Typical results of a version of the KNMI SCM model are illustrated below. Although, of course, results differ from model to model, the results and dependencies presented here are rather typical in the sense that the contain: numerical noise, strong resolution dependencies and (apparent, though we are not certain of that) unphysical feedbacks. The model physics (for which the results are presented here) are based on the ECHAM4 physics (same turbulence scheme, massflux closure, etc.) but differ to the ECHAM4 physics with respect to the cloud/condensation scheme and the entrainment/detrainment coefficients; for latter changes see Siebesma and Cuijpers (1995, J.A.S., p 650-666).

We ran at three different resolutions: 40 meter equidistant (run EU_n96), and the two prescribed resolutions (EU_n40 and EU_n19). Below are results for the time evolution of theta, qt, and cloud cover. Results on the highest vertical resolution are rather good. They show a typical Cu cloud cover (~ 10 %) around midday, a rather well mixed layer below 1000 m, and a conditionally unstable layer between 1000 and 2000 meter (approx.). On the down side, all profiles contain some numerical noise. There are also signs of a too strong inversion at cloud base (look also at qt), which indicates that there the boundary layer turbulence (read K diffusion) does not match well with the turbulence in the Cumulus cloud (read Massflux scheme).

The resolution dependency can be explained roughly as follows. First look at the massflux profile at 19 UTC. It shows that on both high resolutions the massflux is almost identical. However, on the 19-level resolution the massflux start to detrain (at the cloud top) at a much lower level (this may not be too supprising because - due to the coarse resolution - the massflux either detrains above1600 m or above 1300 m (as there is no level inbetween), and the high resolution results already indicate that the transition between positive and negative buoyant just occurs at 1600 m). Now the massflux causes a moistening just below the inversion. Consequently, too high cloud fractions occur. This again introduced additional turbulent mixing (as we use a moist turbulence scheme coupled to the cloud fraction). In a worst case scenario this might even lead to the following positive feedback: enhanched turbulent mixing in the cloud => smaller convergence of moisture in the subcloud => smaller massflux activity (as massflux at cloud base is coupled to moisture convergence). This whole feedback loop changes the system from a massflux dominated Cu regime to a more (turbulent) diffusion dominated Stratocumulus regime.