Philip Ashworth

Three-dimensional sedimentary architecture of a large, mid-channel sand braid bar, Jamuna River, Bangladesh

The three-dimensional subsurface alluvial architecture of a large (approximately 3 km long, 1 km wide, 12 m high), mid-channel sand braid bar in the Jamuna River, Bangladesh is described. Evolution of the bar and its depositional characteristics are assessed from a unique combination of ground-penetrating radar surveys, vibracoring, and trenching that are allied to a series of bathymetric surveys taken during growth of the bar over a 29-month period. This methodology permits identification of the formative processes of different packages of braid-bar sedimentation and provides a facies model for deposition within the entire bar. Mid-channel bar growth occurred in a region of flow expansion and was probably initiated by the stalling and amalgamation of large dunes. These dunes created a bar-core that grew by (i) propagation of a downstream-accreting slipface, (ii) vertical accretion through stacking of dunes on both bar stoss and top, and (iii) lateral accretion on the bar margins during recession of the flood hydrograph. Braid-bar sedimentation is dominated by four radar facies: (1) large-scale, predominantly planar, dipping reflections interpreted as cross-stratification, up to 8 m in height and greater than 100 m in width, that is produced by the cross-channel migration of bar margins, (2) medium-scale, trough-shaped and planar discontinuous reflections interpreted as cross-stratification up to 4 m in height and 300 m wide, that is deposited from large, sinuous-crested sand dunes, (3) discontinuous reflections, up to 2 m high and 30 m wide, interpreted as small-scale trough cross-stratification, that is the product of smaller sinuous-crested dunes, and (4) high-amplitude, undulating reflections interpreted as mud drapes, deposited in regions of slow flow, often in the bar-tail region at low stage. Dune size decreases vertically within the bar, in response to the progressively shallower flows on the emerging bar top. Later evolution of the bar, as one anabranch channel became dominant, created a 1.5 km extension to the bar tail with an 8 m high, angle-of-repose, bar-margin slipface, formed by flow transverse to the long axis of the bar. Seven styles of deposition can be defined that constitute the alluvial architecture: bar-margin slipface, vertical accretion in channel, bar-top vertical accretion, upstream accretion, lateral accretion, downstream accretion, and low-stage mud drapes. A model of braid-bar sedimentation is presented that shares many similarities with previous studies of smaller sand-bed braid bars with the dominance of dune-scale cross-stratification, the presence of large-scale, bar-margin cross-stratification, and the occurrence of lateral, vertical, upstream, and downstream accretion. However, the contribution of the bar-margin facies to the preserved stratigraphy highlighted herein may have been underestimated in previous models of braided rivers in which the braid bars were migrating slowly. This study suggests a scale invariance in several aspects of mid-channel bar sedimentation in sand-bed rivers and proposes a model of braid-bar sedimentation that may be applied widely within studies of braided alluvial architecture.

Sediment slugs: large-scale fluctuations in fluvial sediment transport rates and storage volumes

Variations in fluvial sediment transport rates and storage volumes have been described previously as sediment waves or pulses. These features have been identified over a wide range of temporal and spatial scales and have been categorized using existing bedform classifications. Here we describe the factors controlling the generation and propagation of what we term sediment slugs. These can be defined as bodies of clastic material associated with disequilibrium conditions in fluvial systems over time periods above the event scale. Slugs range in magnitude from unit bars (Smith, 1974) up to sedimentary features generated by basin-scale sediment supply disturbances (Trimble, 1981). At lower slug magnitudes, perturbations in sediment transport are generated by local riverbank and/or bed erosion. Larger-scale features result from the occurrence of rare high- magnitude geomorphic events, and the impacts on water and sediment production of tectonics, glaciation, climate change and anthropogenic influences. Simple sediment routing functions are presented which may be used to describe the propagation of sediment slugs in fluvial systems. Attention is drawn to components of the fluvial system where future research is urgently required to improve our quantitative understanding of drainage-basin sediment dynamics.

MID-CHANNEL BAR GROWTH AND ITS RELATIONSHIP TO LOCAL FLOW STRENGTH AND DIRECTION

Anabranches of braided rivers typically migrate and avulse across the floodplain to produce new channel junctions, scour and subsequent mid-channel bar growth immediately downstream. Few quantitative studies have been made of this bar development process and the link to change in channel geometry and local flow strength and direction. This paper provides data on the spatial and temporal pattern of surface velocity as mid-channel bar growth is initiated downstream of a fixed junction scour in a generic scale flume model. The sequence of channel changes is: (i) development of a confluence scour with flow convergence and maximum velocity in the channel centre; (ii) exceedance of the local transport capacity and initial stalling of coarse sediment in the channel thalweg downstream of the scour; (iii) bar growth through entrapment of all sizes of bedload; (iv) change from velocity maximum to minimum and flow convergence to divergence when the bar height is approximately 55 per cent of the thalweg depth; (v) broadening of the bartop platform, a drop in local competence and bankward migration of the two distributaries whose cross-section and velocity remains approximately constant. These flume data and interpretations are compared to descriptions in the literature of the braiding process with particular reference to the flume work of Leopold and Wolman (1957) and Ashmore (1991, 1993). A new model for mid-channel bar growth is presented which helps explain the long-term development of the confluence–diffluence unit.

Interrelationships of channel processes, changes and sediments in a proglacial braided river

Spatial and temporal variations in channel morphology, velocity and shear stress, bedload transport rate, and bedload and bed material size distribution were measured over five summer weeks in the braided Lyngsdalselva in Arctic Norway. Velocities and shear stresses reflected streamwise convergence and divergence of flow, and in turn influenced bedload transport rates although these depended also on the availability at the surface of appropriate-sized sediment. Large and small particles had almost equal mobility, so transport rate increased rapidly with shear stress as coarse sediment began to move and in so doing exposed finer particles. Size-selective transport is nevertheless indicated by down-reach and down-bar fining. The spatial patterns of shear stress and bedload transport rate within one divided-single-divided channel cycle changed with discharge, as did the resulting channel changes. In high meltwater flows, with a peak transport rate of 0.3 kg m -1 s-1, erosion was localised at two zones of flow convergence and tended to maintain the non-uniform channel geometry. During a higher rainflood, with transport rates up to 3 kg m-1 s-1, a downstream fall in shear stress in the main distributary led to medial deposition and consequent lateral erosion, creating a wider, shallower, more uniform channel. Grain size distributions of bedload at high and low shear stresses match fairly well those of floodplain gravels and falling stage flood deposits respectively.

Relationship between sediment supply and avulsion frequency in braided rivers

The interplay between sediment supply ( S s ), sedimentation rate ( S r ), and the frequency of channel avulsion ( A f ) exerts a primary control on alluvial architecture. In order to investigate the effect of sediment supply on avulsion frequency, four Froude-scale model experiments of an aggrading braided river were undertaken in which the magnitude of S s was progressively increased over an eightfold range. The value of A f increases at a rate slower than the increase in S s , contrary to the trend previously reported by Bryant, Falk, and Paola in their experimental study on alluvial-fan dynamics. These results suggest that the relationship between A f and S s is dependent upon bed slope and that the response of A f to an increase in S s in unconfined braided rivers may be different than that on steep alluvial fans.

Measurements in a Braided River Chute and Lobe: 1. Flow Pattern, Sediment Transport and Channel Change

This paper and its companion (Ashworth et al., this issue) discuss measurements of channel change and associated flow and sediment transport processes in a representative chute-and-bar reach within a proglacial gravel-bed river, the Sunwapta River in the Canadian Rockies. During a week in which water discharge through the reach increased then decreased, a sediment wave passed into and partly along the reach. At first the chute aggraded, then sediment was eroded from the chute and deposited in a prograding lobe to one side of the original bar head between two distributaries. Measurements of velocity, shear stress, and gravel transport rate revealed day-to-day changes in the divergent pattern of flow and sediment transport. The off-center location of the lobe reflected an initial asymmetry in the pattern of flow and sediment transport, but as the lobe grew, the flow and transport gradually switched away to the other distributary. Width-averaged bed load transport rates do not agree well with average rates inferred from bar head deposition or from volumes and spatial patterns of scour and fill; the indirect estimates are considered more reliable than those based on direct sampling for necessarily brief durations. By the end of the study the new lobe had almost accreted onto the original bar head, supporting the idea that most braid bars are of compound origin. Temporal and spatial patterns in grain size distributions of the bed, bed load, and deposited sediment are discussed in the companion paper.

Measurements in a Braided River chute and lobe: 2. Sorting of bed load during entrainment, transport, and deposition

Grain size distributions of bed material, bed load, and bar head deposition were sampled during the evolution of a chute and lobe in the braided gravel bed Sunwapta River, Alberta. Although bed shear stress and total bed load transport rate varied substantially within the reach and from day to day (Ferguson et al., this issue), the median diameter D50 of bed load was remarkably constant. The bed initially fined from chute to bar head but became more uniform. Sediment deposited on the bar head also had a near constant D50, similar to that of the bed load truncated at the same lower size limit. Both load and deposition were finer on average than the bed, suggesting overall selective transport despite little sign of local sorting. In contrast to median diameters, maximum sizes (Dmax) of sediment in motion and deposited on the bar head varied substantially and showed a weak but significant dependence on shear stress. Inferring selective entrainment from this evidence alone would be dubious because Dmax was also found to increase systematically with the mass of sediment sampled. However, two different methods of analysis of fractional transport rates throughout the size range also showed size-selective entrainment and transport. In the marginal transport conditions of this study, size sorting appears to occur but only weakly and mainly in the coarser fractions.

Computational fluid dynamics and the physical modelling of an upland urban river

This paper describes the application of a commercially available, three-dimensional computational fluid dynamic (CFD) model to simulate the flow structure in an upland river that is prone to flooding. Simulations use a rectangular channel geometry, smooth sidewalls and a bed topography obtained from the field site that contains a subdued pool–riffle sequence. The CFD model uses the RNG κ–var epsilon turbulence closure scheme of Yakhot and Orszag (J. Sci. Comput. 1 (1986) 1), as implemented in FLUENT 4.4.4, with a free surface. Results are shown for numerical runs simulating a 1:100 year return interval flood. Output from the numerical model is compared to a physical model experiment that uses a 1:35 scale fibreglass mould of the field study reach and measures velocity using ultrasonic Doppler velocity profiling (UDVP). Results are presented from the numerical and flume models for the water surface and streamwise velocity pattern and for the secondary flows simulated in the numerical model. A good agreement is achieved between the CFD model output and the physical model results for the downstream velocities. Results suggest that the streamwise velocity is the main influence on the flow structure at the discharge and channel configuration studied. Secondary flows are, in general, very weak being below the resolution of measurement in the physical model and less than 10% of the streamwise velocity in the numerical model. Consequently, there is no evidence for a ‘velocity dip’. It is suggested that the subdued topography or inlet morphology may inhibit the development of secondary flows that have been recorded in previous flat-bed, rectangular open channel flows. A significant corollary of these results is that the morphological evolution of the pool–riffle sequence at high discharges may be controlled primarily by the downstream distribution of velocity and sediment transport with little role for lateral sorting and sediment routing by secondary flows. This paper also raises a number of issues that may be of use in future CFD modelling of three-dimensional flow in open channels within the geomorphological community.

Can we distinguish flood frequency and magnitude in the sedimentological record of rivers

Consideration of the origin of alluvial deposits and their paleoenvironmental interpretation has traditionally involved two schools of thought: that they are either the result of processes that, on average, have acted uniformly through time, or that they are related to exceptional events that occur infrequently. Despite the long-running debate of gradualism versus catastrophism within the Earth Sciences, there are surprisingly few quantitative data to assess the magnitude of events that produce alluvial sedimentary successions. This paper reports on a unique ‘natural experiment’ where surface (digital elevation model, DEM) and subsurface (ground penetrating radar, GPR) data were taken immediately prior to, and after, a large (1-in-40 yr) flood event on the sandy, braided, South Saskatchewan River, Canada. Results show that although this high-magnitude flood reworked the entire braidplain, the scale of scour and style of deposition was similar to that associated with lower-magnitude, annual, floods. The absence of a distinct imprint of this large flood within the deposits is related to the fact that as river discharge rises, and begins to flow overbank, channel width increases at a much faster rate than flow depth, and thus the rate of increase in channel bed shear stress declines. Hence, rather than being a product of either frequent or rare events, alluvial deposits will likely be created by a range of different magnitude floods, but discriminating between these different scale events in the rock record may be extremely difficult.

Spatial Pattern of Flow Competence and Bed Load Transport in a Divided Gravel Bed River

A 14-month continuous record of the spatial variation in the incidence and magnitude of bed load transport was provided by nine pit traps sunk into the bed of a 21-m-wide upland gravel bed river. The bed load data are related to a simultaneous record of average bed shear stress derived from the depth-slope product. The study reach is straight and includes a small, low-relief mid-channel bar which represents the divide between two zones of markedly different bed structure and stability. he bed of the left channel and bar tail is loose with an open framework, while that of the bar head and right channel is tightly interlocking and well imbricated. Bed surface grain size is similar throughout the reach. The two zones of contrasting bed structure have a strong influence on the threshold of sediment movement (τ c) and consequently the spatial pattern of bed load transport. Entrainment thresholds for the tightly structured bed are up to seven times higher than those for the loosely structured bed. Sediment transport is therefore most frequent in the left channel and over the bar tail, where flows with excess shear stresses (τ > τc) peaking at 8τc constitute 19% of the flow record. This is in marked contrast to the right channel and bar head, where excess shear stresses of maximum 1.4τc total only 1.3% of the flow record. However, despite the marked difference in the magnitude and duration of competent flows between the two contrasting bed areas, little difference in bed load yield is observed between the nine traps when averaged over the 14-month study period. This study shows that different zones of bed structure and stability in a channel will control the incidence and spatial pattern of bed load transport but not necessarily the medium and long-term bed load yield. This has important implications for local and short-term bed load sampling programs and the numerical modeling of sediment budgets and zone-to-zone bed load transfer.