P-OMIP (Pilot Ocean Model Intercomparison Project)

The GFDL/IARC contribution to P-OMIP

The GFDL/IARC contribution to the Pilot OMIP consists of the MOM4.0 z-coordinate ocean model (Griffies et al. 2003) coupled to the GFDL Sea Ice Simulator (SIS) (Winton 2000). This coupled model is presently being developed at GFDL for use in studying the ocean climate system, and for coupling to land, atmosphere, and ocean biogeochemical models for use in studying the earth system. Note that this contribution follows the P-OMIP protocol, with the exception of using the bulk formulae from the GFDL coupler.

Horizontal resolution and grid

Both the ocean and sea ice models use the Arakawa B-grid. Both also have the same horizontal grid resolution using 180 points in the i-direction and 174 in the j-direction. Their longitudinal grid has a uniform resolution of 2degrees from 78S to 65N. The latitudinal resolution is non-uniform, with refinement towards the poles in a mercator-like fashion, and refinement in the tropics with 2/3 degree resolution within 12 degrees of the equator. North of 65N, the grid switches from spherical to bipolar, using grid factors defined by Murray (1996). A similar grid is used by the HYCOM contribution to P-OMIP. A bipolar Arctic removes the spherical coordinate singularity from the ocean model domain. More details of this non-spherical grid projection can be found at http://www.gfdl.gov/~mw/docs/grid_coupling.php.

Sea Ice Model

The GFDL Sea Ice Simulator (SIS) is a dynamical ice model using the elastic-viscous-plastic technique of Hunke and Dukowicz (1997) to implement the viscous-plastic rheology. The thermodynamic treatment is similar to that of Semtner (1976) with two ice layers and one snow layer. As described by Winton (2000), brine content of the upper ice is simulated rather than parameterized as in Semtner (1976). The model allows for an arbitrary number of ice thickness categories, with 5 chosen here, as in the NCAR CSM sea ice model.

Ocean vertical resolution and topography

There are 50 vertical levels in the ocean, with constant 10m resolution from the ocean surface down to 220m, and increasing grid spacing towards the bottom at 5500m. This grid resolution is motivated by our desire to provide a faithful representation of the pycnocline throughout the World Ocean, especially within the tropics, as well as upper ocean mixed layer processes important for ocean climate phenomena. The bottom topography is represented by the partial cells described by Adcroft et al. (1997) and Pacanowski and Gnanadesikan (1998). Partial cells allow for a faithful treatment of topographic waves, especially those occurring in regions of shallow topographic slope. The topography is a coarsened version of that generated by the Southampton Group for their 1/12 degree OCCAM simulations.

Ocean dynamics and physics

The ocean model is based on the hydrostatic and Boussinesq approximations with a z-coordinate vertical discretization. Time stepping for the inviscid dynamics remains the traditional leap-frog, with a Robert-Asselin time filter applied each time step with a damping coefficient of 0.05. Vertical physical mixing processes are handled implicitly in time so to allow for realistically large diapycnal diffusivities in mixed layer regions. Coupling the ice model to the ocean amplifies power at the inertial frequency in the high latitudes, thus motivating our use of a semi-implicit treatment of the Coriolis force, which helps to maintain numerical stability.

The barotropic mode is split from the baroclinic via an explicit free surface algorithm where the top model grid cell has a time dependent volume. As discussed by Griffies et al. (2001), this approach provides for improved conservation of tracers relative to earlier approaches where the top grid cell volume is fixed in time. Even so, the input of fresh water to the ocean model is through a virtual salt flux, as traditionally employed with rigid lid ocean models. The use of real fresh water fluxes are trivially available with the MOM4 free surface, yet numerical problems associated with the sea ice model, subsequently resolved, led to the salt flux for the present configuration.

Key physical parameterizations include the KPP mixed layer scheme of Large et al. (1994), which computes a vertical diffusivity and vertical viscosity as a function of the flow and surface forcing. A background vertical diffusivity of 0.05 cm²/sec in the upper ocean transitions to 1.0 cm²/sec background in the abyss, with a vertical structure suggested by Bryan and Lewis (1979). Neutral physics consists of Redi (1982) neutral diffusion and Gent-McWilliams (1990) skew-diffusion, and they are implemented according to Griffies et al. (1998) and Griffies (1998). The Redi-GM diffusivities were both set to 0.8 x 10³ m²/sec. The hyperbolic tangent tapering scheme of Danabasoglu and McWilliams (1995) reduces the magnitude of the off-diagonal neutral flux components when the neutral slope steepen greater than 1/100. Horizontal friction within 20 degrees of the equator consists of the anisotropic scheme of Large et al. (2001), which greatly strengthens the equatorial current system. In higher latitudes, the scheme reduces to a traditional isotropic friction with a grid-space dependent background viscosity as well as the Smagorinsky viscosity, implemented according to the ideas in Griffies and Hallberg (2000).

Mixing of tracers between the main ocean basins and artificially enclosed seas, such as the Mediterranean and Red Seas, is implemented by a scheme that allows for the diffusive influence of isolated basins on the large-scale circulation. Such is necessary for models where the grid resolution is too coarse to allow for an explicit connection between, say, the Mediterranean and Atlantic. Tracer diffusion is also prescribed between bottom boxes according to the ideas of Beckmann and Doescher (1997) and Doescher and Doescher (1999). This scheme helps to move dense water from shelves into the abyss, although the signal in the present model remains smaller than Nature. Refined horizontal resolution is needed to remedy this problem, as discussed by Winton et al. (1998).

Ocean spin-up

During the first 80 years of model integration, the ocean tracer time step was set to 3 hours, baroclinic momentum time step to 1 hour, and barotropic time step to 72 seconds. After 80 years of tracer time, the tracer time step was reduced to one hour and the baroclinic and barotropic time steps were unchanged. The model was then run for another 20 years, with the analysis concentrated over years 90-100. A similar spin-up procedure was discussed by Danabasoglu et al. (1996), although they noted the need to run for order thousands of years to fully spin-up the deep ocean. Consistent with the P-OMIP protocol, we ran only for 100 years.

During the spin-up with 3 hour tracer time step, the model took roughly 2 hours per model year on 36 processors using an SGI Origin 3800.

Flux forcing

Turbulent latent and sensible heat flux and evaporation were computed using GFDL's Flexible Modeling System (FMS) flux coupler and its Monin-Obhukov similarity theory based boundary layer module. Net longwave radiation was computed using climatological downward longwave radiation and the sea or ice surface temperature computed by the model. Net shortwave radiation was computed using the climatological downward shortwave with albedo over seaice computed from the ice model, and penetration into the ocean prescribed according to the three-exponential approach of Morel and Antoine (1994). Optical properties of seawater are set according to a monthly climatology of SeaWIFS chlorophyll-a. Fluxes of momentum were taken directly from the climatology, rather than being recomputed using the FMS boundary layer physics package using climatological wind speed. Finally, our ice model required that precipitation be divided into a frozen and liquid contribution. This division was demarcated by a surface air temperature of 273.16 K. No attempt was made to impose closure on the fluxes of heat or salt.

References

Adcroft, A., C. Hill, J. Marshall, 1997: Representation of topography by shaved cells in a height coordinate ocean model, Monthly Weather Review, volume 125, 2293--2315.

Beckmann, A., R. Doescher, 1997: A method for improved representation of dense water spreading over topography in geopotential--coordinate models, Journal of Physical Oceanography, volume 27, 581--591.

Bryan, K, L.J. Lewis, 1979: A water mass model of the world ocean. Journal of Geophysical Research, volume 84, 2503--2517.

Danabasoglu, G., J.C. McWilliams, 1995: Sensitivity of the global ocean circulation to parameterizations of mesoscale tracer transports. Journal of Climate, volume 8, 2967--2987.

Danabasoglu, G., J.C. McWilliams, W.G. Large, 1996: Approach to equilibrium in accelerated global oceanic models. Journal of Climate, volume 9, 1092--1110.

Doescher, R., A. Beckmann, 1998: Effects of a bottom boundary layer parameterization in a coarse-resolution model of the North Atlantic Ocean. Journal of Atmospheric and Oceanic Technology, volume 17, 698--707.

Gent, P.R., J.C. McWilliams, 1990: Isopycnal mixing in ocean circulation models. Journal of Physical Oceanography, volume 20, 150--155.

Griffies, S.M., 1998: The Gent-McWilliams skew-flux. Journal of Physical Oceanography, volume 28, 831-841.

Griffies, S.M., A. Gnanadesikan, R.C. Pacanowski, V. Larichev, J.K. Dukowicz, R.D. Smith, 1998: Isoneutral diffusion in a z-coordinate ocean model. Journal of Physical Oceanography, volume 28, 805--830.

Griffies, S.M., R.W. Hallberg, 2000: Biharmonic friction with a Smagorinsky-like viscosity for use in large-scale eddy-permitting ocean models. Monthly Weather Review, volume 128, 2935--2946.

Griffies, S.M., R.C. Pacanowski, R.M. Schmidt, and V. Balaji, 2001: Tracer Conservation with an Explicit Free Surface Method for Z-coordinate Ocean Models, Monthly Weather Review, volume 129, 1081--1098.

Griffies, S.M., M.J. Harrison, R.C. Pacanowski, and A. Rosati, 2003: A Technical Guide to MOM4, NOAA/Geophysical Fluid Dynamics Laboratory Ocean Group Technical Report. Available on-line from http://www.gfdl.noaa.gov/~fms/

Hunke, E.C. and J.K. Dukowicz, 1997: An elastic-viscous-plastic model for sea ice dynamics. Journal of Physical Oceanography, volume 27, 1849--1867.

Large, W.G., J.C. McWilliams, S.C. Doney, 1994: Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization. Reviews of Geophysics, volume 32, 363-403.

Large, W.G., G. Danabasoglu, J.C. McWilliams, P.R. Gent, F.O. Bryan, 2001: Equatorial circulation of a global ocean climate model with anisotropic horizontal viscosity, Journal of Physical Oceanography, volume 31, 518--536.

Morel, A. and D. Antoine, 1994: Heating Rate within the Upper Ocean in Relation to Its Biooptical State, Journal of Physical Oceanography, vol 24, 1652-1665

Murray, R. J., 1996: Explicit generation of orthogonal grids for ocean models. Journal of Computational Physics, volume 126, 251-273.

Pacanowski, R.C., A. Gnanadesikan, 1998: Transient response in a z-level ocean model that resolves topography with partial-cells, Monthly Weather Review, volume 126, 3248-3270.

Redi, M.H., 1982: Oceanic isopycnal mixing by coordinate rotation. Journal of Physical Oceanography, volume 12, 1154--1158.

Semtner, A.J., 1976: A model for the thermodynamics growth of sea ice in numerical investigations of climate. Journal of Physical Oceanography, volume 6, 379--389.

Winton, M., R.W. Hallberg, A. Gnanadesikan, 1998: Simulation of density-driven frictional downslope flow in z-coordinate ocean models. Journal of Physical Oceanography, volume 28, 2163--2174.

Winton, M., 2000: A reformulated three-layer sea ice model. Journal of Atmospheric and Oceanic Technology, volume 17, 525-531. Also documented at http://www.gfdl.noaa.gov/~mw/