To monitor temporal variations of the Earth’s gravity field and mass transport in the Earth’s system, data from gravity recovery and climate experiment (GRACE) satellite mission and its successor GRACE Follow-On (GFO) are used. To fill in the temporal gap between these missions, other satellites’ kinematic orbits derived from GPS-based high-low satellite-to-satellite tracking data may be considered. However, it is well known that kinematic orbits are highly sensitive to various systematic errors. These errors are responsible for a non-stationary noise in the kinematic orbits, which is difficult to handle. As a result, the quality of the obtained gravity field solutions is reduced. In this research, we propose to apply an epoch-difference (ED) scheme in the context of the classical dynamic approach to gravity field recovery. Compared to the traditional undifferenced (UD) scheme, the ED scheme is able to mitigate constant or slowly varying systematic errors. To demonstrate the added value of the ED scheme, three sets of monthly gravity field solutions produced from 6 years of GRACE kinematic orbits are compared: two sets produced in-house (with the ED and UD scheme), and a set produced with the undifferenced scheme in the frame of the short-arc approach (Zehentner and Mayer-Gürr in J Geodesy 90(3):275–286, 2015. https://doi.org/10.1007/s00190-015-0872-7). As a reference, we use state-of-the-art ITSG-Grace2018 monthly gravity field solutions. A comparison in the spectral domain shows that the gravity field solutions suffer from a lower noise level when the ED scheme is applied, particularly at low-degree terms, with cumulative errors up to degree 20 being reduced by at least 20%. In the spatial domain, the ED scheme notably reduces noise levels in the mass anomalies recovered. In addition, the signals in terms of mean mass anomalies in selected regions become closer to those inferred from ITSG-Grace2018 solutions, while showing no evidence of any damping, when the ED scheme is used. We conclude that the proposed ED scheme is preferable for time-varying gravity field modeling, as compared to the traditional UD scheme. Our findings may facilitate, among others, bridging the gap between GRACE and GFO satellite mission.

The Gravity Recovery And Climate Experiment (GRACE) mission has achieved a quantum leap in knowledge of the Earth's gravity field. However, current gravity field solutions still cannot reach the prelaunch baseline accuracy. One of the reasons for that is the presence of colored noise in GRACE data, which is typically ignored in the classical dynamic approach to gravity field modeling. In this research, we propose to account for colored noise in the classical dynamic approach by applying the frequency-dependent data weighting (FDDW) scheme, so that enhanced estimates of gravity field solutions are produced. The monthly solutions are compared with those produced using the standard least squares adjustment without a data weighting scheme. The comparison is performed in both spectral and spatial domains, showing the positive effect of the FDDW scheme in all considered cases. For instance, the cumulative geoid height errors up to degree 96 are reduced by 18%. In the spatial domain, the FDDW scheme lowers noise level in mass changes over the oceans, Mississippi river basin, and Greenland by 20, 38, and 23%, respectively, when compared to the without a data weighting scheme. In addition, the consistency of mass changes over the Mississippi and Congo river basins with those inferred from the state-of-the-art hydrology model WaterGAP is substantially improved when the FDDW scheme is applied. These results indicate that modeling colored noise in the GRACE data allows to significantly improve the recovered monthly solutions. This finding is likely applicable also to the GRACE Follow-On mission.