H53L-01 INVITED
Persistence and Transience in Bed Surface Texture
The grain size distribution on the surface of stream-channel beds typically displays significant spatial structure. Variations occur at all scales, but once they reach mappable domains they are often referred to as patches or facies of common size distributions. Here we report results of various field and experimental investigations that seek to explain the origin and dynamics of this patchiness. Our observations demonstrate that patches can be persistent through significant flow and sediment transport events, and transient, in that they either migrate downstream or are simply swept away or buried. The persistent patches can be thought of as "fixed" while the moving patches are "free." It appears that fixed patches arise from the interaction of the spatial structure of boundary shear stress and bed topography. For example, in both sand and gravel-bed rivers there can be persistent patterns of coarsening in pools and fining in the downstream tails of bars. Persistence seems tied to stability of channel bed topography. Fixed patches can remain constant in grain-size distribution despite complete particle exchange and large transport rates of dissimilar particle sizes across them. Small, finer- particle patches in coarse-bedded streams can persist during modest sediment transport events. These fine patches, however, may grow, or shrink and disappear depending on sediment supply. These changes in extent have large effects on rates of bedload transport. Free, moving patches emerge from grain interactions, and create a self-propagating perturbation in local boundary shear stress and sediment transport fields. Modeling efforts are underway to explore the predictability of fixed, free and locally transient patches. http://eps.berkeley.edu/~bill/papers/PatchesRCEM2005Dietrich_134.pdf
H53L-02
The Relationship Between Sediment Properties and Sedimentation Patterns on a Macrotidal Gravel Beach over a Semi-lunar Tidal Cycle.
Detailed measurements of profile and sediment dynamics have been obtained from a macrotidal gravel barrier beach in southern England. Surface and sub-surface sediment samples, beach profiles, and disturbance depths were taken from the intertidal zone on consecutive low tides over semi-lunar tidal cycles, along with continuous wave and tide measurements. Results from two separate field surveys are presented, representing 26 and 24 consecutive low tides, respectively. A combination of Canonical Correlation Analysis (CCA) and Empirical Orthogonal Function (EOF) analysis was used to identify a number of consistent relationships in morphological and sedimentological variables not readily apparent using ordinary correlations. The disadvantage of such statistical models is that the relationships obtained cannot be expressed in physically meaningful units, which does limit its utility in physical-numerical modelling. However, the results reveal some interesting relationships between gravel beachface sedimentology and morphological change. For example, beachface morphology and sedimentology are more similar at a given spatial location over time than over space (cross-shore) at any individual time. Subsurface sedimentology over the depth of disturbance indicates that the beach step can be traced through the sediment characteristics. Indeed, the study suggests that gravel beachface sedimentology is 'slaved' to morphological change rather than vice-versa; and that the relationship becomes more evident as secondary morphological features develop on the beachface. The results imply that median sediment size and geometric sorting are suitable parameters for detecting such relationships. Strong hysteresis over space was present in the EOF modes associated with the most variance in the data sets, for both sediment size and sorting. Statistically significant relationships were found between the temporal modes of (absolute) size/sorting and net sedimentation associated with the largest variance in the non-decomposed respective data sets. Finally, significant relationships were found between a suite of measured hydrodynamic time-series and pairs of significantly correlated morpho-sedimentary eigenmodes. The techniques used were thus able to objectively demonstrate linear association between morphological and sedimentological change on a gravel beachface over a semi-lunar tidal cycle; and also that simultaneous changes in each could be linearly correlated to hydrodynamic forcing.
H53L-03
Channel widening due to urbanization and a major flood can alter bed particle organization and bed stability in an urban boulder-bed channel
The Anacostia River is a tributary of the Potomac River north of Washington D.C. that has become progressively more urbanized in the past 50 years. Bankfull discharge and bankfull width in the Anacostia have increased by 3- 4x in the past 50 years. Nearby watersheds of similar size and geology, but without significant urbanization, contain threshold gravel-bed streams. The Anacostia, however, is not a threshold channel; it exhibits break-up of boulder-bed channels in upstream reaches and significant gravel bar formation in downstream reaches. These gravel bars have grown and migrated considerably in the past 10-15 years, contributing significantly to local channel widening that can be twice that of adjacent reaches. The purpose of this study is to determine bedload transport rates and grain size distributions and their relationship to discharge, bed organization and sediment supply. Bed mobility data come from both bedload transport measurements and measurements of channel bed changes. Channel bed changes were obtained from a) repeated channel cross section surveys, b) surface and subsurface size distributions, and c) bed particle organization measurements (measurements of location of particles within reaches). These measurements were made prior to and after the floods of 2006, which equalled the largest floods on record for most parts of the Anacostia River. In some boulder bed reaches, boulders were removed from the center of the channel and deposited along and on the channel banks. The mid-channel boulders were replaced by sheets of gravel and cobbles, significantly altering the bed mobility of the channels.
H53L-04
Implications of Bed-Schematization in Morphological Models with Graded Sediment
For modeling morphology in river with graded-sediment beds, generally methods are used that subdivide the sediment into separate size fractions, and that subdivide the bed into discrete layers. In the past years we have applied and analyzed numerical models using these methods for various river cases, including the full range of time-dependent conditions and spatial variations. It has been found from these applications that particularly the schematization of the bed into discrete layers and the interactions with the transport rate have a serious influence on the outcomes and predictability of these models. The thickness of the active-layer or mixing layer on the bed surface directly determines the time scale of bed- composition changes. Choosing a very thin active layer will lead to very fast adjustments of bed composition to changes or gradients in conditions, such that (generally slow) morphological responses will be cancelled. A more physically sound choice of the layer thickness (e.g., on basis of bed form dimensions), leading to a thicker layer, causes a deceleration of bed-composition changes and propagation. Applications to the Rhine River have shown that this causes the time-scales of morphological changes and bed composition changes to approach each other leading to a stronger interaction between grain size and morphological changes. Tracer experiments in the gravel bed of the German Rhine indicate that a relative thick active layer is required to simulate the slowly migration (2 to 5 km/year) of the granite tracers downstream. Note that both prototype data and model results are showing a dispersion of bed composition disturbances, as propagation speed for different fractions is different. In the model approach sediment-transport rates are coupled to availability of sediment fractions in the active layer. Transport models and hiding and exposure models strongly react to the composition and evolution of the layer, and analysis of the equations reveals some unconventional model responses. For instance, in a case of the Rhine River it has been shown that by removing the coarsest fraction from the bed mixture resulted in halving of mean grain size, a doubling of transport rate, but without a change of the propagation speed or time scale of morphological developments. This feedback between layer composition, transport rate and time scales has been confirmed by theoretical analyses. The feedback mechanism between bed and transport in graded-sediment models complicates further refinement, calibration and validation of the layer-approaches. For instance recent introduction of vertical sorting models (related to bed forms) by means of exchange fluxes between active layer and substratum under-layers result in fining of the active layer and an increase of transport rates in a similar way as shown above. However, the response of morphology and bed composition (sorting evolution) is modified considerably. Further research focuses on unraveling the complex interactions between the bed schematization, transport rates and morphological development. Here fore laboratory and field data are essential ingredients.
H53L-05
Downstream Patterns of Bed-material Grain Size in a Large, Lowland Alluvial River Subject to low Sediment Supply
Downstream patterns of bed-material grain size were investigated over ~380 km of the Sacramento River in northern California, USA. Representative subaqueous bed-material samples were collected from ~125 cross sections spaced ~2 km apart using the 8.2 liter Ponar dredge sampler (for fine beds) and the 23.2 liter Cooper Scooper drag bucket sampler (for mixed and coarse beds), both of which were deployed from a jet boat. Samples were extracted from crossing points in straight reaches that represent the integration of sediment transport through upstream bends. Single samples were collected at simple, narrow cross sections and up to three samples were collected at sections with variable cross-stream topography. Samples were aggregated for each cross section, dried, sieved through eleven phi-sized mesh sieves (0.063-128 mm), and weighed to compute grain size distributions. The largest grain in >95% of all aggregated samples made up <5% of the total sample dry weight, rendering them adequately representative. Sacramento River grain size data suggest departure from widely accepted notions of bed material sorting in rivers in three fundamental ways: 1) subaqueous grain size data differ systematically from nearby data collected from exposed bars; 2) the transition from gravel to sand is not abrupt, but instead extends for several hundred kilometers; and 3) tributaries tend to have a small impact on mainstem downstream grain size sorting. These findings either imply something unique about the Sacramento River dataset compared with prior published data or differences in data collection methods and both possibilities were investigated. The Sacramento River basin is subject to lower sediment supply compared with prior systematically surveyed basins. The impact of decreasing supply (associated with a declining glacial hangover and human impacts) is a decoupling between main channel sediment transport and bar deposition. Thus, in addition to the inherent processes of grain-size censoring that occur on bars, relict bars in a regime of declining sediment supply represent a former state of balance between transport and storage. As such, bar samples are unrepresentative of the currently active processes of sediment movement. Similarly, the low sediment supply probably lengthens the distance over which the bed transitions from gravel to sand. This may occur because bed degradation exposes relatively fine material underlying the gravel bed or because fine sediment supplied from tributaries and bank erosion overwhelms gravel locally in areas of low local slope. Finally, tributaries in the Sacramento Valley do not significantly affect downstream sorting (e.g. by increasing median grain size) because their sediment loads become sorted as they travel across the wide forearc basin of the Central Valley before debouching into the Sacramento River.
H53L-06
Numerical model for analyzing the effects of sediment supply on river morphology and streambed characteristics
We are investigating the effects of different sediment supply regimes on the morphology and aquatic habitat of gravel-bed rivers in mountain basins. Sediment inputs to the river network in these environments range from periodic pulsed events of large magnitude (e.g., landslides and debris flows) to low-level chronic inputs (e.g., road drainage). The pulsed events tend to be natural, while the chronic inputs are commonly anthropogenic. Our goal is to inform managers of the range of sediment supply impacts and to evaluate the common perception that pulsed events are more detrimental to the aquatic ecosystem. Four cases of sediment supply are considered: 1) periodic pulses of fine material (e.g., debris flows in decomposed granite), 2) periodic pulses of coarse material (e.g., landslides in harder bedrock), 3) chronic supplies of fine material (e.g., roads or distributed soil creep and dry raveling) and 4) chronic supplies of coarse material (e.g., mining activity). Currently we are focusing on fine sediment moving primarily as bedload at flows below bank-full over an immobile coarse-grained bed. As part of this work, we evaluated the performance of available surface-based transport models. Predicted transport rates were compared to field measurements from gravel-bed rivers in Idaho. Results show that the Wilcock and Kenworthy (2002) equation performs best. We also have developed a sand conservation model based on filling the voids between coarse particles in a gravel bed. Numerical simulations were conducted coupling the conservation model with the Wilcock and Kenworthy equation using an extension of the MD-SWMS hydraulic model (McDonald et al. 2005). Performance of the transport and conservation models has been tested with data from laboratory experiments. Future modeling will initially compare the effects of cases 1) and 2) (chronic versus pulsed loads of fine sediment), as smaller particles are often implicated in degradation of aquatic habitats.
H53L-07 INVITED
Estimating wind speed from spatial grain size sorting in ripples on Earth and Mars
The landscape initially seen by the Mars Exploration Rover Opportunity at Meridiani Planum is dominated by aeolian (wind-blown) ripples with concentrated surface lags of hematitic spherules and fragments. These ripples exhibit profound spatial grain size sorting, with well-sorted coarse-grained crests and poorly sorted, generally finer-grained troughs - they were the most common bed form encountered by Opportunity in its traverse from Eagle Crater to Endurance Crater. Similar bed features have been sporadically studied in the terrestrial literature, but not in detail. We measured wind speed and sediment flux profiles in White Sands National Monument, New Mexico, during conditions under which such coarse-grained ripples were forming. Data show that these bed features formed by the different transport modes of coarse- and fine-grain fractions in an initially bi-modal sediment distribution. Fine grains were transported via saltation, while coarse grains moved only by creep due to ballistic impacts of finer grains, as originally envisioned by Bagnold. We use this observation to place tight constraints on formative wind conditions of coarse-grained ripples on Mars: wind speed must have exceeded the threshold for saltation of fine grains, but was less than the saltation threshold for coarse grains. Estimated wind speeds are only moderately greater than those associated with modern dust storms. When combined with the observation of sand grains on Opportunity's solar panel following a dust storm, results indicate that modern winds may occasionally be strong enough to cause significant sediment transport on the Martian surface.