Intercomparison of signals and noise in time variable gravity field releases from an ice sheet perspective

The International Combination Service for Time-variable Gravity Fields (COST-G), a product center of IAG’s International Gravity Field Service, aims on the combination of monthly global gravity field models. A consolidated gravity field series is derived from individual releases provided by different analysis centres (AC). COST-G’s Product Evaluation Group (PEG) assesses the combined gravity field series as well as the series provided by the ACs regarding their suitability for studying mass changes in the Earth’s subsystems (e.g. oceans, cryosphere, continental hydrosphere). Here we present results from the PEG’s assessment regarding mass changes of the ice sheets. Our study focuses on the COST-G RL01 series and the contributing releases AIUB RL02, GFZ RL06, ITSG-Grace2018, CSR RL06 and an unconstrained variant of GRGS RL05.

Based on residual variations of the spherical harmonic (SH) coefficients with respect to a long-term and seasonal model, we quantify the noise level of the latest GRACE/GRACE-FO solutions series provided by COST-G and the contributing ACs. This assessment is performed both in the SH domain and in the space domain, focusing on the polar regions. A regional integration approach using tailored sensitivity kernels is applied to derive mass change time series for individual drainage basins and the entire ice sheets in Greenland (GIS) and Antarctica (AIS). A measure for the noise level of the different mass change time series is inferred from the residuals with respect to a climatology, corrected for remaining inter-annual mass changes. In this way we are able to assess if and to which extent mass change products for GIS and AIS can benefit from the combination of different solution series. Furthermore, we also quantify the signal content inherent to the individual mass change time series in terms of the seasonal signal and the linear trend (i.e. mass balance). Differences in the signal content between the releases are further investigated with respect to contributions from different parts of the SH spectrum. In addition to selected SH coefficients (e.g. C21, S21), we systematically attribute the differences to each SH degree and the corresponding SH orders.