We completed a thorough assessment of the present-day accuracies of AOD1B RL06 as represented both by means of synthetic series of true errors suitable for simulations, but also by means of a (stationary) error covariance matrix delivered to IP4 and IP5. Error estimates have been tested with both real GRACE data and with end-to-end gravity mission simulations and were found to be highly realistic.

In cooperation with IP3, AOD1B RL06 has been found to be competitive in an external comparison with a small ensemble of alternative numerical ocean models (Schindelegger et al. (2021). Nonetheless, this study also provided hints for potential further improvements which were then incorporated into the next AOD1B release 07. Insights gained from work in IP2 also contributed to a bias-free estimation of barystatic mean sea-level variations from GRACE and GRACE-FO data (Dobslaw et al., 2020).

IP2 also made progress towards an improved AOD1B by implementing the feedback of self-attraction and loading of surface mass distributions into MPIOM (Shihora et al., 2022a). A preliminary version v71 of a new AOD1B has been distributed to interested colleagues in December 2020 and prompted a novel application of the GRACE-FO LRI data for detecting of rapid surface mass redistributions (Ghobadi-Far et al., 2022). Feedback on this test data has been also incorporated into a revised MPIOM configuration that was finally used for AOD1B RL07. Stokes coefficients of atmosphere-ocean variability covering the whole GRACE mission period have been published in summer 2022 (Shihora et al., 2022b), and AOD1B RL07 for all remaining years from 1975 until present were made available step-by-step until the end of the first funding period of NEROGRAV.