See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/270898495

Sediment Transport and Bedform Development in the Lee of Bars: Evidence

from Fixed- and Partially-Fixed Bed Experiments

Conference Paper · September 2014

DOI: 10.1201/b17133-158

CITATIONS READS

4 154

3 authors:

Arnold Jan Reesink Daniel Roy Parsons

University of Illinois, Urbana-Champaign University of Hull
26 PUBLICATIONS   265 CITATIONS    218 PUBLICATIONS   2,393 CITATIONS   

SEE PROFILE SEE PROFILE

Robert E Thomas
University of Leeds
52 PUBLICATIONS   673 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Influence of Aquatic Vegetation Flexibility on Wave Hydrodynamics View project

HYDRALAB IV View project

All content following this page was uploaded by Arnold Jan Reesink on 16 January 2015.

The user has requested enhancement of the downloaded file.

Sediment transport and bedform development in the lee of bars:
Evidence from fixed- and partially-fixed bed experiments
A. J. H. Reesink
Geography and Environment, University of Southampton, SO17 1BJ, Southampton, UK
D.R. Parsons & R.E. Thomas
Department of Geography, Environment and Earth Sciences, University of Hull, HU6 7RX, Hull, UK

ABSTRACT: The co-existence, interaction, and repeated (re-)establishment of bars, dunes and ripples in nat-
ural channels is responsible for many important flow-form-flow dynamics. Small bedforms are constantly
generated, superimposed on larger ones, particularly in zones affected by large-scale secondary circulation
patterns produced by the larger bedforms. These superimposed bedforms migrate onto the downstream stoss
slope of larger-scale forms where they i) generate additional form-roughness, ii) change sediment transport
dynamics, iii) control bedform splitting and merging, iv) alter the geometry of the host lee slope, and v)
change the resultant sedimentary structures. Our understanding of superimposed bedform development is de-
rived from investigations of bedform development on flat beds in uniform flow and does not adequately de-
scribe bedform development in distinctly non-uniform flows and areas with large-scale secondary circulations.

In order to expand our understanding of bedform initiation, this paper presents fixed-bed and partially-fixed-
bed experiments that investigate the effect of a host-bedform’s separated flow on the development of smaller,
secondary bedforms in its trough. The results show that: 1) scour in the trough of bars increases in depth and
decreases in downstream length with increasing flow velocity over the crest; 2) the point of bedform initiation
moves downstream and the amplitude of the incipient ripples decreases with increasing flow velocity; 3)
crest-trough velocity gradients and coherence of the separated flows in the lee of ripples in bar troughs depend
on their position relative to the separated flow of the larger-scale host bedform and tend to increase down-
stream.

These observations indicate that the development of secondary bedforms is hindered by the host bedform’s
separated flow and is also dependent on the length of the downstream stoss slope. The reduction of bedform
amplification is attributed to the reduced strength and coherence of the separated flows in the lee of the sec-
ondary bedforms as a result of the stronger separated flow of the host bedform. Thus, this study presents a step
towards a fuller understanding of bedform initiation and development in areas with complex topography and
local variability in the flow field.

1 INTRODUCTION damental and opposing challenges: to expand our

knowledge of diverse flow-form-flow feedback
Bedforms of different sizes and shapes co-exist and mechanisms, and, to quantitatively test the increas-
interact via a broad range of flow-form-flow feed- ing number of known feedbacks in controlled and
back mechanisms that significantly affect many as- reproducible experiments over a broad range of con-
pects of their natural environment, such as water ditions. This paper presents an experiment that re-
depths, flow velocities, and the nature of preserved produces bedform trough scour, and the initiation
strata. Detailed knowledge of hydro-dynamic inter- and amplification of secondary, or superimposed,
actions is sparse and our overview of feedback bedforms within a narrow set of boundary condi-
mechanisms remains incomplete because of the ex- tions. Thus, the methods reproduce a very limited set
tensive range of bed shapes, substrate characteristics, of naturally occurring conditions but effectively iso-
flow characteristics, and time-space scales that are late a limited range of processes for analysis, whilst
possible. As a consequence of this natural diversity, simultaneously providing a template that is easily
research on fluvial morphodynamics faces two fun- reproduced and modified for further research.
1.1 Interaction of different scales of bedforms gradation rates and low migration rates enhance bed-
form preservation (Allen 1982; Rubin & Hunter,
The coexistence of different types and sizes of bed-
1983; Paola and Borgman, 1991; Bridge 1997; Le-
forms is both the cause and effect of a broad range of
clair and Bridge 2001; Reesink and Bridge 2009,
important non-linear water-sediment dynamics in
2011).
rivers, estuaries and coastal seas: the product of
morphodynamic feedbacks (Bridge 2003). Distribu-
tions of sizes of bedform types (e.g. dunes) generally 1.2 What controls the existence of superimposed
resemble log-normal or gamma-distributions that are bedforms?
skewed to the smaller forms (Paola & Borgman
1991; Leclair & Bridge 2001; Bridge 2003). This As outlined above, bedform co-existence and super-
skewness implies that superimposition of bedforms imposed bedform dynamics underpins many funda-
is ubiquitous, and that interactions between forms mental process dynamics. However, our understand-
are abundant because individual migration rates will ing of the initiation of new bedforms in non-uniform
vary according to bedform volume (Bridge, 2003). flow conditions generated by larger bedforms is in-
The lower limit of identifiable superimposed bed- ferred from bedform development on flat beds
form sizes is often prescribed by the resolution of the (Raudkivi & Witte 1990; Coleman & Melville 1994;
measurements. High-resolution measurements have Baas, 1994). The existence of superimposed bed-
shown that millimetre-thin sediment waves on the forms has been interpreted as an expression of dise-
backs of dunes are a fundamental mechanism by quilibrium between the dunes and an unsteady flow
which bedload sediment moves over dunes, and (Allen 1982), as an expression of the naturally oc-
therefore a primary mechanism for dune migration curring variability in dune size (Bridge 2003), and as
(Venditti et al, 2005a). Superimposed bedform dy- an inherent characteristic of the re-establishment of a
namics enable dunes to grow and decay in response boundary layer on bedform stoss slopes (Rubin &
to changes in discharge and water depth (Kleinhans, McCulloch 1980). Together, these views provide a
2002; Venditti et al. 2005b; Martin and Jerolmack, fairly coherent picture of the key controls on bed-
2013). In particular, superimposed bedform genera- form superimposition. However, a critical assess-
tion is prerequisite to the splitting of bedforms that is ment of their relative significance is prevented by a
needed to reduce bedform length (Raudkivi & Witte lack of understanding of the mechanics of superim-
1990; Gabel 1993; Coleman & Melville 1994; posed bedform initiation, amplification, and dissipa-
Warmink et al. 2013). As such, bedform superimpo- tion. Rubin & McCulloch’s (1980) hypothesis on the
sition is a fundamental mechanism by which dunes internal boundary layer has been used to support the
interact when they re-establish a (more) stable pat- comparison of superimposed-bedform development
tern (Venditti et al. 2005b; Kochurek et al. 2010). to development of bedforms on flat beds (Fernandez
Superimposition of bedforms changes the flow et al. 2005; Warmink et al. in press). However, the
field by increasing bed roughness (Van Rijn, 1994) observed alteration of a host’s separated flow by a
and by changing the spatial patterns of momentum superimposed bedform (Fernandez et al. 2005) ought
exchange and energy dissipation within the flow be larger in the inverse: where a host’s separated
(Bennett & Best 1995; Best, 2005). Superimposed flow affects small bedforms in its trough. A logical
bedforms that migrate across the host crest change hypothesis that follows this line of thought is that
the strength and geometry of the host’s separated flow separation, which is considered essential to
flow (Fernandez et al. 2005; Best, 2013). Passage of bedform growth, does not occur in the lee of small
successive bedforms therefore causes systematic forms in the host reattachment zone. Thus, the lo-
fluctuations in the separated flow of the host bed- cality of superimposed bedform initiation lies
form. The superimposition of relatively large super- somewhere in the co-flow region (Jopling, 1961;
imposed bedforms (approx. 25-30% of the host bed- Boersma et al. 1968), and is affected by the charac-
form height) causes reduction of the host bedform teristics of the host’s separated flow. At a larger
lee slope and the creation of reactivation surfaces. scale, flume-like uniform flows, on which our cur-
Both upstream trapping of bedload and active ero- rent understanding of bedform stability is based, are
sion of the host slope by the superimposed bed- highly unusual in natural channels. Non-uniform
form’s separated flow are responsible for the nature flows and secondary circulations are abundant in na-
of such reactivation surfaces (McCabe & Jones ture. Multiple scales of secondary circulation gener-
1977, Allen 1982; Collinson et al. 1980; Reesink & ated by large bedforms, bends, and tributaries are
Bridge 2007, 2009, 2011). Bar-scale deposits com- known to affect bedform stability. It is clear that
monly contain series of reactivation surfaces that are gaps in our understanding of the mechanics of bed-
identified as stacks of inclined sets (Rubin & Hunter form interaction and superimposition undermine a
1982; Haszeldine 1982) with thicknesses that exceed broad range of qualitative and quantitative research
those of near-horizontal sets (Reesink & Bridge on bedforms.
2009, 2011). Such increased thicknesses are ex-
pected in areas of flow deceleration, where high ag-
2 METHODS ment transport and trough scour also decreased with
increasing flow velocities.
In order to improve our understanding of bedform The depth-length characteristics of the trough
development in areas affected by secondary flows, scour varied with flow velocity (Figure 1, label T).
this paper investigates the influence of the separated The scour in the trough of bars increased in depth
flow in the lee of a bar on the hydro-dynamics of and decreased in downstream length with increasing
ripples in its trough. A bar form and was made of flow velocity. Near the re-attachment zone, the di-
plywood with 300 μm sand glued onto it to create a rection of sediment transport associated with distinct
realistic roughness, with a height of 0.15 m and a lee turbulent events is visually isotropic: with sediment
slope of 30 degrees was installed in a 8 m long 0.3 m displacement in all directions appearing equal in
wide Armfield recirculating flume with an adjustable amount and distance. Conversely, in the co-flow
slope. The slope was set at such that the flow over zone, downstream-directed currents and shear visu-
the 2.2 m long, flat bar-top was uniform. Flow ve- ally cause sediment transport to become strongly an-
locity was measured using a Nortek Vectrino II isotropic in its transport direction.
acoustic velocity profiler. The raw velocity meas- The maximum accumulation of the return flow
urements were filtered as outlined in Thomas et al. (Figure 1, label R) is located at a distance away from
(2013). Bed elevations were measured after the wa- the toe of the host lee slope. The decrease in sedi-
ter had been drained from the flume using a laser ment transport towards the toe of the host lee slope
scanner mounted on a robotic traverse. may be attributed to the decrease in transport by tur-
Two sets of experiments are presented here (Table bulent eddies hitting the bed in the reattachment
1) with: (1) a fixed 2D bar form with a mobile zone, by the relatively weak current of the host’s
trough and; (2) a fixed 2D bar form with a fixed rip- separated flow (a clockwise current in Figure 1), and
pled trough. The first set of experiments reproduces by the potential development of an additional current
the development of the scour and ripples in the in the corner where the lee slope joins the horizontal
trough of a fixed 2D bar from flat bed under differ- trough bed (a counter clockwise current in Figure 1).
ent flow conditions. In each experiment, the flume Notably, return-flow ripples that were migrating to-
was run until the deepest part of the trough scour wards the host lee slope did not migrate onto the lee
was about 20 mm below the initial, flat bed level. slope. Observations suggest that the separated flows
The second set of experiments, with a fixed rippled in the lee of ripples migrating towards the host lee
trough, allows the comparison of ripple hydrody- slope increase in strength as they approach the host
namics between locations where they naturally exist lee slope. Instead of migrating onto the lee slope, re-
and locations where they are ordinarily attenuated. turn-flow ripple crests stay in place while aggrada-
tion continues. Note that no sediment was introduced
Table 1. Overview of the experiments presented in into the system from the crest or slope of the bar.
this paper
______________________________________________ The reduced migration of ripples in the return flow
Experiment Morphology Boundary conditions in an area of ongoing deposition leads to the for-
____________ _____________
bar trough Hcrest (m) Ū (m/s) mation of climbing ripples strata that are, in most
______________________________________________ natural situations, eventually buried by the prograd-
Semi-mobile 1 fixed 300µm 0.1 0.6 ing bar lee slope. It is likely that cross-lamination
Semi-mobile 1 fixed 300µm 0.1 0.7
Semi-mobile 1 fixed 300µm 0.1 0.8 formed by ripples climbing onto a bar lee slope
Fixed with ripples fixed fixed 0.1 0.6 (Basumalick et al. 1966; Reesink & Bridge 2011) are
_____________________________________________
only possible in oblique currents.

3 RESULTS 3.2 Fixed bar with a fixed rippled trough

Informed by our first experiments, we installed fixed
ripples between the reattachment zone where no rip-
3.1 Fixed bar with a mobile trough ples are found and the zone in the co-flow where
As flow velocities and sediment transport rates in- ripples were found. Somewhat contrary to what
creased, the location where ripples develop in the might be expected, the geometries of the separated
trough (Figure 1, label S) was located further down- flows in the lee of the ripples were comparable for
stream. This downstream shift in the location of bed- all but one ripple (Figure 2B-G, left column). Thus,
form development contrasts with the stable position the geometry of the time-averaged separated flow
of the trough (Figure 1, label T) and the maximum cells does not appear to correlate with observations
return-flow accumulation (Figure 1, label R). In ad- of ripple development. The only significant excep-
dition to the shift in bedform development, amplifi- tion in the ripples’ separated flow volumes is that of
cation of the incipient ripples relative to the sedi- the second ripple (label C), located directly down-
stream from the reattachment point, which had a
separated flow with a reduced volume. The ripple 4.2 Return-flow accumulation
nearest the bar and the reattachment point (Label B) In contrast to the stable positions of the trough and
had a stronger return-flow compared to the other return-flow accumulation, the width-depth character-
ripples. Here, the ripple crest likely provides more istics of the trough and the location of bedform initi-
sheltering from downstream-directed currents in ation did vary with flow velocity. In the case of
comparison to upstream-directed currents, which in- trough shape, morphological change depends direct-
creases the magnitude of the time-averaged return- ly on sediment transport, and hence on the magni-
flow in the lee of the ripple. Return-flow currents tude of (excess) shear stresses relative to the thresh-
had very low velocities (exceeding -0.03 m/s) and old for motion (Van Rijn, 1994). Thus, the results
are expected to be near or below the threshold for suggests that an increase in velocity over the bar
motion. However, the contrast between the separated crest increases the erosivity of the reattaching flow,
flows and the current over the crest did significantly which locally increases scour and creates a narrow,
increase between successive ripples. Since sediment deep trough. Decreased velocity over the crest is
deposition depends on the spatial gradient of veloci- then associated with a decrease in the erosivity of the
ty, such an increased velocity contrast is a likely reattaching flow and a relatively larger role of sedi-
driver of bedform amplification. Observations of ment transport by coherent flow structures that
sediment transport motion indicated that turbulence sweep sediment across over a larger area.
and sediment transport (an-) isotropy change dramat-
ically between the reattachment zone and the co-
flow where ripples form. The directionality of the 4.3 Isotropy and diffusion in sediment transport
sediment transport is visualised here by the coher-
The observations in these experiments highlight
ence in the directionality of the flow (Figure 2B-G,
how the nature of sediment transport changes from
right column). The measurements show that the di-
isotropic sediment motion in the reattachment zone,
rectionality in the reattachment zone is split evenly
to distinctly anisotropic bursts and sweeps in the co-
(≈ 0.5), and that the flow recovers gradually over
flow zone. This simple observation on changes in
successive ripples. Thus, bedform development de-
the directionality of erosion, transport, and deposi-
pends strongly on gradients of flow velocity and sed-
tion implies that sediment diffusion varies across
iment transport, which require the current to be co-
bedforms and notably so in areas affected by sec-
herent in its direction.
ondary currents. The nature of the transition from
isotropic to anisotropic sediment dispersion between
the trough and stoss slope has a potential role in con-
trolling the three-dimensionality of dune forms.
4 DISCUSSION

4.4 Amplification of bedforms in secondary flows

4.1 Scour and deposition in bar troughs The fixed bed experiment indicates that turbulent
fluctuations in flow direction may play a role in re-
The geometry of the separated flow has proven rela- ducing the coherence of the separated flows in the
tively stable for a range of velocities given a fixed lee of the ripples. No clear trends were found that
bedform geometry (Paarlberg et al. 2007; Unsworth correlate the geometry of the separated flows in the
2013). The results of this investigation show that the lee of ripples directly to bedform amplification.
maximum depth of trough scours and return-flow However, distinct changes were observed in the ve-
accumulations did not vary with flow velocity. The locity gradients between the crests and troughs and
stable position of these maxima can therefore be at- in the coherence in flow direction (Figure 2). These
tributed to the relatively stable geometry of the re- results suggest that turbulent fluctuations reduce
turn-flow, as opposed to its velocity characteristics. bedform stability by reducing the coherence of the
The return-flow accumulation did not attach to, or
separated flows in the lee of the bedforms, whereas
merge with, the host lee slope. This observation downstream-directed shear increases bedform stabil-
complicates traditional views (e.g. Jopling 1961, Al- ity. Thus, at the scale of secondary bedforms, sedi-
len 1982) in which tangential strata can be generated ment transport gradients and directional coherence
by significant sediment transport to the host lee are important for bedform amplification.
slope by the return-flow current. Such tangential The partially-mobile bed experiments (Figure 1)
strata are more likely a product of oblique flows show that the distance to the point of secondary bed-
(Reesink & Bridge 2011). form initiation increases with increasing velocity.
Although the host bar geometry was fixed, the
change in bedform development implies that, at the
host bedform scale, the amplification of superim-
posed bedforms depends on the relative magnitudes
of bedform length (given a fixed height) and the ve- Publishing Company, Developments in Sedimentology 30B,
locity over the crest. 593 p.
The reduction of the separated flows in the lee of Baas, J.H. 1994. A flume study on the development and equi-
the ripples due to the larger-scale secondary flow librium morphology of small -scale bedforms in very fine
patterns also provides some insights into larger-scale sand. Sedimentology, 41, 185-209
interactions of bedforms. For example, the reduc- Basumallick, S. 1966. Size differentiation in a cross-stratified
tions in dune size that are common in meander bends unit. Sedimentology 6 (1) p. 35-68
could be a partial consequence of the effect of the Bennett, S. J. and Best J. L. 1995. Mean flow and turbulence
down-welling helical flow on the separated flows in structure over fixed, two-dimensional dunes: implications
the lee of the dunes. However, the actual magnitude for sediment transport and bedform stability. Sedimentology
of separated flow modification by larger-scale sec- 42(3) p. 491-513
ondary currents remains to be fully quantified. Best J.L. 2005. The fluid dynamics of river dunes: A review
and some future research directions. Journal of Geophysical
Research. 110, F04S02
Best J.L., Blois G., Barros J. & Christensen K. 2013. The dy-
5 CONCLUSIONS namics of bedform amalgamation; new insights from a very
thin flume. Proceedings of the MArine and RIver Dunes
Conference 2013, Bruges, Belgium
A series of simple, adaptable, and easily reproduci- Boersma, J.R., Van de Meene, E.A., Tjalsma, R.C., 1968. Intri-
ble experiments were conducted with i) a fixed bar cated cross stratification due to interaction of a mega ripple
and a mobile trough, and ii) a fixed bar with fixed with its leeside of backflow ripples (upper point bar depos-
ripples in its trough. The fixed bar with mobile its, Lower Rhine). Sedimentology 11, 147–162.
trough experiments show that the distance between Bridge, J.S. 2003, Rivers and Floodplains; Forms, Processes,
the crest and the point of bedform initiation increas- and Sedimentary Record Oxford, U.K., Blackwell Publish-
es with flow velocity. This suggests that bedform ing, 600 p.
superimposition occurs when host bedforms are long Coleman, S.E., Melville, B.W., 1994. Bed form development.
in comparison to the velocity over the crest. Obser- Journal of Hydraulic Engineering 120, 544–560.
vations indicate a change from isotropic sediment Collinson, J.D., 1970. Bedforms of the Tana River, Norway.
transport in the reattachment zone, where sediment Geografiske Annaler, Series A: Physical Geography, v. 52,
transport is equal in all directions, to anisotropic p. 31–56.
transport in the co-flow areas, is a key mechanism in Fernandez, R., Best, J., & Lopez, F., 2006. Mean flow, turbu-
controlling bedform dimensionality. The experi- lence structure, and bed form superimposition across the
ments with a fixed bar and fixed ripples in its trough, ripple–dune transition. Water Resources Research, 42(5), p.
placed in the reattachment zone into the co-flow ar- 948–963
ea, where ripples are formed in the mobile-bed ex- Gabel, S.L., 1993. Geometry and kinematics of dunes during
periments, show that most geometries of the separat- steady and unsteady flows in the Calamus River, Nebraska,
ed flows in the lee of the ripples are reasonably USA. Sedimentology 40, 237–269.
similar. Flow velocity measurements over successive Haszeldine, R.S., 1982. Descending tabular cross-bed sets and
ripples indicate that bedform amplification is more bounding surfaces from a fluvial channel in the Upper Car-
likely controlled by the downstream-increasing ve- boniferous of North-East England. Journal of Sedimentary
locity gradients between crests and troughs of the Petrology V53, p. 1233–1247.
smaller bedforms, and downstream increasing co- Hunter, R.E., & Rubin, D.M., 1983. Interpreting cyclic cross-
herence in the direction of flow and sediment bedding, with an example from the Navajo Sandstone. In
transport. Turbulent fluctuations in flow velocity and Brookfield, M.E., and Ahlbrandt, T.S., eds., Aeolian Sedi-
flow direction are also a likely cause of bedform dis- ments and Processes: Amsterdam, Elsevier, p. 429–454.
sipation. Thus, this study presents a step towards a Jopling, A.V., 1961, Origin of regressive ripples explained in
fuller understanding of bedform initiation and devel- terms of fluid-mechanic processes: U.S. Geological Survey,
opment in areas where complex topography gener- Professional Paper, 424D, p. 15–17.
ates secondary circulations, coherent flow patterns, Kleinhans, M.G., 2002. Sorting out sand and gravel: sediment
and enhanced turbulence. deposition in sand-gravel bed rivers. PhD Thesis Universi-
teit Utrecht, NGS 293, 227 p.
Kocurek, G., Ewing, R.C. & Mohrig, D. 2010. How do bed-
form patterns arise? New views on the role of bedform in-
teractions within a set of boundary conditions. Earth Sur-
6 REFERENCES face Processes and Landforms, 35, 51-63.
Leclair S.F. & Bridge J.S. 2001. Quantitative interpretation of
Allen, J.R.L. 1982. Sedimentary structures, their character and sedimentary structures formed by river dunes: Journal of
physical basis, Volume 2: Amsterdam, Elsevier Scientific Sedimentary Research, v. 71, p. 713–716.
Martin R.L. & Jerolmack D.J., 2013., Origin of hysteresis in
bed form response to unsteady flows. Water Resources Re-
search
McCabe, P.J., & Jones, C.M. 1977. Formation of reactivation
surfaces within superimposed deltas and bedforms: Journal
of Sedimentary Petrology, V47, p. 707–715.
Paarlberg, A.J., Dohmen-Janssen, C.M., Hulscher, S.J.M.H., &
Termes, P., 2007. A parameterization of flow separation
over subaqueous dunes: Water Resources Research, V43,
W12417.
Paola, C. & Borgman, L., 1991. Reconstructing random topog-
raphy from preserved stratification: Sedimentology, v. 38, p.
553–565.
Raudkivi, A.J. & Witte, H.-H. 1990. Development of bed fea-
tures. Journal of Hydraulic Engineering 116, 1063–1079.
Reesink, A.J.H. & Bridge, J.S. 2007 Influence of superimposed
bedforms and flow unsteadiness on formation of cross strata
in dunes and unit bars: Sedimentary Geology, v. 202, p.
281–296.
Reesink, A.J.H. & Bridge, J.S. 2009. Influence of bedform su-
perimposition and flow unsteadiness on the formation of
cross strata in dunes and unit bars—Part 2, further experi-
ments. Sedimentary Geology, v. 222, p. 274–300.
Reesink A.J.H. & Bridge J.S. 2011. Evidence of bedform su-
perimposition and flow unsteadiness in unit bar deposits,
South Saskatchewan River, Canada. Journal of Sedimen-
tary Research
Rubin, D.M. & Hunter, R.E. 1982. Bedform climbing in theory
and nature. Sedimentology, v. 29, p. 121–138.
Rubin, D.M. & McCulloch, D.S. 1980 Single and superim-
posed bedforms—a synthesis of San Francisco Bay and
flume observations. Sedimentary Geology, v. 26, p. 207–
231.
Thomas, R.E. McLelland S.J. Frostick L.E. 2013. The Impact
of Macroalgae on Mean and Turbulent Flow Fields. Pro-
ceedings of 2013 IAHR World Congress.
Unsworth, C.A., Parsons D.R., Reesink A.J.H., Best J.L., Ash-
worth P.J., & Hardy R.J. 2013 Flow structures over fixed
2D bedforms in transient states. Proceedings of the MArine
and RIver Dunes Conference 2013, Bruges, Belgium
Van Rijn, L.C. 1994. Principles of fluid flow and surface waves
in rivers, estuaries, seas and oceans. Aqua Publications,
Amsterdam.
Venditti, J.G., Church, M.A. & Bennett, S.J. 2005a. Morpho-
dynamics of small-scale superimposed sandwaves over mi-
grating dune bedforms. Journal of Geophysical Research
110, F01009.
Venditti, J.G., Church, M.A. & Bennett, S.J. 2005b. On the
transition between 2D and 3D bedforms, Sedimentology,
52, 1343–1359,
Warmink, J.J., Dohmen-Janssen, C.M., Lansink, J. Naqshband,
S., Van Duin, O.J.M., Paarlberg, A.J., Termes, A.P.P., &
Hulscher S.J.M.H. In Press. Understanding river dune split-
ting through flume experiments and analysis of a dune evo-
lution model. Earth Surface Landforms and Processes
Figure 1: Digital Elevation Models (axis in metres, elevation contours in millimeters) of three experimental runs with a fixed bar,
mobile trough, and flow velocities over the crest of approximately (A) 0.6, (B) 0.7, and (C) 0.8 m/s. The DEM’s are coloured by the
amount of deposition (in mm). The grey profiles visualise the 0.3 x 0.3 m average scour and deposition relative to the initial hori-
zontal bed. Note the difference in trough shapes. S denotes the visible start of the co-flow ripple formation, which varies between the
runs, T is the trough, R is the maximum return-flow accumulation.
Figure 2: Fixed bar with fixed ripples placed in the reattachment zone, where ripple development is normally suppressed, of the co-
flow zone where ripples formed in the mobile bed experiments. Figures B-G are flow profiles over the crests and troughs of the suc-
cessive ripples and represent the velocity profiles (left column, coloured by downstream velocity) and the coherence in current direc-
tion (proportion of velocities directed downstream).

View publication stats