Fig. 1: Multiplot of backscatter and Doppler velocity of one 24h example time series, as was already shown in the project presentations; we should also add ARSCL cloud top and MWR LWP. This plot gives the reader a feeling of what type of data we are relying on.

3.) Results

Fig. 2: One 2x2 plot of the shallow cumulus clouds (time vs. w = @ 1 bin below cloud base and @ cloud center, just left and right) showing completely different situations: a) cloud with very narrow subsiding shell (accelerating), b) cloud with rather broader subsiding shell (decelerating), c) cloud with no updrafts below base (dissolving), d) ??? This plots shows that the methodology principally works. However, it is important to calculate here the distance to cloud front or back rim in meters.
Fig. 3: 2x3 plot showing w as funtion of distance to front and back rim of the cloud and as a function of height related to cloud base.  Maybe the double plots shown in the project presentation can be combined to one, so that the velocity below cloud base is also shown? Then we would reduce this figure to a multi-plot of 3 figures above eachother a) (top) all clouds, b(middle) accelerating and decelerating  clouds, c(bottom) dissolving clouds. 
Fig. 4: Show sub-sets of data analysis; still have to produce the analyses here.. We were thinking of classifying the subsiding shell as a function of a) adiabaticity - is the subsiding shell stronger and more narrow when the cloud is still more adiabatic in the beginning of it's lifetime (i.e. LWP from MWR is close to LWPad)? LWPad can be calculated by assuming cloud base temperature and cloud depth - this would be the maximum LWP you can expect. b) cloud depth - is cloud vertical extent related to the strength of the subsiding shell?  c) what is the dependency of the subsiding shell on wind speed below cloud base and maybe even on horizontal wind shear?
Additional questions to discuss & answer
-  Why is the subsiding shell stronger on the back rim of the cloud?
- Why does the subsiding shell "diverge" laterally from the cloud below cloud base - continuity reasons?
- Can we find an objective way to describe the width of the subsiding shell in a quantitative way? Such a metric would allow model comparisons in a more easy way.
- Any way of finding out if our time-height plot "hit" the cloud edge or went though in the cloud center? Via cloud top analyses?
- IWV (integrated vapor fluctuations associated to subsiding shell (MWR highly enough resolved?)
- Do LES show similar signatures (LASSO - link via Andy Vogelmann)? Specifically concerning the asymetries between front and back rim?

4.) Summary and Outlook