Maarten Kleinhans | Utrecht University (original) (raw)

Papers by Maarten Kleinhans

Fluvial meander belt sediments form some of the most architecturally complex reservoirs in hydroc... more Fluvial meander belt sediments form some of the most architecturally complex reservoirs in hydrocarbon fields due to multiple scales of heterogeneity inherent in their deposition. Currently, characterization of meander belt bodies largely relies on idealized vertical profiles and a limited number of analogue models that naively infer architecture from active river dimensions. Three-dimensional architectural data are needed to quantify scales of grain-size heterogeneity, spatial patterns of sedimentation and bar preservation in a direct relationship with the relevant length scales of active river channels. In this study, three large flume experiments and a numerical model were used to characterize and construct the architecture (referred to as 'archimetrics') and sedimentology of meander belt deposits, while taking reworking and partial preservation into account. Meander belt sandbody width-to-thickness ratios between 100 and 200 were observed, which are consistent with reported values of natural meander belts. For the first time, the relief of the base of a meander belt is quantified, enabling improved estimates of connectedness of amalgamated meander belts. A key observation is that the slope and number of lateral-accretion packages within natural point bar deposits can be well predicted from fairly basic observ-ables, a finding subsequently tested on several natural systems. Probability curves of preserved architectural characteristics for three dimensions were quantified allowing estimates of bar dimensions, baffle and barrier spacing distributions and container dimensions. Based on this, a set of rules were identified for combining reservoir parameters with the identified probability curves on sand-body dimensions and character, to help create more realistic geomodels for estimating exploration success on the basis of seismic and core data.

Journal of Marine Science and Engineering, 2015

PLOS ONE, 2015

River biota are affected by global reach-scale pressures, but most approaches for predicting biot... more River biota are affected by global reach-scale pressures, but most approaches for predicting biota of rivers focus on river reach or segment scale processes and habitats. Moreover, these approaches do not consider long-term morphological changes that affect habitat conditions. In this study, a modelling framework was further developed and tested to assess the effect of pressures at different spatial scales on reach-scale habitat conditions and biota. Ecohydrological and 1D hydrodynamic models were used to predict discharge and water quality at the catchment scale and the resulting water level at the downstream end of a study reach. Long-term reach morphology was modelled using empirical regime equations, meander migration and 2D morphodynamic models. The respective flow and substrate conditions in the study reach were predicted using a 2D hydrodynamic model, and the suitability of these habitats was assessed with novel habitat models. In addition, dispersal models for fish and macroinvertebrates were developed to assess the re-colonization potential and to finally compare habitat suitability and the availability/ability of species to colonize these habitats. Applicability was tested and model performance was assessed by comparing observed and predicted conditions in the lowland Treene River in northern Germany. Technically, it was possible to link the different models, but future applications would benefit from the development of open source software for all modelling steps to enable fully automated model runs. Future research needs concern the physical modelling of long-term morphodynamics, feedback of biota (e.g., macrophytes) on abiotic habitat conditions, species interactions, and empirical data on the hydraulic habitat suitability and dispersal abilities of macroinvertebrates. The modelling framework is flexible and allows for including additional models and investigating different research and management questions, e.g., in climate impact research as well as river restoration and management.

Coastal Engineering 2008 - Proceedings of the 31st International Conference, 2009

Existing sand transport models are primarily based on data from oscillatory flow tunnel (OFT) exp... more Existing sand transport models are primarily based on data from oscillatory flow tunnel (OFT) experiments. However, theory and former experiments indicate that flow differences between full scale surface waves and OFT's may have a substantial effect on the net sand ...

Coastal Dynamics 2005 - Proceedings of the Fifth Coastal Dynamics International Conference, 2006

ABSTRACT

Proceedings of the 5th IAHR Symposium on River, Coastal and Estuarine Morphodynamics, Enschede, NL, 17-21 September 2007, 2007

The prediction of suspended sediment concentration and transport by combined currents and waves i... more The prediction of suspended sediment concentration and transport by combined currents and waves is hampered by a lack of data in energetic conditions and in sediment mixtures. Our aim is to improve predictors for the entire range of no motion to high wave and current energy. We measured hydrodynamics, sediment transport and bedform dimensions and shapes on an ebb-tidal delta with current velocities up to 1 m/s and orbital velocities up to 1.2 m/s over a very poorly sorted sandy gravel sediment with a mean grain size of 0.3-0.6 mm. In varying conditions symmetrical vortex ripples or asymmetrical current dunes occurred. Our proposed model consists of shear stress predictors, diffusivity profiles for currents and waves, and a reference concentration predictor. In wave-only conditions the concentration decays exponentially as expected, but in current-only conditions the diffusivity is much larger than expected. This could not be explained by finer sediment suspended higher above the bed. We hypothesise that the effective mixing is enhanced by the short disequilibrium bedforms. A correction for this effect considerably improved concentration predictions. In combined currents and waves the concentrations were best predicted for a linear addition of wave-and current-related diffusivities in contradistinction to the common squared addition verified for lower-energy conditions. The reference concentrations were best reproduced by the method of Van Rijn 1984 which has a more conservative estimate of orbital bed shear stress than alternative formulations.

Proceedings of the 5th IAHR Symposium on River, Coastal and Estuarine Morphodynamics, Enschede, NL, 17-21 September 2007, 2007

We present a proof of principle for using a recent model for grain size-sorting over a bed form l... more We present a proof of principle for using a recent model for grain size-sorting over a bed form lee face to reconstruct the bed form height from a partially preserved dune or delta deposit. The lee face sorting model was developed to describe the grain size-selective deposition of particles over a bed form lee face for use in a new type of stochastic sediment continuity model for conditions with nonuniform sediment and river dunes. The lee face sorting model is applied to variable elevation ranges of the measured sorting in dunes and deltas of which the height is known but not used in the model predictions. We find that reliable estimates of the bed form height and the composition of the lee face deposit can be derived, under the conditions that more than 50% of the original bed form height has been preserved and sufficient samples of the bed composition over this range are taken. The estimate of the bed form height is a lower limit, as the method shows a systematic underestimation of the original bed form height due to the sorting being stronger over the lower part of a bed form than over its upper part. The sorting model can thus be applied to estimate bed form height and bed load sediment composition from vertical sorting profiles collected from individual cross-sets of partially preserved bed form deposits. Bed form height and sediment composition are robust indicators for palaeoflow conditions.

Proceedings of the International Conference on Fluvial Hydraulics, Lisbon, Portugal, 6-8 September 2006, 2006

River dunes, alternate bars, and subaqueous deltas are often characterised by a lee face at about... more River dunes, alternate bars, and subaqueous deltas are often characterised by a lee face at about the angle of repose. Transport of mixed-size sediment over the crest of such a bed form results in grain sizeselective deposition of particles over the lee face. The complex interaction between two different deposition mechanisms, i.e. grain fall from suspension and grain flows at nearly the angle of repose, results in coarse particles preferentially depositing at the lower elevations of the lee face, and fine particles at the upper ones. In this study we develop a simple empirical model describing the sorting of particles over the lee face of a bed form. This lee face sorting model is calibrated and verified on data from flume experiments with deltas and dunes. Field data of lee face sorting over dunes in an ebb-tide dominated mouth bar of the river Teign, UK, are used to validate the model. The lee face sorting model describes the grain size-selective deposition over a lee face well, although the model does not reproduce the deposition of fines from suspended load which slightly modifies the sorting by producing a fine deposit at the base of the bed forms.

Chaos suggest large scale collapse and subsidence (1500 m) of the entire area, which is consisten... more Chaos suggest large scale collapse and subsidence (1500 m) of the entire area, which is consistent with a massive expulsion of liquid water from the subsurface in one single event. Morphological analysis of Aram Valley suggests a complex process starting with the outflow of water from two small channels, continuous groundwater sapping and headward erosion ending with catastrophic lake rim collapse and carving of the Aram Valley, synchronous with the 2.5 Ga stage of Ares Vallis formation. The water volume and the formative time scale needed to carve the Aram channels indicate that a single, rapid (maximum tens of days) and catastrophic (flood volume of 9.3·10 4 km 3 ) event carved the outflow channel. We conclude that the subice lake collapse scenario can explain the features of Aram Chaos -Valley system as well as the time scale needed for its formation.

Bifurcations (also called diffluences) are as common as confluences in braided and anabranched ri... more Bifurcations (also called diffluences) are as common as confluences in braided and anabranched rivers, and more common than confluences on alluvial fans and deltas where the network is essentially distributary. River bifurcations control the partitioning of both water and sediment through these systems with consequences for immediate river and coastal management and long-term evolution. Their stability is poorly understood and seems to differ between braided rivers, meandering river plains and deltas. In particular, it is the question to what extent the division of flow is asymmetrical in stable condition, where highly asymmetrical refers to channel closure and avulsion. Recent work showed that bifurcations in gravel bed braided rivers become more symmetrical with increasing sediment mobility, whereas bifurcations in a lowland sand delta become more asymmetrical with increasing sediment mobility. This difference is not understood and our objective is to resolve this issue. We use a one-dimensional network model with Y-shaped bifurcations to explore the parameter space from low to high sediment mobility. The model solves gradually varied flow, bedload transport and morphological change in a straightforward manner. Sediment is divided at the bifurcation including the transverse slope effect and the spiral flow effect caused by bends at the bifurcation. Width is evolved whilst conserving mass of eroded or built banks with the bed balance. The bifurcations are perturbed from perfect symmetry either by a subtle gradient advantage for one branch or a gentle bend at the bifurcation. Sediment transport was calculated with and without a critical threshold for sediment motion. Sediment mobility, determined in the upstream channel, was varied in three different ways to isolate the causal factor: by increasing discharge, increasing channel gradient and decreasing particle size. In reality the sediment mobility is mostly determined by particle size: gravel bed rivers are near the threshold for sediment motion whereas sand bed rivers have highly mobile sediment at channel-forming conditions. For sediment transport without a critical threshold for motion, bifurcations become more asymmetrical with increasing sediment mobility. In contrast, sediment transport prediction including the threshold for motion leads to highly asymmetrical bifurcations for low sediment mobility, more symmetrical bifurcations for higher mobility and again decreasing symmetry for higher mobility where results of transport with and without the threshold converge. Thus, the general trend is more asymmetrical bifurcations for higher sediment mobility, but the presence of the threshold for motion leads to an optimum in symmetry. Results were similar for the different options used to vary mobility, excluding first-order effects of backwater adaptation length and hydraulic roughness. We conclude that the seemingly conflicting results between gravel-bed and sand-bed rivers in literature are well explained by the difference in sediment mobility.

The sedimentary dynamics of a lowland river system can be defined as the whole of processes that ... more The sedimentary dynamics of a lowland river system can be defined as the whole of processes that involve erosion, transport or deposition of sediment in the systemincluding the floodplains -on every possible spatial and temporal scale. The sedimentary dynamics of the river Rhine in the Netherlands have been influenced by many human activities in the past. A major human intervention has been the embankment around 1300 AD, by which the so far free meandering river was confined between dikes. Previous studies analyzed the impact of human interventions in the period after the embankment. The present study focuses on the sedimentary dynamics of the natural river Rhine. We reconstructed the amount of sedimentation and erosion that is involved with lateral migration of the river, and estimated the residence time of the sediment on the floodplain. Lateral migration was reconstructed with a one-dimensional bank erosion model. Lateral migration rates of meander bends were modeled based on channel parameters (e.g. bend radius, discharge) and bank material, all derived from detailed borehole reconstructions of sub-recent channel belts from previous studies. Model results will be used to compare the sedimentary dynamics of the natural river Rhine, with dynamics in human-influenced periods, and hence to predict present and future sedimentation and erosion processes. Moreover, results of reworked sediment volumes can be included in existing large-scale sediment budgets of the river Rhine to evaluate the dynamics within these slices.

Bedform development and distribution on the lower shoreface (14-18 m LLWS) of the central Dutch c... more Bedform development and distribution on the lower shoreface (14-18 m LLWS) of the central Dutch coast differ markedly between fair-weather and storm conditions. Modern observations of bedforms on multibeam sonar images confirm that unidirectional tide-dominated currents rework the lower shoreface during fair-weather conditions with significant waveheights below 2.5 m. Several Spring-Neap cycles produced straight-crested 2D-megaripples as the dominant bedform in two research areas. The megaripples are very similar to those generally observed below wave base on the inner shelf. Even minor seasonal storms (significant waveheights 3-4 m) produce an entirely different type of bedform distribution with round-crested 3D 'hummocky' bedforms (wavelength 20-40 m) and 3D-megaripples (spacing 5-8 m and 12-14 m). The bedforms observed here in a seabed of medium sands differ from those characteristic of fine sand(stone)s. Spatial differences in bedform development are partly attributed to feedback mechanisms between hydrodynamics and wave damping by the tube worm Lanice conchilega.

Geophysical Research Abstracts, 2014

For the prediction of river bed destabilisation and fractional sediment transport of mixtures, we... more For the prediction of river bed destabilisation and fractional sediment transport of mixtures, we aim to solve two problems that are poorly understood. First, the flow and pressure fluctuations surrounding both the embedded and exposed particles must be parameterised for hydraulically smooth to rough flow. Second, an adequate relation between particle size and particle exposure should be based on the particle size distribution and the (water-worked) bed structure. We use a recently developed force balance model for the threshold of motion of uniform sediments incorporating the effects of particle exposure, pressure fluctuations into the bed, very shallow flow and bed slope. The flow module is extended to non-uniform roughness of sediment mixtures. Our extended model predicts the critical Shields values of arbitrary mixtures directly as function of exposure and no longer needs empirical hiding-exposure relations. Several empirical and geometrical relations between particle size and exposure were tested. The results are compared to extensive datasets from the literature of incipient fractional transport rates. The modelled hiding-exposure relations are very sensitive to the relation between particle size and exposure, which differ for unimodal, skewed and bimodal mixtures. This is explained by the pore structure of these sediments. The existing relations fail particularly for the smaller particles in bimodal and skewed distributions. These small particles percolate through the pores so their exposure or embedding strongly depends on the fractional content and pore structure, in agreement with empirical data. We are working on a universal relation for exposure containing particle size distribution, pore structure and water-working. The model reproduces data of uniform sediments well for the entire physically possible range of particle exposures and for hydraulically rough to nearly smooth conditions. Trends in existing data for mixtures are also reproduced but depend strongly on exposures that were not measured.

ABSTRACT The evaluation of the water volume and the formative time scale needed to carve the outf... more ABSTRACT The evaluation of the water volume and the formative time scale needed to carve the outflow channels repre- sents a fundamental process for the validation of their evolutive models. We calculate these attributes for the Aram channels and we compared the results with the volume of liquid water that was produced in a single chaotization event of the Aram Chaos. The analysis suggests that a single rapid and catastrophic event is sufficient to carve the channel and the volume of flood is compatible with the volume of liquid water release in a single chaotization event of the Aram Chaos.

Icarus, 2016

The quasi-circular collapsed landforms occurring in the Chryse region of Mars share similar morph... more The quasi-circular collapsed landforms occurring in the Chryse region of Mars share similar morpholog- ical characteristics, such as depth of collapse and polygonally fractured floors. Here, we present a statis- tical analysis of diameter, maximum and minimum depth, and amount of collapse of these features. Based on their morphometric characteristics, we find that these landforms have a common origin. In par- ticular, the investigated landforms show diameter-depth correlations similar to those that impact craters of equivalent diameters exhibit. We also find that the observed amount of collapse of the collected fea- tures is strongly correlated to their diameter. Furthermore, the linear relation between minimum filling and pristine depth of craters, the constant ratio between collapse and the amount of filling and the frac- tured and chaotic aspect of the filling agree with melting and subsequent collapse of an ice layer below a sediment layer. This interpretation is consistent with a buried sub-ice lake scenario, which is a non-climatic mechanism for producing and storing abundant liquid water under martian conditions.

Proceedings of the 5th IAHR Symposium on River, Coastal and Estuarine Morphodynamics, Enschede, NL, 17-21 September 2007, 2007

Journal of Marine Science and Engineering, 2015

PLOS ONE, 2015

Coastal Engineering 2008 - Proceedings of the 31st International Conference, 2009

Coastal Dynamics 2005 - Proceedings of the Fifth Coastal Dynamics International Conference, 2006

ABSTRACT

Proceedings of the 5th IAHR Symposium on River, Coastal and Estuarine Morphodynamics, Enschede, NL, 17-21 September 2007, 2007

Proceedings of the International Conference on Fluvial Hydraulics, Lisbon, Portugal, 6-8 September 2006, 2006

Geophysical Research Abstracts, 2014

Icarus, 2016

Proceedings of the 5th IAHR Symposium on River, Coastal and Estuarine Morphodynamics, Enschede, NL, 17-21 September 2007, 2007

Meandering rivers have partly been explained from theories that predict bend and bar formation an... more Meandering rivers have partly been explained from theories that predict bend and bar formation and migration, whilst assuming a simplistic erosion formulation for the bank opposite to where
bars form. However, floodplains are highly heterogeneous in composition and vegetation density and bank erosion consists of a set of complicated processes. Our objective is to understand the formation and heterogeneity of experimental self-formed floodplains and test a forced bar theory on the self-formed meandering channel. We did experiments in a 1.25x7.5 m flume with an initially straight and narrow channel in a bed of river sand and silica flour. We measured the evolving channel location and the resulting stratification within the floodplain. Our experimental and theoretical alternate bar wave length agree within 25 percent. Despite the simplicity of the setup and the limited dynamics of meandering, the stratification in the floodplains was quite complicated, reflecting a rich history of stratification, suspended sediment deposition onto the floodplains and unexpected erosion processes in the cohesive floodplain cover. The latter affected channel planform formation and success of chute cut-offs. We conclude that the bar theory provides a reasonable general estimate of bar length, but the exact channel location and curvature were affected by the heterogeneity of bank erodibility.

River floods are known to have large impacts on fluvial morphology as the capacity to carry water... more River floods are known to have large impacts on fluvial morphology as the capacity to carry water and rework sediment during these events is large. However, recent experimental findings are conflicting: some suggest that varying discharge contributes to a more single-thread pattern whereas others suggest that discharge variations cause multiple threads to be active, and yet others show no significant effect on the morphology. Our objective is to study the effect of varying discharge on experimental river patterns with otherwise similar conditions, and to quantitatively compare the resulting morphology and deposits. Our experiments were conducted in a flume of 10x6 meter, which was split up into two separate fluvial plains (each 10x3 m). Fluvial landscape evolution was recorded by high-resolution line-laser scanning and digital Single Lens Reflex (SLR) camera used for channel-floodplain segmentation and particle size estimation. The bed sediment consisted of a poorly sorted sediment mixture ranging from fine sand to fine gravel. First, a braided and meandering river pattern evolved for identical and constant boundary conditions, except that slightly cohesive silt-sized silica flour was added to the feed sediment of the meandering channel. A second set of experiments had an identical cycled discharge regime with a long-duration low flow and a short-duration high flow.The varying discharge largely affected the fluvial landscape by biasing the morphology towards the high flow conditions. This was reflected by an increase of the bar wave length with nearly a factor 2. Also, the depth of maximum erosion increased, which affects the preservation potential. The meandering and braided patterns responded differently to the floods. The noncohesive sediment combination with varying discharge results in a higher degree of braiding when compared to constant discharge. This was observed as a higher number of re-activating channels during high flow. In contrast, the silica flour acted as floodplain builder, which was more efficiently distributed during floods. As a result, the system with slightly cohesive sediment remained mostly confined to one migrating meandering channel that developed scroll bars, channel fills, splays and levees. We conclude that the response to varying discharge depends on the availability and cohesion of fine floodplainforming sediment in combination with the potential of high flows to re-activate residual channels.

Preservation is the link between fluvial surface morphodynamics and what is recorded in the fluvi... more Preservation is the link between fluvial surface morphodynamics and what is recorded in the fluvial sedimentary. Reconstruction of the original channel morphology from stratification can provide important information about paleoflow conditions. To infer the original dimensions of paleomorphological features such as river channels from the fluvial record, a detailed understanding of the relation between morphodynamics and preservation is needed. So far, theories to reconstruct the original morphology from preserved stratification have not been tested for meandering river channels for lack of detailed bathymetry.

We report on a series of controlled flume experiments and Delft3D physics-based numerical model runs with the objectives to i) test the prediction of set thickness as a function of the morphology formed by a meandering river channel, and ii) explore and explain spatial and temporal set thickness variations in the resulting channel belt. High-resolution measurements of time-dependent surface elevation were used to quantitatively relate the preserved stratification to the river morphology. Experimental design corresponds to the predicted hydraulic geometry for a non-cohesive gravel-bed river, and the width-depth ratio is chosen such that alternate bars form.

We find that the mean set thickness agrees well with the theoretical prediction from channel morphology. The mean preserved set thickness is 30% of the mean channel depth. Finally, there is much systematic spatial variation in set thickness related to repetitive point bar growth and chute cutoff. We find undisturbed and thick sets close to channel belt margins and more irregular stratification with stacked thinner sets in the channel belt center. We conclude that set thickness statistics can be used to provide quantitative error bounds for the reconstruction of paleochannel dimensions.

Strong feedbacks exist between river channel dynamics, floodplain development and riparian vegeta... more Strong feedbacks exist between river channel dynamics, floodplain development and riparian vegetation. Several experimental studies showed how uniformly sown vegetation causes a shift from a braided river to a single-thread and sometimes meandering river. The objective of this study is to test what the effect of fluvially distributed seeds and vegetation settling is on channel pattern change and channel dynamics.

The experiments were carried out in a flume of 3 m wide and 10 m long. We tested where the vegetation deposited in a braided and meandering river and how the morphology changed. We used a simple hydrograph of 0.25 hour high flow and 3.75 hour low flow, where alfalfa seeds were added during high flow. The bed sediment consisted of a poorly sorted sediment mixture ranging from fine sand to fine gravel. The evolution was recorded by a high-resolution laser-line scanner and a Digital Single Lens Reflex (DSLR) camera used for channel floodplain segmentation, water depth approximation and vegetation distribution.

In an initially braided river, vegetation settled on the higher banks and stabilized the banks. In an initially meandering river, vegetation settled in the inner scrolls, and also on the outer banks when water level exceeded bankfull conditions. In agreement with earlier work, the outer bank was stabilized; erosion rate decreased and bends became sharper. The inner bend vegetation stabilized a part of the point bar and hydraulic resistance of the vegetation steered water in the channel and to the non-vegetated part of the inner bend. As result the meander bend became braided as water flows along the vegetation. Vegetation formed patches that grew over time and reduced channel dynamics. We conclude that self-settling vegetation decreased local bank erosion and that vegetated islands leads to a multi-thread system instead of single-threaded.

In dynamically meandering rivers with relatively high stream power such as the river Allier (Fra... more In dynamically meandering rivers with relatively high stream power such as the river Allier (France) chute cutoffs occur frequently. The conditions for chute initiation and closure are not necessarily related to floods as meander experiments with constant discharge exhibit cyclic chute cutoff formation and simultaneous bend abandonment (Van Dijk et al., 2012). Our objective is to determine the necessary conditions for initiation and formation of a chute cutoff. We compare Delft3D physics-based model results with observations of chute formation in the Allier and flume experiments.

We used a rectangular grid with two immersed boundary polygons representing the river banks. Along these bank lines, bank erosion is computed (Jagers et al., 2011). Three bends in the river Allier were modeled with and without initial chute channels across the point bars, in order to test the stability of the chute channels. The model results for flow and sediment partitioning at the inlet of the chute channel were analyzed in the framework of river bifurcation stability.

Aerial photographs of the river Allier show that the chute cutoff was rapidly abandoned when upstream bend curvature changed. This is consistent with numerical simulations which initial calculations suggest that chute formation depends on the direction and curvature of the flow just upstream of the potential chute location. Furthermore, the relative channel lengths and gradients along the meander and along the chute channel are important. In agreement with flume experiments, we observed that upstream bend curvature and dynamics were important to develop and maintain a single-thread channel after the occurrence of chute cutoffs.

Jagers, B., Spruyt, A. and Mosselman, E. (2011), How to include steep bank retreat in 2D/3D morphological models? , AGU Fall Meeting Abstracts, 1, 0655.
Van Dijk, W.M., Van de Lageweg, W.I. and Kleinhans, M.G. (2012), Experimental meandering river with chute cutoffs. J. Geophys. Res. doi:10.1029/2011JF002314, in press.

"1. Introduction Rivers have distinctive patterns such as multi-channel braided and single-chann... more "1. Introduction
Rivers have distinctive patterns such as multi-channel braided and single-channel meandering. Why these different river patterns emerge is only qualitatively understood. Many fundamental fluvial processes are relatively well understood, but their combined effect on river patterns remains hypothetical. Braided rivers emerge when banks are weak, whereas meandering rivers emerge for stronger banks (Kleinhans 2010). Vegetated banks and cohesive floodplains provide bank strength, which affects the channel width-depth ratio. Rivers self-organize their morphology through interactions between channels, floodplain and vegetation. In order to predict the migration and bank/floodplain development of rivers, we try to understand these interactions and to understand how this ultimately leads to different river patterns. The objective of this study is to obtain insights on interactions between channel dynamics, floodplains and vegetation by conducting flume experiments.

2. Effect of channel dynamics
The development of alternate bars and meander bends can be explained from simplified analytical physics, i.e. linear stability analysis. The analytical solution can be divided in a bar theory and a bend theory, which was tested in several experiments.

2.1 Bar and bend theory
According to the linear stability analysis, meandering channels can theoretically arise from a straight channel (Blondeaux and Seminara, 1985). The unstable river bed leads to formation of alternate bars when width-depth ratios are low, while mid-channel bars form for high width-depth ratios. The alternate bars induce alternate erosion of the banks and deposition at the opposite side of the channel. In experiments starting with a straight channel, we indeed observed the development of alternate bars, bending of the channel and lateral expansion towards maximum amplitude followed by chute cutoff and channel straightening.

2.2 Upstream perturbation
The dynamics of a meandering river are determined by the nature of the bend instability. Theoretically, this instability could be absolute or convective. Lanzoni and Seminara (2006) found that bars and channel curvature for meandering rivers mostly convect in one direction, usually downstream due to low width-depth ratios. This implies that, to sustain meandering dynamics, the upstream boundary must continuously be perturbed, so that instabilities continue to propagate from the upstream boundary in downstream direction. In an experiment we showed that a static perturbation led to low amplitude meanders that were static in the sense of no lateral migration, while an experiment with a transverse moving boundary produced series of scroll bars and infrequent bend cutoffs in downstream migrating sinuous meanders (Van Dijk et al., 2012).

3. Effect of floodplain construction and destruction
Channel dynamics determined the development of meander bends in the channel, while lack of floodplain formation resulted in chute cutoffs. We tested the effect of vegetation as well as cohesive fines on floodplain construction and destruction.

3.1 Vegetated floodplains
Bar and bend theory and empirical relations showed that bank stability is important for the development of a meandering river. Bank strength should be sufficient to keep low-width depth ratios, in order to prevent the channel widening and braiding. Riparian vegetation has the effect of hydraulic resistance and bank strength and it increases bar sedimentation. Prior experiments showed that the addition of vegetation (Alfalfa) on a braided river resulted in bank stabilization and the development of a single-thread channel (Tal and Paola, 2010). The addition of vegetation has also been shown to result in sedimentation of fine material in chutes and decreased the number of cutoffs (Braudrick et al., 2009). We performed experiments in which vegetation was uniformly distributed on the floodplain of a meandering channel. This resulted in bank stabilization and the development of meander bends that were tighter compared to a control experiment without vegetation (Van Dijk et al., submitted b). Furthermore, we tested the effect of vegetation seeds that are distributed by the flow during floods. The vegetation developed on lower areas and hydraulic resistance of the stems resulted in flow diversion, which resulted shallow water depths over bars and floodplains and decreased channel dynamics. Vegetation in the experimental setup was successful in bank stabilization, but the development of new vegetation is limited to continuous manual distribution of seeds and not dependent on the feedbacks between channel dynamics and floodplain formation.

3.2 Cohesive floodplains
Bank erosion by bend migration and cutoff is balanced by deposition of sediment forming new floodplains. The process of floodplain shaving and channel extension results in local differences between erosion and deposition (Lauer and Parker, 2008). Overbank deposits are an important floodplain forming process that balances these local differences. We conducted an experiment where the initial bed consisted of poorly sorted sand. The experiment had a simple hydrograph of bankfull discharge and discharge exceeding bankfull. During the higher discharge slightly cohesive fines were added in the feed, so that these fines could deposit in the lows of the floodplain. The addition of fines in the experiment and the simple hydrograph led to an experimental meandering river with higher sinuosity and less chute cutoffs. The fines deposited in the lows of the point bar in the inner bend and formed a levee/ splay on the outer bank where flow diverted from the channel to the floodplain (Van Dijk et al., 2013). The cohesive floodplain in this experiment decreased bank erosion and chute incision compared to the experiment without the addition of fines.

4. Conclusions
The formation of meanders is controlled by feedbacks between channel dynamics and floodplains. We conclude that channel dynamics are controlled by the bankfull conditions, while overbank flow is important in constructing and destructing floodplains. Results show that:
• In-channel morphology depends on the width-depth ratio that determines the initiation of alternate bars.
• With a static perturbation bend growth lead to low amplitude bend and are static in lateral migration.
• Experiments show that when upstream perturbation is dynamic a richer morphology than hitherto is produced, including cycles of meander growth and chute cutoff.
• Riparian vegetation stabilizes banks, sustaining low width-depth ratios, which results in local erosion and tight bends.
• The interaction of flow exceeding bankfull discharge is important in developing floodplains that balance bank erosion.
• Cohesive sediment results in distinctive floodplains that sustain dynamic meandering, whereas vegetation increases hydraulic resistance and reduces channel dynamics.
"

Flume experiments have shown that uniformly distributed vegetation causes a shift from a braided ... more Flume experiments have shown that uniformly distributed vegetation causes a shift from a braided river to a single-thread and sometimes meandering river. The vegetation reduces channel width-depth ratios by increasing floodplain flow resistance and bank strength which changes the bar pattern from braid bars to alternate bars. In nature, many riparian species are not uniformly distributed and in addition to stabilising the banks, patches of vegetation also steer the flow on the floodplain. The objective of this study was to experimentally test the effect of vegetation, in particular distribution, density and growth duration, on the pattern and dynamics of an experimental meandering river. The experiments were carried out in a flume of 3 m wide and 10 m long. We used additional small-scale tests to quantify the effect of vegetation on flow resistance and resultant flow velocities. Our earlier work isolated the effects of vegetation density and age on bank erosion. The small tests showed that flow resistance of vegetation decreased floodplain flow which resulted in higher flow velocities in the channel. In agreement with earlier work the uniformly distributed vegetation (representing wind dispersed seeds) decreased bank retreat, increased flow resistance, stabilized banks and resulted in tighter bends. However, in the experiment with seeds distributed by the flow during floods (hydrochory), point bars were stabilized and vegetation patches diverted the flow which resulted in shallower floodplain flow and less bank erosion. We conclude that patchy vegetation reduces flow velocity in the channel as well as on the floodplain, which reduces the dynamics of the channel and modifies the river pattern. This implies that the combined variables of discharge variation and vegetation settling behavior have a large effect on morphology and dynamics of rivers.

Meandering rivers have partly been explained from theories that predict bend and bar formation an... more Meandering rivers have partly been explained from theories that predict bend and bar formation and migration, whilst assuming a simplistic erosion formulation for the bank opposite to where bars form. However, floodplains are highly heterogeneous in composition and vegetation density and bank erosion consists of a set of complicated processes. Our objective is to understand the formation and heterogeneity of experimental self-formed floodplains and test a forced bar theory on the self-formed meandering channel. We did experiments in a 1.25x7.5 m flume with an initially straight and narrow channel in a bed of river sand and silica flour. We measured the evolving channel location and the resulting stratification within the floodplain. Our experimental and theoretical alternate bar wave length agree within 25 percent. Despite the simplicity of the setup and the limited dynamics of meandering, the stratification in the floodplains was quite complicated, reflecting a rich history of stratification, suspended sediment deposition onto the floodplains and unexpected erosion processes in the cohesive floodplain cover. The latter affected channel planform formation and success of chute cut-offs. We conclude that the bar theory provides a reasonable general estimate of bar length, but the exact channel location and curvature were affected by the heterogeneity of bank erodibility.

Through bar theory it is known that channel width-depth ratio affects bar pattern and that relati... more Through bar theory it is known that channel width-depth ratio affects bar pattern and that relatively narrow channels with strong banks are required for meandering. Riparian vegetation is able to alter the channel width-depth ratio and therefore the channel pattern. However, determination of the underlying mechanisms and processes has remained scarce and qualitative and hence these effects are not yet fully understood. The objective of this study is to experimentally determine the effects of riparian vegetation on bank strength, channel pattern and meandering dynamics. Bank erosion and -failure experiments were performed with four different vegetation species to quantify the strength introduced by plant roots at the experimental scale. To study pattern evolution and morphodynamics we used a 1.25x7.5 m flume with a constant discharge and sediment feed. The introduction of vegetation in experiments results in narrower and deeper channels. Higher vegetation density leads to static channels with occasional sharp bends. Vegetation increases the threshold for erosion and failure of bank material as indicated by higher measured cohesion values and lower bank erosion rates. We conclude that a low vegetation density is crucial to maintain a dynamic fluvial system.

Braided rivers are relatively easily formed in the laboratory, whereas self-formed meandering riv... more Braided rivers are relatively easily formed in the laboratory, whereas self-formed meandering rivers in the lab have proven very difficult to form, indicating a lack of understanding of the necessary and sufficient conditions for meandering. Our objective is to create self-formed dynamic meandering rivers and floodplains in a laboratory. Early experiments attempted to initiate meandering with upstream inflow at a fixed angle different from the general flow direction. The resulting bends were fixed at one position, which is not the dynamic meandering observed in nature. Another important condition for meandering is to have banks stronger than the non-cohesive bed sediment, which has been attained by growing vegetation. Furthermore, finer or light-weight sediment has been used to let chute channels fill up where otherwise multi-thread channels would have evolved, which is braiding. Yet the fixed-angle inflow kept meander migration and channel belt width and complexity limited. We accomplished dynamic meandering in the laboratory by using an upstream migrating boundary, which simulates a meander migrating into the flume. Our experiments were conducted in a circulated flume of 11x6 meter, with a constant discharge and sediment feed consisting of a sediment mixture ranging from silt to fine gravel (Kleinhans et al., 2010, this conference). The downstream boundary is a lake into which the river built a branched fan delta (Van de Lageweg et al., 2010, this conference). The morphology was recorded by high-resolution (0.5 mm) line-laser scanning and digital Single Lens Reflex (SLR) camera used for channel-floodplain segmentation and particle size estimation, at an interval of 8 hours. Furthermore a large number of smaller-scale auxiliary experiments were conducted to explore meandering tendency in a large range of parameters. Initial alternate ‘forced’ bars were formed at fixed positions with low sinuosity when the upstream boundary was at one fixed position. Migration of the upstream boundary caused further erosion of the outer banks and formation of point bars in inner bends, so that sinuosity increased to about 1.25. When the upstream boundary reversed migration direction chute cut-offs formed and meander bends reformed in the opposite direction. Hence in the first meander sweep the reworked floodplain showed nodes and antinodes at a wave length in agreement with linear bar stability analysis. After 260 hours experimental time the floodplain had become much more complex, exhibiting meandering channels, point bars, chutes, abandoned and partially filled channels, and slightly cohesive floodplains similar to natural meandering gravel-bed rivers such as the Allier near Moulins (France) and the Rhine near Emmerich (Germany). The flow became even more confined to a single-thread channel when pulses of silica flour were fed during short flood events, which significantly enhanced cohesive floodplain formation. The strengthened floodplains decreased channel mobility, however. We conclude that the necessary and sufficient conditions for meandering are a dynamic upstream boundary and active floodplain formation by fines.

Sedimentary records provide comprehensive and detailed information on the sedimentary history of ... more Sedimentary records provide comprehensive and detailed information on the sedimentary history of a basin. As these systems are the result of a complicated interplay between many processes, distinguishing between pure autogenic and allogenic forcings remains challenging. We performed a series of experiments in which we first had a fluvial system grade itself to the shelf edge during a low-stand sea level scenario. By maintaining the extrinsic variables (discharge, sediment supply, sea-level and basin relief) constant, the resulting morphodynamics and architecture are the expression of pure autogenic behavior. A subsequent sea-level forcing allowed us to study the fluvial-deltaic morphodynamics and architecture resulting from the interaction between autogenic fluvial processes and allogenic sea-level forcing. The experiments were conducted in the Eurotank facility (11 x 6 m) and river flow measurements involved channel width, depth, flow velocity, sediment transport and time lapse monitoring. Landscape evolution was recorded by a high-resolution (0.5 mm) line laser-scanner and Single-Lens-Reflex (SLR) camera. Results from the autogenic fluvial experiment indicate that an ongoing upstream disturbance is needed to maintain an active meandering pattern. Also, the addition of fines (silica flour) is effective in filling up chute channels and thereby promotes a single-thread river. No major changes in fluvial pattern were observed during sea level rise. Stratigraphical analysis shows the presence of a clear bend sorting effect with a distinctive ridge and swale topography. The delta which formed during constant base level is build up from multiple fining-upward sequences with clear erosion surfaces in between. During sea level rise a back-stepping delta formed across the active fluvial plain.

Traditionally, rivers were downscaled to the laboratory through similarity of the Froude, Shields... more Traditionally, rivers were downscaled to the laboratory through similarity of the Froude, Shields and Reynolds numbers. This has worked well for rivers with fixed banks and for braided gravel-bed rivers. However, for self-formed dynamic meandering rivers in experiments, Froude scaling is incomplete without a constrained width-depth ratio. This aspect ratio should be small enough to obtain alternate bars. Bank erosion and bar migration have to be limited by somewhat cohesive or vegetated self-formed floodplains. Our objective is to determine the conditions that lead to river meandering in the laboratory. We developed an experimental scaling strategy for meandering gravel-bed rivers that reduces scale problems or quantifies scale effects. A sediment mixture ranging from silt to fine gravel produces subcritical to critical flow, a hydraulically rough boundary without scour holes or current ripples. Furthermore the mixture leads to richer morphodynamics with measurable sorting trends, narrower channels and cohesive self-formed floodplains. We cycle the inflow point of constant flow discharge and sediment feed in transverse direction at the upstream boundary to perturb an initially straight channel and simulate a meander migrating into the flume (Van Dijk et al., this conference). The downstream boundary is a lake into which the river progrades a branched fan delta (Villiers et al., Cheshier et al., van de Lageweg et al., this conference). Morphology was recorded by high-resolution line-laser scanning and digital photography allowed image segmentation and particle size estimation through an entropy method. In agreement with earlier work, the experimental river initially evolves from alternate bars to a fully braided river without significant floodplain building. With silica flour added to the feed, a transitional river between braided and meandering evolves with frequent chute cut-offs but mostly single-thread. During chute cut-offs the water and bed levels upstream of the cut-off location rises, so modest levees and crevasse splays are built and former channels are flooded, though usually not reactivated. After applying floods laden with silica flour to build the floodplain; the river evolved to, on average, a much less mobile single channel with infrequent chute cut-offs. The lateral channel migration rate decreases with increased floodplain building in agreement with natural systems (Lavooi et al., this conference). Ongoing work quantifies the effect of migration rate and amplitude of the inflow position on the morphodynamics and channel pattern. We conclude that necessary and sufficient conditions for a self-formed, dynamically meandering channel in the laboratory are the presence of floods and floodplain-building sediment and a dynamic boundary condition simulating meander migration. In smaller-scale pilot experiments we use different vegetation species and controlled conditions allow adjustable growth rate, bank strength and hydraulic resistance (van Breemen et al., this conference). We will apply this in the large-scale experiment next year to be able to use less silica flour yet inhance floodplain formation.

There are strong feedbacks between river channels, floodplains and riparian and floodplain vegeta... more There are strong feedbacks between river channels, floodplains and riparian and floodplain vegetation. We study the effect of experimental vegetation on channel pattern. Through linear bar theory it is known that channel width-depth ratio affects bar pattern and relatively narrow channels with strong banks are required for meandering. Riparian vegetation is able to alter the channel width-depth ratio and therefore the channel pattern through strengthening of the banks. Floodplain vegetation adds hydraulic resistance so the flow is more focused into the channels. However, determination of the underlying mechanisms and processes has remained scarce and qualitative and hence these effects are not yet fully understood. The objectives of this study are 1) to develop a controllable and scalable method to reproduce vegetation effect in experimental self-formed channels, and 2) to experimentally determine the effects of riparian vegetation on bank strength, channel pattern and meandering dynamics. Sprouts of three plant species were systematically subjected to different seeding densities and to various growing conditions, including light intensity, submergence and nutrient starvation. Denser seeding reduced sprout growth after about a week. Stronger light increased plant growth and plant strength. Nutrient starvation caused different branching intensity of the root system. Tens of small-scale bank erosion experiments and bank failure experiments (see Kleinhans et al., this conference) were performed to quantify the strength of banks reinforced by plant roots at the experimental scale, demonstrating that bank strength is strongly determined by seeding density, rooting density and depth relative to channel depth. To study pattern evolution and morphodynamics we used a 1.25x7.5 m flume with a constant discharge and sediment feed. The introduction of vegetation in experiments results in narrower and deeper channels. Higher vegetation density leads to static channels with occasional sharp bends. Vegetation increases the threshold for erosion and failure of bank material as indicated by higher measured cohesion values and lower bank erosion rates. We conclude that a low vegetation density is crucial to maintain a dynamic fluvial system, that bank strength and hydraulic resistance can be controlled through manipulation of vegetation species, density and growing conditions. Ongoing experiments quantify the effect of matric suction on vegetation development and bank cohesion, as well as flooding magnitude and frequency on the channel pattern and vegetation development.

The creation of small scale deltas in the laboratory has proven challenging but significant progr... more The creation of small scale deltas in the laboratory has proven challenging but significant progress has been made in several labs in recent years. Various types of river and fan deltas have been created, ranging from fine- to coarse-grained, non-cohesive to cohesive sediment, and from sheet flow dominated to strongly channelised. These experiments have elucidated a number of important factors and autogenic behaviours under constant boundary conditions, such as channelisation and mouth bar formation and backsedimentation leading to avulsion. However, there has been very little systematic investigation of scale effects (Kleinhans et al., this conference) and of experimental parameter space, particularly the link between the sediment mixture and autogenic delta behaviour. Here we describe systematic pilot experiments designed to investigate the link between sediment mixture properties, such as grain size, sorting, addition of fines or polymers, and delta morphodynamics, described by channel dimensions, network pattern, planform delta shape, and bifurcation mechanisms. These experiments were designed to be small, fast and repeatable, so that large quantities of data can be collected for different mixtures. They are run in conjunction with other small-scale tests on the same sediment mixtures to quantify the link between the mechanical properties of delta substrate, such as bank erodibility, and delta organisation. Discharge and sediment feed rate were kept constant and the same in all experiments. Delta evolution was recorded by high-resolution time-lapse photography, emphasising the channels by coloured flow and showing sorting patterns in cases with sediment mixtures. Topography was measured at sub-mm resolution at a few time steps for each delta. For sediment we used various combinations of well-sorted fine sand or poorly sorted sand with some fine gravel, and silica flour and/or polymer with benthonite. The uniform fine sand invariably exhibited sheet flow or formed unwanted scour holes because of hydraulic smooth conditions. In contrast, the poorly sorted sand formed well-channelised deltas without scour holes, because the coarser particles cause hydraulically rough conditions.and bifurcations were temporarily stabilised by gravel deposition. Both silica flour and polymer increased channelisation and channel depth in the poorly sorted sand. At the largest polymer concentrations knick points and backward steps appeared similar to those in deltas on cohesive lake beds or bedrock. The polymer percolated to underlying layers which became stronger over time. We conclude that poorly sorted sediment removes the unwanted scale effect of scour holes in fine sand and facilitates sorting-related bifurcation and bar formation. Silica flour increases channelisation by strengthening the banks, whilst polymer increases bank strength and channelisation much more but causes unwanted backward step erosion and channel bed hardening. Strong banks and hardened beds prevent dynamic meandering; hence a major challenge in landscape experimentation remains to scale the properties of the fluvial system and the deltaic system at the same time.

Abstract For various river channel patterns, the necessary formative conditions differ, but how i... more Abstract For various river channel patterns, the necessary formative conditions differ, but how is not entirely understood. Furthermore, not only the morphology and dynamics differ, but also the resulting stratigraphy differs, of which understanding is required to infer past environmental conditions and predict reservoir behaviour. Our objective is to identify the necessary and sufficient conditions for forming dynamic meandering and braided rivers.

"Meandering rivers have proven difficult to reproduce experimentally, while braided rivers are re... more "Meandering rivers have proven difficult to reproduce experimentally, while braided rivers are relatively easily formed in the laboratory. Our objective is to create self-formed dynamic braided and meandering rivers in a laboratory, and to understand and quantify the necessary and sufficient conditions for meandering. In addition, we have tested existing theories to reconstruct the original morphology from preserved stratification as stratification is the key to reconstruct formative channel dimensions and palaeoflow conditions from the fluvial record.

The experiments were carried out in a flume of 6 m wide and 11 m long, which was split up into two separate fluvial plains (each 3x10 m). We designed experimental conditions that minimize the most important scaling issues. The morphological and stratigraphical evolution was recorded by high-resolution line-laser scanning and Digital Single Lens Reflex (DSLR) camera used for channel-floodplain segmentation, water depth approximation, silica flour distribution and particle size estimation.

Key findings include that the experimental river without silica flour evolves from alternate bars to a multi-thread braided river. Addition of silica flour to the feed resulted in a single-thread meandering river with chute cutoffs. Large bends developed with scroll bar complexes and sinuosity reached a maximum of 1.4. Furthermore, we find that the mean set thickness agrees well with the theoretical prediction from channel morphology. The mean preserved set thickness is 30% of the mean channel depth. Finally, there is much systematic spatial variation in set thickness related to repetitive point bar growth and chute cutoff. We find undisturbed and thick sets close to channel belt margins and more irregular stratification with stacked thinner sets in the channel belt center."