This research is focused on new technology development. The NERC's Technologies Theme Action Plan, has identified the area of new numerical model development as a critical area where UK skills and expertise should be developed. An important goal of NERC-funded research is 'tackling the key issue of climate change', and as such 'identifying the limitations of a particular model is an important part of stimulating further improvements, and advancing our understanding' (http://www.nerc.ac.uk/research/issues/climatechange/predict.asp). The proposed research focuses on this goal in relation to atmospheric flows. Contemporary numerical models used in the simulation of stratified rotating atmospheric flows are predominantly based on structured computational meshes, with rigid connectivity of a Cartesian grid. For some problems (e.g., hurricanes and flows in long winding valleys), mesh adaptivity has a potential to achieve solutions not obtainable by other methods. However, existing unstructured mesh models are still in their infancy compared to both established structured-grid codes and state-of-the-art engineering advancements with unstructured meshes. Furthermore, their implementation tends to emphasize small-scale convective phenomena and emergency responses, which are relatively easy to model because of the large noise-to-signal ratio, and because of the proximity of events to the excitation region. Insofar as the full-range of wave dynamics are concerned -- including such subtleties as wave-wave and wave-mean-flow interactions, as well as large-amplitude events occurring far from the excitation region -- the potential of unstructured-mesh technology remains unknown. In order to prove the competence and competitiveness of unstructured-mesh technology for simulating all-scale flows in the atmosphere and oceans, there is a pressing need for developing an advanced, fully non-hydrostatic model for simulating accurately rotating stratified flows in a broad range of Rossby-, Froude-, and Reynolds-number regimes. In this work we propose to develop a novel code operating on hybrid (arbitrary polyhedra) meshes, for solving a number of optional forms of non-hydrostatic equations of atmospheric fluid dynamics with flexible mesh-adaptivity capabilities. The proposed model will mirror stratified, rotating turbulence-simulation capabilities of the structured-grid model EULAG (EUlerian/LAGrangian), the proven record of which includes direct and large-eddy simulations of complex fluid problems from laboratory-, to meso-, up to the planetary scale. Additionally, we shall perform rigorous studies and comparisons, by applying both the new model and EULAG to complex benchmarks and research problems combining wave dynamics and turbulence generation on scales relevant to weather, climate and extreme events. To the best of our knowledge, the proposal offers the first ever in-depth study of the relative performance of structured and unstructured/adapted meshes for stratified turbulent flows which involve practical computations of inertia-gravity-wave dynamics. Deliverables: 1) Novel technology --- a high-resolution non-hydrostatic unstructured mesh based model. 2) Method validation and first ever demonstrations of unstructured meshes on advanced test cases, which will deliver information about the applicability of such meshes to realistic atmospheric problems. 3) Quantitative study identifying performance properties of the mesh adaptivity technologies.