Stratospheric gravity waves over the mountainous island of South Georgia: testing a high-resolution dynamical model with 3-D satellite observations and radiosondes

Atmospheric gravity waves (GWs) play an important role in atmospheric dynamics but accurately representing them in general circulation models (GCMs) is challenging. This is especially true for orographic GWs generated by wind flow over small mountainous islands in the Southern Ocean. Currently, these islands lie in the “grey zone” of global model resolution, where they are neither fully resolved nor fully parameterised. It is expected that as GCMs approach the spatial resolution of current high-resolution local-area models, small-island GW sources may be resolved without the need for parameterisations. But how realistic are the resolved GWs in these high-resolution simulations compared to observations? Here, we test a high-resolution (1.5 km horizontal grid, 118 vertical levels) local-area configuration of the Met Office Unified Model over the mountainous island of South Georgia (54∘ S, 36∘ W), running without GW parameterisations. The island's orography is well resolved in the model, and real-time boundary conditions are used for two time periods during July 2013 and June–July 2015. We compare simulated GWs in the model to coincident 3-D satellite observations from the Atmospheric Infrared Sounder (AIRS) on board Aqua. By carefully sampling the model using the AIRS resolution and measurement footprints (denoted as model sampled as AIRS hereafter), we present the first like-for-like comparison of simulated and observed 3-D GW amplitudes, wavelengths and directional GW momentum flux (GWMF) over the island using a 3-D S-transform method. We find that the timing, magnitude and direction of simulated GWMF over South Georgia are in good general agreement with observations, once the AIRS sampling and resolution are applied to the model. Area-averaged zonal GWMF during these 2 months is westward at around 5.3 and 5.6 mPa in AIRS and model sampled as AIRS datasets respectively, but values directly over the island can exceed 50 mPa. However, up to 35 % of the total GWMF in AIRS is actually found upwind of the island compared to only 17 % in the model sampled as AIRS, suggesting that non-orographic GWs observed by AIRS may be underestimated in our model configuration. Meridional GWMF results show a small northward bias (∼20 %) in the model sampled as AIRS that may correspond to a southward wind bias compared to coincident radiosonde measurements. Finally, we present one example of large-amplitude (T′≈15–20 K at 45 km altitude) GWs at short horizontal wavelengths (λH≈30–40 km) directly over the island in AIRS measurements that show excellent agreement with the model sampled as AIRS. This suggests that orographic GWs in the full-resolution model with T′≈45 K and λH≈30–40 km can occur in reality. Our study demonstrates that not only can high-resolution local-area models simulate realistic stratospheric GWs over small mountainous islands but the application of satellite sampling and resolution to these models can also be a highly effective method for their validation.