GFDL - Geophysical Fluid Dynamics Laboratory

Topic: Grid resolution & model features

Info associated with a Journal of Climate paper (Delworth et al., 2011, in press) and a poster presented at the October 2011 WCRP Open Science Conference (Dixon et al, 2011).

[thumbnail of Grid Resolution table]






(beta level – Oct 2011)

50 km cubed-sphere grid (C180)

32 levels

square grid: ~11km at Equator to 5km at 65°N, etc.

tri-polar north of 65°N

50 z* levels

LM3 – for land water, energy, and carbon balance


(Delworth et al. 2011, J Climate, in press)

50 km cubed-sphere grid (C180)

32 levels

square grid: ~28km at Equator to 12km at 65°N, etc.

tri-polar north of 65°N

50 z* levels

LM3 – for land water, energy, and carbon balance


(Delworth et al. 2006, J Climate)

2° longitude x 2.5° latitude

24 levels

1° longitude x 1° latitude, reducing to 0.33° latitude in equatorial region

tripolar north of 65°N

50 levels

LM2 – a previous generation land surface model

Based on MOM 4.1 (Griffies., 2010), the GFDL CM2.5 model?s ocean resolution is nominally one-quarter of a degree. The CM2.6 model?s ocean has horizontal resolution that is nominally one-tenth of a degree (see table above). While the global CM2.5 ocean model can be considered ?eddy-permitting?, the CM2.6 model?s ocean is ?eddy-resolving?. Both global climate models employ an atmospheric model with cubed sphere geometry having approximately 50km horizontal resolution (C180) and 32 vertical levels. All three models use a tri-­?polar grid (Murray, 1996), in which there are displaced poles located over Northern Canada and Russia at 65°N  to avoid a singularity at the North Pole.


Horizontal grid for the ocean component of GFDL CM2.1

Zooming in on the Atlantic to the east of the northeastern United States, we see the GFDL CM2.1 model’s ocean grid has grid spacing of 1 degree latitude and longitude in this region (~111 km north-south and 84 km east-west at 41°N, the latitude of New York City). The size of individual ocean grid cells is indicated by the rectangles that make up the checkerboard blue pattern. For reference, the observed coastline is drawn in white.

? [larger CM2.1 grid image]

Horizontal grid for the ocean component of GFDL CM2.5

In the newer, eddy-permitting GFDL CM2.5 model, we see ocean grid cell sizes that are ~21 kilometers (13 miles) on a side at the latitude of New York City.  Though sometimes referred to as having a nominal resolution of 0.25°, the use of a square grid in CM2.5 maintains a square aspect ratio across most of the globe, resulting in grid spacing varying by the cosine of latitude. (The Arctic Ocean north of 65°N is an exception.)

? [larger CM2.5 grid image]

Horizontal grid for the ocean component of GFDL CM2.6

The eddy-resolving GFDL CM2.6 model has ocean grid spacing of 8.4 km (5.2 miles) at the latitude of New York City, allowing the model to capture much of the detail of the northeastern US coastline. With additional grid resolution comes additional computational expense – thus, for certain applications, the CM2.5 model will serve as our workhorse model starting in 2011 and CM2.6 will be run for a more limited number of experiments.

? [larger CM2.6 grid image]

CM2.6 and CM2.5: similar configurations

As part of this model development effort,  a conscious decision has been made to build the CM2.5 and CM2.6 models in a similar fashion. In addition to being coupled to the same atmospheric, sea ice, and land surface model componenets, the two modes’ ocean components are configured to allow them to maintain sharp gradients and to simulate very energetic flows, including intense boundary currents. We also made configuration choices that minimize parameterization differences between the two models, thus allowing cleaner assessments of the role of ocean horizontal resolution on the model simulations.

CM2.6 and CM2.5 share the following configuration characteristics:

  • No parameterization for the effect of meso-scale eddies.?
  • No explicit lateral tracer diffusion.
  • No prescribed background vertical diffusion.
  • Vertical mixing is determined by K-profile parameterization (KPP) scheme (Large et al., 1994).
  • Schemes for internal tide mixing (Simmons et al., 2006) & coastal tide mixing (Lee et al., 2006).
  • Use of fifth-order conservative, monotonic advection scheme (White and Adcroft, 2008).
  • Very low, scale-selective viscosity following the Smagorinsky biharmonic formulation (Griffies and Hallberg, 2000).
  • All straits connecting bodies of water (such as the Atlantic and Mediterranean Sea) have explicit flow, rather than a parameterized exchange as in CM2.1
  • The ocean and atmosphere model components exchange updated surface fluxes once an hour.

? Though not optimal for an eddy-permitting model such as CM2.5, omitting a meso-scale eddy parameterization such as Gent-McWilliam (Gent and McWilliams, 1990) facilitates comparisons of CM2.5 with CM2.6?s eddy-resolving ocean simulation.


  • Delworth, T.L., et al., 2006: GFDL’s CM2 Global Coupled Climate Models. Part I: Formulation and Simulation Characteristics. Journal of Climate, 19, 634-674.
  • Delworth, T.L., et al., 2011: Simulated climate and climate change in the GFDL CM2.5 high resolution coupled climate model. Journal of Climate, (in press).
  • Dixon, K.W, et al., 2011, Ocean circulation features of the GFDL CM2.6 & CM2.5 high-resolution global coupled climate models,a poster presented at the WCRP Open Sciences Conferece, Denver Colorado. [available online at]
  • Gent, P.R., and J.V. McWilliams, 1990: Isopycnal Mixing in Ocean Circulation Models. J. Phys. Oceanogr., 20, 150-155.
  • Griffies, S.M., 2010, Elements of MOM4p1, GFDL Ocean Group Technical Report No. 6., NOAA/Geophysical Fluid Dynamics Laboratory, 444 pages. [ LINK ]
  • Griffies, S.M., and R.W. Hallberg, 2000: Biharmonic Friction with
    a Smagorinsky-Like Viscosity for Use in Large-Scale Eddy-Permitting
    Ocean Models. Mon. Wea. Rev., 128, 2935?2946.
  • Large, W.G., et al., 1994: Ocean vertical mixing: A review and a model with a nonlocal boundary layer parameterization. Rev. of Geophysics, 32, 363-403.
  • Lee, H.C., et al., 2006: Barotropic tidal mixing effects in a coupled climate model: Oceanic conditions in the N. Atlantic. Ocean Modelling, 11, 467-477.
  • Murray, R.J., 1996: Explicit generation of orthogonal grids
    for ocean models, J. Comput. Phys., 126, 251-273.
  • Simmons, H., et al., 2006: Tidally driven mixing in a numerical model of the ocean general circulation. Ocean Modelling, 6, 245-263.
  • White, L, and A. Adcroft, 2008: A high-order finite volume remapping scheme for nonuniform grids: The piecewise quartic method (PQM). J. Comp.Phys, 227, 7394-742.