Improving the recovery of monthly regional water storage using one year simulated observations of two pairs of GRACE-type satellite gravimetry constellation

Increasing the spatial sampling isotropy is a major issue in designing future missions dedicated to continue the task of the Gravity Recovery And Climate Experiment (GRACE) mission. From various possible future satellite gravimetry scenarios, the two-pair multi-orbit satellite configuration (Bender-type in the sequence), consisting of a coupled semi-polar pair (the same as GRACE) and an inclined pair of satellites seems to be an optimal mission choice.
This contribution examines the performance of a Bender-type scenario at altitudes of 335 km and 352 km and inclinations of 89° and 63°, respectively, for improving the regional recovery of hydrological signals. To this end, we created one full year of simulated observations of the GRACE and Bender-type configurations. Our investigations include: 1) evaluating the feasible spatial resolution for the recovery of terrestrial water storage (TWS) changes in the presence of realistic instrumental noise and errors in the background models; 2) assessing the influence of aliasing errors in the TWS recovery and its separation from instrumental noise and introduced hydrological signals; and 3) analyzing the regional quality of the gravity-derived TWS results by assessing water storage changes over the 33 world major river basins.
From our simulations, the Bender-derived spectral error curves indicate that, in spite of the instrumental noise, aliasing errors still contaminate the gravity fields above geopotential spherical harmonic coefficient (SHC) degree and order (d/o) 80 till 100. Regarding to the TWS recovery, we found notable improvements for the Bender-type configuration results in medium and small-scale basins, such as the Brahmaputra, Euphrates, Ganges, Indus, Mekong basins in Asia and the Yellow and Orange basins in South Africa. These results were achieved without applying post-processing, which was unachievable using simulations of one pair of GRACE-like configuration. Comparing the magnitudes of errors in the Bender-derived solutions with those of GRACE indicate that the accuracy derived from the Bender-type fields is about two times better than that of GRACE, specifically at medium spatial resolutions of 250 km (SHC d/o 80). We truncated the TWS recovery up to SHC d/o 80 in the spectral domain, whereas all comparisons are demonstrated in the spatial domain after a truncation of the solutions and WGHM field at d/o 60, since beyond this range; a relatively strong instrumental and aliasing errors contaminate the solutions.
Our numerical results indicate that the spatial resolution of the Bender-type TWS recovery can be even higher for the basins with strong temporal water storage variations such as the Amazon basin. Short wavelength mass variations in basins with relatively weaker temporal TWS magnitude, such as the Murray basin, might still need the application of a filter with small averaging kernel.