Notwithstanding the large number of studies on bedforms such as dunes and antidunes, performing quantitative predictions of bedform type and geometry remains an open problem. Here we present the results of laboratory experiments specifically designed to study how sediment supply and caliber may impact equilibrium bedform type and geometry in the upper regime. Experiments were performed in a sediment feed flume with flow rates varying between 5 l/s and 30 l/s, sand supply rates varying between 0.6 kg/min and 20 kg/min, and uniform and non-uniform sediment grain sizes with geometric mean diameter varying between 0.22 mm and 0.87 mm. The analysis of the experimental data and the comparison with datasets available in the literature revealed that the ratio of the volume transport of sediment to the volume transport of water Q_s/Q_w plays a prime control on the equilibrium bed configuration. The equilibrium bed configuration transitions from washed out dunes, to downstream migrating antidunes for Q_s/Q_w between 0.0003 and 0.0007. For values of Q_s/Q_w greater than those typical of the downstream migrating antidunes, the bedform wavelength increases. Equilibrium bed configuration for fine sands is characterized by upstream migrating antidunes or cyclic steps, and significant suspended sand. At these high values of Q_s/Q_w the equilibrium bed for coarse sands is plane with bedload transport in sheet flow mode. Standing waves form at the transition between downstream migrating antidunes and bed configurations with upstream migrating bedforms or bedload transport in sheet flow mode. 

Experimental Overview

Laboratory experiments were conducted in the Hydraulics Laboratory at the Department of Civil and Environmental Engineering, University of South Carolina. The main objective was to observe and quantify how bedform type and geometry are affected by sediment supply rate and grain size. Sediment supply and flow discharge were chosen with the intended goal to obtain equilibrium bed configurations that evolve from lower regime to upper regime bedforms, with upstream bedforms migrating in the upstream and downstream directions. Sediment geometric mean grain sizes ?? ranged between 0.22 mm and 0.87 mm, flow rates ?? varied between 5 l/s and 30 l/s, and sediment feed rates varied between 0.5 kg/min and 20 kg/min. Sediment sizes were chosen based on the range where sand waves occur and what material is locally available.

Experimental Setup

Experiments were conducted in a glass wall sediment feed flume. The flume is 13 m long, 0.5 m wide and 0.9 m deep. A custom sediment trap is placed 9 m downstream of the flume entrance and a tailgate controls the downstream water surface level. A calibrated orifice plate and a Dwyer series 490 wet-wet manometer were used to measure the flow rate from the head tank. To decrease the sediment supply needed for experiments and the occurrence of three-dimensional bedforms, the cross-section of the test reach was narrowed to 0.19 m with the use of marine plywood.  In the first 2 m of the flume, the cross-section is gradually narrowed to 0.19 m to obtain a 7 m long experimental test reach.

Experimental Procedure

All experiments started from a net-depositional or net-erosional (disequilibrium) condition and continued until the bed level averaged over a series of bedforms did not change in time (equilibrium). At equilibrium, suspended sediment concentration was measured at the downstream end of the test reach and the experiment terminated. Experiment 1-SS started with a 10 cm thick flat layer of sediment with ?? equal to 0.43mm. The equilibrium bed of one experiment was used as the initial condition for the next experimental run. For example, the initial deposit in experiments 2-SS and 5-SS, was the equilibrium bed of experiments 1-SS and 4-SS respectively. After experiment 3- SS, a ~3 cm layer of sediment mix with ?? equal to 0.34 mm was sprinkled over the existing deposit to perform experiments with a finer sediment grain size. This process was repeated once more before experiment 7-SS to use coarser sediment with ?? equal to 7 0.62 mm. The duration of an experiment varied between 45 minutes to two hours depending on the initial condition and the feed rate. As the sediment size decreased, the time it took to reach equilibrium conditions decreased.

On the glass wall and at the top of the flume there are vertical and horizontal rulers which indicate the distance from the flume entrance and the elevation above the flume bed. Bed and water surface elevations were measured at every 10 cm interval moving downstream with ruler readings.  The first measurement was recorded at 3.50 m from the flume entrance and the last measurement was taken at 8.90 m from the flume entrance. During the experiments, measured values were reported in a spreadsheet, plotted, and the slopes of the best fit lines were computed to estimate the bed slope and the water surface slope. When the bed and water surface slopes did not significantly change in time, the flow and the sediment transport were deemed to be at equilibrium.

Pictures and Videos

The folder entitled ‘Pictures’ contains pictures and videos from each respective experiment in .JPEG and .MOV files.  These files are unprocessed.

Suspended Sediment Concentration and Velocity Profiles Measurements

Raw suspended sediment concentration data is presented in EXCEL .xlsx files within the ‘Suspended_concentration’ folder.  This folder also contains raw data from velocity probes as ‘.xlsx’ files for supplemental information.   

Experimental Data

‘CumulativeRunTable’ Excel file .xlsx, contains all variables found in the laboratory setting and ‘GSD_Design Mix,’ contains the sieve analysis for all experiments, suspended sediment, and parent material used for the dataset.  ‘Engelund phase diagram’ Excel file .xlsx, contains the current dataset with respect to datasets performed in the same laboratory facility.