A new empirically derived source parametrization for convectively-generated subgrid scale gravity waves in the Met Office GCM

Despite a drive for Earth system models (ESMs) to adopt enhanced horizontal and vertical resolutions, they are still unable to represent the middle atmosphere momentum budget to sufficient accuracy without the inclusion of parametrizations for small-scale gravity waves. The current generation of ESMs are able to represent key features such as the mean strength of polar night jets or the period of the quasi-biennial oscillation in the equatorial stratosphere with the inclusion of non-orographic gravity wave schemes that are focused on zonal mean budgets. Such models have a characteristic tendency, however, to underestimate tropical variability in the middle atmosphere with the quasi-biennial and semi-annual oscillations appearing to be more regular than observational evidence suggests they should be. Improved variability is relevant for detection and attribution of climate change signals against the 'natural' control background of the ESM and for prediction of diagnostics of interest to mitigation and adaptation that measure the onset date at which given thresholds of change are reached.

The non-orographic gravity wave scheme used in the Met Office general circulation model (GCM) currently has a source which is spatially and temporally invariant. However, various non-orographic sources for gravity waves in the atmosphere have been recognised, of which arguably the most important relates to convection. The inhomogeneity of convective activity both spatially and over time raises an obvious prospect that the representation of gravity wave sources generated by convection is a 'missing' process that can help to address the lack of variability in the ESM. Furthermore, by including a process neglected by many current state-of-the-art gravity wave parametrizations, the physical realism of ESMs may be improved and a better match to observations obtained in cases where this process occurs. And, in the climate change context, alterations in pattern and occurrence of convection arising from global warming in the future will have the potential to affect the generation of gravity waves and hence feedback on large-scale phenomena such as the Brewer-Dobson Circulation.

Results will be presented from simulations with a new simple parametrization of convective gravity wave sources within the Met Office model (MetUM) that is based on an empirical relationship with total precipitation. This was derived from analysis of results obtained from a convection-resolving configuration of the model. The impact due to such a scheme on the MetUM middle atmosphere evolution will be explored, with particular attention paid to forcing of the simulated quasi-biennial oscillation. An initial focus will be on performance assessment of schemes at the process level to characterize distribution patterns, seasonal responses and statistics of the gravity wave momentum fluxes generated and propagated in both a free-running version of the global model and a nudged version, in which the temperature and wind conditions between 1km and 65km were constrained towards those of the European Center for Medium-Range Weather Forecasting Re-Analysis. Output will be evaluated against a combination of local (e.g. balloon) and global (e.g. satellite) observation derived data and output from high resolution simulations. As variability at levels above the upper troposphere is influenced not just by variability in the original gravity wave sources but also by the effect of filtering due to wind shear encountered as the waves propagate vertically, the relative contributions of these different aspects will be investigated.