| Summary: | The purpose of this study is to gain some first insights into the role of burst-like turbulent motions in sediment suspension over a sandy channel bed, during typical conditions of strong sediment transport with active bedforms. The focus is the suspension mechanism that maintains sizeable sediment concentrations away from the bed, where much of the downstream transport occurs, rather than entrainment at the sediment boundary itself.
Flow components downstream and normal to the mean boundary, along with the output of an optical suspended sediment sensor, were monitored 1 m above the bed. The main study data were collected in a 10 m deep channel of the Fraser River near Mission, British Columbia, Canada. Velocities averaged 1.4 m/s at the surface and 0.9 m/s at the sensors, where mean suspended sediment concentrations were 500 mg/l; decimetre height small dunes on the backs of larger, metre amplitude dunes covered the channel bed. Many hours of data were recorded at 5 Hz, allowing multi-second scale turbulent motions as well as multi-minute oscillations to be resolved in both the velocity and turbidity records.
Burst-like "ejection and inrush" motions were identified, producing a high degree of intermittency in momentum exchange: 80% of the mean Reynolds stress at the 1 m level is produced during 12% of the record duration. The burst recurrence period appears to be significantly greater than predicted by applying the conventional outer flow scaling in this environment. It is hypothesised that the non-uniform shear and pressure gradient conditions over the various scales of bedforms on the river floor may somehow affect mean burst periodicity, modifying the recurrence scaling developed over flat boundaries. The determination of a burst recurrence timescale from one-point data is inherently imprecise however and, as elsewhere, a continuous variation of return periods with relative magnitude of extreme (u'v') events is observed.
The optical turbidity (OBS) time series reveals that these intermittent burst-like motions are, as expected, very important in vertically mixing sediments across the 1 m level in the flow; for example violent ejections, occurring only 1% of the time and contributing some 10% to mean turbulent momentum flux, appear to account for 6% of the total vertical sediment flux. The statistical association between the momentum and sediment mixing efficiencies of any ejection appears to be only moderately strong, however; very intense suspension can be associated with rather "weak" ejections (in terms of stress), and vice-versa. Differences between momentum and sediment mixing effects of a given ejection may partly be related to the "crossing trajectories effect"; sand grains continually fall out of the eddies that bear them, so the momentum and sediment "contents" of an eddy at 1 m off the bed are not perfectly linked.
Turbulent sediment suspension is, like momentum exchange, a highly intermittent process in itself. After selecting turbulent events only for suspension efficiency, the largest ones, occupying only 5% of the time, contribute approximately one half of the total vertical sediment flux. There is no indication that the conventional scaling of burst recurrence corresponds to the occurrence of any distinctive event level for suspension. Interestingly, burst-like turbulent motions are not the only flow oscillations contributing to suspension in the high flow conditions of the study. Multi-minute period flow perturbations at 1 m off the bed significantly assist burst-scale turbulent motions in driving the upward sediment mixing.
In summary, turbulent mixing of both momentum and sediment at 1 m over a typical sandy river bed is dominated by intermittent, intense "burst-like" events. However, the extrapolation of intermittent "bursting" concepts and structural constants from small-scale laboratory flows to the larger fluvial environment may be misleading. === Arts, Faculty of === Geography, Department of === Graduate
|
|---|
