coupled with AVIM enhances the simulative capability for interannual variability and makes the annual cycle variability more apparent. Using Singular Value Decomposition （SVD） analysis, the relationship between the ecosystem and the atmospheric circulation in East Asia is explored. The result shows that the strengthening and weakening of the East Asian monsoon, characterized by the geopotential heights at 500 hPa and the wind fields at 850 hPa, correspond to the spatiotemporal pattern of the NPP. The correlation between NPP and the air temperature, precipitation and solar radiation are different in interannual variability because of the variation in vegetation types.
A basin-wide ocean general circulation model （OGCM） of the Pacific Ocean is employed to estimate the uptake and storage of anthropogenic CO2 using two different simulation approaches. The simulation （named BIO） makes use of a carbon model with biological processes and full thermodynamic equations to calculate surface water partial pressure of CO2, whereas the other simulation （named PTB） makes use of a perturbation approach to calculate surface water partial pressure of anthropogenic CO2. The results from the two simulations agree well with the estimates based on observation data in most important aspects of the vertical distribution as well as the total inventory of anthropogenic carbon. The storage of anthropogenic carbon from BIO is closer to the observation-based estimate than that from PTB. The Revelle factor in 1994 obtained in BIO is generally larger than that obtained in PTB in the whole Pacific, except for the subtropical South Pacific. This, to large extent, leads to the difference in the surface anthropogenic CO2 concentration between the two runs. The relative difference in the annual uptake between the two runs is almost constant during the integration processes after 1850. This is probably not caused by dissolved inorganic carbon （DIC）, but rather by a factor independent of time. In both runs, the rate of change in anthropogenic CO2 fluxes with time is consistent with the rate of change in the growth rate of atmospheric partial pressure of CO2.
