Figure Legends:
Figure 1 . (a) Poyang Lake is located at the south bank of Yangtze River. It receives inflows from five tributary rivers (i.e., Xiushui, Ganjiang, Fuhe, Xinjiang, and Raohe) and discharges into Yangtze River at Hukou. The study area (marked as black square) was located at the southeast of Poyang Lake. (b) Two sites, named A and B, were chosen at floodplains formed in Ganjiang River to measure the flow velocity. (c) and (d) Positions of all measurement cases in sites A and B. Black arrows show the flow direction.
Figure 2 . (a) Water level variation (in days) in Duchang Hydrological Station in the year of 2015. Velocity measurements were conducted from August 18 to August 24, which was marked as the red square in the figure. (b) Wind speed (maximum in hours) measured by Poyang Lake Wetland Observation Station, Chinese Academy of Sciences, located in Xingzi from Aug. 18 to 24. Blue arrows denote the direction toward which the wind is blowing, with northward up and eastward to the right.
Figure 3 . (a) Power spectral density (PSD) of the instantaneous velocityuu measured near the water surface (z = 55 cm with z = 0 representing the bottom) of case A4. Grey circles show the domain of wave signal with its boundaries labeled asfl and fh . Outside the wave domain (marked as black line) is the spectra of turbulence. Red dashed line presents the best linear fit for the amplitude within the frequency range of fL ~fl and fh~ fH . (b) Time series of the instantaneous velocities demonstrating the decomposition of the original velocity (black line) into time-averaged (green line), wave (red line), and turbulent (blue line) velocities using the method of spectral decomposition.
Figure 4 . Vertical distributions of the time-averaged velocity (U ), wave energy percentage (Pw ), wave orbital velocity (Uw ), and turbulent kinetic energy (TKE ) for the bare-bed case S0. Dashed line in (b) presentsPw = 10%. The red and blue dashed curves in (c) show the predictions of linear wave theory using eqs. (6) and (7), respectively.
Figure 5. Vertical distributions of the time-averaged horizontal velocity,Uhoriz , for cases with vegetation. The upper five plots and lower five plots present velocity profiles measured in sites A and B, respectively. Blue and green dashed lines show the positions of water surface and the top of vegetation, respectively. The absence of green dashed lines in some plots indicated that the vegetation was under emergent conditions. Sketch of the vegetation was also present with the velocity profile for each case.
Figure 6 . Vertical distributions of wave velocity (Uw ) for vegetated cases. The horizontal (Uw_horiz ) and vertical (Uw_vert ) components were considered separately and marked as grey circles and blue triangles, respectively, in the figure. Measurement points with Pw_u < 10% in the lower part of water column were excluded from each profile. For each plot, the red and blue dashed curves show the prediction ofUw_horizand Uw_vert by eqs. (6) and (7) with the wave amplitude estimated by fitting eq. (6) to measuredUw_horiz at the highest three measurement points. Predictions of Uw_horiz andUw_vert were overlap for cases A5 (e), B4 (i), and B5 (j). Please note that the scale of x -axis was different between cases.
Figure 7 . Vertical distribution of the turbulent kinetic energy (TKE ) for vegetated cases. Upper five plots and lower five plots presentTKE profiles measured in sites A and B, respectively. TheTKE for the measurement points with Pw_u< 10% near the bed was calculated by eq. (8) using original instantaneous velocity. Blue and green dashed lines show the positions of water surface and the top of vegetation, respectively. Sketch of the vegetation for each case was present with the TKE profile.