Quote from: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2008JC004736

  From September 2002 to May 2003, fifteen bottom-mounted, acoustic Doppler current profilers measured currents of the northern Adriatic basin. Tidal fluctuations at all seven of the major Adriatic frequencies were synthesized from a response tidal analysis of these measurements. Most observed tidal current ellipses were nearly reversing, but near the bottom, tidal current ellipses all shortened and broadened, semidiurnal currents led upper water column currents, and diurnal tidal current ellipse orientations rotated counterclockwise toward the bottom. Theoretical solutions for a tidally forced, bottom Ekman layer with vertical eddy viscosity of the form Az = βz + k were least squares fit to the observations. Average values were β = 3 · 10−4 m/s and k = 5 · 10−4 m2/s. The value of k was important in matching tidal orientation and phase changes, and a nonzero β was important in matching tidal amplitude changes. The Navy Coastal Ocean Model (NCOM) and the Quoddy model were also compared to the observations. The average RMS errors for the bottom Ekman layer were 0.22 cm/s for the best fit theory, 0.35 cm/s for NCOM, and 0.36 cm/s for Quoddy. Az structures from NCOM and Quoddy show that time variation in Az is relatively unimportant for Adriatic tides. The bottom shear stresses from theory were larger in magnitude than those from the bottom drag formulations in NCOM and Quoddy.