Measuring turbulence from wave-following platforms

Autonomous surface platforms equipped with pulse-coherent high-resolution (HR) ADCPs are a promising tool for measuring turbulence and estimating turbulent dissipation rates, ϵ(z), close to the air-sea interface. However, surface gravity waves generate significant bias in ϵ(z) if not sufficiently separated from the turbulent signal. In a previous study, the authors developed a method of isolating wave orbital velocities from the data using empirical orthogonal functions (EOFs). Low-mode EOFs had characteristics of surface gravity waves, while higher-mode EOFs had characteristics of turbulence. After filtering empirical wave profiles constructed from the low-mode EOFs from the data, resultant ϵ(z) were in close agreement with law-of-the-wall scaling during quiescent conditions. In this study, we further validate the EOF-filtering technique by comparing EOFs of the HR ADCP data with those computed from synthetic wave data which does not contain turbulence. As expected, low-mode EOFs of the synthetic data are in strong agreement with those of the real data, while high-mode EOFs reflect only noise due to the absence of turbulence. Wave profiles constructed from the low-mode EOFs are then used to quantify the potential for bias in ϵ(z) if wave velocities are not sufficiently filtered from the data.