Climate-driven and intrinsic oceanic excitation of low-frequency polar motion variability

Earth rotation is fundamental for defining reference systems and applications such as navigation on Earth and in space. In addition to lunisolar torques, internal processes such as mantle deformations or mass redistributions in geophysical fluids affect the rotation of our planet. Along with lunisolar torques, variations in the atmosphere, ocean, terrestrial hydrology, and cryosphere cause small fluctuations, reported as Earth Orientation Parameters and including polar motion and changes in length-of-day (ΔLOD). Since the total angular momentum is conserved, it can only be exchanged between the subsystems of the Earth. The ocean, with its large-scale mass redistributions, induces rotational fluctuations, quantified as excitation functions previously unexplored termed ocean angular momentum (OAM). This thesis aims to shed light on the processes involved in the oceanic excitation of polar motion on interannual time scales. Specifically, the excitation induced by (i) the El Niño–Southern Oscillation (ENSO) and (ii) intrinsic oceanic (i.e., internally generated) variability is quantified. In addition, OAM estimates from three ocean reanalyses are tested for their value in Earth rotation studies, with an eye on potential impacts of the reanalyses' sequential data assimilation (DA) schemes. For periods <120 days and after correction of atmospheric effects, the three reanalyses explain 40–50% of equatorial and 30–40% of ΔLOD excitation variance, comparable to a widely used ocean state estimate. Uncertainties remain at longer time scales, reflecting an increased sensitivity to the chosen DA scheme. With the reliability of the OAM series varying across products and time scales, a statistical combination is a good middle ground by suppressing uncertainties. Quantifying ENSO-induced polar motion excitation is the second contribution of this thesis. Previous studies have shown that ENSO modulates ΔLOD through changes in atmospheric winds. Here, the hypothesis is tested whether ENSO excites polar motion through changes in OAM. The ENSO-induced polar motion is derived from four coupled climate models. The mass term is dominant over the motion term (five orders of magnitude) mainly along 90°E with amplitudes of ±4 mas, particularly via a Pacific-Indian basin-wide mass exchange. During three ENSO cycles (1997/98, 2009/10, 2015/16), OAM changes explain 40–50% of the residual observed polar motion excitation variance, but co-occurrence with other broadband signals complicates a clear attribution to ENSO. The final goal is to investigate the excitation signals associated with intrinsic ocean variability, which emerges from non-linear local processes and can attain large (e.g., basin-wide) scales. Ensemble simulations (1995–2015) indicate that intrinsic OAM fluctuations account for ~46% of interannual oceanic excitation, quantified here for the first time. Mass variability is dominated by a single mode of intrinsic bottom pressure fluctuations, which emerges from Drake Passage and exhibits a see-saw pattern between the Atlantic and Southern/Pacific Oceans. Overall, the results in this thesis provide a valuable contribution to understanding ocean-induced low-frequency polar motion variability. Accurate OAM estimates and consideration of model uncertainties due to, e.g., intrinsic variability, are crucial when the goal is to scrutinise rotation data for more subtle effects from other geophysical fluids. Ocean model limitations could be mitigated by assimilating gravity field data or satellite altimetry in future work. In addition, knowledge of the ENSO-induced oceanic excitation signals—as worked out in this thesis—creates an opportunity to improve the accuracy of both OAM estimates and long-term polar motion predictions.