An evaluation of cloud size distributions using LES

We study LES results of different shallow cumulus cases: BOMEX with and without shear, and cases based on SCMS (Florida) with various values of the sensible surface heat flux. Visual inspection of the simulated cloud fields immediately reveals dramatic differences in the `typical cloud size' between the cases. We calculate the cloud size histograms in each case, and study which cloud size exerts a dominant influence on cloud cover and mass-flux. A priori it is not easy to predict this dominant cloud size: large clouds obviously have a large influence on the mentioned quantities but they are a minority in the total ensemble; conversely, the smallest clouds have a small effect per cloud but their number is very large. Is there a dominating size in between?