Jürgen Kusche (Institute of Geodesy and Geoinformation, University of Bonn, Germany)

The determination of time-variable gravity field from DORIS is typically restricted to the very first degrees of the spherical harmonic spectrum. In order to better attribute the orbit perturbations to spatially localized mass redistributions, we apply in this contribution a novel parametrization by representing the mass changes by a limited number of spatial patterns which were identified in a previous step by a principal component analysis applied to the gravity field time series from GRACE and GRACE Follow-On. These spatial patterns serve as a kind of large-scale mascons tailored within the GRACE/GRACE-FO data period. This approach was developed in a recent publication for time-variable gravity field modelling with the spherical satellites tracked by SLR. It is shown here that it is highly suitable also for the DORIS system. The large-scale mass variations found from ten DORIS satellites at altitudes up to 1000 km agree well with those from SLR, though the solutions are based in some early years on one single satellite. Substantial synergies arise when merging DORIS and SLR at the normal equations level. Using observations from both techniques not only reduces the noise in the time series, but allows for solving for more spatial patterns which increases the amount of recoverable signal and, thus, the spatial resolution. Over the Ocean, in Greenland and for several large river basins, the DORIS/SLR mass change time series are very close to GRACE and GRACE-FO while, for the pre-GRACE era, some confirmation can be found from hydrometeorological reconstructions of land water storage. The same agreement is, of course, not achieved in smaller regions where both the sensitivity of SLR and DORIS and the applied set of spatial patterns reach their limits.
The time series presented here will contribute to long-term mass studies concerned with sea-level change, ice mass loss at high latitudes or water storage variability for very large river basins. It might also be useful as a background force model for precise orbit determination at higher altitudes. A dynamic reconstruction of Jason-3 orbits from GNSS-derived positions reveals that the DORIS/SLR solution performs as well in this task as the frequently applied EIGEN-GRGS.RL04 model, with a slight advantage in the later years when the EIGEN model provides pure predictions.