Characterization of Internal Tide Incoherence: Eulerian versus Lagrangian Perspectives

Abstract

The Lagrangian and Eulerian surface current signatures of a low-mode internal tide propagating through a turbulent balanced flow are compared in idealized numerical simulations. Lagrangian and Eulerian total (i.e. coherent plus incoherent) tidal amplitudes are found to be similar. Compared to Eulerian diagnostics, the Lagrangian tidal signal is more incoherent with comparable or smaller incoherence timescales and larger incoherent amplitudes. The larger level of incoherence in Lagrangian data is proposed to result from the deformation of Eulerian internal tide signal induced by drifter displacements. Based on the latter hypothesis, a theoretical model successfully predicts Lagrangian autocovariances by relating Lagrangian and Eulerian autocovariances and the properties of the internal tides and jet. These results have implications for the separation of balanced flow and internal tides signals in the sea level data collected by the future Surface Water and Ocean Topography (SWOT) satellite mission.

Validating the spatial variability of the semidiurnal internal tide in a realistic global ocean simulation with Argo and mooring data

“…However, the Doppler effect and spatial decorrelation induced by the Lagrangian sampling of the floats both act as decorrelating processes causing the autocovariance to decay at longer time lags (Geoffroy and Nycander, 2022). Following Caspar-Cohen et al (2022) we call this mecanism apparent decorrelation, as it is unrelated to the intrinsic decorrelation of the propagating IT. Geoffroy and Nycander (2022) estimated the apparent decorrelation timescale to be longer than that of the observed Lagrangian autocovariance function, on a global average.…”

“…In other words, IT lose coherence by interacting with the eddying ocean. They decorrelate: the autocovariance of a time series representing the internal tide variability at a fixed position (away from the source) inevitably decays with time lag (Caspar-Cohen et al, 2022;Geoffroy and Nycander, 2022).…”

“…the autocovariance at short time lags, before any loss of coherence occurs. On the other hand, the Lagrangian sampling of the drifting floats results in decorrelating effects on top of the intrinsic decorrelation of the IT (Zaron and Elipot, 2021;Caspar-Cohen et al, 2022;Geoffroy and Nycander, 2022). The latter decorrelation cannot be disentangled from the former using Argo data only.…”

Using Lagrangian Filtering to Remove Waves From the Ocean Surface Velocity Field

“…Pinkel (2008), Shakespeare and Hogg (2017), and Caspar‐Cohen et al. (2022) show that both balanced flows and IGWs are Doppler shifted by the large scale flow field. This means that fixed‐location frequency filtering may be inaccurate, particularly in regions with fast background flows.…”

“…Using this method, motions are measured at a fixed location on the Earth, with motions at frequencies lower than the inertial frequency labeled as balanced, and motions at frequencies higher that the inertial frequency labeled as wave-like. Pinkel (2008), Hogg (2017), andCaspar-Cohen et al (2022) show that both balanced flows and IGWs are Doppler shifted by the large scale flow field. This means that fixed-location frequency filtering may be inaccurate, particularly in regions with fast background flows.…”