How do General Circulation models work ?

The Mars Climate database has been constructed based on output from a set of general circulation models (GCM) developed jointly at LMD and AOPP. (Forget et al., 1999) GCMs are a widely used tool in terrestrial weather forecasting, climate forecasting and meteorological research. The basic idea is the following: from an initial state (defined by temperature, pressure, wind fields, etc.), the model uses well known physical laws (e.g. fluid mechanic equations, radiative transfer laws) to compute the evolution of the system timesteps after timesteps.

The General Circulation Models can be schematically divided into two parts.

• The first part is an hydrodynamical code dedicated to the time and spatial integration of the equations of hydrodynamics to compute the large scale atmospheric motions. This part, based on a rather general formulation of the equations of fluid mechanics (primitive equations in meteorological dialect) can be used without any change for the other terrestrial planets. Two different kind of hydrodynamical code have been developed so far: 1. finite-differences or grid-points models (on use at LMD) and 2. spectral solver based on a decomposition of the horizontal fields on spherical harmonics (on use at AOPP).

• The second part of the code consists of a set of physical parametrizations which are used to force the general circulation and compute the details of the local climate at every grid point. LMD and AOPP use a common set of parameterizations which includes :
  • Radiative transfer.

    This is the main source of atmospheric motions (principally the latitudinal variations of the absorption of solar energy). The effects of gaseous carbon dioxide (absorption and emission of radiation) and suspended dust (absorption, emission and scaterring of radiation) are included in the model at solar and thermal infrared wavelengths. To construct the current version of the database and extend the top of the model up to 120 km, a major effort was put in the calculation of the radiative transfer in the upper atmosphere in collaboration with the experts from IAA in Granada. At low pressure above 70 km, one have to consider the fact that the assumption of Local Thermodynamic Equilibrium (LTE) is not valid and that non-LTE effects must be taken into account when computing the radiative heating and cooling budget for the upper CO$_2$ atmosphere.

  • Surface Processes.

    The temperature of the surface is computed from the radiative, sensible and latent heat fluxes at the surface using an 11-level model of thermal diffusion in the soil. Surface properties, i.e. albedo and thermal inertia, are based on Mars Global Surveyor (MGS) and Viking observations. The model also use the accurate topography map obtained by the MOLA altimeter aboard MGS.

  • Sub-grid scale dynamics.

    The physical parametrizations must also include representation of small scale motions not explicitly represented by the hydrodynamical code. On the one hand, this includes the representation of vertical exchange of momentum and energy between the surface and the atmosphere due to turbulence and convection, mostly in the so-called Planetary Boundary Layer. On the other hand, the small mountains which are not represented in the GCM smooth grid can influence the model scale flow by (a) producing a form drag on the flow at low levels, and by (b) exciting internal gravity waves which can propagate in the vertical, break, and decelerate the flow far away from the mountains themselves.

  • CO$_2$ condensation.

    a specific parametrization has been developed for Mars to account for the condensation of the CO$_2$ atmosphere in the polar region during the polar night seasons. CO$_2$ can condense directly on the surface or up in the atmosphere. In that last case, the CO$_2$ ice particles formed in the atmosphere can affect the infrared emissivity of the system as suggested by the oabservations.