Reduced bed material stability and increased bedload transport caused by foraging fish: a flume study with juvenile Barbel (Barbus barbus)

The plants and animals that inhabit river channels may act as zoogeomorphic agents affecting the nature and rates of sediment recruitment, transport and deposition. The impact of benthic‐feeding fish, which disturb bed material sediments during their search for food, has received very little attention, even though benthic feeding species are widespread in rivers and may collectively expend significant amounts of energy foraging across the bed. An ex situ experiment was conducted to investigate the impact of a benthic feeding fish (Barbel Barbus barbus) on particle displacements, bed sediment structures, gravel entrainment and transport fluxes. In a laboratory flume changes in bed surface topography were measured and grain displacements examined when an imbricated, water‐worked bed of 5.6 to 16 mm gravels was exposed to feeding juvenile Barbel (on average, 0.195 m in length). Grain entrainment rates and bedload fluxes were measured under a moderate transport regime for substrates that had been exposed to feeding fish and control substrates which had not. On average, approximately 37% of the substrate, by area, was modified by foraging fish during a four‐hour treatment period, resulting in increased microtopographic roughness and reduced particle imbrication. Structural changes by fish corresponded with an average increase in bedload flux of 60% under entrainment flows, whilst on average the total number of grains transported during the entrainment phase was 82% higher from substrates that had been disturbed by Barbel. Together, these results indicate that by increasing surface microtopography and undoing the naturally stable structures produced by water working, foraging can increase the mobility of gravel‐bed materials. An interesting implication of this result is that by increasing the quantity of available, transportable sediment and lowering entrainment thresholds, benthic feeding might affect bedload fluxes in gravel‐bed rivers. The evidence presented here is sufficient to suggest that further investigation of this possibility is warranted. Copyright © 2014 John Wiley & Sons, Ltd.     

Tree root mounds and their role in transporting soil on forested landscapes

The growth and decay of tree roots can stir and transport soil. This is one process that contributes to the mass‐movement of soil on hillslope. To explore the efficiency of this process, we document the mounding of soil beside Ponderosa and Lodgepole pine trees in the forests that dominate the mid‐elevations of Colorado's Boulder Creek watershed. Mounds are best expressed around Ponderosa pines, reaching vertical displacements above the far‐field slopes of order 10–20 cm, fading into the slope by roughly 100 cm distance from the trunks with common diameters of 30 cm. Positive mounding occurs on all sides of trees on slopes, indicating that the mounding is not attributable to deflection of a creeping flow of soil around the tree, but rather to the insertion of root volume on all sides in the subsurface. Mounding is commonly asymmetric even on cross‐slope profiles. Significant variation in the mound sizes results in no clear relationship between tree diameter and root volume displaced. These observations motivated the development of a discrete element model of tree root growth using the LIGGGHTS model, in which grains we specified to be ‘root cells’ were allowed to enlarge within the simulated granular matrix. Mounding could be reproduced, with the majority of the vertical displacement of the surface being attributable to reduction of the bulk density due to dilation of the granular matrix during root enlargement. Finally, we develop a previous analysis of the role of roots in transporting soil during growth and decay cycles. We find that even in shallow soils, the root‐cycle can drive significant soil transport down forested montane slopes. Copyright © 2013 John Wiley & Sons, Ltd.     

The role of channel morphology on the mobility and dispersion of bed sediment in a small gravel‐bed stream

This study uses a unique 10‐year tracer dataset from a small gravel‐bed stream to examine bed mobility and sediment dispersion over long timescales and at a range of spatial scales. Seasonal tracer data that captured multiple mobilizing events was examined, while the effects of morphology on bed mobility and sediment dispersion were captured at three spatial scales: within morphological units (unit scale), between morphological units (reach scale) and between reaches with different channel morphologies (channel scale). This was achieved by analyzing both reach‐average mobility and travel distance data, as well as the development of ‘mobility maps’ that capture the spatial variability in tracer mobility within the channel. The tracer data suggest that sediment transport in East Creek remains near critical the majority of the time, with only rare large events resulting in high mobility rates and grain travel distances large enough to move sediment past dominant bedforms. While a variable capturing both the magnitude and frequency of flow events within a season yielded a better predictor to sediment mobility and dispersion than peak discharge alone, the distribution of events of different magnitude within the season played a large role in determining tracer mobility rates and travel distances. The effects of morphology differed depending on the analysis scale, demonstrating the importance of scale, and therefore study design, when examining the effect of morphology on sediment transport. Copyright © 2016 John Wiley & Sons, Ltd.     

Transport and storage of bed material in a gravel‐bed channel during episodes of aggradation and degradation: a field and flume study

The dynamics of sediment transport capacity in gravel‐bed rivers is critical to understanding the formation and preservation of fluvial landforms and formulating sediment‐routing models in drainage systems. We examine transport‐storage relations during cycles of aggradation and degradation by augmenting observations of three events of channel aggradation and degradation in Cuneo Creek, a steep (3%) gravel‐bed channel in northern California, with measurements from a series of flume runs modeling those events. An armored, single‐thread channel was formed before feed rates were increased in each aggradation run. Output rates increased as the channel became finer and later widened, steepened, and braided. After feed rates were cut, output rates remained high or increased in early stages of degradation as the incising channel remained fine‐grained, and later decreased as armoring intensified. If equilibrium was not reached before sediment feed rate was cut, then a rapid transition from a braided channel to a single‐thread channel caused output rates for a given storage volume to be higher during degradation than during aggradation. Variations in channel morphology, and surface bed texture during runs that modeled the three cycles of aggradation and degradation were similar to those observed in Cuneo Creek and provide confidence in interpretations of the history of change: Cuneo Creek aggraded rapidly as it widened, shallowed, and braided, then degraded rapidly before armoring stabilized the channel. Such morphology‐driven changes in transport capacity may explain the formation of flood terraces in proximal channels. Transport‐storage relations can be expected to vary between aggradation and degradation and be influenced by channel conditions at the onset of changes in sediment supply. Published in 2011. This article is a US Government work and is in the public domain in the USA.     

Further perspectives on the evolution of bed material waves in alluvial rivers

In one‐dimensional mathematical models of ?uvial ?ow, sediment transport and morphological evolution, the governing equations based on mass and momentum conservation laws constitute a hyperbolic system. Succinctly, the hyperbolic nature excludes dispersion or diffusion operators, which is well known in the context of differential equations. There is no doubt that the so‐called ‘dispersion’ argument for bed material wave evolution is questionable, as we have explicitly asserted. Surprisingly, in a recent communication, the authors of the ‘dispersion’ argument suggest that dispersion is not precluded in hyperbolic systems. We provide herein further perspectives to help explain that the dispersion argument is neither appropriate nor necessary for interpreting bed material wave evolution. Also the continuity equations involved are addressed to prompt wider understanding of their signi?cance. In particular, the continuity equation of the water–sediment mixture proposed by the authors of the ‘dispersion’ argument is proved to be incorrect, and inevitably their reasoning based on it is problematic. Copyright © 2005 John Wiley & Sons, Ltd.     

On evolution of bed material waves in alluvial rivers

Field and laboratory measurements have shown distinct characteristics of bed sediment waves under differing conditions, whilst their theoretical interpretation has emerged to be equivocal. This note aims to clarify the interpretation of evolution of bed material waves. The complete set of governing equations for the flow–sediment–morphology system is deduced to demonstrate its universally hyperbolic nature, irrespective of the sediment transport functions implemented to close the equations. The hyperbolic nature can admit not only attenuating bed material waves, but also shock‐like waves that are not unusual in the real world. It is suggested that the theory of dispersion/diffusion is not universally appropriate for evolution of bed material waves. Copyright © 2003 John Wiley & Sons, Ltd.     

Ecosystem engineering by hummock-building earthworms in seasonal wetlands of eastern South Africa: Insights into the mechanics of biomorphodynamic feedbacks in wetland ecosystems

This paper resolves the origin of clay hummock micro-topography in seasonal wetlands of the Drakensberg Foothills, providing a review and appraisal of previously-suggested mechanisms of hummock formation in the context of new field and laboratory data. Field surveys revealed neo-formation of clay hummocks in a river channel that had been abandoned in c.1984. Fresh earthworm castings were located atop hummocks protruding from inundated abandoned channel margins. Earthworm castings, and sediment cores taken in hummocks and adjacent hollows, were analysed for soil-adsorbed carbon and nitrogen using an HCN analyser, and for ²¹⁰Pb activity using alpha-geochronology. ²¹⁰Pb activity profiles suggest relative enrichment of the isotope in hummocks, and relative depletion in adjacent hollows. Earthworm castings are characterised by very high ²¹⁰Pb activity, as well as high C and N contents. Hummocks have significantly higher C and N contents than adjacent hollows. Results suggest that it is the foraging activity of earthworms in litter-rich seasonal wetland hollows, and repeated excretion of castings atop adjacent hummocks, that is responsible for the elemental enrichment observed. The paper presents a conceptual model of hummock formation in wetlands through interactions between hydrogeomorphology and earthworm activity, and illustrates a mechanism of biogeomorphic inheritance through which ordered patterns of preferential flow can emerge in ecosystems. Further implications of hummock formation and nodal accumulation of nutrients are considered in relation to wetland resilience and regulatory ecosystem service provision.© 2018 John Wiley & Sons, Ltd.     

Bed material transport estimate in large gravel‐bed rivers using the virtual velocity approach

This paper reports on a first attempt of using the virtual velocity approach to assess sediment mobility and transport in two wide and complex gravel‐bed rivers of northern Italy. Displacement length and virtual velocity of spray‐painted tracers were measured in the field. Also, the thickness of the sediment active layer during floods was measured using scour chains and post‐flood morphological changes as documented by repeated survey of channel cross‐sections. The effects of eight and seven floods were studied on the Tagliamento and Brenta Rivers, where 259 and 277 spray‐painted areas were surveyed, respectively. In the Tagliamento River 36% of the spray‐painted areas experienced partial transport, whereas in the Brenta River this accounted for 20%. Whereas, full removal/gravel deposition was observed on 37% and 26% of these areas on the Tagliamento and Brenta Rivers, respectively. The mean displacement length of particles, the thickness of the active layer and the extent of partial transport are well correlated with the dimensionless shear stress. The virtual velocity approach allowed calculation of bed material transport over a wide range of flood magnitudes. Annual coarse sediment transport was calculated up to 150 for the Tagliamento, and 30 × 10³  m³ yr^?1 for the Brenta. The outcomes of this work highlight the relevance of partial transport condition, as it could represent more than 70% of the total bed material transported during low‐magnitude floods, and up to 40% for near‐bankfull events. Results confirm that bed material load tends to be overestimated by traditional formulas. Copyright © 2016 John Wiley & Sons, Ltd.     

More on the evolution of bed material waves in alluvial rivers

Sediment waves or pulses can form in rivers following variations in input from landslides, debris flows, and other sources. The question as to how rivers cope with such sediment inputs is of considerable practical interest. Experimental, numerical and field evidence assembled by the authors suggests that in mountain gravel‐bed streams, such pulses show relatively little translation, instead mostly dispersing in place. This research has recently been the subject of discussion. In particular it has been suggested that (a) the equations of flow and sediment mass balance used in the analyses, and in most other morphodynamic analyses, require correction; (b) the dominance of dispersion appears only because the hyperbolic nature of the governing equations has not been adequately considered; and (c) the sediment transport equation used in the analyses does not lead to generalizable results. Here we suggest that (a) the relations for mass balance do not require the indicated correction; (b) the hyperbolic nature of the governing equations does not preclude the result of dispersion dominating translation in mountain streams; and (c) the general behaviour of an appropriate hyperbolic model of sediment waves (pulses) includes the relative roles of dispersion and translation, and is not affected by the precise choice of a sediment transport relation (as long as the choice is reasonable for the case in question). Copyright © 2005 John Wiley & Sons, Ltd.     

Sediment mobility and bed armoring in the St Clair River: insights from hydrodynamic modeling

The lake levels in Lake Michigan‐Huron have recently fallen to near historical lows, as has the elevation difference between Lake Michigan‐Huron compared to Lake Erie. This decline in lake levels has the potential to cause detrimental impacts on the lake ecosystems, together with social and economic impacts on communities in the entire Great Lakes region. Results from past work suggest that morphological changes in the St Clair River, which is the only natural outlet for Lake Michigan‐Huron, could be an appreciable factor in the recent trends of lake level decline. A key research question is whether bed erosion within the river has caused an increase in water conveyance, therefore, contributed to the falling lake level. In this paper, a numerical modeling approach with field data is used to investigate the possibility of sediment movement in the St Clair River and assess the likelihood of morphological change under the current flow regime. A two‐dimensional numerical model was used to study flow structure, bed shear stress, and sediment mobility/armoring over a range of flow discharges. Boundary conditions for the numerical model were provided by detailed field measurements that included high‐resolution bathymetry and three‐dimensional flow velocities. The results indicate that, without considering other effects, under the current range of flow conditions, the shear stresses produced by the river flow are too low to transport most of the coarse bed sediment within the reach and are too low to cause substantial bed erosion or bed scour. However, the detailed maps of the bed show mobile bedforms in the upper St Clair River that are indicative of sediment transport. Relatively high shear stresses near a constriction at the upstream end of the river and at channel bends could cause local scour and deposition. Ship‐induced propeller wake erosion also is a likely cause of sediment movement in the entire reach. Other factors that may promote sediment movement, such as ice cover and dredging in the lower river, require further investigation. Copyright © 2012 John Wiley & Sons, Ltd.     

High frequency stream bed mobility of a low‐gradient agricultural stream with implications on the hyporheic zone

Little Kickapoo Creek (LKC), a low‐gradient stream, mobilizes its streambed–fundamentally altering its near‐surface hyporheic zone–more frequently than do higher‐gradient mountain and karst streams. LKC streambed mobility was assessed through streambed surveys, sediment sampling, and theoretical calculations comparing basal shear stress (τ_b) with critical shear stress (τ_c). Baseflow τ_b is capable of entraining a d₅₀ particle; bankfull flow could entrain a 51·2 mm particle. No particle that large occurs in the top 30 cm of the substrate, suggesting that the top 30 cm of the substrate is mobilized and redistributed during bankfull events. Bankfull events occur on average every 7·6 months; flows capable of entraining d₅₀ and d₈₅ particles occur on average every 0·85 and 2·1 months, respectively. Streambed surveys verify streambed mobility at conditions below bankfull. While higher gradient streams have higher potential energy than LKC, they achieve streambed‐mobilization thresholds less frequently. Heterogeneous sediment redistribution creates an environment where substrate hydraulic conductivity (K) varies over four orders of magnitude. The frequency and magnitude of the substrate entrainment has implications on hyporheic zone function in fluid, solute and thermal transport models, interpretations of hyporheic zone stability, and understanding of LKC's aquatic ecosystem. Copyright © 2008 John Wiley & Sons, Ltd.     

Transport and dispersal of organic debris (peat blocks) in upland fluvial systems

This paper assesses the mechanisms and pathways by which peat blocks are eroded and transported in upland fluvial systems. Observations and experiments from the north Pennines (UK) have been carried out on two contrasting river systems. Mapping of peat block distributions and appraisal of reach‐based sediment budgets clearly demonstrates that macro‐size peat is an important stream load component. In small streams block sizes can approximate the channel width and much of the peat is transported overbank. Local ‘peat jams’ and associated mineral deposition may provide an important component of channel storage. In larger systems peat blocks rapidly move down‐channel and undergo frequent exchanges between bed and bank storage. Results of peat block tracing using painted blocks indicate that once submerged, blocks of all sizes are easily transported and blocks break down rapidly by abrasion. Vegetation and bars play an important role in trapping mobile peat. In smaller streams large block transport is limited by channel jams. Smaller blocks are transported overbank but exhibit little evidence of downstream fining. In larger rivers peat blocks are more actively sorted and show downstream reduction in size from source. A simple model relating peat block diameter (D_p) to average flow depth (d) suggests three limiting transport conditions: flotation (D_p < d), rolling (d < D_p > d/2) and deposition (D_p > d/2). Experiments demonstrate that peat block transport occurs largely by rolling and floating and the transport mechanism is probably controlled by relative flow depth (d/D_p ratio). Transport velocity varies with transport mechanism (rolling is the slowest mode) and transport lengths increase as flow depth increases. Abrasion rates vary with the transport mechanism. Rolling produces greater abrasion rates and more rounded blocks. Abrasion rates vary from 0 to 10 g m^?1 for blocks ranging in mass from 10 to 6000 g. Copyright © 2001 John Wiley & Sons, Ltd.     

The use of short‐lived radionuclides to quantify transitional bed material transport in a regulated river

We investigate the use of the short‐lived fallout radionuclide beryllium‐7 (⁷Be; t_1/2 = 53·4 days) as a tracer of medium and coarse sand (0·25–2 mm), which transitions between transport in suspension and as bed load, and evaluate the effects of impoundment on seasonal and spatial variations in bed sedimentation. We measure ⁷Be activities in approximately monthly samples from point bar and streambed sediments in one unregulated and one regulated stream. In the regulated stream our sampling spanned an array of flow and management conditions during the annual transition from flood control in the winter and early spring to run‐of‐the‐river operation from late spring to autumn. Sediment stored behind the dam during the winter quickly became depleted in ⁷Be activity. This resulted in a pulse of ‘dead’ sediment released when the dam gates were opened in the spring which could be tracked as it moved downstream. Measured average sediment transport velocities (30–80 metres per day (m d^?1)) exceed those typically reported for bulk bed load transport and are remarkably constant across varied flow regimes, possibly due to corresponding changes in bed sand fraction. Results also show that the length scale of the downstream impact of dam management on sediment transport is short (c. 1 km); beyond this distance the sediment trapped by the dam is replaced by new sediment from tributaries and other downstream sources. Copyright © 2006 John Wiley & Sons, Ltd.     

The effect of bed mobility on resistance to overland flow

When sediment grains are transported as bed load in overland flow, there is a net transfer of momentum from the flow to the grains. When these grains collide with other grains, whether on the bed or in the flow, streamwise flow velocity decreases and resistance to flow increases. Resistance to flow generated in this manner is termed bed‐load transport resistance. Resistance to flow f over a plane bed may be partitioned into grain resistance f_g and bed‐load transport resistance f_bt. We use the symbols f_btf and f_btm to denote f_bt for flows over fixed beds and over mobile beds, respectively, and we compute the effect of bed mobility on flow resistance f_mob by subtracting f_btf from f_btm. The data for this study come from 54 flume experiments with fixed beds and 38 with mobile beds. On average f_mob is approximately equal to half of f_btm, which is about one‐quarter of f. Hence, f_mob is about one‐tenth of f. Predictive equations are developed for f_btf, f_btm and f_mob using dimensional analysis to identify the relevant independent variables and regression analysis to evaluate the coefficients associated with these variables. Values of f_mob are always positive which implies that mobile beds offer greater resistance to flow than do fixed beds. Evidently bed‐load grains colliding with mobile beds lose more momentum to the bed than do grains colliding with fixed beds. In other words, grain collisions with mobile beds are less elastic than those with fixed beds. Copyright © 2005 John Wiley & Sons, Ltd.     

Experimental study on the bank erosion and interaction with near-bank bed evolution due to fluvial hydraulic force

Bank erosion is a typical process of lateral channel migration,which is accompanied by vertical bed evolution.As a main sediment source,the failed bank soil may directly cause the increase of sediment concentration and considerable channel evolution in a short time.The paper presents an experimental study on non-cohesive and cohesive homogenous bank failure processes,influence of the failed bank soil on bank re-collapse,as well as the interaction between bank failure and near-bank bed evolution due to fluvial hydraulic force.A series of experiments were carried out in a 180° bend rectangular flume.The results reveal the iteration cycle between bank erosion and bed deformation:undercutting of the riverbank,slip failure of the submerged zone of the bank,as well as cantilever failure of the overhang,failed bank soil staying at bank toe temporarily or hydraulic transportation,exchange between the failed bank soil and bed material,bed material load being re-transported either as bed load or as suspended load,and bed deformation.Same as bank failure,the mixing of failed bank soil and bed material is more severe near the curved flow apex.Moreover,non-cohesive bank failure tends to occur near the water surface while cohesive bank failure near the bank toe.For non-cohesive dense(sandy) soil,the bank erosion amount and residual amount of failed bank soil on the bed increase with the near-bank velocity or bed erodibility.But for cohesive soil,only bank erosion amount follows the above rule.The results are expected to provide theoretical basis for river management and flood prevention.     

Comparison of fractional bed‐material load ­computation methods in sand‐bed channels

This paper presents a comparison of different fractional bed‐material load computation methods in sand‐bed channels. These methods include the direct computation by size fraction approach of Einstein, Laursen and Toffaleti; the bed material fraction (BMF) approach using equations of Engelund and Hansen, Ackers and White, Yang, and Karim's modified BMF method; and the transport capacity fraction (TCF) approach using the transport capacity distribution functions of Karim and Kennedy, Li, and Wu and Molinas. Over 150 sets of flume and field data in the sand size range containing a total of 1007 data points are used to evaluate and compare the accuracy of these methods. Statistical analysis and graphical comparison are utilized to demonstrate the performance and variations in different methods. Overall, the Einstein method underpredicts the transport rate for finer sizes and overpredicts for the coarser sizes, while the other methods overestimate the finer fractions and underestimate the coarser fractions. The Wu and Molinas method, which was developed to account for these deficiencies, is shown to significantly improve fractional bed‐material load computations. The Karim and Kennedy method is also found to be applicable. Copyright © 2000 John Wiley & Sons, Ltd.     

The impact of pro‐glacial lakes on downstream sediment delivery in Norway

Glacier recessions caused by climate change may uncover pro‐glacial lakes that form important sedimentation basins regulating the downstream sediment delivery. The impact of modern pro‐glacial lakes on fluvial sediment transport from three different Norwegian glaciers: Nigardsbreen, Engabreen and Tunsbergdalsbreen, and their long‐term development has been studied. All of these lakes developed in modern times in overdeepened bedrock basins. The recession of Nigardsbreen uncovered a 1.8 km long and on average 15 m deep pro‐glacial lake basin during 1937 to 1968. Since then the glacier front has been situated entirely on land, and the sediment input and output of the lake has been measured. The suspended sediment transport into and out of the lake averaged 11 730 t yr^?1 and 2340 t yr^?1 respectively. Thus, 20% remained in suspension at the outlet. The measured mean annual bedload supplied to the lake was 11 800 t yr^?1, giving a total transport of 23 530 t yr^?1 which corresponds to a specific sediment yield of 561 t km^?2 yr^?1. A 1.9 km long and up to 90 m deep pro‐glacial lake basin downstream from Engabreen glacier was uncovered during 1890 to 1944. The average suspended sediment load delivered from the glacier during the years 1970–1981 amounted to 12 375 t yr^?1and the transport out of the lake was 2021 t yr^?1, giving an average of 16% remaining in suspension. The mean annual bedload was 8000 t yr^?1, thus the total transport was 20 375 t yr^?1, giving a specific sediment yield of 566 t km^?2 yr^?1. For Tunsbergdalsbreen glacier, measurements in the early 1970s indicated that the suspended sediment transport was on average 44 000 t yr^?1. From 1987 to 1993 the recession of the glacier uncovered a small pro‐glacial lake, 0.3 km long and around 9 m deep. Downstream from this, the suspended sediment load measured in 2009 was 28 000 t yr^?1, indicating that as much as 64% remained in suspension. Flow velocity, grain size of sediment, and morphology of the lake are important factors controlling the sedimentation rate in the pro‐glacial lakes. A survey of the sub‐glacial morphology of Tunsbergdalsbreen revealed that there are several overdeepened basins beneath the glacier. The largest is 4 km long and 100 m deep. When the glacier melts back they will become lakes and act as sedimentation basins. Despite an expected increase in sediment yield from the glacier, little sediment will pass these lakes and downstream sediment delivery will be reduced markedly. Beneath Nigardsbreen there was only a small depression that may form a lake and the sediment delivery will not be significantly affected. © 2014 The Authors. Earth Surface Processes and Landforms published by John Wiley & Sons Ltd.     

Dune mobility and vegetation cover in the Southwest Kalahari desert

As part of a wider project investigating the palaeoenvironmental significance of partially vegetated linear dunes in the southwest Kalahari, data collected in the latter part of 1992 concerning dune movement and vegetation cover suggest that sediment transport is occurring on some dune surfaces, and that the majority of surface activity occurs on the crests and upper slopes of the dunes. The data suggest that the limiting variables on surface sediment movement vary on different parts of a dune. On interdunes and lower dune slopes the primary limiting variable is available wind energy, while on dune crests and upper slopes it is vegetation cover. Ground cover by litter has much greater importance in protecting the surface sediment from erosion than rooted vegetation. From individual data points, no evidence is found to support a threshold vegetation cover below which sediment movement occurs. Rather, a gradient of activity is suggested whereby a reduction in vegetation cover increases the potential for sediment movement and surface change. However, dunes with differing amounts of mean vegetation cover display differing degrees of surface activity, and at this scale, a vegetation cover threshold in the region of 14 per cent may be recognized.     

Effect of bed topography on soil aggregates transport by rill flow

After its formation, a rill may remain in the field for months, often receiving lower flow rates than the formative discharge. The objective of this work was to evaluate the rill flow transport capacity of soil aggregates at discharges unable to erode the rill, and to analyse the influence of the rill macro‐roughness on this transport process. A non‐erodible rill was built in which roughness was reproduced in detail. In order to assess only the rill macro‐roughness, a flat channel with a similar micro‐roughness to that in the rill replica was built. Rill and channel experiments were carried out at a slope of 8 and at six discharges (8·3 × 10^?5 to 5·2 × 10^?4 m³ s^?1) in the rill, and eight discharges (1·6 × 10^?5 to 5·2 × 10^?4 m³ s^?1) in the channel. Non‐erodible aggregates of three sizes (1–2, 3–5 and 5–10 mm) were released at the inlet of the rill/channel. The number of aggregates received at the outlet was registered. The number and position of the remaining aggregates along the rill/channel were also determined. The rill flow was a major sediment transport mechanism only during the formation of the rill, as during that period the power of the flow was great enough to overcome the influence of the macro‐roughness of the rill bed. At lower discharges the transport capacity in the previously formed rill was significantly less than that in the flat channel under similar slope and discharge. This was determined to be due to local slowing of flow velocities at the exit of rill pools. Copyright © 2006 John Wiley & Sons, Ltd.