How to build a meandering river in your basement

A post by Anne JeffersonMeandering rivers are characterized by regularly spaced bends that grow and cutoff and generally march downstream in a fairly orderly fashion. Click the image below to watch a movie of meander migration on the Allier River near Chateau de Lys, France
Movie 1. Meander bend migration and cut off using aerial photos and maps from: 1945,1960,1971,1980,1982, 1992, 1995, and 1997 on the Allier River, France. Created by A. Wilbers, originally found here.
Though meandering rivers are by far the most common river form on Earth, building a meandering river in a laboratory flume eluded scientists for decades. The conditions necessary to support self-maintaining meandering rivers were not known well enough to recreate in the laboratory. Flumes, or experimental channels, are a really important tool for understanding river processes, because sediment and water influxes can be tightly controlled and high precision measurements made.
Sand and gravel, the most common sediments in river banks, have low cohesion. In flumes, channels through sand and gravel, even if initially forced into a meander form, inevitably end up as wide channels with active braid bars. Solving the bank cohesion problem, by replacing sand and gravel with silt and clay, results in flume channels that have lots of curvature (sinuousity) but do not maintain their geometry through multiple meander cut-offs. Over the last 10 years, graduate students Karen Gran and Michal Tal working with Chris Paola at the University of Minnesota figured out how to make a self-sustaining single channel in coarse sediment. The key to creating a single channel was to plant alfalfa seedlings to give the banks some cohesion. You can see the results of alfalfa growth in a Quicktime video of Tal’s experiments. (Click the image below.)

Movie 2. Tal and Paola’s experiments with alfalfa seedlings and channel form. More movies of these experiments here.
If you watched the video, you’ll notice that while the channel is indeed single thread and it does move around, the meanders don’t move downstream in the relatively orderly fashion of a natural river. So the insight of alfalfa sprouts from Gran and Paola (2001) and Tal and Paola (2007) got geomorphologists a long way towards understanding the controls on meander self-maintenance in coarse-bedded rivers, but they didn’t quite reach the finish line.
Now, a paper in the Proceedings of the National Academy of Sciences by UC Berkeley graudate student Christian Braudrick, his advisor Bill Dietrich and collaborators Glen Leverich and Leonard Sklar from San Francisco State University reports that they have succeeded where so many others have failed. In a 17-m long, 6.7 m wide flume, Braudrick and colleagues created a self-sustaining meandering channel. Their work was featured on National Public Radio’s Science Friday show, which produced the following video giving the basics of Braudrick’s process.

Movie 3. Science Friday’s video about Braudrick et al’s experiments.
One of the key things mentioned in the video, but not explained is why the lightweight sediment was plastic. In slimming down a river to fit within a laboratory, researchers have to take into account all of the possible scaling effects. That’s why alfalfa seedlings are used to simulate the grasses and trees of a normal riparian zone, for instance. The power of the water, or its shear stress, is a function of depth, slope, fluid density, and gravity. Since the depth of flume channels is so much smaller than real rivers, it means that the shear stress available to move sediment is much lower. This means flumes can’t move fist sizes pieces of gravel and the size of the sediment in the study must be scaled down accordingly. Gravel scales down reasonably well to coarse sand, but sand scales down to silt, and silt has much different cohesive properties than sand. This is where the plastic came in, because the researchers wanted to create meanders using the alfalfa to create cohesive banks not by adding cohesive sediment. The plastic beads were the size of very fine sand and they lacked cohesion. Thus, the researchers created laboratory conditions of that mimicked natural rivers – channel banks where there was a mixture of sizes of non-cohesive sediment held together by roots.
When the flume was turned on, the little plastic beads moved both along the channel bed and suspended within the water column, much as sand would do in a natural channel. With a small initial curvature at the upstream end of the flume, meanders propogated downstream and began to grow and cut off. In previous alfalfa-only experiments ( Tal and Paola, 2007), each time meanders were cut off, a trough was left on the upstream side of the abandoned meander. In natural systems, these troughs get plugged with fine sediment and create oxbow lakes that eventually fill in. In the alfalfa-only, the troughs persisted, opening the possibility of islands developing in the channel. In Braudrick’s alfalfa+plastic experiments, the little plastic beads moving in suspension filled in the troughs at the upstream end of the abandoned meander, blocking future flow through that old pathway.
From Braudrick and colleagues’ results, it appears that sand and fine sediment have an important role to play in reinforcing and maintaining the meandering pattern of river channels. Out in the real world, such fine sediment is often regarded as an undesirable pollutant of coarse-bedded rivers, so these results have the potential to change the goals of river restoration and management. Plus, now that geomorphologists have a way to simulate realistic meandering rivers in the flume, new insights into the controls and behavior of meandering rivers are likely to start pouring in.

Categories: by Anne, geomorphology
Tags: , , , , ,

Comments (15)

  1. BrianR says:

    Nice summary, very cool stuff.
    Check out the blog Riparian Rap – is from some researchers/educators working on similar experimental work:
    (also, would be cool if you guys put some indication of who the author of post is up top that is visible in a feed reader … see the Sciencewomen site for an example. I can usually guess, but it’s just nice to know :))

  2. Monado says:

    Lovely. I was going to say that I have one already, but it’s really a straight trickle. 🙂

  3. --bill says:

    Check out the University of Minnesota’s Outdoor Stream Lab .

  4. Erik says:

    Wow, this was confusing at first glance. They need to refer to their plastic beads as synthetic sediment, or something similar. “Plastic” already has a pretty specific meaning when talking about soil, and it’s the opposite of how their synthetic sediment behaves. 🙂

  5. Chris Rowan says:

    So could you say that meandering rivers were probably quite rare before the evolution of land plants?

  6. BrianR says:

    Chris … that’s a fascinating question. My gut-instinct reaction is that they were probably not the dominant morphology like today, but is difficult to know just how rare. I suppose one way to address that is see if there are any meandering (i.e., with actual migrating barforms as opposed to simply sinuous planform) rivers on Earth in plantless areas. Could mud w/out plants have enough bank strength? Maybe, but my guess is probably not enough for robust and long-lived meandering rivers. But, I could be wrong. If I were to research this I’d start with the great work of Schumm in the 1960s.

  7. Robert says:

    The basic principle is that water flows down the steepest path. So don’t you need to take erosion of the land into account also because that’s part of opening up steeper paths for the water? Or are you saying the erosion is too slow to be relevant?
    I also thought meanders would have a strong relation to flooding. A flood jumps the river over the banks where it finds a new path that is steeper, so it can’t easily get back to the old path.
    Why would you care about a model that neglects erosion and flooding?

  8. BrianR – Riparian Rap does have a nice write-up. Those guys also have some nice educational youtube clips available. (And, I’ll bug Chris about seeing if we can come up with little author headers.)
    Bill – Yep, the Outdoor Stream Lab is really amazing, as is all of St. Anthony Falls Lab (SAFL). I did a little bit of my post-doc work in the indoor main channel flume at SAFL, I have been scheming and trying to figure out a way to get involved with the outdoor lab.
    Erik – I never thought about that, and I bet the authors didn’t either. We meant plastic in its noun defintion but then used it as an adjective. Sorry for the confusion.
    Chris – Interesting question, but I might approach it by considering the types of channels found on other planets. Dietrich and Perron had a Nature cover story in 2006 on the (lack of) a topographic signature of life. I’ll have to go back and reread the article.
    Robert – Erosion is definitely part of it, because it is intimately involved in the meander migration and cutoffs. Or are you referring to the long term (10s to 100s of thousands of years) lowering of the land surface (denudation)? If so, that’s a much longer time scale than that involved in migration of meanders, so its OK to neglect.
    Flooding! Now that’s an interesting question and one I opted not to include in the body of the post. While Tal and Paola argued that flooding was necessary for meanders based on their work, Braudrick et al. argue that flooding is not necessary because they still got meander migration during 72 hours of running the flume without floods. However, they did have some floods earlier in the same flume run. I think the role of floods in meander formation and migration may still be a bit of an open question, so now that we can build meandering rivers in flumes, I expect to see more related research.
    Thanks for all the great questions!

  9. GeoDTurner says:

    This old man has three comments/questions:
    1.Tao and Paola’s work, according to their own words, was about braided streams, not meandering ones.
    2 Braudrick, et al’s work looks like a case of fiddling with the parameters until you get the right answer-the one you want. How does this fit into a falsification paradigm?
    3. In what way does any of this work make a significant advance over what was done in the 1940’s in Mississippi, as reported in Gilluly, et al, Prin. of Geology, 2nd ed, p.204-5.There seems to be a popular rationale that escapes me.
    Can anybody help out this old man?

  10. Tal and Paola’s work was about dynamic single thread channels, of which meandering channels are the most common natural form.
    Braudrick’s work was aimed at replicating in the lab what we see in nature. His team, and its predecessors, eliminated lots of potential explanations for meandering channels before they fiddled into the right parameters. How is that not falsification?
    I’m sorry I’m not familiar with those pages of Gilluly’s book, and my university library does not have a copy. Perhaps you’d care to elaborate?
    Chris and Ole – Of course you are. Brilliant!

  11. GeoDTurner says:

    Being an old dog, new tricks come hard to me. I know there is an HTML way to send you the reference, but it was simpler for me to plop it on my wordpress blog. Could you pick it up from there?
    Maybe I’m just a bug about the topic of falsification, but it seems to me we need to use the ideas more than just in arguments about intelligent design and so forth.
    George D. Turner

  12. Steve Gough says:

    Anne, thanks for a great post. Though I’ve visited their lab, I wasn’t that familiar with Tal and Paola’s work. As you’ll see from the videos we make, our pure plastic-media channels do like to braid, but I think braided channels are not that uncommon–I’ve done a lot of work in the Ozarks, and most of the channels there are braided, though it’s somewhat hidden by young sycamore/willow stands that never get very old.
    The key to Braudrick et al.’s work, I think, is plugging of the sub-dominant channels and chutes that cut off bars from floodplains. I’ve seen this a lot in the field. Single-thread gravel bed channels are not as common as most imagine, though, at least not in my experience.
    I’m giving a talk at GSA-Portland next week on the plastic media on Sunday morning. Very exciting stuff!
    Thanks for the props to Riparian Rap. Besides the YouTube video, we have a DVD collection of nearly 60 clips available for cost at, and maybe for free at Portland next week. We (Little River R&D) will have one of our models in the exhibit hall so you can see plastic media firsthand.

  13. coconino says:

    At the risk of starting off on the wrong foot by bringing up Rosgen, I’m going to do just that. Leaving the philosophy of Rosgen-type restoration projects aside, if you look at Rosgen’s classification system, the E and C streams are more characterized by finer grained sediments (especially E). They are also the ones with well-established meanders (I’m ignoring entrenched G streams for now). I think the research is great, but my first thought was, well, look at what you see in the field. The most meandering streams have the finest-grained sediments, and are ususally characterized by herbaceous banks as opposed to shrub/tree layers. This is clearly what I see in NM, AZ, and CO.
    Also, in response to Steve Gough’s comment above, I would look at the surrounding land uses to determine why streams in the Ozarks would be braided and aggrading. Why do they have such a high sed load in a wooded, well-vegetated environment? What are the stream gradients and other factors that generate those kind of sed loads?