Stratospheric gravity waves in a post-limb sounder era: can GNSS-RO be used to extend the SABER QBO-driving record?

The quasi-biennial oscillation (QBO) is a slowly repeating cycle of winds which dominates tropical lower-stratospheric dynamics and has been described as the “heartbeat of the stratosphere”. However, it is challenging to represent in weather and climate models because its periodicity and magnitude are controlled by small-scale gravity waves (GWs) that cannot be resolved on model grids. To quantify this GW driving we require high-resolution measurements, ideally from satellites to ensure full spatial coverage. Since 2002, the SABER instrument on the TIMED satellite has provided such data, facilitating long-term studies of QBO GW driving. However, SABER is expected to be decommissioned later this year, and no replacement is planned. Here, we assess the possibility of using GNSS-RO data to extend the 23-year SABER record. GNSS-RO cannot be used as a simple replacement for long-term studies because data volumes are too low before 2006 and for most of the late 2010s, and thus, we ideally wish to supplement rather than replace the SABER record. However, while the two datasets have broadly similar lower stratospheric resolutions when compared to the full Earth observation constellation, GNSS-RO measurements are higher resolution in all three dimensions than SABER and are oriented differently in 3D space. As a result of this, GNSS-RO GW measurements exhibit much larger GW potential energies (GWPE) and shorter vertical wavelengths than those from SABER. To understand these differences, we use a high-resolution run of the GEOS model to produce synthetic GW measurements, then systematically vary the measurement characteristics between those of the two real instruments. This allows us to identify the key drivers of the different GW properties they measure. We demonstrate that the differences between QBO-driving GW properties measured by the two instruments are primarily due to vertical resolution, with horizontal resolution (either along or across line of sight) and orientation angle playing a negligible role. We further demonstrate that, with a simple vertical smoothing of the GNSS-RO data in the vertical before analysis for GWs, the measured GW properties become near-identical, allowing us to use SABER and GNSS-RO data near-equivalently for this use case. Since GNSS-RO data are now a crucial component of the global numerical weather prediction constellation and are hence highly likely to be available in the long term, this allows us to produce a consistent long-term record of QBO GW forcing from 2002 onwards without key gaps which would be otherwise present in the early 2000s and late 2010s.