Model Properties: Dynamics and Numerics, Physics, Data Assimilation, and more ...

The COSMO Model is a limited-area model developed within the framework of the Consortium for Small-Scale Modelling (COSMO).

 

A general description of the COSMO model is available from the COSMO website. More detailed information on the COSMO Model can be found in the COSMO Newsletters or from the scientific documentation (see box "further reading").

MeteoSwiss uses the COSMO Model in two resolutions: COSMO-7 with a grid spacing of 6.6 km, and COSMO-2 with a grid spacing of 2.2 km.

COSMO-7 is integrated three times a day out to 72 hours (at 00, 06 and 12 UTC), the higher resolution COSMO-2 is updated eight times a day out to 24 hours. Both models run on a Cray XT4 parallel computer at the Swiss National Supercomputing Centre (CSCS) in Manno (TI). During the operational forecasting slots, the Cray XT4 works in dedicated mode. About 50 minutes are required for a full COSMO-7 and COSMO-2 forecast cycle.

For more details on the operational COSMO Models, refer to the operational applications page of the COSMO website, which also lists the current namelist settings used for the COSMO Models.

The COSMO Model is based on the primitive hydro-thermodynamical equations describing compressible nonhydrostatic flow in a moist atmosphere without any scale approximations. The model equations are solved numerically on a rotated latitude-longitude grid, with terrain-following coordinates in the vertical, using an Eulerian finite difference method.

 

The dynamical and numerical key features of the COSMO Model are summarized on the dynamics and numerics page.

Physical processes not be resolved by the 3-dimensional numerical grid need to be parameterized. In numerical weather prediction, "parameterization" is the technical term for the art of estimating the properties of sub-grid scale processes (e.g. condensation: is sub-grid scale condensation taking place?) and their feedback on the resolved model variables (e.g. temperature and humidity: does condensation heat the air in the grid box under consideration and hence alter it's temperature?) . The COSMO Model uses parameterization schemes for the following sub-grid scale physical processes: vertical diffusion (turbulence), cloud and precipitation formation (condensation), convection (not needed for resolutions better than few kilometers), radiation, and finally soil processes in a soil model.

The COSMO physical parameterizations page provides more details about the parameterization schemes of the COSMO Model.

To provide initial conditions for the COSMO Model forecasts, a four-dimensional data assimilation system based on the observation nudging technique is used. This nudging or Newtonian relaxation relaxes the model's prognostic variables (i.e., pressure, temperature, wind, and mass fractions of different water species) towards observations within a predetermined space- and time-window. The output of this procedure is called analysis and - besides providing initial conditions for COSMO Model forecasts - is the best estimate of the state of the atmosphere (and a 3-dimensional hydro-thermo-dynamically consistent snapshot) for any given time.

More information is available on the COSMO data assimilation page.

Besides initial conditions, the limited-area COSMO Model forecasts also need lateral boundary conditions for the entire forecasting period. For COSMO-7, these are provided by the coarser global model (IFS) of the ECMWF, whose fields drive COSMO-7 using a 1-way nesting technique. For COSMO-2, the 1-way nesting is used to run it with lateral boundary conditions from COSMO-7.