H54E-01 INVITED
Formation and Maintenance of Sand-Mud Transitions
The sand-mud transition (SMT) is a common boundary on continental shelves where the mean diameter of seabed sediments drops abruptly from ~ 100 μm to ~ 10 μm. This transition is important because sands and muds host different benthic communities, have different acoustic properties, and carry different loads of heavy metals and biogenic particles. The transition from sand to mud is also a basic facies change that can be used to infer properties of ancient environments of deposition recorded in sedimentary rocks. The SMT in general occurs because of decreasing shear stress on the seabed. In short, sands erode and deposit at higher shear stresses than muds. This simple notion, however, does not explain the abruptness of the transition, which often occurs over a depth range of only a few meters. Recent observations of in situ particle size distributions show that the size, settling velocity, and abundance of suspended flocs increase rapidly with decreasing shear stress. Because individual mud particles sink very slowly, deposition in flocs is the primary way of transferring mud from suspension to the seabed. The increases in floc abundance and settling velocity with decreasing shear stress therefore lead to formation of an SMT by causing a rapid increase in the depositional flux of mud. Observations of size sorting during controlled erosion of mixed grain size beds demonstrate that SMTs are accentuated by the cohesive effects of mud. If the mud fraction is low in sediment composed of both sand and mud, then muds will be winnowed from the sands at low shear stresses. If the mud fraction is high, then many grain sizes are bound together. Winnowing of muds from sands is reduced or eliminated. By this mechanism, sediment on the sand side of an SMT will be cleaned by subsequent resuspension, while sediment on the mud side will not, thereby retaining its poor sorting and small mean grain diameter. The processes described here need to be incorporated into sediment transport models. Floc settling velocity and abundance need to be defined in terms of boundary shear stress. The critical erosion shear stress of muds needs to be defined in terms of the percentage of fine sediment in the seabed. These changes to current models should allow them to reproduce the abruptness of sand-mud transitions.
H54E-02
A fine-scale turbidity record as a view of fine bed sediment supply, transport, and dynamics
Fine bed sediments in gravel-bedded rivers are detrimental for salmonid reproduction, ecosystem productivity, groundwater-surface water exchange, and streambank pumping operations. However, the quantity and grain size of fine bed sediments are generally unknown. Direct measurements are temporally and spatially sparse, valid for only a short length of time, and often lack volumetric, subsurface, and sediment quality analyses. California's Russian River is impaired for both turbidity and sedimentation of the bed by fines. Bed sedimentation has been relatively unquantified; we hypothesize that it is possible to extract information about the quantity and grain size of bed sediment from the extensive record of streamflow and turbidity data available in the basin. A unique database has been assembled by joining all US Geological Survey (USGS) daily and 15-minute monitoring data from the basin (22,000,000 data points) with USGS water quality field-sampling data, NOAA atmospheric data, and ancillary data collected by the USGS, the California Department of Fish and Game, the Sonoma and Mendocino County Water Agencies, and academic and private researchers. This database has been organized with a data cube, which allows for quick retrieval of information organized by different dimensions (e.g. by water year, frequency, site, etc.) Analyses made thus far have focused on six years of 15-minute turbidity and streamflow data collected at two gauging stations (drainage areas 900 and 3500 km2) on the main stem. Differences in the relationship between turbidity and suspended sediment concentration during different flow phases and the progression of turbidity/streamflow hysteresis loops over series of storms suggest that early in the initial rising limb, turbidity is largely controlled by local mobilization of fines in the bed. Farther into the discharge event, fine material loads are interpreted to become source-dependent (i.e. sediment mobilized from well upstream.) To build on our analysis, we have created additional databases of individual storms and turbidity spikes. The database/data cube structure permits both expansion of the study (e.g. into long-term trends of sediment storage and flux by inclusion of lower-frequency data collected since the 1960's), and sharper focus on the question of bed sediment (through incorporation of short, high-frequency turbidity datasets at additional sites and estimation of bed shear stress from ancillary data.) In general, a data cube built with high frequency flow and turbidity data from multiple years and gauging stations provides unique opportunities for testing models for fine sediment dynamics in gravel-bedded streams.
H54E-03
Variation and Distribution of Sediments in a Mixed Glacifluvial-Aeolian System in West Greenland
There is a clear association between the distribution of wind-blown sediments and the former extent of ice sheets and glaciers. Glacial erosion processes produce significant quantities of fine sediments that are washed out from beneath glaciers by meltwater. Once deposited and desiccated, aeolian processes may transport them across the landscape resulting in the formation of sand dunes and loess, and adding dust to the atmosphere. This research reports the use of digital imaging and laser sizing to obtain the grain size distribution and textural attributes of sand and dust in Sandflugtdalen, a valley adjacent to the West Greenland ice sheet. An initial assessment of the rates of sand and dust transport, made using semi-isokinetic directional sediment samplers, indicate that the flux of aeolian sediment comprises clays, silts and sand-sized particles. Digital imaging of the surficial sediment deposits provides a rapid means of sampling the large, spatially and temporally variable, proglacial valley. Sediments were initially photographed during June 2007 and then resampled after a 9-week interval. The grain size distribution and surface texture were computed using a calibrated autocorrelation method. It is estimated that individual particles may be resolved down to a size of 0.045 mm. The regions of aeolian entrainment, transport and deposition are directly linked to the development and distribution of sediments on the proglacial floodplain, which varies considerably in terms of surface roughness. On the floodplain close to the ice sheet, aeolian flux is controlled by sediment supply and lag formation and the total surface roughness is determined by the combination of grain-scale roughness and topography. Further down valley, recycling of sediments by aeolian and fluvial activity is significant and wind speed becomes an important controlling factor. Within the dunefields, surface roughness is principally determined by topography and vegetation. Close to the ice sheet, significant aeolian sediment fluxes occur during the summer, whenever freshly deposited and desiccated sediments are available. However, within the dunefields, the maximum flux occurs in early summer prior to the development of the annual vegetation cover.
H54E-04
Modeling the Role of Small Scale Physics in Sediment Transport From Grain Size to Grain Shape
In recent years work has focused on the detailed physics of sediment transport at or near the grain scale. Although computational resources often restrict the domain size, deterministic models for sediment motions can prove useful in improving our understanding of sediment dynamics. Using a discrete particle model (DPM), we have performed computer simulations that describe the collective and individual motions of sediment grains immersed in fluid in an effort to emulate the physics of the sea floor, at the fluid-sediment interface, in shallow water under forcing from waves and currents. Examples of our DPM (briefly described) are shown for research applications at a range scales from millimeters to meters involving fluid flow models from simple one- dimensional eddy viscosity up to three-dimensional direct numerical simulation. Based on hundreds of different simulations over the past decade, our findings have shown: how a parameterization of pressure gradients or equivalently fluid accelerations on particle motions under waves influences sand bar migration in the surf zone; how grain shape changes bulk bedload transport rates; how efforts to model sediment particle motions in the swash zone can yield insight toward models for shoreline erosion and accretion; how recently simulated bedload transport using bimodal size distributions has uncovered a new power law; how upcoming work focuses on simulating the role of grain size distributions in small-scale sand ripple dynamics. Good agreement is found between comparisons of model output for both bulk transport rates and time dependent concentration profiles with laboratory data. Likewise, parameterizations obtained from simulation results have demonstrated skill in hindcast applications to both field and laboratory measurements. Conclusions will discuss the future role of reductionism in sediment transport modeling.
H54E-05
Determining a minimum parameter set for one-dimensional modeling of sediment transport
One-dimensional modeling of sediment transport in natural rivers has been performed for over two decades. However, there is still no consensus on the formulation of the governing equations of sediment transport. In particular, various researches have proposed multiple formulations for the bed sorting processes. Also, various method to separately model suspended and bed load transport have been proposed, but combined modeling of suspended and bed load is also commonly used. We perform an analysis to determine the minimum set of calibration parameters that can be used to reliable predict sediment deposition and erosion with a one- dimensional model. Multiple simulations using various models of bed mixing methods are performed. The parameters of each model are calibrated to experimental results. We also evaluate the sensitivity of the model to each parameter. We determine that a simpler model of bed sorting can simulate the processes of sediment deposition and downstream fining. The model also requires a smaller number of calibration parameters. We also found the there is no need to separate bed load and suspended load in the context of a one-dimensional model. The bed load and suspended load can be simulated with the same governing transport equation with no loss of accuracy. http://www.usbr.gov/pmts/sediment
H54E-06
Experimental Study of Different Sized Tracer Particles in a Gravel-Bed Laboratory Flume
The prediction of particle movement in bedload transport is complicated by the fact that particles tend to segregate by size. Well-known examples in gravel-bed rivers include armouring, where the surface layer is coarser than the subsurface layer, and grain size patchiness, where relatively well-sorted gravel patches form along the river bed. The mechanics that drive these sorting effects have important implications for river restoration efforts and the management of river systems. We study how bedload transport movements vary according to the size of the sediment particles in a simple experimental system. We use a gravel bed sediment-feed flume to perform controlled experiments under conditions similar to the bank-full discharge in natural gravel-bed rivers. To minimize spatial and temporal variations that might otherwise obfuscate the results, we operate the flume under plane-bed equilibrium transport conditions, with well-sorted unimodal gravel as bed material. Tracer particles, colored according to their size and initial position in the bed, indicate how streamwise and vertical movements of particles depend on particle size. We found that even under plane-bed equilibrium transport conditions, particles very close in size have measurably different streamwise and vertical transport. A simple scaling relationship models the distribution function along the bed. A stochastic model we developed incorporating each phase of the bedload transport describes this behavior well. We discuss these results, as well as their implications for size sorting in alluvial systems.
H54E-07 INVITED
Downstream Sediment Sorting as a Fractionation Process
Downstream size segregation in net depositional systems can be thought of as a fractination process in which a well mixed, heterogeneous input is unmixed based on its relative mobility. Although we are accustomed to thinking of the segregation process as hydraulically driven and rather complex, we argue that at large time and length scales size segregation can be substantially simplified. The main controls are the downstream distribution of sediment extraction, which is typically controlled externally (e.g. by subsidence) and the size distribution of the sediment supply. Hydraulics plays a secondary role because of the tendency for river channels to self organize to a shape that maintains a limited range of dimensionless shear stress on the channel bed. The end result of this line of reasoning is a simple method for calculating downstream size segregation in depositional systems that is in good agreement with the limited data available. In terms of local dynamics, we introduce evidence that topographic roughness plays an important role. This is not explicitly incorporated in our analysis, and the best ways to characterize roughness for this purpose are yet to be determined. Finally, to estimate the importance of abrasion effects at large scales, we re-introduce a dimensionless parameter to describe the relative importance of abrasion, which sorts material by durability, and selective transport, which sorts by transportability.