An investigation of mass changes in the Bohai Sea observed by GRACE

The Gravity Recovery and Climate Experiment (GRACE) satellite mission has profoundly advanced our knowledge of contemporary sea level change. Owing to the coarse spatial resolution and leakage issue across the land–ocean boundary, it is challenging (even impossible) for GRACE to detect mass changes over a region smaller than its spatial resolution, especially a semi-enclosed basin (e.g., the Bohai Sea) that is adjacent to land with significant mass variation. In this contribution, the causes for the GRACE RL06 mass changes in the Bohai Sea are investigated using a reconstruction technique that is implemented with multisource data, including altimeter observations, steric estimates, and land mass changes from GRACE RL06 mascon solution. Our results by the reconstruction technique demonstrate that the GRACE annual cycles are primarily caused by water mass changes rather than sediment changes. On the other hand, the mass trends from both reconstructed signals and those observed by the GRACE RL06 spherical harmonic coefficients (SHCs) are small, ranging from − 0.38 mm/year to 0.51 mm/year (depending on different data sources). Given that our estimated accuracies are > 0.8 mm/year (the real accuracies should be larger), our reconstructed results cannot directly confirm the presence of sediment accumulation or water mass increase; however, analysis of only the altimetry data suggests the mass trends are due to water mass increase, which would amount to ~ 0.44 Gt/year. Further investigation suggests that the mass trends in the Bohai Sea suffer from a − 2.9 mm/year leakage-in effect from groundwater depletion in the North China and about 2.5 mm/year signal attenuation (resulting in a ~ 2.5 mm/year remaining trend that is roughly equivalent to the leakage-in trend, consequently leading to the small mass trend in the Bohai Sea). Our reconstruction results exemplify that elaborate data processing is necessary for specific cases. We also test whether the recently released RL06 mascon solutions that are resolved with constraints and require no further processing would improve the agreement with altimeter observations. We find that the seasonal cycles are improved relative to the RL06 SHCs; however, the rates derived from the mascon solutions cannot properly represent the altimeter-derived ocean mass estimates for the Bohai Sea, probably because the mascon solutions underestimate the rates or contain some processing artifacts. Nevertheless, the mascon solutions show enh