Since 2002, time-lapse satellite gravimetry missions have successfully observed global time-variable mass transport. The GRACE (Gravity Recovery And Climate Experiment; Tapley et al., 2004, 2019) mission and its successor, GRACE-FO (GRACE-Follow On; Chen et al., 2022; Landerer et al., 2020), have almost continuously delivered monthly observations of the gravity field for more than two decades. As GRACE-FO approaches the end of its lifetime, new satellite gravity missions are planned for launch. For the continuation of the record, GRACE-C (GRACE-Continuity by NASA and the German Space Agency at DLR with support from GFZ, BMWK, BMFB, HGF and MPG ) is planned to be launched in 2028 in a near-polar orbit at an altitude of ~500 km. GRACE-C will be followed by the Next Generation Gravity Mission (NGGM) launched by the European Space Agency (ESA) in 2032 in an inclined orbit of 65–70 degrees at an altitude of ~400 km. In their overlapping period, these two satellite pairs form the Mass-change And Geoscience International Constellation (MAGIC).
In the frame of the ESA SING (Studying the Impact of the NGGM and MAGIC Gravity missions) project (SING project website, 2026), extensive simulations have been performed to evaluate the added value of extended gravity field measurements in time with enhanced spatial and temporal resolution, and, reduced latency in data availability. Synthetic observations of the gravity field have been generated at Levels 2 and 3 for GRACE-C-like, NGGM and MAGIC satellite configurations using a closed-loop numerical simulator integrating instrument noise, background model errors, and realistic satellite orbits. The simulations utilize target Earth signals from the ESA Earth System Model ESM 2.0 (Dobslaw et al., 2015), including hydrology, ice, and solid Earth components. Two parameterisation strategies used in this study yield simulated gravity solutions of mean fields at 5-day and 30-day resolutions. The other two strategies result in direct estimation of the trend and annual signal (trendannual) and direct estimation of the long-term trend (trendonly). For each parameterisation strategy, the data products are separated into three levels (L2, L2P, and L3). L2 are Stokes coefficients of the simulated Earth’s potential provided separately for each mission scenario and expressed in the spherical harmonic basis in ICGEM format. L2p and L3 synthetic data represent simulated surface mass anomalies provided separately for each mission scenario and expressed in equivalent water heights over regular 1°*1° grids in NetCDF format. The L3 data were corrected for Glacial Isostatic Adjustment (GIA), while the L2p data were not. . The full description of the data and methods is provided in the data description publication (Schlaak et al. in prep.), and the file structure of this data set is fully described in the file inventory.
The resolution of all data products (L2a, L2b, L2P, L3) for each mission scenario (GRACE-C-like, NGGM, MAGIC) depends on the simulation type described in Schlaak et al. (in prep.). For 5-daily solutions, the spatial resolution corresponds to ca. 285 km (d/o 70). For monthly solutions, the temporal resolution is 30 days, with a spatial resolution of ca. 166 km (d/o 120). Trend and annual signals (trend-and-annual solutions) have been estimated simultaneously over a period of 12 years (with 1-year increments) and a spatial resolution of ca. 150 km (d/o 130). For trend-only solutions, the spatial resolution is increased to 125 km, with trend estimates over 5 and 12 years.
Additionally, empirical Variance-Covariance Matrices (VCMs) are provided for L2a and L3b data for the 5-daily and monthly simulation types, computed from Monte Carlo simulations (Schlaak et al., in prep.).
The new unconstrained static gravity field HUST-Grace2026s is recently developed at Huazhong University of Science and Technology. During retrieving our model, the reprocessed GRACE L1b RL03 data and GRACE-FO RL04 data are used, and the newly de-aliasing product AOD1B RL07 is applied in the time span from 2002-04 to 2025-03. In addition, a hybrid processing chain is applied to improve the quality of final solutions. Further details are presented in Zhou et al. (2024) and Zheng et al. (2026).
GRACE/GRACE-FO Level-3 products based on GFZ RL06 Level-2B products (Dahle & Murböck, 2019) representing ice-mass changes for the Antarctic Ice Sheet (AIS) and the Greenland Ice Sheet (GIS). The ice-mass changes are provided both as basin average product and as gridded product.
Basin-average ice-mass changes are obtained using the inversion procedure based on a forward modelling approach as described in Sasgen et al. (2013) for the AIS and Sasgen et al. (2012) for the GIS.
Gridded ice-mass changes are provided at polar-stereographic grids with a grid spacing of 50 x 50 km^2. The applied algorithm is based on tailored sensitivity kernels (Döhne et al., 2023; Groh & Horwath, 2021), and has also been used to generate gravimetric mass balance products within the ESA Climate Change Initiative (CCI) projects for the AIS and the GIS.
These Level-3 products are visualized at GFZ's web portal GravIS (https://gravis.gfz.de).
---------------------------------------------------------------------------------------------
Version History:
13 March 2026:
Release of Version 0005. This is an update of Version 0004 of the same data set (see changelog).
16 January 2025:
Release of Version 0004. This is an update of Version 0003 of the same data set (see changelog).
21 April 2023:
Release of Version 0003. This is an update of Version 0002 of the same data set (see changelog).
09 June 2020:
Release of Version 0002. This is an update of Version 0001 of the same data set (see changelog).
All changes and updates are documented in the changelog available via the data download section. Previously released versions of this data set are available in the "old_versions" subfolder of the data download folder.