Effects of Anisotropy on Observation of Turbulent Dissipation in Bottom Boundary Layers

Turbulent mixing in the ocean is often quantified using observations of kinetic energy dissipation (e.g., the Osborn model). However, because direct measurement of dissipation remains a challenge, it is commonly estimated from acoustic Doppler velocity data using methods such as inertial fitting or the structure function. These methods employ Kolmogorov constants and Taylor’s frozen turbulence hypothesis, which are appropriate for steady homogeneous and isotropic turbulence. For boundary layer flows, these assumptions breakdown when the velocity observations occur near the buffer layer and – in particular – the flow becomes strongly anisotropic. To address this issue, we performed well-resolved numerical simulations of turbulent channel flows at geophysically relevant Reynolds numbers. The numerical data were sub-sampled and processed using both the inertial and structure function methods. However, the dissipation was computed a priori directly from the simulation data, allowing us to evaluate the effects of anisotropy on the Kolmogorov constants. Our results suggest that using the canonical Kolmogorov constants, within the oceanic buffer layer, may result in significant errors (50% or more) in measurement of turbulent dissipation from the inertial and structure function methods. Moreover, the convection velocity, required to convert frequency spectra into wave number spectra, was found to be larger than the commonly used local mean velocity by approximately a factor of 2 near the bed, due to advection-dominated sweep-type structures.