Over the past 3 to 4 years, there have been two climate models based at NCAR: the Climate System Model (CSM) and the Parallel Climate Model (PCM). The first versions of both models use the same atmosphere and land components, namely the CCM3 and the LSM. However, the ocean and sea-ice components are different in the CSM and the PCM.
The CSM ocean component uses a spherical grid and has filtering in the Arctic Ocean. The resolution is 2.4° x 1.2° – 2.3° x 45 levels, with 4 levels in the upper 50 m. This component uses the KPP vertical mixing scheme of Large et al. (1994) and the mesoscale eddy parameterization of Gent and McWilliams (1990). The PCM ocean component uses the POP code with a displaced northern pole, so that no filtering is required. The resolution is 2/3° – 1.2° x 0.5° – 1° x 32 levels, with 2 levels in the upper 50 m. The vertical mixing scheme is from Pacanowski and Philander (1981), and the mesoscale eddy parameterization is biharmonic horizontal mixing of potential temperature and salinity.
The CSM sea-ice component uses the cavitating fluid rheology of Flato and Hibler (1992) and is on the same grid as the ocean component. The PCM sea-ice component uses the viscous-plastic rheology of Zhang and Rothrock (1997), but is on a polar projection grid that is not the same as the ocean component grid.
In discussions between the CSM and PCM groups at NCAR about the second versions of their respective models, it was realized that there was considerable convergence in their choices of component models. Both wanted to use the same upgraded atmosphere and land components. In addition, both wanted to use the same sea-ice component that was under development. It is based on the elastic-viscous-plastic rheology of Hunke and Dukowicz (1997) and the ice thickness distribution of Bitz et al. (2000). This component works with a displaced pole grid, and so can be used on the same grid as a POP ocean component.
Both CSM and PCM wanted their ocean components to be based on the POP code and to use the displaced pole grid to avoid filtering. This left the only discrepancy between the two ocean components being the different choices of vertical mixing and eddy parameterization schemes. Thus, it was decided to compare these schemes in an ocean alone setting with the finer horizontal grid used in the PCM-1. The forcing used is described in Large et al. (1997) and uses NCEP reanalyses and other observed quantities. Four different integrations of at least 100 years were performed starting with initial conditions from Levitus observations.
They are:
a) Case 1.06 PP(81) vertical mixing and biharmonic mixing of T,S.
b) Case 1.07 KPP(94) vertical mixing and biharmonic mixing of T,S.
c) Case 1.09 KPP(94) vertical mixing and GM(90) parameterization
d) Case 1.10 KPP(94) vertical mixing and Visbeck et al. (1997)
The Visbeck et al. (1997) parameterization is a modification of the GM90 scheme, where the coefficient is not a constant, but varies in the horizontal depending on the local baroclinicity of the ocean flow. Results from these calculation are shown in Figures 1 – 3. In ocean alone integrations like these, the globally-averaged SST remains very close to the observed value. The reason is that the heat budget is balanced using observed SSTs, and the negative feedbacks in the heat flux laws keep the averaged SST very close to the observed value. The mixing parameterizations then set the ocean temperature profile below the surface, and in 100 years this will have come into equilibrium over about the upper kilometer of the ocean. Thus, the quality of the mixing parameterizations can be tested by how far the globally-averaged temperature profile drifts away from the initial condition, which comes from the Levitus observations.
FIGURE 1 shows the globally-averaged temperature at 290 m depth from the four cases as a function of time, and FIGURE 2 shows the same quantities but at a depth of 520 m. It is clear at both depths that cases 1.09 and 1.10 have much smaller drifts than cases 1.06 and 1.07. Using the Visbeck et al. (1997) scheme does not make very much difference to these globally-averaged curves, but may improve regional diagnostics of the same quantities. FIGURE 3 shows the global, meridional heat transport at the end of the four 100 year integrations. The largest difference occurs in the southern hemisphere, and the difference at 40°S would be larger if the transport due to the eddy parameterization was added to the case 1.09 and 1.10 curves. These cases would then show poleward heat transport in the southern hemisphere, which is more in accord with observational estimates. The conclusion from these ocean alone integrations is that the KPP and Visbeck et al. parameterizations should be used with the finer horizontal grid used in the PCM.
Once the four model components were agreed upon, then all that was left was to agree that the CSM and PCM couplers should be merged. This process is now also underway, with the goal of producing a next generation coupler that combines the advantages of the present CSM and PCM couplers. Thus, agreement was reached as to merging of the CSM and PCM efforts. In order to reflect this, and the input from a large number of people outside NCAR, it was decided that the CSM effort should be renamed the Community Climate System Model effort. This is the reason for the change in the name of this workshop. I am happy to see many people who have worked on the PCM and its data here at this workshop, and would like to welcome them.
References
Bitz, C. M., M. Holland, A.J. Weaver, and M. Eby, 2000:
Simulating the ice-thickness distribution in a coupled
climate model, J. Phys. Oceanogr., submitted.
Flato, G. M., and W. D. Hibler, 1992: Modeling pack ice as a cavitating fluid. J. Phys. Oceanogr., 22, 626–651.
Gent, P. R., and J.C. McWilliams, 1990: Isopycnal mixing in ocean circulation models. J. Phys. Oceanogr., 20, 150–155.
Hunke, E. C., and J. K. Dukowicz, 1997: An elastic-viscous-plastic model for sea ice dynamics. J. Phys. Oceanogr., 27, 1849–1867.
Large, W.G., G. Danabasoglu, S. Doney, and J.C. McWilliams, 1997: Sensitivity to surface forcing and boundary layer mixing in a global ocean model: Annual-mean climatology, J. Phys. Oceanogr., 27, 2418–2447.
Large, W.G., J.C. McWilliams, and S.C. Doney, 1994: Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization. Reviews of Geophysics, 32, 363–403.
Pacanowski, R.C., and S.G.H. Philander, 1981: Parameterization of vertical mixing in numerical models of the tropical oceans. J. Phys. Oceanogr., 11, 1443–1451.
Visbeck, M., J. Marshall, T. Haine, and M. Spall, 1997: Specification of eddy transfer coefficients in coarse-resolution ocean circulation models, J. Phys. Oceanogr., 27, 381–402.
Zhang, J., and D.A. Rothrock, 1997: On an efficient numerical method for modeling sea ice dynamics. J. Phys. Oceanogr., 102, 8691–8702.