|
home
people
research
publications
teaching
movies
|
 |
|
Research Projects
Understanding the impact of bedforms on surface coherent structures
Postdoc: Mike Barad
This work investigates the connection of free-surface flow signatures to bedforms in shallow,
slow moving and possibly tidal waters, and explores methods to extract bedform shape from
free-surface observations. Our focus is on field-scale flows with 2.5-5 m depths, where the
bedforms are less than 1m in amplitude, and the Froude number is on the order of 0.1.
Large-eddy simulations were conducted using a curvilinear-grid, finite-volume, incompress-
ible Navier-Stokes solver.
For the first part of this effort the bedform shape and amplitude are varied, and the
sensitivity of the results to mesh spacing is tested. Instantaneous free-surface signatures
of horizontal velocities and vertical vorticity show minimal dependence on the bedforms.
However, several time-averaged variables at the free surface have a strong dependence on
bedform, including (in order of importance): pressure, streamwise velocity, and root-mean-
square vertical velocity fluctuation. Bedform spectra are transmitted to the free surface,
with time-averaged signals having the best transmission. Upwelling and especially vorticity
magnitude patches can be used to estimate bedform roughness, though the connections are
weak. Coherent velocity structures are persistent for at least a downstream distance of
eight flow depths, and this persistence is modulated by bedforms. Time-averaged pressure
correlations with local depth are not significantly altered during transition from laminar
to fully-developed turbulent flow, while time-averaged streamwise velocity correlations are.
This reduction in correlation during transition to turbulence implies that highly turbulent
flows actually degrade the detectability of bedforms, particularly because the background
turbulent field associated with a flat-bed channel flow overwhelms signatures that may result
from bedforms. Figure 1 presents a visualization of flow past a 3D Gaussian bump.
Because bedform identification is more plausible before transition to turbulence, it is likely
that bedforms can be inferred from free-surface signatures in transitioning flows, such as tidal
rivers. For the second part of this effort, our focus is on visualizing and understanding co-
herent flow structures generated in a 200 by 100 m river reach using observed high-resolution
bathymetry, where the flow depths are on the order of 5 m, the bedforms are less than 1m in
amplitude, and the Froude number is on the order of 0.1. Results from parallel simulations
show the development of coherent structures in rivers that are strongly affected by the tides.
The tidal river flow transitions from a state of near-rest to an accelerating state, thereby
enabling possible bedform identification early in the accelerating phases of the tide when the
flow transitions from a near-laminar to a fully turbulent flow, see Figure 2.
Figure 1: Simulation of flow past a centered 3D Gaussian bump. Isosurfaces are of u/U =
1, vertical wall slices are colored by the streamwise velocity, and the clipped horizontal
pseudocolor plot is of the vertical velocity for the control volume closest to the surface (red
is upwelling, blue downwelling). Mean flow is from bottom right to upper left, parallel to
the x-axis, as indicated by the arrow.
Figure 2: Simulation of a dye release for a river scale domain, with idealized 4-meter wavelength, uniform bedforms (simulations using multibeam bathymetry are in process). The
point dye source is in the lower left. Flow is from the lower left to the upper right, and
approximates the start of an ebb tide. Semi-opaque isosurfaces are presented of the loga-
rithm of dye concentration, along with the dye surface signature. Surface signature colors
are shown in green (low concentration) and red (high concentration). Time series of dye
evolution at 50-second intervals from 300 seconds past slack.
This project is funded by a MURI grant from the ONR
Coastal Geosciences and Physical Oceanography programs.
For more information please visit Mike's web page.
Related publications:
M.F. Barad and O.B. Fringer, 2009, "Identifying bedforms from free-surface observations: a numerical study", Submitted to Physics of Fluids.
M.F. Barad and O.B. Fringer, 2009, "Simulation of coherent flow structures in a river. In preparation.
main
| next
|
