Review of: New Outdoor Experimental River Facility to Study River Dynamics

Authors: Mahmoud, Dickson, Renault, Trudel, Biron, Sklar and Lacey

Reviewer: Maarten G Kleinhans, d.d. 20 Sept 2025

This is a well-written manuscript presenting a pilot experiment in a new and promising outdoor facility. However, there are two major issues and one minor issue with the manuscript related to the design for meandering and the underlying assumptions on the meandering process and related to the interaction between sediment sorting and morphological response. The less major issue is a poor development of reasons to do this kind of experiment, which is biased towards the outdated dogmas of engineering scale models. As such, the necessary recommendation is reject. I strongly suggest the authors to reconstruct a paper with much less focus on the meandering and much more focus on the armouring processes.

MAIN POINTS

* I congratulate the authors on obtaining a great experimental facility. In my opinion these experiments remain necessary complements to numerical modelling especially for the channel-bank interactions and for sediment sorting dynamics that are the focus of this paper. 

* The arguments for doing physical modelling are poorly developed and need nuance. 

* The experimental conditions conducive to meandering, which is apparently the purpose of the experiment, are not met. Literature suggestions are given. 

* The interesting interaction between sediment sorting and morphological development is in part presented, but the essential point (the different timescales for armouring and for morphological response) is underdeveloped. Literature suggestions are given. 

DETAILED POINTS (referring to line numbers)

8 the abstract would be more readable in words and without unexplained symbols

18-19 a recent review is found in my paper Kleinhans et al. 2024

24 a more appropriate reference is Schumm's ten ways to be wrong

26 this is far too simplistic and needs more nuance. Of course no physical model can ever produce the full complexity of natural river systems, because all models are simplified, namely on the controlled initial and boundary conditions, in the processes and mechanisms that are allowed by the operators, and possibly in certain scale effects (also see my 2024 paper for development of a complex systems view). Physical and numerical models are representations of a target system, in which the aim is to include relevant respects and to obtain a sufficient degree of similarity. The whole point of these models is to simplify otherwise science would be impossible for humans. 

35 "full spectrum" likewise oversimplifies the matter.

38 other references are needed here, including the (in)famous Paola et al 2009 paper and my 2014 paper, both in ESR. Besides, some indoor facilities are quite large (for instance the USACE wave tank and the BAW river scale model in Germany), so the 'indoor' is irrelevant here. 

41 Again simplistic. Why not representative? why not representative to a certain degree? Such representation serves a purpose, so if you want to argue that having low Re is not representative then a specific process is needed here for which that is relevant. As Paola and Metivier and I and many others have demonstrated, the Reynolds number is not relevant for many phenomena. The authors here provide no arguments but basically repeat the dogma from the engineering scale model world, which has been outdated in geomorphology for about a quarter of a century. To put it as clear as possible without the intention to be blunt: the fact that  our meandering and estuarine and deltaic and debris flow experiments violate these rules is the reason that we obtained spectacular results where the classic scale models (including those shown in the recent Crosato and Mosselman paper) fail. This point was well put by Paola et al in 2009 in ESR.

44 "Inaccurate predictions" implies a certain purpose, such as engineering scale modelling of a certain target system where, for instance, measured water levels need to approximate those in the target within a predetermined accuracy range. However, for the study of many phenomena such a scaling is not needed and instead other aspects of the experiments need to be realized. 

46 This depends very much on the dimensions of a natural system that one can have in mind. For example, the Ganges river is about 1000 times wider than the channel in this facility, so from the engineering dogma the scale numbers are off the chart. However, the van Dijk et al. 2012 paper cited by the authors, was the first to produce sustained dynamic meandering with chute cutoffs in the lab, and the occurrence and mechanisms of the chute cutoffs are arguably adequate representations of the same phenomenon in small and large rivers regardless of the geometric scale. By the implicit standards of the authors, however, their own experiment is a terrible representation of large meandering rivers. There are plenty of reasons to disagree with this position: this new facility is suitable and a great addition for a large number of phenomena as argued in Paola et al. 2009 and Kleinhans et al. 2014. 

135 This reference is unfortunately not an acceptable reference unless the Dickson thesis is published open online for the forseeable future. Alternatively, the authors provide these descriptions or a summary thereof in an online supplement. 

144 I don't understand how the W/D=12 can be reconciled with the objective of the authors to obtain incipient meandering. The formation of alternate bars requires at least double that W/D plus a dynamic perturbation on the upstream boundary, especially since this flume has about the same length/W ratio as that in the van Dijk et al experiments. This necessary W/D follows from bar theory and from empirical and experimental evidence. The Crosato and Mosselman paper cited by the authors does not provide this number of W/D=10. Our application of the theory (Kleinhans and van den Berg 2011) indicates that W/D>25 at least. 

156 this history of perturbations is not correct (see Kleinhans et al 2024 for the full story). In the first place, Christian Braudrick based his static perturbation on Friedkin 1945. In the second place, van Dijk et al did not build on this idea of static topographic forcing. Instead, the idea was that in a dynamic meandering river, the topographic dynamics come from upstream, AND nature is full of perturbations in the inflow, sediment influx and directions thereof. We also discovered after conducting the experiments that Lanzoni and Seminara had developed some theory and a numerical meandering model showing the need of such instabilities. This was later mathematically proven by Weiss and Higdon (2022). Critically for the author's paper, this is solid evidence that dynamic meandering in a relatively short flume such as this one requires a large and dynamic upstream perturbation. 

258 Given the tendency to armouring, the low mobility and the low W/D, and the initial flume runs to create a water-worked bed, it does not come as a surprise that no bars developed. This was also a result in Lanzoni/ 2000 experiments in a 1.5 m wide 50 m long flume, and he needed to increase the mobility to obtain bars, even though he used sediment recirculation rather than the lack of sediment supply in these experiments which inevitably leads to a static state. The extremely interesting behaviour of sediment mixtures is that an initially flat bed responds in two different manners, each with their own timescale: a sedimentological response, in this case armouring, and a morphological response, in this case alternate bars (but similarly in the context of bifurcations in bends that the Dutch did a lot of work on, see review on the effects of armouring in Kleinhans et al. 2012). So I agree with 284 that the sediment mobility needs to be tuned. 

287 If the purpose of these experiments is to represent small streams, then the bank strength is indeed one way to avoid braiding and promote meandering. However, as we demonstrated in our experiments and models, starting with van Dijk et al 2012 and reviewed extensively in the aforementioned 2024 paper, the more likely and more important mechanism needed for meandering is surface cover by vegetation or cohesive sediment on the inner bend. 

292 I agree with the suggestion to reconsider the sediment recirculation issue. A sediment feed setup has greatly different results and at low sediment mobility the system progressively develops towards a static stable state of armoured bed and possibly supply-limited bedforms. See Parker and Wilcock 1993 and Kleinhans 2005. 

307 I agree that LSPIV is excellent for this kind of experiments and an appropriate reference is needed here (perhaps work by Wim Uijttewaal from whom we learned it?)

310 I suggest to attempt structure for motion applied to a set of fixed cameras, which worked well in our past experiments. 

312 I suggest to use biological pest controls such as presented in our papers (Weisscher et al 2022, protocol in the supplementary material)

REFERENCES

Kleinhans, M. G. (2005), Upstream sediment input effects on experimental dune trough scour in sediment mixtures, J. Geophys. Res., 110, F04S06, doi:10.1029/2004JF000169

Kleinhans, M. G., McMahon, W. J., & Davies, N. S. (2024). What even is a meandering river? A philosophy-enhanced synthesis of multilevel causes and systemic interactions contributing to river meandering. Geological Society Special Publication, 540(1), 43-74, https://doi.org/10.1144/SP540-2022-138, open access copy here: https://research-portal.uu.nl/ws/files/243459455/What_even_is_a_meandering_river_A_philosophy-enhanced_synthesis_of_multilevel_causes_and_systemic_interactions_contributing_to_river_meandering.pdf

Kleinhans, M. G., & van den Berg, J. H. (2011). River channel and bar patterns explained and predicted by an empirical and a physics-based method. Earth Surface Processes and Landforms, 36, 721-738 doi:10.1002/esp.2090

Kleinhans, M. G., Ferguson, R. I., Lane, S. N., & Hardy, R. J. (2012). Splitting rivers at their seams: bifurcations and avulsion. Earth Surface Processes and Landforms, 38(1), 47-61, https://doi.org/10.1002/esp.3268

Lanzoni, S. (2000). Experiments on bar formation in a straight flume: 2. Graded sediment. Water Resources Research, 36(11), 3351-3363.

Parker, G. P., and P. R. Wilcock (1993), Sediment feed and recirculating flumes: Fundamental difference, J. Hydraul. Eng., 119, 1192–1204.

Weiss, S.F. and Higdon, J.J.L. 2022. Dynamics of meandering rivers in finite-length channels: linear theory. Journal of Fluid Mechanics, 938, A11, https://doi.org/10.1017/jfm.2022.131

Weisscher, S. A. H., Van den Hoven, K., Pierik, H. J., & Kleinhans, M. (2022). Building and raising land: mud and vegetation effects in infilling estuaries. Journal of Geophysical Research: Earth Surface, 127(1), 1-24. Article e2021JF006298, https://doi.org/10.1029/2021JF006298

Review of New Outdoor Experimental River Facility to Study River Dynamics

Summary. This study reviews larger scale river experimental facilities and introduces a new experimental facility designed to look at river morphodynamic problems. The manuscript details an attempt at evolving a meandering channel through the use of a static inlet bar and documents shortcomings and potential next steps.

General comments. Overall, I found the manuscript a bit underwhelming in terms of the new science that it presents. However, given the stated goal of introducing the new experimental facility, I have no major reservations with the paper being accepted for publication. Perhaps the largest issue with the current manuscript is that the introduction and initial experimental purpose are somewhat disjoint with the reported results and especially the focus within the results of the velocity measurements. Currently, much of the focus of the initial experiments is on creating a self-formed sinuous or meandering channel, however, this didn't really work out (that's fine, many of my experiments did not) and so the results focus on observed channel change, bed grain size coarsening, and the velocity measurements. It would be worthwhile, in my opinion, to downplay the focus on meandering and reframe the introduction towards what actually happened in the experiment. I am not suggesting rewriting it wholesale, but adding more about what you ended up doing and less about what you wanted to do.

I would really like to know, from the authors perspective, what they felt they learned that we didn't necessarily already know. The measurements that were taken are very nice and the change in grain size distribution and DEMs of difference are interesting, but overall I am left with the feeling that we'd be better off doing smaller scale experiments because we still don't know enough about channels to properly set up and initiate a larger scale experiment. I remember a conversation with the engineers at SAFL that can be summed up as 'big experiments come with big logistical problems', that seems to be true here. This isn't to downplay what is presented, but instead i hope the authors will share more about what they did learn from their experiments from the science perspective in addition to the lessons they have already offered. Rather than failing to make a meandering channel, what did we learn about a mixed grain size no sediment feed channel that we didn't already know? In a sense, what does this new experiment at large scale say about the previous fixed wall experiments with mixed sediment transport (particularly, some of the large flume experiments done at SAFL, CSU in Fort Collins with bars, experiments by Hassan's or Eaton's group in British Columbia or even Peter Wilcock's original experiments).

One of the large challenges with doing mobile channel experiments is in setting the inlet and outlet conditions. Given the considerable knowledge gained with the preliminary experiments, can the authors provide additional comments/thoughts on how they would envision the inlet condition for the sediment recirculation case especially under the condition of a variable flow regime? 

Some minor Specific Comments.

Ln. 46 - I am all for larger experiments, but can you enlighten the reader here what we stand to gain through more realistic reynolds numbers? Other than increased secondary flows, how is a higher reynolds expected to change the morphodynamics. A follow up, how much does coarse sediment care about secondary flows?

Ln. 51 - Within this paragraph, would the authors be amenabe to adding the year that the facilities became operational?

Ln. 67 - Not clear to me why Wilcock and Crowe is the citation here. Mixed grain size transport as developed in this reference has already been shown to work well at the field scale. You might consider the work from the IPGP group here that builds on Metivier et al., (2017) as they develop substantial theory at very small scales and would require an erodible bank for larger scale testing. Popovic et al., (year) seems like a better fit here.

Popović, P., Devauchelle, O., Abramian, A., & Lajeunesse, E. (2021). Sediment load determines the shape of rivers. Proceedings of the National Academy of Sciences, 118(49), e2111215118. https://doi.org/10.1073/pnas.2111215118

Ln. 88 - Can you provide some input on how long it takes to change out the bed conditions and what types of labor is required given the facilicity constraints? For example, does the sediment bed need to be prepped by hand or can some sort of heavier machinery be involved.

Section 2.1 - Is there any ability to change the elevation of the inlet or outlet? Weir adjustment etc.

Section 2.1 - Could you describe how is the outlet condition handeled to keep the sediment bed from eroding or depositing with variable flows?

Table 1 - Consider adding the shields stress and transport capacity ratio to this table given the sediment transport focus of the initial experiment.

Ln. 221 - What is the basis for the critical shields value of 0.045? Is this visually confirmed or validated from transport measurements?

Ln. 308 - Given the difficulties with drones, and the relative constrained spatial scale with good access, do the authors recommend continuing to use drones or would a ground based TLS unit on a platform be a wiser investment?

Conclusions - Can you comment (if appropriate) on whether this facility is expected to be open to external collaborations, seeking collaboration or is it primarily for internal use?