Enhanced terrestrial water storage change estimation by joint inversion of GNSS and GRACE data with spatiotemporal constraints

Summary Global Navigation Satellite System (GNSS) technology and Gravity Recovery and Climate Experiment (GRACE) satellite gravimetry provide essential observational constraints for monitoring terrestrial water storage (TWS) changes. In this study, we develop a spatiotemporally constrained joint inversion model that inverts GNSS and GRACE observations to estimate reliable TWS changes over Brazil from January 2008 to July 2016. We evaluate the performance of the spatiotemporally constrained joint inversion model using both closed-loop simulation and independent GNSS surface displacement observations. The simulation results indicate that the joint inversion results using spatiotemporal constraints outperform GNSS-only solutions, GRACE-only solutions, yielding improved accuracy and reliability than joint inversion results with spatial constraints. The corresponding standard deviations are 45.94 mm, 57.44 mm, 52.53 mm, and 48.88 mm, respectively. Analysis from the measured data shows that vertical displacement time series simulated from TWS changes derived from spatiotemporally constrained joint inversion indicate higher consistency with GNSS-observed surface displacement time series at nine GNSS stations compared with the GNSS-only, GRACE-only (from CSR-M), and the joint inversion results with spatial constraints. The corresponding standard deviations and correlation coefficients are 1.93 mm & 0.92, 2.54 mm & 0.87, 4.45 mm & 0.74, and 1.94 mm & 0.91, respectively. Meanwhile, the joint inversion results from measured GNSS and GRACE data using spatiotemporal constraints show lower uncertainty and higher stability than those of GNSS-only and joint inversion results with spatial constraints. The proposed inversion model provides an alternative means to fully exploit the potentials of GNSS and GRACE technologies for integrated monitoring of TWS changes.