H41G-01 08:00h
Channel Geometry of Mountain Rivers Within the Debris-Flow Process Domain of the Idaho Batholith
Channel width {\it w} and depth {\it d} exhibit systematic downstream changes with increasing drainage area {\it A} and hydraulic discharge ({\it w = cA$^{b}$, d = eA$^{f}$}). These relationships vary regionally as a function of physiography (climate, topography, and geology), but exhibit grossly similar trends across many orders of magnitude in drainage area, and across different channel types (e.g., alluvial, bedrock and, in some instances, tidal channels). The consistency of these relationships likely reflects the fact that each of these channel types is formed by hydraulic forces of flowing water. However, headwater rivers in mountain basins experience a combination of fluvial discharge and periodic passage of debris flows and hyperconcentrated flows. Consequently, channel geometry of headwater rivers may reflect a competition between fluvial and mass-wasting processes. We conducted field surveys of channel characteristics in headwater basins of the Idaho Batholith to examine this issue and to determine if channel geometry can be used to infer either the recency or dominance of these competing geomorphic processes (fluvial vs. mass wasting). Mass-wasting events are common occurrences in the study area, resulting from summer thunderstorms and winter rain-on-snow events. Moreover, the magnitude and frequency of mass wasting increases in the study area following wildfires that alter vegetation, soil permeability and consequent basin hydrology. Results indicate that headwater channels recently impacted by debris flows tend to be wider, but have width-drainage area exponents similar to lowland alluvial channels ({\it b} = 0.5).
H41G-02 08:15h
Debris Dams, Sediment Impoundment, and the Relief of Headwater Streams
In forested, mountain landscapes where debris flows are common, valley-spanning debris dams formed by debris-flow deposition are a common feature of headwater valleys. In this paper, the effects of debris dams on the evolution of headwater valley profiles over geologic time are examined. The main channels of three small (approximately 2 sq. km) watersheds in the Oregon Coast Range, USA, were surveyed, with special care given to capturing the relief of each significant step and noting the step-forming agent, whether wood, boulders, a complex mixture of the two, or bedrock (although the last is not treated herein). Channel and valley widths were also measured, and surface bed material measurements (pebble counts) were taken at one of the sites. The amount of relief in wood and/or boulder steps is highly variable within and among sites, reaching a maximum of 58% at one location. The surveyed steps comprise 9.8%, 19%, and 6.4% of total basin relief in the three sites, respectively. A model of valley profile evolution is derived that accounts for the facts that (a) only a fraction of the valley width is occupied by the channel at any given time and (b) bedrock is often shielded from erosion by sediment impounded behind debris dams. Under the assumption of steady-state incision, the equation for incision rate is solved for valley gradient as a function of contributing area, parameter values are estimated or supplied by field data, and the results are compared to plots of gradient vs. contributing area for the field sites. These comparisons suggest--but do not demonstrate--a strong effect of network structure, which varies significantly among the sites, on profile shape and relief because of the different susceptibilities to debris-flow deposition and, therefore, debris-dam formation. Specifically, in a reach where the ratio of step height to spacing drops abruptly, stream gradient also falls more steeply in a way consistent with model predictions. Finally, for one site, steady-state model profiles with and without the effects of steps are constructed. For these model profiles, which only extend as far as the surveyed profile at that site, 55% of the profile relief is due to steps and the concomitant shielding of the bed by impounded sediment. These results suggest that a significant fraction of the relief of such forested, mountain landscapes is due to the effects of relatively immobile wood and boulders deposited in the valleys by debris flows.
H41G-03 08:30h
Observations on Alluvial Fans with Relevance to Recent Sediment Transport
Steep (slope $>$ 0.01) alluvial fans are widely written about, but there are almost no quantitative field studies of the fluvial channels moving water and sediment across them. For this reason, it is difficult to apply existing mechanistic models of fan evolution (e.g., Parker et al., 1998). It follows that we have little ability to quantitatively predict the effects of regional changes in water flux, sediment supply, or vegetation from ongoing anthropogenic and climatic changes on fans, particularly in arid lands. We have begun a field program to quantify fluvial sediment transport across alluvial fans by measuring the hydraulic geometry and bed texture of channels. The goal of the measurements is to parameterize sediment transport in a way that predicts the observed pattern of slope reduction, often from ~0.07 at fan heads to ~0.02 at downfan margins. In the Mojave Desert of California, we find that alluvial fan channel bankfull depths are largely 0.4-1.0 m at fan heads, decreasing to 0.1-0.2 m at distal fan margins. Contrary to many previous studies, we find that median gravel diameter does not change systematically along the upper 60-80% of active fan channels, and thus downstream gravel fining cannot explain most of the observed channel slope reduction. However, as slope declines, surface sand cover increases systematically downfan from values of $<$20% above fan heads to distal fan values in excess of 70%. Plots of reach gradient versus sand content from these channels agree with experimental flume data from Ikeda and Iseya (1988) with similar hydraulic geometries, suggesting a general relation where similar loads can be transported at lower slopes because of the role that increased fines play in reducing the threshold for sediment transport. However, our attempts to model this role using reduced nondimensional critical shear stress values at higher sand concentrations (Wilcock & Crowe, 2003) with conventional excess shear stress sediment transport formulas result in underpredictions of the observed downfan slope decline. Although the relation between reach slope and sand cover suggests a strong role for sand cover in reducing fan slope, calculations indicate a strong component of bedload deposition is also necessary to match observed slope patterns. These uncertainties about the relative roles of sediment load and grainsize illustrate that much remains to be understood about alluvial fans before we can predict even first-order responses to climatic or anthropogenic change.
H41G-04 08:45h
Relations Between Sediment Transport and Storage During Aggradation and Degradation in a Gravel Bed River
Sediment routing models typically transfer sediment through a channel network using fixed values of transport capacity. However, transport capacity is dynamic at the sediment pulse time scale. The purpose of this research is to develop relations between transport capacity and volume of sediment in storage (transport-storage or T-S relation) that would replace fixed values of transport capacity applied to channel segments within sediment routing models. The form and stability of T-S relations were studied in a series of flume experiments and compared to a T-S relation observed in a single reach (1 km) in a steep (3%) gravel-bed river in northern California. Flume experiments were designed to simulate cycles of aggradation and degradation, similar to those documented at the field site. T-S relations for full cycles of aggradation and degradation showed hysteresis. For a given storage volume, higher output rates were observed during channel degradation than aggradation due to opposite longitudinal trends in transport and changes in bed texture that increased bed mobility during early stages of degradation before armoring intensified. The stability of T-S relation was examined by conducting flume experiments with variations in sediment feed, initial channel armor development, adjustment to sediment supply before and after equilibrium was reached, and sequential episodes of aggradation and degradation. Results of these experiments showed that the T-S relation during aggradation is stable and is not sensitive to initial channel armor or equilibrium status. In contrast, the form of the T-S relation during degradation is sensitive to sediment feed rate during degradation and whether equilibrium was reached before sediment feed was reduced and degradation began. Transport rates during channel degradation were correlated with local stored sediment volume and a reach scale measure of the concentration of finer particles on the bed surface. Multiple aggradation-degradation cycles increased residual storage volume, but did not affect the slope of T-S relations. Results from these flume experiments were used to constrain the form of the T-S relation developed from limited field data. Collectively, this set of flume experiments provides insight into the form of a T-S relation under full aggradation-degradation cycles within a single reach, as well as the application of T-S relations to a drainage network.
H41G-05 09:00h
The Effect of Large Roughness Elements on Local Flow and Bedload Transport
Most mountainous drainage networks contain large roughness elements (boulders, large woody debris) that alter reach-scale flow and sediment transport. The obstacles create large spatial deviations in the flow shear, so bedload transport equations based on the total boundary shear stress do not apply in these channels. Thus it is difficult to determine, for a given roughness configuration, the effect of sediment supply on channel morphology. We have previously developed a transport equation that partitions the total shear stress between immobile bed elements and the finer, more mobile sediment. This approach improves predictions of sediment flux, but does not explain the mechanics of sediment transport around large obstacles. We investigated the mechanics of sediment transport in a flume set at a constant gradient and water discharge. Sand was transported over two beds: mobile sand grains and regular arrays of immobile spheres. A high-speed video camera, mounted above the flume, recorded the variation in transport rate around the spheres. Particle imaging velocimetry (PIV) was used to determine 2-D velocity fields (downstream and vertical) around the roughness elements. The large spherical particles generated a horseshoe vortex flow pattern with downward flow along the lower half of the front and side of the particle. Turbulence structures in the main flow (probably sweep events) intermittently interacted with this vortex pattern to produce high near-bed downstream velocities that generated the bulk of the transport. Preliminary results show that the downstream transport rate varied by an order of magnitude from immediately adjacent (123 grains/cm/s) to halfway between the immobile grains (14 grains/cm/s). This large variation in sediment flux and shear stress led to significant morphologic changes; scour holes and sand deposits developed adjacent to and midway between the immobile elements, respectively. Furthermore, the immobile grains increased local stresses and therefore caused sediment transport that did not occur on a bed of only sand. In contrast, previous steeper gradient (10%) experiments showed that large grains trap sediment that would otherwise be highly mobile. In these very shallow and steep flows, the total boundary shear stress was much greater than the critical stress of the mobile fraction. Thus, the effect of roughness elements on sediment transport and bed morphology may vary depending on their position in the channel network. We are currently testing our stress partitioning model and the correlation between local sediment flux and stress.
H41G-06 09:15h
Transported Sediment in Comparison with Channel Bed Material in Low-Order Alluvial Channels
Sediment transported in alluvial channels consists of mobilized channel bed material combined with that detached from channel banks and sediment contributed through overland flow. Unconsolidated bed material provides an almost limitless sediment source. Although the particle distributions of the bed material and the transported sediment are often assumed to be the same, few dataset are available to test this assumption. Detailed measurements of channel geometry, bed material, and spatial patterns of coarse particle distribution were made on a low-order channel within the USDA-ARS Walnut Gulch Experimental Watershed. Sediment transported during the 2002 monsoon season was collected with a traversing slot sediment sampler and a pit trap. Particles size distributions of channel bed sediment are compared with the particle distributions of sediment collected during flash-flows. Initial results indicate that the bed material contains less fine sediment than the transported load, and sediment particles larger than 4 mm make up as much as 15% of the total sediment load transported during the measured events.
H41G-07 09:30h
Experimental Bedrock Channel Incision: Scaling, Sculpture and Sediment Transport
Abrasion by sediment in turbulent flows often sculpts bedrock channels into dramatic forms; quantifying the feedbacks between fluid flow, sediment impacts, and channel morphology is needed to refine models of fluvial incision into bedrock. We present data from laboratory flume experiments funded by the National Center for Earth-Surface Dynamics and conducted at St. Anthony Falls Laboratory, University of Minnesota that show how the spatial and temporal distribution of erosion is strongly coupled to the evolving topography of the bed. These experiments focus on the high Froude number and tool-starved end of parameter space, where bed cover tends to be negligible. Independent variables include flume slope, water flux and sediment flux and size distribution. Sediment moves energetically as bedload, suspended load, or locally transitional between transport modes. Quantitative measurements of the evolving bed topography show that the synthetic brittle "bedrock" in the flume (cured sand-cement mixture) eroded to form narrow incised channels with tight scoops and potholes. The experimental erosional forms are similar in morphology, and sometimes in scale, to those observed in natural bedrock rivers in southeast Utah and other field settings. The experiments demonstrate that both the mean and distribution of measured erosion rates change as the bed topography evolves, even with constant water and sediment discharges. Even starting with a plane bed geometry, erosion and sediment transport very quickly become localized in interconnected topographic lows. Positive feedback develops between the evolving topography and the fluid velocity and sediment transport fields, resulting in the incision of an inner channel. Once formed, the erosion rate in the axis of the inner channel decreases as local bed shear stresses and fluid velocities are reduced by increasing wall drag, and sediment fluxes through the channel but causes less incision (no deposition). Decreasing the sediment flux (all else held equal) causes renewed incision, but of an even narrower inner channel; increasing the sediment flux leads to inner channel deposition. Where erosion is most vigorous, sediment generally moving as saltating bedload becomes locally suspended by upward-directed mean flow. For example, swirling clouds of "bedload" particles are continuously suspended by vortices developed within potholes such that the upward flux of particles out of the potholes balance the total sediment flux through the flume. Potholes spontaneously form where average bed slope and fluid velocities were highest, dramatically accelerating the local erosion rate. Our experimental potholes are smaller in scale but morphologically strikingly similar to many observed in the field, and include features such as corkscrew grooves down the outside walls and a protruding horn at the pothole center. More generally, abrasion becomes focused in places where the flow is spatially accelerated, such as in scoops and bends with high curvature. The knife-edge margins and spatial distribution of erosional forms indicate abrupt transitions in erosional efficiency that are tightly coupled to near-bed fluid flow patterns, which in turn are strongly influenced by the erosional forms themselves. Our experiments suggest that, in highly sculpted bedrock channels, naturally developed bed roughness presents a physical length scale that is important to controlling the interaction between sediment impacts and the bed, rather than a length scale based explicitly on sediment transport and average flow conditions such as the saltation hop length.
H41G-08 09:45h
Mechanics of Horseshoe Waterfalls
The interaction between flow, sediment transport, and channel morphology is very poorly known for step units in bedrock rivers. Unlike dams and weirs, natural steps have complex 3D morphologies. A detailed study of the fluid mechanics of horseshoe falls was performed using a scaled model with a 0.91-m vertical drop in a 2.75-m wide flume with flows up to 122 cfs. Five non-dimensional upstream energy levels with 3-5 non-dimensional tail depths were assessed for the resulting 3D water surface topography via digital elevation modeling, flow dynamics via digital videography, and overall energy dissipation via an energy and momentum conservation model. Regardless of tail depth, the horseshoe step was found to have 3 distinct zones beyond the brink: 1) a nappe whose degree of convergence depends on upstream energy input and brink configuration, 2) a convergence zone whose features vary strongly with upstream energy input, brink configuration, and tail depth, and 3) a downstream tailwater region whose dynamics primarily depend on tail depth. The centerline nappe profile and brink velocity were reasonably predicted using Rouse's jet trajectory equations when (H+P)/H$>$3. Peripheral profiles were not predictable using existing equations. For any arbitrary broad-crested step brink configuration, maximum energy dissipation was found to occur when no jump was present and downstream tail depth was exactly critical. DEM and process data including direct shear stress and sediment transport measurements from 4 natural horseshoe waterfalls in the Pacific Northwest will be presented for comparison against the flume study.