Jackson, D.R., Gadian, A., Hindley, N.P., Hoffmann, L., Hughes, J., King, J., Moffat-Griffin, T., Moss, A.C., Ross, A.N., Vosper, S.B., Wright, C.J., Mitchell, N.J., Jackson, D.R., Gadian, A., Hindley, N.P., Hoffmann, L., Hughes, J., King, J., Moffat-Griffin, T., Moss, A.C., Ross, A.N., Vosper, S.B., Wright, C.J., and Mitchell, N.J.
New observations at South Georgia (an extremely intense source of gravity waves) are used to evaluate model simulations of gravity waves and wakes, paving the way for improved understanding of these important phenomena. Gravity waves (GWs) play an important role in many atmospheric processes. However, the observations-based understanding of GWs is limited, and representing them in numerical models is difficult. Recent studies show that small islands can be intense sources of GWs, with climatologically significant effects on the atmospheric circulation. South Georgia, in the South Atlantic, is a notable source of such “small island” waves. GWs are usually too small scale to be resolved by current models, so their effects are represented approximately using resolved model fields (“parametrization”). However, the “small island” waves are not well represented by such parametrizations and the explicit representation of GWs in very high resolution models is still in its infancy. Steep islands such as South Georgia are also known to generate low-level wakes, affecting the flow hundreds of kilometers downwind. These wakes are also poorly represented in models. We present results from the South Georgia Wave Experiment (SG-WEX) for 05 July 2015. Analysis of GWs from satellite observations are augmented by radiosonde observations made from South Georgia. Simulations were also made using high-resolution configurations of the Met Office Unified Model (UM). Comparison with observations indicates that the UM performs well for this case, with realistic representation of GW patterns and low-level wakes. Examination of a longer period simulation suggests that the wakes generally are well represented by the model. The realism of these simulations suggests they can be used to develop parametrizations for use at coarser model resolutions.