Are Ocean Reanalyses Useful for Earth Rotation Research?

Abstract

AbstractOceanic circulation and mass‐field variability play important roles in exciting Earth's wobbles and length‐of‐day changes (ΔΛ), on time scales from days to several years. Modern descriptions of these effects employ oceanic angular momentum (OAM) series from numerical forward models or ocean state estimates, but nothing is known about how ocean reanalyses with sequential data assimilation (DA) would fare in that context. Here, we compute daily OAM series from three 1/4° global ocean reanalyses that are based on the same hydrodynamic core and input data (e.g., altimetry, Argo) but different DA schemes. Comparisons are carried out (a) among the reanalyses, (b) with an established ocean state estimate, and (c) with Earth rotation data, all focusing on the period 2006–2015. The reanalyses generally provide credible OAM estimates across a range of frequencies, although differences in amplitude spectra indicate a sensitivity to the adopted DA scheme. For periods less than 120 days, the reanalysis‐based OAM series explain ∼40%–50% and ∼30%–40% of the atmosphere‐corrected equatorial and axial geodetic excitation, similar to what is achieved with the state estimate. We find mixed performance of the reanalyses in seasonal excitation budgets, with some questionable mean ocean mass changes affecting the annual cycle in ΔΛ. Modeled excitations at interannual frequencies are more uncertain compared to OAM series from the state estimate and show hints of DA artifacts in one case. If users are to choose any of the tested reanalyses for rotation research, our study points to the Ocean Reanalysis System 5 as the most sensible choice.