The dimensionless critical shear stress (
Estimates of the onset of sediment motion are integral for flood protection and river management but are often highly inaccurate. The critical shear stress (
- Award ID(s):
- 1734752
- NSF-PAR ID:
- 10449141
- Publisher / Repository:
- DOI PREFIX: 10.1029
- Date Published:
- Journal Name:
- Journal of Geophysical Research: Earth Surface
- Volume:
- 123
- Issue:
- 12
- ISSN:
- 2169-9003
- Page Range / eLocation ID:
- p. 3308-3322
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract τ *c) needed for the onset of sediment motion is important for a range of studies from river restoration projects to landscape evolution calculations. Many studies simply assume aτ *cvalue within the large range of scatter observed in gravel‐bedded rivers because direct field estimates are difficult to obtain. Informed choices of reach‐scaleτ *cvalues could instead be obtained from force balance calculations that include particle‐scale bed structure and flow conditions. Particle‐scale bed structure is also difficult to measure, precluding wide adoption of such force‐balanceτ *cvalues. Recent studies have demonstrated that bed grain size distributions (GSD) can be determined from detailed point clouds (e.g. using G3Point open‐source software). We build on these point cloud methods to introduce Pro+, software that estimates particle‐scale protrusion distributions andτ *cfor each grain size and for the entire bed using a force‐balance model. We validated G3Point and Pro+ using two laboratory flume experiments with different grain size distributions and bed topographies. Commonly used definitions of protrusion may not produce representativeτ *cdistributions, and Pro+ includes new protrusion definitions to better include flow and bed structure influences on particle mobility. The combined G3Point/Pro+ provided accurate grain size, protrusion andτ *cdistributions with simple GSD calibration. The largest source of error in protrusion andτ *cdistributions were from incorrect grain boundaries and grain locations in G3Point, and calibration of grain software beyond comparing GSD is likely needed. Pro+ can be coupled with grain identifying software and relatively easily obtainable data to provide informed estimates ofτ *c. These could replace arbitrary choices ofτ *cand potentially improve channel stability and sediment transport estimates. -
more » « less
Abstract Steep, boulder bed streams often contain sediment patches, which are areas of the bed with relatively well‐defined boundaries that are occupied by distinct grain size distributions (GSD). In sediment mixtures, the underlying GSD affects the critical Shields stress for a given grain size, which is commonly predicted using hiding functions. Hiding functions may vary with reach‐wide bed GSD, but the effect of local GSD on relative sediment mobility between sediment patches is poorly understood. We explore the effects of patch‐scale GSD on sediment mobility using tracer particles combined with local shear stresses to develop hiding functions for different patch classes within a steep stream. Hiding functions for all tested patch classes were similar, which indicates that the same hiding function can be used for different patches. However, the critical Shields stress for a given grain size generally decreased with lower patch median grain size (
D 50) suggesting that patches control the relative mobility of each size through both the underlying GSD and local shear stresses. The effects of the underlying GSD partly depend on grain protrusion, which we measured for all grain sizes present on each patch class. Protrusion was generally greater for larger grains regardless of patch class, but for a given grain size, protrusion was increased with smaller patchD 50. For a given grain size, higher protrusion results in greater applied fluid forces and reduced resisting forces to partly explain our lower critical Shields stresses in finer patches. Patches therefore can importantly modulate relative sediment mobility through bed structure and may need to be included in reach‐scale sediment transport and channel stability estimates. -
more » « less
Abstract To explore the causes of history‐dependent sediment transport in rivers, we use a 19‐year record of coarse sediment transport from a steep channel in Switzerland. We observe a strong dependence of the threshold for sediment motion (
τ c ) on the magnitude of previous flows for prior shear stresses ranging from 104 to 340 Pa, resulting in seasonally increasingτ c for 10 of 19 years. This stabilization occurs with and without measureable bedload transport, suggesting that small‐scale riverbed rearrangement increasesτ c . Following large transport events (>340 Pa), this history dependence is disrupted. Bedload tracers suggest that significant reorganization of the bed erases memory of previous flows. We suggest that the magnitude of past flows controls the organization of the bed, which then modifiesτ c , paralleling the evolution of granular media under shear. Our results support the use of a state function to better predict variability in bedload sediment transport rates. -
more » « less
Abstract Flow resistance in mountain streams is important for assessing flooding hazard and quantifying sediment transport and bedrock incision in upland landscapes. In such settings, flow resistance is sensitive to grain‐scale roughness, which has traditionally been characterized by particle size distributions derived from laborious point counts of streambed sediment. Developing a general framework for rapid quantification of resistance in mountain streams is still a challenge. Here we present a semi‐automated workflow that combines millimeter‐ to centimeter‐scale structure‐from‐motion (SfM) photogrammetry surveys of bed topography and computational fluid dynamics (CFD) simulations to better evaluate surface roughness and rapidly quantify flow resistance in mountain streams. The workflow was applied to three field sites of gravel, cobble, and boulder‐bedded channels with a wide range of grain size, sorting, and shape. Large‐eddy simulations with body‐fitted meshes generated from SfM photogrammetry‐derived surfaces were performed to quantify flow resistance. The analysis of bed microtopography using a second‐order structure function identified three scaling regimes that corresponded to important roughness length scales and surface complexity contributing to flow resistance. The standard deviation
σ z of detrended streambed elevation normalized by water depth, as a proxy for the vertical roughness length scale, emerges as the primary control on flow resistance and is furthermore tied to the characteristic length scale of rough surface‐generated vortices. Horizontal length scales and surface complexity are secondary controls on flow resistance. A new resistance predictor linking water depth and vertical roughness scale, i.e.H /σ z , is proposed based on the comparison betweenσ z and the characteristic length scale of vortex shedding. In addition, representing streambeds using digital elevation models (DEM) is appropriate for well‐sorted streambeds, but not for poorly sorted ones under shallow and medium flow depth conditions due to the missing local overhanging features captured by fully 3D meshes which modulate local pressure gradient and thus bulk flow separation and pressure distribution. An appraisal of the mesh resolution effect on flow resistance shows that the SfM photogrammetry data resolution and the optimal CFD mesh size should be about 1/7 to 1/14 of the standard deviation of bed elevation. © 2019 John Wiley & Sons, Ltd. -
more » « less
Abstract The impacts of aquatic vegetation on bed load transport rate and bedform characteristics were quantified using flume measurements with model emergent vegetation. First, a model for predicting the turbulent kinetic energy,
k t , in vegetated channels from channel average velocityU and vegetation volume fractionϕ was validated for mobile sediment beds. Second, using data from several studies, the predictedk t was shown to be a good predictor of bed load transport rate,Q s , allowingQ s to be predicted fromU andϕ for vegetated channels. The control ofQ s byk t was explained by statistics of individual grain motion recorded by a camera, which showed that the number of sediment grains in motion per bed area was correlated withk t . Third, ripples were observed and characterized in channels with and without model vegetation. For low vegetation solid volume fraction (ϕ ≤ 0.012), the ripple wavelength was constrained by stem spacing. However, at higher vegetation solid volume fraction (ϕ =0.025), distinct ripples were not observed, suggesting a transition to sheet flow, which is sediment transport over a plane bed without the formation of bedforms. The fraction of the bed load flux carried by migrating ripples decreased with increasingϕ , again suggesting that vegetation facilitated the formation of sheet flow.
