Mean flow and turbulence scaling in a tidal boundary layer

Calculations based on data from 10 triplets of orthogonally oriented ducted impeller current meters vertically spaced near the bottom of an estuary in Puget Sound show that: (1) 10 min averaging reduces the uncertainties in the value of the various means to±3.6% for the velocity,± 8.8% for the friction velocity,±50% for the extrapolated zero-velocity height (roughness length), and±40% for the Reynolds stress (95% confidence); (2) because of the magnitude of the lowfrequency (tidal) variability, lengthening of the averaging period does not reduce the uncertainties; and (3) a quasi-stationary model suffices to relate the Reynolds stress to the mean velocity profile within the above uncertainties. Including the dynamic effects of temporal variability in the model does not produce a measurably better prediction.