Due to increasing computer facilities, GCMs can be run with significantly higher resolution and/or higher complexity than a decade ago. Nevertheless, the closure problem of unresolved dynamical scales remains a major issue. Regarding conventional parameterizations of gravity waves (GWs), which are based on the assumption of a stationary and horizontally uniform mean flow, the issue is even more important since the scales of resolved and parameterized waves may become comparable. One solution to this problem is a continuous description of the unresolved wave field on the basis of an adequate radiative-transfer equation. Alternatively, one may dispense with parameterization of GWs, apply a high resolution, and represent unresolved scales by a macro-turbulent diffusion such as to account for damping of resolved waves in a self-consistent fashion. This is the approach of currently two GCMs that extend into the lower thermosphere and simulate a realistic GW drag in the upper mesosphere: The Kanto model (e.g., Watanabe and Miyahara, 2009, JGR) and the Kuehlungsborn Mechanistic general Circulation Model (KMCM; e.g., Becker, 2012, Space Sci. Rev.). In this presentation we describe a new version of the KMCM (including radiation and moisture cycle) with particular emphasis on 1) the turbulent diffusion scheme, 2) the way that energy deposition is formulated with full respect to energy conservation and the second law, and 3) the generation of secondary waves in regions of strong GW drag. We shall also discuss the fact that for resolutions that are typically feasible in middle atmosphere GCMs, the resolved GW scales are strongly resolution-dependent.

---

*email: becker@iap-kborn.de
*Preference: Oral