Validation of ocean mass variability derived from the Gravity Recovery and Climate Experiment - Studies utilizing in-situ observations and results from a Finite Element Sea ice - Ocean Model

Abstract

The Gravity Recovery and Climate Experiment (GRACE) provides estimates of the Earth's time-variable and static gravity field with an unprecedented accuracy. As fluctuations of the gravity field on sub-annual time scales are mainly induced by mass redistribution on the Earth's surface, GRACE is potentially able to monitor oceanic mass variability and redistribution. Current gravity data products, however, suffer from aliasing effects due to insufficient accuracy of background models. This study compares different filter mechanisms and develops a new filtering approach which uses information on ocean circulation patterns derived from model simulations with the Finite Element Sea - Ice Ocean Model (FESOM). To obtain a realistic representation of the ocean mass budget on interannual time scales in the FESOM model, the computation of sea surface height is extended in order to consider the effect of surface freshwater fluxes and internal mixing. The river runoff forcing is modified to balance the net evaporation on long time scales. For validation, we utilize in-situ ocean bottom pressure (OBP) data from a global bottom pressure data base. For the correction of tides, variations derived from the tidal model (FES2004), also used as a background model in the GRACE data processing, are subtracted from the in-situ data. In general, the validation of GRACE-derived anomalies against in-situ time series indicates a good agreement between the two data sets. Especially, for the high latitude arrays GRACE captures a considerable part of the observed oceanic variability. Largest errors in the GRACE data are found in the tropical Atlantic, where the GRACE-derived OBP data from all data centers feature spurious variability which probably goes back to tidal aliasing and/or the large hydrological cycle over the Orinoco/Amazon river basin. Filtering the GRACE data with the new pattern-based approach improves the correlations between GRACE and in-situ OBP anomalies compared to conventional isotropic Gauss filtering. Focussing on the ocean domain, the filter reduces land leakage effects and introduces valuable information on the ocean circulation to the GRACE data, which helps to reduce errors and to identify geophysical signals in the gravity field solutions. As a first application, the improved GRACE data sets are utilized to study the relation between oceanic transport variability and cross-flow gradients of OBP anomalies. Maps of correlations between the Southern Annular Mode (which is the dominant mode of atmospheric variability in the Southern Ocean) and GRACE OBP anomalies reveal a circumpolar band of highly negative correlation around Antarctica closely following f/H contours. Although this spatial structure is already found for data filtered with an isotropic Gauss filter, correlations are further enhanced when using the new filter approach. Based on these results, this study supports previous findings indicating that SAM affects ACC transport variability at least on a month-to-month time scale.