GLOFI – A methodology and toolbox for scale-separation of satellite observations for analysis of gravity waves

Abstract. The direct analysis of atmospheric gravity waves (GWs) in temperature observations is difficult since the much stronger signal of large-scale temperature perturbations such as planetary waves obscure the perturbations due to GWs. The small-scale GW perturbations need to be isolated from the measurements by removing the large-scale temperature background, thereby revealing the object of analysis. In this study, the scale-separation via 2D spectral decomposition, which has the advantage of removing physical wave modes of zonal wavenumber up to 7 and wave frequency up to one cycle per day, is discussed. The technical implementation of this technique in a scale-separation Python-based toolbox, GLOFI (GLObal wave FIt), is detailed and demonstrated on a simulated satellite dataset for the ESA Earth Explorer 11 candidate CAIRT incorporating ECMWF ERA5 temperature data. Planetary wave spectra for the specified wavenumbers and frequencies are obtained by using a 28 day sliding window. These spectra are subsequently used to remove perturbations due to planetary waves from the measurements. This is followed by the removal of tides in a similar way but using a shorter 5-day sliding window and a fit of only stationary waves for ascending and descending orbits separately.

For the considered dataset, the variances of the difference between reference and GLOFI-generated temperature background are an order of magnitude smaller than GW temperature variances, which suggests that the method removes the large-scale waves to a degree that enables the separation of the GW perturbations. Furthermore, the obtained spectra can be used to generate a global temperature background grid which approximately resembles the actual global temperature field. More importantly, the temperature background estimated by GLOFI at the satellite track coordinates is almost identical to the actual reference temperatures along the tracks. Regarding the performance on data including GW perturbations, the isolated small-scale temperature perturbations are virtually identical to the actual reference GW perturbations from the model.

The GLOFI toolbox for scale separation of satellite observations is published as open access along this article.