Behavior indicative of stratified turbulence was found in a large variety of large-domain nonhydrostatic numerical experiments involving many patterns of specified heating functions. The purpose of the experiments was to investigate convective mesovortices, however the results are also of interest here because of the upscale layered circulations generated, along with interesting results regarding coupling with gravity waves in some experiments (low inertial stability) and clean separation in other experiments (high inertial stability). The results are broadly similar to Vallis, et.al. (1997, QJRMS pp1621-1652), but using controlled heating or cooling functions instead of dynamically generated precipitation, and using much better vertical resolution here (around 300 m). Heating functions included ensembles of cloud-scale heating, lines of heating, broad mesoscale regions of cooling (simulating MCC anvil melting), and a considerable integration period after heating is terminated. Environments include ambient cyclonic rotation, anticyclonic rotation, no initial flow, latitudes of 35 degrees, 15 degrees, and equator. Details of the adjustment process yielding the rotational mesoscale flows have been examined, as well as potential vorticity generation diagnostics. The picture is consistent with the stratified turbulence hypothesis, thus explaining how larger scale flows can be generated by arrangements of smaller scale heating, with upscale development enabling the flow to exceed the local Rossby radius. The local Rossby radius for a mesoscale vortex and the Froude number for application to stratified turbulence are both related (differently) to ratios of wind perturbation to gravity wave phase speed.
Note: I am the sole author of this and another modification to this conference; I realize some conferences restrict first authors to one presentation - if this is a problem I would hope you could give me a poster presentation for one of them (while still retaining oral presentation for the other, or possibly do poster for both if that's how it happens to work out; I have not presented at any conference in several years and I am not co-author on any modifications)