Specification for GCM cold-air outbreak Case

Below we describe the setup for the global model intercomparison, and suggest different sensitivity experiments. We are trying to be as inclusive as possible and will accept contributions also in case they deviate in some aspects from the provided requirements.

Time period

Start date of the forecast: 12Z 30th January 2010. The simulation should be run for 36 hours (i.e. until 00Z 1st February 2010).


We encourage the groups to run their global models at the highest possible resolutions. In order to be able to investigate the scale-dependency of the parameterizations, the models should be run at at least two different resolutions. Global resolutions of 10 km and 5 km are recommended, but higher resolutions are welcome. In case the global model allows for grid refinement (one-way nesting or two-way nesting) or grid stretching, we encourage the groups to also perform 1km and 2km resolution simulations for the region of interest. In case you use grid refinement by one-way nesting and submit the output of the refined grid to the LAM intercomparison, you do not need to submit the same data to the global model intercomparison.

Forcing Data

To be consistent with the LAM intercomparison it is recommended that the GCM is initialised using ECMWF analysis.

Parametrization switches

For each of the global model grid resolutions we would like to receive results from the two permutations of the convection off and on. There might be other parameterizations that should be switched on or off. For instance your global model might have a separate shallow and deep cumulus convection scheme. Please feel free to experiment with different combinations of these parameterisations switched on and/or off. In case you have different convection or boundary layer parameterizations implemented in your model, it is very welcome if you produce separate experiments with different parameterizations. Could participants please provide details regarding the used parameterisations, and references.


In case of questions please contact Lorenzo Tomassini (lorenzo.tomassini@mpimet.mpg.de) or Axel Seifert (axel.seifert@dwd.de).