Acquiring Long-Term Turbulence Measurements from Moored Platforms Impacted by Motion

For measurements from either profiling or moored instruments, several processing techniques exist to estimate the dissipation rate of turbulent kinetic energy epsilon, a core quantity used to determine oceanic mixing rates. Moored velocimeters can provide long-term measurements of epsilon, but they can be plagued by motion-induced contamination. To remove this contamination, two methodologies are presented that use independent measurements of the instrument's acceleration and rotation in space. The first is derived from the relationship between the spectra (cospectra) and the variance (covariance) of a time series. The cospectral technique recovers the environmental (or true) velocity spectrum by summing the measured spectrum, the motion-induced spectrum, and the cospectrum between the motion-induced and measured velocities. The second technique recovers the environmental spectrum by correcting themeasured spectrum with the squared coherency, essentially assuming that the measured signal shares variance with either the environmental signal or the motion signal. Both techniques are applied to moored velocimeters at 7.5 and 20.5m above the seabed in 105m of water. By estimating the orbital velocities from their respective spectra and comparing them against those obtained from nearby wave measurements, the study shows that the surface wave signature is recovered with the cospectral technique, while it is underpredicted with the squared coherency technique. The latter technique is particularly problematic when the instrument's motion is in phase with the orbital (environmental) velocities, as it removes variance that should have been added to the measured spectrum. The estimated epsilon from the cospectral technique compares well with estimates from nearby microstructure velocity shear vertical profiles.