Analysis of Inertial Oscillations in the Global Surface Drifter Dataset

An analysis of the characteristics of inertial oscillations in a high-temporal resolution version of the global surface drifter dataset is presented, based on inferences from a complete stochastic model for the structure of Lagranian velocity spectra. A change in the Argos satellite tracking of the surface drifters in January 2005, with five or six instead of two satellites available for tracking, led to a dramatic improvement in the time resolution of the surface drifters to nearly hourly resolution. This enhanced dataset is analyzed using a stochastic modeling approach, in which the various physical components contributing to the drifters' Fourier spectra are represented as stochastic processes. Fitting for the values of the parameters controlling these processes allows the properties of inertial oscillations to be extracted and separated from the background flow and tidal signals. A key ingredient is to correctly account for the semidiurnal tide, which emerges as regionally important, and is associated with locations of strong internal tide generation. Accounting for the fact that wind spectra are sloped, rather than flat, turns out to be important for correctly inferring inertial damping timescales, which otherwise tend to be substantially overestimated. High-resolution global maps of inertial oscillation amplitude and damping time scales are then presented, together with an assessment of errors sources a comparison to previously published results from the literature.