…because nothing cheers me up like playing with water ;)
…because nothing cheers me up like playing with water ;)
I am a huuuge fan of Wlodek Brühl’s liquid art: Pictures of water sculptures that are created with focus on the tiniest of details, that only persist for milliseconds, but that are captured forever in all their fragile beauty. And I think these pictures are an awesome tool in science communication — I see so much physics in them (some of which I wrote about here already), and even if you come to an exhibition for the art, who wouldn’t love to learn some physics while they were there, too?
Well, we are about to find out! There is a new exhibition being opened (with brand new pictures!) on March 3rd in Preez. And I will actually give the opening speech for the liquid art half of the exhibition! I haven’t seen all the pictures yet so I can’t tell you exactly what I will be talking about, but whatever I say will definitely have to involve lots of fun physics :-)
It has been a long time in the making, but finally we have a nice fjord circulation in our tank!
Pierre and I tried to improve it 6 years ago, Steffi, Ailin and I have been working on it for a couple of days last August, then finally this morning, Steffi and I tried again — and it worked beautifully right away!
We now have an experiment that shows how a fresh, yellow inflow (representing the freshwater input into fjords close to their heads by rivers) flows over a initially stagnant pool of salt water. As the freshwater plume flows out of the fjord, it entrains more and more salt water from below, thus thickening and setting up a return flow that brings in more salt water from the reservoir (representing the open ocean) on the right.
We drop dye crystals to visualize the surface current going out of the fjord and the return flow going in, and draw the profiles on the tank to be able to discuss them later.
Here is a movie of the whole thing:
But there is more to see!
When tipping the tank to empty it, a lot of turbulence was created at the sill (see movie below). While a fjord typically isn’t tipped very often, what we see here is basically what tides do on the sill (see the waves that keep going back and forth over the sill after the tank is initially lifted? Those are exactly like tides). This could purposefully integrated in teaching rather than only happen by accident, those waves could be created just by surface forcing rather than by tipping the tank. That’s a very nice demonstration to explain high mixing rates in the vicinity of steep topography!
And then there is also the issue of very low oxygen concentrations in some Norwegian fjords, and one proposed solution is to bring the river inflow deep down into the fjord. The idea is that the less dense river water will move up to the surface again, thereby creating mixing and oxygenating the stagnant deep water that, in some cases, hasn’t been renewed in many years.
We model this by putting the inflow (the hose) down into the tank and see the expected behaviour. What we also see: Since the water has a quite strong downward component as it enters the fjord, it stirs up a lot of old dye from the bottom. So possibly something to be aware of since there might be stuff dumped into fjords that you might not necessarily want to stir up…
And last, not least, a bonus picture: This is how we measure temperature at GFI. You would think it should be possible to find a smaller thermometer that isn’t an old reversing mercury one? But in any case, this worked very well, too :-)
Inspired by the article “Affordable Rotating Fluid Demonstrations for Geoscience Education: the DIYnamics Project” by the Hill et al. (2018), I spent a fun Sunday afternoon with my friends Joke and Torge in their kitchen, playing with Legos, water and food dye.
Turns out building a rotating table isn’t as easy as we had hoped, because my Lazy Susan’s axle is unfortunately really off centre (how did I never notice before?), which makes it pretty difficult to drive it with a grinding wheel, and the LEGO motor we were using only has one speed (which would have to be regulated by changing the diameter of the gears). That makes it really difficult to spin up a tank at rest if you go at it zero to full force…
But we got it to spin! Look at the cool paraboloid surface!
Next issue, though: my awesome glass vase which looks like it should work well as a tank has a really irregular bottom, which makes it very difficult to have anything stand in the centre without too much of a wobble. Also, for the Hadley Circulation experiment we were trying to set up here, when do you add in the cooling in the center? Would be best to do it after the tank is spun up, but that is such a pain to do! And I messed up the dye here, too.
But at least you can see a little bit of what it will be like when we are done, right?
And then it will be almost ready to be used in teaching. Well, almost…
Funny how tank experiments that you think should be quick and easy to set up & run always take sooo much longer than expected. But it’s so much fun that I really don’t mind! :-)
Let’s talk about zonal jets! They keep popping into my life all the time right now, and that has got to mean something, right?
Zonal jets, for all that are not quite familiar with the term, are fast-flowing currents (i.e. “jets”) that move along lines of constant latitude (therefore “zonal”). The occur in the ocean (e.g. the Antarctic Circumpolar Current, or the Gulf Stream after separating from the coast) and in the atmosphere (e.g. the subtropical jets stream). And you might be familiar of pictures of Saturn with all the belts around it? Yep, zonal jets!
In December I went to the Science and Industry Museum in Manchester (a.ma.zing place!) and they had one exhibit there that shows zonal jets: A sphere sitting inside a transparent sphere with some sort of fluid between the two. You can put the outer sphere in rotation and, through friction, this puts the fluid in motion. But instead of all the fluid moving with the outer sphere, there is of course also friction with the inner sphere, so a shear flow develops, which breaks up into those zonal jets (which then break up into all the eddies when the outer sphere slows down again).
Please excuse the crappy video. You see the largest part of the upper half of the sphere, but I was filming with one hand and turning the thing with the other… And I didn’t plan on writing anything about it, but then this happened: My friend Judith (check out her Instagram!) and I went on a mini cruise (all the way across Kiel canal!) in freeeeezing temperatures, and therefore obviously ended up with this:
And this is where kitchen oceanography comes in. What do you think happens when you drop in that yummy chocolate and start stirring? This!
Do you see how the fluid doesn’t move solid body-ish, but how there are jets and then more stagnant areas? Doesn’t this make you want to have a hot chocolate, and Right Now? For scientific purposes, of course…
I’d love your input: If your student lab for GFD tank experiments had to downsize, but you had to present a “wish list” for a smaller replacement, what would be on that list? Below are my considerations, but I would be super grateful for any additional input or comments! :-)
The awesome towing tank that you have come to love (see picture above) will have to be removed to make room for a new cantina. It might get moved into a smaller room, or possibly replaced all together. Here are some external requirements, as far as I am aware of them:
Here are a couple of things I think need to be definitely considered.
If the tank was to be replaced by a smaller one, how small could the smaller one be?
The dimension of the new tank depend, of course, on the type of experiment that should be done in the tank. Experiments that I have run in the tank that is to be replaced and that in my opinion should definitely be made possible in the new location/tank include
If we want to be able to continue running these experiments, here is why we should not sacrifice the dimensions of the tank.
The first reason for keeping the length of the tank is that the “mountains” being towed to create the lee waves are already 1 and 1.5m long, respectively. This is a length that is “lost” for actual experiments, because obviously the mountain needs space inside the tank on either end (so in its start and end position). Additionally, when the mountain starts to move, it has to move for some distance before the flow starts displaying the features we want to present: Initially, there is no reservoir on the “upstream” side of the mountain and it only builds up over the first half meter or so.
The second reason for keeping the length of the tank are wave reflections once the ship or mountain comes close to the other side of the tank. Reflected surface waves running against the ship will set up additional drag that we don’t want when we are focussing on the interaction between the ship and the internal wave field. Reflected internal waves similarly mess things up in both experiments
The third reason for keeping the length of the tank is its purpose: as teaching tank. Even if one might get away with a slightly shorter tank for experiments when you just film and investigate the short stretch in the middle of the tank where there are no issues with either the push you gave the system when starting the experiment or the reflections when you get near the end, the whole purpose of the tank is to have students observe. This means that there needs to be a good amount of time where the phenomenon in question is actually present and observable, which, for the tank, means that it has to be as long as possible.
In the experiments mentioned above, with exception of the “dead water” experiment, the tank represents a “slice” of the ocean. We are not interested in changes across the width of the tank, and therefore it does not need to be very wide. However, if there is water moving inside the tank, there will be friction with the side walls and the thinner the tank, the more important the influence of that friction will become. If you look for example at the surface imprint of internal wave experiment, you do see that the flow is slowed down on either side. So if you want flow that is outside of the boundary layers on either side, you need to keep some width.
Secondly, not changing the tank’s width has the advantage that no new mountains/ships need to be built.
Another, practical argument for a wide-ish tank (that I feel VERY strongly about) is that the tank will need to be cleaned. Not just rinsed with water, but scrubbed with a sponge. And I have had my hands inside enough tanks to appreciate if the tank is wide enough that my arm does not have to touch both sides at all times when reaching in to clean the tank.
The first reason for keeping the height is that for the “dead water” experiment, even the existing tank is a lot shallower than what we’d like from theory (more here). If we go shallower, at some point the interactions between the internal waves and the ground will become so large that it will mess up everything.
Another reason for keeping the depth is the “waves running up a slope” experiment. If you want waves running up a slope (and building up in height as they do), you have the choice between high walls of the tank or water spilling. Just sayin’…
And last not least: this tank has been used in “actual” research (rather than just teaching demonstrations, more on that on Elin’s blog), so if nothing else, those guys will have thought long and hard about what they need before building the tank…
Without getting too philosophical here about models and what they can and cannot achieve (and tank experiments being models of phenomena in the ocean), the problem is that scaling of the ocean into a tiny tank does not work, so “just use a mountain/boat half the size of the existing ones!” is actually not possible. Similarly to how if you build the most amazing model train landscape, at some point you will decide that tiny white dots are accurate enough representations of daisies on a lawn, if you go to a certain size, the tank will not be able to display everything you want to see. So going smaller and smaller and smaller just does not work. A more in-depth and scientific discussion of the issue here.
When building a new tank or setting up the existing tank in a new spot, there are some features that I consider to be important:
The “source” reservoirs (plural!) are the reservoirs in which water is prepared before the tank is filled. It is crucial that water can be prepared in advance; mixing water inside the tank is not feasible.
There should be two source reservoirs, each large enough to carry half the volume of the tank. This way, good stratifications can be set up easily (see here for how that works. Of course it works also with smaller reservoirs in which you prepare water in batches as you see below. But what can happen then is that you don’t get the water properties exactly right and you end up seeing stuff you did not want to see, as for example here, which can mess up your whole experiment)
Both reservoirs should sit above the height of the tank so that the water can be driven into the tank by gravity (yes, pumps could work, too, more on that below).
Depending on the kind of dyes and tracer used in the water, the water will need to be collected and disposed of rather than just being poured down the drain. The reservoir that catches the “waste” water needs to
If the waste water is just plain clear tap water, it can be reused for future experiments. In this case, it can be stored and there need to be…
If reservoirs cannot be located above and below tank height to use gravity to fill and empty the tanks, we need pumps (plural).
Preferable the first and the latter are not the same, because changing settings between calibrating the pump for an experiment, setting it on full power to empty the tank, and calibrating it again will cause a lot of extra work.
Sometimes it would be extremely convenient if there was a possibility to insert dyes into the tank for short, distinct periods of time during filling to mark different layers. For this, it would be great to be able to connect syringes to the inlet
I’ve worked for years with whatever hoses I could find, and tons of different adapters to connect the hoses to my reservoir, the tap, the tank. It would be so much less of a hassle if someone thought through which hoses will actually be needed, bought them at the right diameter and length, and outfitted them with the adapters they needed to work.
The tank needs to be accessible from the back side so the experimenter can run the experiment without walking in front of the observers (since the whole purpose of the tank is to be observed by students). The experimenter also needs to be able to get out from behind the tank without a hassle so he or she can point out features of interest on the other side.
Also, very importantly, the experimenter needs to be able to reach taps very quickly (without squeezing through a tight gap or climbing over something) in case hoses come loose, or the emergency stop for any mechanism pulling mountains in case something goes wrong there.
There needs to be enough room to have a class of 25ish students plus ideally a handful of other interested people in the room. But not only do they need to fit into the room, they also need to be able to see the experiments (they should not have to stand in several rows behind each other, so all the small people in the back get to see are the shoulders of the people in front). Ideally, there will be space so they can duck down to have their eyes at the same height as the features of interest (e.g. the density interface). If the students don’t have the chance to observe, there is no point of running an experiment in the first place.
Ideally, when designing the layout of the room, it is considered how tank experiments will be documented, i.e. most likely filmed, and there needs to be space at a sufficient distance from the tank to set up a tripod etc..
Both for direct observations and for students observing tank experiments, it is crucial that the lighting in the room has been carefully planned so there are minimal reflections on the walls of the tank and students are not blinded by light coming through the back of the tank if a backlighting solution is chosen.
In my experience, even though many instructors are extremely interested in having their students observe experiments, there are not many people willing to run tank experiments of the scale we are talking about here in their teaching. This is because there is a lot of work involved in setting up those experiments, running them, and cleaning up afterwards. Also there are a lot of fears of experiments “going wrong” and instructors then having to react to unexpected observations. Running tank experiments requires considerable skill and experience. So if we want people using the new room and new tank at all, this has to be made as easy as possible for them. Therefore I would highly recommend that someone with expertise in setting up and running experiments, and using them in teaching, gets involved in designing and setting up the new room. And I’d definitely be willing to be that person. Just sayin’ ;-)
Can you do a bottom Ekman layer demonstration without a rotating table? That’s the kind of challenge I like :-)
The way I’ve previously showed bottom Ekman layers is by spinning up a cylindrical tank on the rotating table until it reaches solid body rotation, adding dye crystals to visualise the circulation later, and then stopping the tank to create friction at the bottom (and the sides, but we are mainly interested in the bottom since we want a bottom Ekman layer) as the water continues moving but comes under the influence of friction. But what if we just invert the whole thing?
My initial idea was to use a Lazy Susan (you know, the kind of tray on a swivel base that you can use for your jam and honey etc on your breakfast table, but which you shouldn’t turn too rapidly (ask me how I know)) and to have a cylindrical vase sit on it, which will then be put in rotation and will rotate around and under the (initially still stagnant) water. The friction with the moving vase will then lead to a bottom-intensified circulation.
Problem here: While I have a Lazy Susan at home as well as a vase that would work as “tank”, I am currently in Bergen where I don’t have access to my own equipment. Instead, though, I have access to a rotating table in GFI’s basement which I used to simulate my Lazy Susan idea (Cool, eh? Simulating a non-rotating-table situation on a rotating table ;-)).
That worked quite well, didn’t it?
This, btw, is what the setup looked like:
The physics themselves obviously work in this setup. However, they will be really difficult to observe for several reasons:
So all in all: it was a nice idea, but either I haven’t thought it through well enough, or it is a whole lot easier to do with a rotating table. I would imagine that it’s quite hard to observe when you don’t know very well what you are looking for, so it is unfortunately not useful as a demonstration to introduce people to the topic. What do you think? Any suggestions on how to improve this and make it work at home?
How will lee waves look differently if we are using the asymmetrical mountain instead of the symmetric one? And is symmetry actually important at all or are we just looking at different slopes downstream while the upstream slope doesn’t have an influence on the wave field?
After admitting I had only ever used the symmetrical mountain to generate lee waves in the long tank in the GFI basement, I had to try the asymmetrical one!
There are a couple of reasons why I had not done that before:
But then it would be really interesting to see what the difference would be, right?
I tried two different stratifications.
Since I just wanted a quick idea of what this mountain would do, I used leftover water I had prepared for the moving mountain experiment. Since there wasn’t a lot left, I ended up with 11.5 cm fresh water, but only 4 cm salt water at approximately 20 psu (since I stretched the 35 psu a little).
What I noticed: A LOT more mixing than with the other mountain! Stratification is pretty much destroyed after the first run, usually we run back and forth a lot. This can be for several reasons:
And the shape of the “reservoir” that is being built up upstream of the mountain is different to what I have observed before: Running in either direction, the reservoir didn’t built up smoothly, but as a hump that was pushed in front of the mountain. Maybe because the internal wave speed in this case was very close to the speed of the mountain, something like 7cm/s, so the disturbance created by the mountain couldn’t propagate upstream. Is that an upstream hydraulic jump we are seeing there?!
What’s also cool: Lee waves are now not only happening as internal waves, but you see a very clear signature in surface waves! Usually all we see are surface convergences and divergences, adjusting the surface layer to the internal waves underneath. That we now see surface waves is, I am assuming, mainly due to the shallow water relative to the height of the obstacle.
Since I was not satisfied with this at all, I ran a second experiment:
First, I tried to set up the same stratification as for this lee wave experiment with the symmetrical mountain because I thought that would be easiest to compare. But I aborted that after having moved the mountain just a little because it was mixing so much that there stratification was destroyed completely and nothing could be seen. I ended up putting more dense water in and ended up with 12 cm pink (35 psu) and 4 cm clear freshwater. And this is what this looked like:
You now see a wave train with wave lengths longer than in the symmetrical case. Probably due to the longer length of the obstacle (even thought the waves are still shorter than the obstacle)? Or what sets the wavelength?
This time, with a faster internal wave speed of around 10cm/s while the mountain is still pulled with 7cm/s, we don’t see the “hump” in the upstream reservoir — the signal can propagate faster than the mountain and thus smoothes out.
So that is what I think is going on here. While the first experiment mainly showed effects of the stratification compared to previous experiments, the second one might provide some insight on the different slopes of the mountain, although I am not sure in what way. Do you see something I didn’t observe? How would you expect the different slopes to influence the lee waves?
I am so glad I tried this and I’m looking forward to thinking about this more! :-) Any insights you’d care to share with me?
This blog post is meant as guidelines if someone other than me might have to set up this demonstration at some point… Have fun! :-)
If I am working fast and nothing goes wrong, this takes almost 2.5 hours. Make sure you have enough time to set this up! Filling the tank takes time and there is not much you can do to speed up the process once you’ve started…
Here is a sketch of the contraption that pulls the boat:
Ask students to observe:
Below are movies of a couple of experiments which you could use in teaching instead of running the experiment for real (if for some reason running the experiment is not possible. But I would totally 100% recommend doing the experiment!). For a fun video, watch the one above (the ones below are cut to only show the tank so might be a little
boring less exciting ;-))
Ship pulled with 5g in the bag
Ship pulled with 4g in the bag (for a repeat, see experiment 4!)
Ship pulled with 3g in the bag
Ship pulled with 4g in the bag (again, because we like repeat experiments ;-))
This blog post is meant as guideline if someone other than me might have to set up this demonstration at some point… Have fun! :-)
Lee waves are the kind of waves that can be observed downwind of a mountain in the clouds, or downstream of an obstacle in a current as a series of undulations with crests parallel to the disturbance.
Students sometimes find it hard to imagine that a moving mountain should be equivalent to flow across a ridge. It helps to discuss how it would be really difficult to set up a flow in a tank: A huge amount of water would need to be moved without too much turbulence. Instead, it’s a lot easier to imagine the water is moving by moving a mountain through the tank, so the water is moving relative to it if not relative to the lab.
The size of the tank is 60×1.5×5 dm, so it can hold a total of 450l of water.
The mountain we use is 10.5 cm high and 1 m long and it’s symmetric, so pulling it either way shows similar lee waves (which is why I’ve always used it). There is a second, asymmetrical mountain on the shelf that I have never used*.
The stratification that we’ve found works well is 10 cm at 35 psu (here dyed pink) and 9 cm at 0 psu. This leads to an internal wave speed of approximately ~11cm/s.
Prepare the dense water in a barrel that sits high enough so gravity will bring the water down into the tank (see picture below). For the 80l barrel, you need 2.8kg of salt and 1/3 tea spoon of dye MAX.
You achieve the stratification by filling in the fresh water first through the bottom left inlet, moving the mountain over it, and filling in the dense water. That way the mixing is contained to the volume underneath the mountain which will be a lot better for your nerves (believe me!).
The system that pulls the mountain can go at two speeds: “fast” and “slow”, “slow” meaning 5m in 1:11min (7cm/s) and “fast” meaning 5m in 0:36min (14cm/s).
Here is where you run the mountain from:
Troubleshooting if the mountain doesn’t move:
As you see in the pictures above (or the movie below), there is a lot to observe!
Here is a movie of the lee wave experiment. Feel free to use it in teaching if you like! And let me know if you need the movie in a higher resolution, I am happy to share!
*Yes, this was true at the time of writing. But I am setting up that experiment as we speak. Write. Read. Whatever. Will post movies tomorrow!