Model CSIRO: Elaborations

Participation

Spinup/Initialization

• The atmosphere model was coupled to the dynamic sea ice model and integrated for many decades (without restoring forces toward climatology or surface stress adjustments) to ensure equilibrium of the ice.  The surface fluxes from the final 10 years of the integration were averaged to obtain a monthly climatology for use in calculating the flux adjustments for the coupled run.

• The ocean model started from rest and with temperature and salinity at each level specified from global average values after Levitus (1982).  It was forced at the surface with the climatological wind stress data of Hellerman and Rosenstein (1983).  In the first phase, the ocean model was integrated for 1000 years with annual-mean forcing using asynchronous time stepping (2 days for T,S and 20 minutes for U,V); then it was integrated for an additional 100 years with annual mean forcing a reduced asynchronicity (1 day for T,S and 20 minutes for U,V); then it was integrated for an additional 1070 years with monthly mean forcing and further-reduced asynchronicity (1 day for T,S and 40 minutes for U,V); finally it was integrated for 710 years with monthly mean forcing using synchronous time stepping (1 hour for T,S,U,V).

• Flux adjustments for heat, salinity, and surface stresses were calculated following the method of Sausen et al. (1988) and the ocean model's SST and SSS were further corrected at each time step before being passed to the atmospheric model in order to reduce the drift in coupled mode (cf. Gordon and O'Farrell 1997 for details).

• The atmosphere/sea ice and ocean models then were coupled and run for 105 years with these flux adjustments and corrections.  The resulting climate drift was modest (~ -0.4 deg  K in 2m surface air temperature and ~ -0.3 deg K in SST).

Land Surface Processes

• The scheme includes a single-layer vegetation canopy that acts as a source or sink of water vapor and sensible heat. Monthly-varying values of local fractional vegetation type, unconstrained stomatal resistance, roughness length, and albedo are specified from data of Dorman and Sellers (1989). The canopy shields the ground from solar radiation and intercepts precipitation and dew up to a maximum depth that depends on the leaf area index (LAI); the intercepted moisture subsequently evaporates at the potential rate. Following the approach of Noilhan and Planton (1989), evapotranspiration from the dry portion of the canopy is regulated by a variable stomatal resistance that depends on the unconstrained value modified by factors associated with incident radiation, air temperature, availability of water in the root zone, and atmospheric vapor pressure deficit. Cf. Kowalczyk et al. (1991,1994) for further details.

• Precipitation falls as rain and, if the surface air temperature is below 0 deg C, freezes at the surface with release of its latent heat. The presence of snow modifies the roughness length and the radiative/thermal properties of the surface. The latter depend on the fractional snow depth: the surface albedo is assigned a maximum value of 0.8 when a snow scaling depth of 0.14 m is exceeded and the surface temperature is calculated using a combined snow/soil layer. Changes in the total snow mass are calculated from a budget equation.

• Soil temperature is calculated following McGregor et al. (1993). Three layers are used, their thicknesses (0.03, 0.23, and 2.5 m) chosen to enable a reasonable representation of both diurnal and seasonal temperature waves in the soil. The temperature of the i-th layer is incremented by the diffusive heat flux into and out of the layer. Soil moisture is calculated using two layers, as described by Kowalczyk et al. (1994). The chief source of soil moisture is surface infiltration which depends on precipitation, snowmelt, surface runoff, and soil hydrological properties; moisture sinks include evaporation from bare soil, root extraction associated with canopy transpiration, and gravitational drainage when the local field capacity is exceeded. Surface runoff and gravitational drainage do not contribute to the flux of freshwater to the ocean model, however.

Sea Ice

• In a departure from common practice, the representation of sea ice (O'Farrell 1998) is associated with the atmospheric component of the coupled model: physics (radiation, clouds, moisture and heat budget, etc.) are calculated separately for ice and open-water fractions of each grid box, and then are combined as an area-weighted average before computing the dynamics (advection, mixing, etc.). The ice concentration and the percentage of open water in each grid box are determined from the energy budget of the ocean mixed layer adjacent to the ice: once this water reaches the saltwater freezing point (-1.85 deg C), new ice forms at increments of 4 percent of the grid box. If ice advecting outside the existing area encounters a mixed layer temperature close to the freezing point, the advected ice is allowed to remain.

• Both vertical and lateral accretion of ice are treated.   If the existing ice floe is less than 0.25 m thick, new ice is added to its base; otherwise, ice is added to its side. Additional ice also is added to the base of the floe when the ice concentration is greater than 98%  in the Antarctic or 99.5% in the Arctic.   In addition, "white ice " (cf. Ledley  1985) forms from the overlying snow pack when it is massive enough to depress the floe below  the ocean surface.  Maximum ice thickness is constrained to 5 m, but only when ice  is trapped in embayments where grid resolution inhibits the calculation of ice velocities.

• Vertical and lateral ice ablation also are represented.  Surface melting is calculated after the thermodynamic model of Semtner (1976)  which divides the ice system into three layers--two for ice and one for overlying snow, with constant albedos specified for the two media; simulated brine pockets also provide an internal heat source.  No lateral melting occurs for a system temperature between -1.85 and -1.0 deg C.  For a temperature between -1.0 and 0.5 deg C, any additional heat is partitioned between the ocean mixed layer and lateral ice melt.  All additional heat goes to lateral ice melt when the temperature is greater than 0.5 deg C.  Basal melt also results from ocean-to-ice heat transfer, which  is proportional to the difference between the surface temperature of the ocean and the ice bottom at the saltwater freezing temperature.  (The diffusion coefficient K = 1.5*10^-6 m² s^-1 permits the ocean temperature to rise 1 to 2 degrees above the saltwater freezing point without excessive melting of ice.) In the Antarctic an additional fixed ocean-to-ice flux of 10 Wm^-2 is applied to represent the higher transfer rates that occur during convection and the passage of subgrid-scale warm eddies.

• The model also incorporates the cavitating fluid rheology of Flato and Hibler (1990),  wherein the sea ice has a finite resistance to compression, but can diverge freely.   Convergence of ice, with thickening by the formation of ridges, occurs when the internal ice pressure is exceeded.  Divergence increases the area of open water in leads (i.e. reduces the ice compactness) rather than affecting the ice thickness.  The ice also is assumed to offer no resistance to shear stress; its motion is determined by a dynamical balance between the internal ice presssure, the atmospheric and oceanic surface stresses, and the Coriolis force. Ice advection is treated numerically with an upstream differencing scheme.

Ocean Component

• The ocean component of the coupled model is based on the Bryan-Cox code (cf. Cox 1984), and is described in Hirst et al. (2000). Temperature/salinity grid points are coincident with the physics grid of the atmospheric model, hence the model has a horizontal resolution of 3.2° latitude by 5.2° longitude. The model features three "islands" (Antarctica, Australia-New Guinea, New Zealand), as well as the major Eurasian-African-American land mass. There is no Bering Strait throughflow. Nonzero net circulation is permitted around each island. The Mediterranean Sea and Hudson Bay are included, and the effect of water exchange between these features and the North Atlantic is parameterized via a mixing of tracers across the (unresolved) Strait of Gibraltar and Hudson Strait.

• The ocean component has 21 levels in the vertical, with level spacing grading from 25 m near the surface to 450 m in the deep ocean.

• Tracers (temperature, salinity) are diffused strongly along the neutral density surfaces, and weakly in the vertical. The isoneutral diffusion is performed by the Cox (1987) implementation of the Redi (1982) scheme using a uniform isoneutral tracer diffusivity of 1* 10³ m² s^-1. The vertical diffusivity is specified in terms of local static stability according to the formula of Gargett (1984), however with a minimum allowed default value of 3 * 10^-5 m² s^-1.  Larger default minima of 2 * 10^-3 m² s^-1 and 1 * 10^-4 m² s^-1 for vertical diffusivity are set between the first and second, and second and third model levels, respectively, to roughly simulate some wind forced mixing effect. Convection is simulated by applying an enhanced vertical diffusivity, of 10³ m² s^-1, in regions of static instability.

• The model includes the advective form of the Gent and McWilliams (1990) scheme for eddy-induced transport of tracers (cf. Gent et al. 1995). The isoneutral thickness diffusivity is set at 1 * 10³ m² s^-1 below 270 m depth, above which this diffusivity is tapered linearly to zero at the surface. Horizontal tracer diffusivity is set to zero, as is common in models employing the Gent and McWilliams (1990) scheme.

• The horizontal viscosity is set at 9 * 10⁵ m² s^-1 and the vertical viscosity is set at 2 * 10^-3 m² s^-1, which are standard choices for the present resolution (e.g., cf. Bryan et al. 1975).

Details of the CMIP Integrations

• The 100 years of output featured in CMIP 1 is for model year 111 through model year 210 of the 1,000 year simulation described in Hirst et al. (2000). The model version employed is the same as that described in Gordon and O'Farrell (1997), except that the Gent and McWilliams (1990) scheme has been used in the ocean component and the number of oceanic levels has been increased from 12 to 21. Hirst et al. (2000) give a detailed comparison of this 1,000 year simulation (which they refer to as "GM") with the earlier control simulation presented in Gordon and O'Farrell (1997).

• The CMIP 2 integration is performed using the same model version as for CMIP 1, and is initialized from the model fields at the end of year 130 of the CMIP 1 simulation. Thus the 80 years of the transient CMIP 2 integration corresponds in time to the final 80 years of the period included in CMIP 1 from the control integration. The response of the model to the 1% per annum compounding CO2 concentration in the CMIP 2 integration is similar to that shown in Hirst et al. (1996) and Hirst (1999) for the same model version forced according to the IPCC IS92a scenario. See also Gordon and O'Farrell (1997) for a detailed discussion of an earlier 1%-compounding response experiment performed with the CSIRO model (prior to inclusion of the Gent and McWilliams 1990 scheme).

Chief Differences from the Closest AMIP Model

Bryan, K., S. Manabe, and R.C. Pacanowski, 1975: A global ocean-atmosphere climate model. Part II: The ocean circulation. J. Phys. Oceanogr., 5, 30-46.

Cox, M.D., 1984: A primitive equation, 3-dimensional model of the ocean. GFDL Ocean Group Tech. Rep. 1, GFDL/Princeton University, Princeton , N.J., 141 pp. [Available from Geophysics Fluid Dynamics Laboratory/NOAA, Princeton University, Princeton, NJ 08542.]

Cox, M.D., 1987: Isopycnal diffusion in a z-coordinate ocean model. Ocean Modelling, 74, 1-15.

Dorman, J.L., and P.J. Sellers, 1989: A global climatology of albedo, roughness length and stomatal resistance for atmospheric general circulation models as represented by the Simple Biosphere Model (SiB). J. Appl. Meteor., 28, 833-855.

Flato, G.M., and W.D. Hibler 1990: On a simple sea-ice dynamics model for climate studies. Ann. Glaciol., 14, 72-77.

Gargett, A.E., 1984: Vertical eddy diffusivity in the ocean interior. J. Mar. Res., 42, 359-393.

Gent, P.R., and J.C. McWilliams, 1990: Isopycnal mixing in ocean circulation models. J. Phys. Oceanogr., 20, 150-155.

Gent, P.R., J. Willebrand, T.J. McDougall and J.C. McWilliams, 1995: Parameterising eddy-induced tracer transports in ocean circulation models. J. Phys. Oceanogr., 25, 463-474.

Gordon, H.B., and S.P. O'Farrell, 1997:  Transient climate change in the CSIRO coupled model with dynamic sea ice, Monthly Weather Review, 125, 875-907.

Hellerman, S., and M. Rosenstein, 1983: Normal monthly wind stress over the world ocean with error estimates. J. Phys. Oceanogr., 13, 1093-1104.

Hirst A.C., H. B. Gordon, and S.P. O'Farrell, 1996: Global warming in a coupled climate model including oceanic eddy-induced advection. Geophys. Res. Lett., 23, 3361-3364.

Hirst A.C., 1999: The Southern Ocean response to global warming in the CSIRO coupled ocean-atmosphere model. Environ. Modeling and Software: Special Issue on Modelling Global Climatic Change, 14, 227-242.

Hirst, A.C., S.P. O'Farrell and H.B. Gordon, 2000: Comparison of a coupled ocean-atmosphere model with and without oceanic eddy-induced advection. Part I: Ocean Spinup and control integrations. J. Climate, 13, 139-163.

Kowalczyk, E.A., J.R. Garratt, and P.B. Krummel, 1991: A soil-canopy scheme for use in a numerical model of the atmosphere--1 D stand-alone model. CSIRO Division of Atmospheric Research Technical Paper No. 23, 56 pp.

Kowalczyk, E.A., J.R. Garratt, and P.B. Krummel, 1994: Implementation of a soil-canopy scheme into the CSIRO GCM-Regional aspects of the model response. CSIRO Division of Atmospheric Research Technical Paper No. 32, 59 pp.

Ledley, T.S., 1985: Sea ice: Multiyear cycles and white ice.  J. Geophys. Res., 90, 5,676-5686.

Levitus, S., 1982: Climatological atlas of the world's oceans. NOAA Professional Paper 13, 173 pp.

McGregor, J.L., H.B. Gordon, I.G. Watterson, M.R. Dix, and L.D. Rotstayn, 1993: The CSIRO 9-level atmospheric general circulation model. CSIRO Division of Atmosheric Research Technical Paper No. 26, 89 pp.

Noilhan, J., and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117, 536-549.

O'Farrell, S.P., 1998: Investigation of the dynamic sea ice component of a coupled atmosphere sea-ice general circulation model. J. Geophys. Res. (Oceans), 103, 15751-15782.

Redi, M.H., 1982: Oceanic isopycnal mixing by coordinate rotation. J. Phys. Oceanogr., 12, 1154-1158.

Sausen, R., K. Barthel, and K. Hasselman, 1988: Coupled ocean-atmosphere models with flux correction. Climate Dyn., 2, 145-163.

Semtner, A.J., 1976: A model for the thermodynamic growth of sea ice in numerical investigations of climate.  J. Phys. Oceanogr., 6, 379-389.
 

Top

CMIP Documentation Directory

Last update 15 May, 2002.  This page is maintained by Tom Phillips (phillips@pcmdi.llnl.gov).
 

LLNL Disclaimers

UCRL-MI-126431