S. Borchert, U. Achatz, S. Remmler, S. Hickel, U. Harlander, M. Vincze, K.D. Alexandrov, F. Rieper, T. Heppelmann, S.I. Dolaptchiev (2014)
Meteorologische Zeitschrift 23: 561-580.  doi: 10.1127/metz/2014/0548  

The differentially heated rotating annulus is a classical experiment for the investigation of baroclinic flows and can be regarded as a strongly simplified laboratory model of the atmosphere in mid-latitudes. Data of this experiment, measured at the BTU Cottbus-Senftenberg, are used to validate two numerical finite-volume models (INCA and cylFloit) which differ basically in their grid structure.

Both models employ an implicit parameterization of the subgrid-scale turbulence by the Adaptive Local Deconvolution Method (ALDM). One part of the laboratory procedure, which is commonly neglected in simulations, is the annulus spin-up. During this phase the annulus is accelerated from a state of rest to a desired angular velocity. We use a simple modelling approach of the spin-up to investigate whether it increases the agreement between experiment and simulation. The model validation compares the azimuthal mode numbers of the baroclinic waves and does a principal component analysis of time series of the temperature field. The Eady model of baroclinic instability provides a guideline for the qualitative understanding of the observations.