Effect of the particle to fluid density ratio on bedform development: An application of PTV
The particle to fluid density ratio plays a key role in sediment transport and strongly governs the relative
importance of the transport mode. In aeolian systems, this ratio is three orders of magnitude larger than for the
transport of sedimentary particles in water, such that saltation is the dominant mode for diameters (250
microns) commonly found in ripples and dunes. The partitioning of fluid momentum to saltators, and therefore
to the surface upon impact, is extremely important to the entrainment of sediment, the maintenance of
transport, and the scaling of aeolian bedforms. This paper demonstrates the use of Particle Tracking
Velocimetry in measuring the partitioning of momentum associated with particle collisions on beds of quartz
sand (2630 kg m-3) typical of aeolian dunes, and acrylic particles (1210 kg m-3) similar to blowing snow (920
kg m-3). The experiments were carried out in the boundary layer wind tunnel at Trent University on full beds that
were 13.8 m in length and 0.71 m in width. In the majority of experiments, the wind speeds were either at or
just above the threshold for saltation so that we could distinguish discrete particle trajectories. Surface ripples
formed in the majority of experiments and passed through the camera's field of view so
that the height, length and rate of migration could be measured in relation to the distributions of particle impact
speed and angle, as well as those for the number, speed and angle of the particles ejected. Although similar
in height, the ripples comprised of acrylic particles were 2 to 4 times longer, much more asymmetric, and
migrated significantly faster than those in sand. The particle impact and ejection speeds were very similar,
although the sand particles approached and left the bed at substantially larger angles than observed for the
lighter acrylic particles of similar diameter. In a separate experiment, glass beads were flung onto each bed
material at 4 ms-1 in still air. It was discovered that 90 per cent of the impact energy was lost to the acrylic bed,
as compared to 78 per cent for the sand bed. This evidence suggests that at smaller density ratios than
investigated here, ballistic ripples likely cannot be maintained in air.
Challenges in Measuring and Predicting Medium Term (Weeks to Annual) Aeolian Sediment Transport in Beach-Dune Systems
Coastal dune budgets depend on sediment input by wind from the beach. Calculation of aeolian transport is thus a primary factor to understand coastal dune evolution and beach-dune coupled dynamics. However, measuring aeolian sediment transport in coastal areas presents fundamental technical and conceptual limitations that make numerical modeling difficult. Wind tunnel experiments isolate and reduce the number of variables to study, which is a necessary procedure to clearly manifest mechanistic relationships between cause and effect. But even with refinement and inclusion of new variables, traditional sediment transport formulas derived from wind tunnel experiments do not usually work well in natural areas. Short-term experiments may include precise instrumentation to obtain high frequency, detail time series of variables involved in aeolian transport, but inferring information at larger scales is problematic without knowledge of the timing and magnitude of particular transport events. There are two primary problems in attempting to predict sediment inputs to coastal dunes over periods of weeks, months or years: 1) to determine an appropriate set of predictive equations that incorporate complexities such as surface moisture content, beach width and the presence of vegetation; and 2) to provide quantitative data on these variables for input into the model at this time scale. Remote sensing techniques and the use of GIS software open the possibility to monitor key parameters regulating sediment transport dynamics at high spatial and temporal resolution over time scales beyond short-term experiments. These were applied at Greenwich Dunes, Prince Edward Island National Park (Canada), in an attempt to measure factors affecting aeolian sediment input to the foredune at a medium scale. Three digital cameras covering different sections of the beach and foredune provide time series on shoreline position, fetch distances, vegetation cover, ice/snow presence, or superficial moisture content. The rectification of oblique images to UTM maps allows to keep the spatial variability of these factors, and thus to perform detailed analysis on their complex evolution. Auxiliary instrumentation such as anemometers, safires, or erosion-deposition pins completes the basic set up. Data is processed using ArcGIS 9.2 and PCI Geomatica 9.1, and managed by an ArcCatalog Geodatabase. The coupling of new technologies (digital imagery) with traditional instrumentation (e.g. anemometers), and the extensive GIS capabilities both in the spatial and temporal domain, permits a new set of questions in aeolian coastal research. The overall goal is to obtain information on what is the frequency and magnitude of transport events at the beach or what are the key parameters regulating them. Challenges remain in improving methodologies to measure sediment transport rates. Ironically enough, we are able to obtain high quality time series on the factors affecting aeolian transport at the beach, but actual transport rates are measured with rather rudimentary techniques or instrumentation not adapted to meso-scale monitoring. This information is needed to test new approaches in modeling and understanding aeolian sediment input from the beach to the foredunes.
Using 3-D Laser Imaging to Track Topographic Changes on the Slipface of an Avalanching Aeolian Dune
Sediment dynamics on the lee slope of an avalanching aeolian dune were studied using changing topography
as an indicator of spatially heterogeneous sediment transport rates. Experiments were conducted in a Dune
Simulation Wind Tunnel, housing a 1:1 scaled transverse dune (1.2 m height, 1 m width, and 9.3 m length) of
medium sand (d̅=2.0φ).
Sediment transport on the lee slope of an aeolian dune may occur through settling of grains transported
across the dune's brink, reptation of grains on the slope, or avalanching of a portion of the lee slope. This
study measured settling flux rates (QS) with a passive array of traps on the dune surface. Reptation
transport rates (QR) were observed through passive traps on the surface. 3-D scanning of the surface
topography was conducted using a Konica-Minolta Vivid 9i laser scanner (vertical accuracy 0.096 mm,
horizontal resolution <1 mm). Sets of three digital elevation maps were obtained for each run, allowing for
determination of the rate of change of height (dzxy / dt) with and without avalanching. These observations
allow the utilization of a spatially integrated mass balance equation,
∬ ρ dzxy/dt- QS dx dy = ∬ dQR/dx + dQA/dx dx dy,
where ρ is the bulk density of the sediment, to describe the sediment transport required for the dune. Comparison of the localized settling and reptation rates reveals areas where settling rates are on par with, or even lower than reptation rates. This indicates that reptation is a significant form of transport on the slipface of an aeolian dune. It is expected that by redistributing sediment on the slipface, reptation alters the topography and changes the local slope. Altering the location of the maximum slope is thought to modify the initiation point of the following avalanche, in turn changing the volume of the material to be avalanched and thus the frequency-magnitude relationship.
Uncertainty in spatial distributions of ADCP apparent bed load velocity
Apparent bed load velocity (v) is the bias in Acoustic Doppler Current Profiler (ADCP) bottom tracking (Doppler sonar) due to particle motion near the bed surface. Apparent bed load velocity provides a measure of relative bed load magnitude, and can be obtained by comparing boat velocity from bottom tracking and high precision Global Positioning System (GPS) data. Spatially intensive asynoptic four-beam ADCP surveys were performed in a 6 km long, 500 m wide reach of wandering gravel-bed Fraser River. Spring freshet surveys were performed in 2006 near the threshold of bed particle motion, and in 2007 when flow discharge exceeded bankfull conditions and most of the gravel bed surface should have been mobile. Transects were spaced an average of 130 m apart in 2007, and 110 m in 2006, less than a quarter of the average channel width. The observed mean v was the same for both surveys (0.055 m/s), but higher maximum and standard deviation of v (σv) were observed in the higher flow conditions of 2007 [ σv = 0.128 m/s (2007); σv = 0.094 m/s (2006)]. Spatial distributions of v were mapped for each survey, using kriging to interpolate the irregular data onto a uniform grid. High bed load transport was observed along the entire channel thalweg during the high flow in 2007. In the lower flow of 2006, maximum transport was observed adjacent to a rapidly eroding bank downstream of a channel confluence, where it is possible that transport along the river bed of relatively fine material from the channel bank dominated the v signal. Uncertainty was assessed using a measurement variance model, which combined uncertainty due to ADCP single-ping error, GPS error, and estimated real temporal variability of bed load. The measurement variance model produced results similar to observed local measurement variance. The uncertainty in estimated spatial distributions of v was evaluated using both the measurement variance model and interpolation errors as indicated by kriging standard deviations. It appears that uncertainty in the estimated bed load spatial distribution was dominated by real bed load temporal variability in the relatively low bed load transport conditions of 2006, and by interpolation error in the high bed load transport conditions and wider transect spacing of 2007. In either case, reduction of uncertainty will require more intensive surveying.
Monitoring Large-Scale Sediment Transport Dynamics with Multibeam Sonar
Multibeam Echo-Sounder systems have developed rapidly over recent decades and are routinely deployed to
provide high-resolution bathymetric information in and range of environments. Modern data handling and
storage technologies now facilitate the logging of the raw acoustic back-scatter information that was previously
discarded by these systems. This paper describes methodologies that exploit this logging capability to quantify
both the concentration and dynamics of suspended sediment within the water column. This development
provides a multi-purpose tool for the holistic surveying of sediment transport dynamics by imaging suspended
sediment concentration, the associated flows and providing concurrent high-resolution bathymetry.
Results obtained a RESON 7125 MBES are presented from both well constrained dock-side testing and full
field deployment over dune bedforms in the Mississippi. The capacity of the system to image suspended
sediment structures is demonstrated and a novel methodology for estimating 2D flow velocities, based on
frame cross-correlation methods, is introduced.
The results demonstrate the capability of MBES systems to successfully map spatial and temporal variations
in suspended sediment concentration throughout a 2D swath and application of the velocity estimation
algorithms allow real-time holistic monitoring of turbulent flow processes and suspended sediment fluxes at a
scale previously unrealisable. Turbulent flow over a natural dune bedform on the Mississippi is used to
highlight the process information provided and the insights that can be gleaned for this technical development.
Development of the use of in-situ magnets for detection of bed load movement through magnetic induction.
Over the last 100 years, a large number of samplers and bed load measuring devices have been developed in order to estimate sediment transport in rivers. In spite of this work, the geomorphic and engineering communities do not have a reliable method to estimate sediment transport. To better model and understand river dynamics there is an urgent need for a method that can yield detailed estimates of sediment transport in rivers. Available techniques suffer from insufficient temporal and spatial resolution to capture the variability inherent in bed load movement. Data collection from large events can also be dangerous and suffers from problems such as over-filling of pit traps. The use of in-situ magnetic detection devices shows promise as a method that may be able to overcome many of these limitations. These systems have seen limited development by several researchers, the most advanced deployment by Tunnicliffe in 2000 at O'Ne-ell Creek. The sensors work by inducing a magnetic dipole in naturally magnetic stones via magnets installed in the bed of the channel. These stones then pass over a coil of wire, inducing a small voltage, which is recorded. An analysis of the most recent system showed that the sensor had several limitations: the torrid shaped magnet caused double peaks and variable magnetic strength near the face of the sensor and the spacing between the sensors allowed particles to pass over with out detection. To address these issues the new system features a continuous bar magnet, polarized through the thickness, to provide a continuous, uniform field. The previous sensor also produced a narrow range of response for a large range in particle size. This has been improved by using a stronger magnet and coils with a higher induction. Improvements in data storage and computing power since the installation at O'Ne-ell Creek will allow for a higher density of sensor across the channel and higher sampling frequencies. Laboratory experiments have shown that the new design performs significantly better than its predecessor. This system is being deployed in the field at East Creek, an established research stream, where the system will be calibrated using existing pit traps.