Importance of numerical ocean modelling and in situ ocean bottom pressure observations for satellite gravimetry from GRACE and GRACE-FO

• Main Author: Poropat, L.
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: Freie Universität Berlin 2020
• Subjects: Antarctic; Southern Ocean; The Antarctic; Antarc*; Ice Shelves
• Online Access: https://gfzpublic.gfz-potsdam.de/pubman/item/item_5004445

The GRACE and GRACE-FO satellites observe the redistribution of mass in terrestrial water storage, ice sheets, oceans, atmosphere, and solid Earth. Because GRACE data is typically accumulated into monthly-mean gravity fields, an a priori background model, namely the Atmosphere and Ocean Dealiasing Level 1B (AOD1B) product, is applied to remove non-tidal variability that would otherwise alias into the monthly solutions. The main disadvantage of AOD1B RL06 compared to its previous release is that it does not simulate the dynamics beneath the Antarctic ice shelves, which can have a strong influence on global ocean circulation. The primary motivation for this work is the development of the new release of AOD1B, but the performed model experiments can also provide useful insight into the influence processes in the Southern Ocean have on global ocean dynamics. To be able to test various model experiments, as well as to compare GRACE gravity field solutions, validation against in situ measured ocean bottom pressure (OBP) is used. The validation is somewhat better suited for submonthly variability of the ocean models than for long-term signals measured by GRACE because the in situ time series are affected by the errors in trend and drift removal on longer temporal scales. The difference between the pointwise in situ and the area-averaging GRACE measuring technique also influences the comparison. It is shown that post-processing choices can severely impact the results of the validation of GRACE fields, so if different solutions are compared, their post-processing needs to be identical. Validation against in situ OBP is used to compare the EGSIEM combined GRACE solution with its five contributing datasets. It is shown that the combined solution is very close to the leading CSR RL05 and ITSG-Grace2016 solutions, outperforming the others. To investigate whether GRACE is able to detect submonthly signals, the ITSG-Grace2016 daily Kalman solution, from which the submonthly atmospheric and oceanic variability has been removed ...