Sensing individual components of atmospheric mass variability with future satellite gravimetry missions

Observations of the time-variable gravity field by means of GRACE and GRACE-FO are long known to be impacted by atmospheric non-tidal mass variability. To avoid aliasing artifacts from high-frequency mass variations into the Level-2 monthly gravity fields that are being calculated routinely out of the mission data, the impact of atmospheric mass variations is reduced with a time-variable background model based on ECMWF operational and reanalysis data (i.e., AOD1B). Future satellite constellations based on low-low satellite-to-satellite tracking in a double pair configuration as currently jointly explored by NASA and ESA within the MCM/MAGIC mission concept have the ability to directly observe atmospheric mass variations down to much shorter time-scales so that (possibly erroneous) background models might not be necessary anymore in the future. 

By means of two modern global atmospheric reanalyses ERA5 and MERRA2, we will assess the impact of non-tidal atmospheric mass variability for both the total atmospheric signal as well as its individual components (i.e., dry atmospheric mass; water vapor partial pressure; atmospheric moisture flux divergences, etc.). Results will be presented for different frequency bands (< 3days; 3-10 days; 10-30 days) and contrasted against the differences between ERA5 and MERRA2 as a measure of the present-day uncertainties in large-scale atmospheric modelling. This analysis will allow to explore the potential contribution of gravimetric satellite observations to future atmospheric reanalysis efforts -- or eventually even operational numerical weather prediction as long as the latency of satellite gravimetric data can be drastically improved.