A common, matter-efficient science-fiction habitat is a hollow cylinder or ring in space that is spun to simulate the pull of gravity on its interior surface. These habitats have been imagined as small as a spaceship, mere meters in radius, up to a ringworld, 1 AU in radius.

Say we have a rotating space habitat designed to mimic Earth’s gravity and atmosphere at sea level. Assume that the habitat has been rotating for however long it takes to reach whatever equilibrium state can be achieved. Will the rotational motion of the habitat generate winds in its atmosphere? In which direction will they prevail, spinward (with the direction of rotation) or anti-spinward (against it)? Will they circle the ring in a single direction or will there be antiparallel winds at different altitudes?

Ideally answers would be applicable to rotating habitats of various size, however, if the answers would vary widely between habitats please use a ring 10,000 km in radius for the answer.

Edit: There have been some previous questions about weather on rotating habitats such as my own What would the weather be like in an asteroid habitat? and Weather on a mini-ringworld/Banks Orbital. These questions are quite broad as weather is a complex subject, to put it mildly, and so were not very answerable. Here I have attempted to narrow the focus of the question to a single aspect of weather, the wind, in the hopes of generating more complete answers.

  • 2
    $\begingroup$ @Aify It's funny being told my question is a duplicate of my own question from almost 4 years ago. I think part of why that old question didn't generate many useful responses is that it was far too broad. I've intentionally tried to narrow this question to only discuss wind which I think is much more answerable and certainly wasn't covered in any of the answers to the previous question. $\endgroup$ Commented Jul 31, 2018 at 22:22
  • 1
    $\begingroup$ A dupe is a dupe. The last paragraph of the accepted answer on that question specifically includes a line stating that depending on which way the "wind" was blowing, the weather (thus the air, and therefore the wind) would either be stagnant ("wind" moving with the rotation) or travel in loops (against the rotation). $\endgroup$
    – Aify
    Commented Jul 31, 2018 at 22:27
  • 6
    $\begingroup$ @RonJohn, Larry Niven's ringworld is really big. As far as alien megastructures go, this is just a golf ball. 😝 $\endgroup$
    – JBH
    Commented Jul 31, 2018 at 22:34
  • 6
    $\begingroup$ I disagree that this is a duplicate. My answer only briefly talks about wind patterns, and definitely doesn't describe in what directions they move. I used the line "Either it would be stagnant - not a good thing - or it would travel in loops, depending on which way the wind (artificial?) was blowing.", but I don't say explicitly that the wind would travel in loops. It was a guess, an unsupported guess, and not really elaborated on in much detail. I don't think my answer there answers this question, and so I don't think this question is a duplicate. $\endgroup$
    – HDE 226868
    Commented Aug 1, 2018 at 14:54
  • 2
    $\begingroup$ @Upper_Case Let's say it's open like a Ringworld but not nearly as large. I would like it to be small so that the angular velocity changes significantly as you go up in altitude. In this case, based on its size it would be more like a Bishop Ring or a Banks' Orbital. $\endgroup$ Commented Aug 6, 2018 at 23:52

5 Answers 5


The real answer is we don't know, we can make some guesses though.

Lets start with ground effects; the range of effect of pure physical topography on the atmosphere, not including heating effects, is only about half a kilometer. On Earth this is purely vertical, as are a number of other effects that we'll talk about, but on a Ringworld type structure the effects are also horizontal because of the side walls that keep the atmosphere in. So near the floor and walls the winds are going to be relatively turbulent but the prevailing wind pattern will be a gentle anti-spinward breeze as the air is dragged at by the ground topography, and the walls, but doesn't move quite as fast as the ground. The ground will be moving at roughly $ 10000m/s$ to supply $ 1g $ of pseudogravity, if the air immediately above it is only doing an average of say $9990m/s $ local winds could be quite strong due to turbulent flow but the air mass as a whole will be doing $ 10m/s $, or $36km/h $, anti-spinward; while I'm not sure of the magnitude I would expect that to be the case.

Then there's heat effects; the floor of the ring will be hotter than the walls due to solar angle effects this is going to create a wind system very similar to a Hadley Cell with the centre of the ring as the equator/tropics and the walls as the subtropical desert zone. The reason I asked about the width of the ring is that over millions of kilometres, like Niven's ring, this effect wouldn't be so singular but a "narrow" 200km ring is going to see an almost perfect singular formation.

Net effect; surface winds are going to be turbulent but blow prevailingly from the edge inwards along the floor of the ring with an anti-spinwards twist, similar to the Coriolis Effect but with less deflection, high altitude winds are going to be fairly laminar and travel from the centre of the ring out towards the walls with ever more anti-spinward drift as they gain altitude. The wall areas will be relatively dry as most rain will fall as the air initially rises close to the centre of the ring.

Note this answer assumes reasonably vertical walls rather than some smooth bowl shape, not sure how that would go, probably broadly similar. It also assumes that the ring floor is "flat" not meaning particularly smooth but without any full-width profiling to either a "^" or "v" shape. I have not included the effects of a Shadow Square system in this model at all, I can try but they'll be extremely complex and vary considerably according to proportional day length and a number of other variables.

  • $\begingroup$ Excellent answer, your assumptions about ring shape are all reasonable. I was wondering if you had considered the effect of the system's rotation on rising and falling hot and cold air currents. As air near the surface of the ring heats and rises it will have a greater spinward velocity than the cool air it displaces which will, in turn, have a large anti-spinward velocity compared to the air it is descending in to. I think this would add another dimension to your theorized air cells deflecting rising warm air spinwards and cool air anti-spinwards. $\endgroup$ Commented Aug 7, 2018 at 16:29
  • $\begingroup$ Another interesting aspect of this model that you could incorporate if you cared to do so is the Eötvös effect. This is simply the idea that spinwards winds will be effectively heavier than "stationary" air anti-spinward winds will be lighter. The effects are relatively small on Earth but I imagine would be magnified on the rapidly spinning ring. $\endgroup$ Commented Aug 7, 2018 at 16:40
  • 1
    $\begingroup$ @MikeNichols With regard to the velocity differences in rising and falling air I think you'd find that compared with the turbulence introduced by the land and wall friction and the way that the surface winds interface along the central strip of the ring floor the temperature based mixing effects would be overwhelmed and obscured by larger effects. The Eötvös effect may be magnified considerably, since Ω is effectively doubled, but I'm not sure that the Eötvös effect model holds for the atmosphere in this situation, out of contact with the ground on a ringworld things get a bit odd. $\endgroup$
    – Ash
    Commented Aug 7, 2018 at 17:10
  • $\begingroup$ a great answer, i might add a smooth bowl shape would probably make the hadley cell more prominent as the continuous temperature gradient would help stabilize it more $\endgroup$ Commented Aug 7, 2018 at 21:20
  • $\begingroup$ @taylorswift It may but I think the relatively low difference in angle of incidence between the floor and walls of a bowl would actually act to weaken the overall effect. $\endgroup$
    – Ash
    Commented Aug 8, 2018 at 17:02

Convection is the answer.

Imagine you are inside a car, traveling at (just to say any speed) 65 mph. You, your clothes, your seat, and (of course) the air inside the car will be traveling at the same speed: 65 mph. And you will not feel any air current inside despite the fact you are moving fast.

If you have an isolated gas inside a rotating cylinder or a ring (no matter the size), the gas will rotate at the same speed as the other objects inside the cylinder or ring, so you will not feel any wind at all. BUT (here comes the interesting part): what really causes the wind currents in the atmosphere are the thermal differences. And here goes a further explanation:

If your cylinder has a "night" part and a "day" part, the part in the daylight will be hotter than the part in the dark. Therefore, you will then have the first "convection circuit": The hot ground in the day will heat the air nearby, and the air will raise (as a hot air balloon), dragging cold air from the night side. If you have water involved (lakes, oceans) the water acts as a heat reservoir, and includes additional convection circuits in the loop.

And of course there are the cities. The cities generate a lot of hot air that goes upward, and creates additional air currents.

And finally, now that you have "moving air" then yes, your moving air is affected by the rotation of your space station the same way the sea and air currents are affected by the Earth rotation.

The more "realistic" model you want your air currents to be, the more variables you need to include in your model. And, well... it is actually a science by itself. But you can have a very good starting point defining your day and night regions, and your water regions.

  • 2
    $\begingroup$ An interesting note (though I don't know if it will change your answer), but the air inside your car does in fact move (or slosh around) when you accelerate or decelerate. You can think of air as a very thin liquid. In a rotating cylinder you have air close to the axis that will not be moving, and air near the surface that will be moving fast, and along with convection I think you could also get some wind from the differences in these layers. Of course the size of the station does matter. You might not notice this effect on the Ringworld. $\endgroup$
    – AndyD273
    Commented Aug 1, 2018 at 15:11
  • 1
    $\begingroup$ Also, this effect would be a lot different on a ring than a cylinder. In an air filled cylinder you have a spot in the middle where things aren't moving, and the whole outer surface is. In a closed ring the ceiling would also be moving, and so the lower air and higher air would be pulled along almost equally. This is also different than a large open ring like the Ringworld, where there is nothing but centripetal force holding the air in, but there is still no still air at the axis, since the axis is in vacuum. $\endgroup$
    – AndyD273
    Commented Aug 1, 2018 at 15:33
  • $\begingroup$ this really isn’t an answer, though it might be a starting point for a simulation that could provide the answer $\endgroup$ Commented Aug 7, 2018 at 21:17
  • $\begingroup$ @AndyD273 The rotating habitat doesn't accelerate or decelerate though $\endgroup$
    – endolith
    Commented Oct 5, 2020 at 14:29
  • $\begingroup$ @endolith That was just to illustrate that air sloshes. The point was that you're going to have areas of more or less air movement, depending on how close to the center of rotation you are. A ring would have less of this than a cylinder, since it has both a floor and ceiling to drag the air along, as well as any internal structures. A cylinder will have that near the "ground", but less so up near the axis of rotation. This could create turbulence and potentially wind. $\endgroup$
    – AndyD273
    Commented Oct 5, 2020 at 15:08

Example with an O´Neill cylinder

Because this is a question science based, let´s take a standard O´Neill cylinder. This was a space settlement design proposed by American physicist Gerard K. O'Neill in his 1976 book "The High Frontier: Human Colonies in Space". You can find more information here:

This design is a cylinder with six stripes along. 3 of them for settlement (land) and 3 of them as windows. Each of the windows has a movable mirror that simulates the day-night cycle. There is plenty of documentation about this concept. In this link you can find a nice diagram (I am also including the image): https://www.artstation.com/artwork/28NNB
ONeill Space Colony

And this is how it may look from the inside: enter image description here

That space colony, viewed from the top of the cylinder will look like this:
enter image description here

And the sun will enter into the space colony this way (remember the mirrors open and close to simulate day and night): enter image description here

Now here comes the answer to the wind question:
The 3 land stripes will be heated by the direct sunlight, and the 3 glass windows will be cold (just like the glass window in your house when you touch it on a very cold day. Space is quite cold when not in the sunlight, and the windows must be transparent so the sun can enter. Therefore, the windows will be cold and the land will be hot. That creates the first convection circuit (the first winds in your cylinder): enter image description here

However, the cylinder has to rotate to generate gravity. And the wind particles (as mentioned above in previous answers) will be rotating faster in the positions away from the center of the cylinder, and motionless in the axis along the center of the cylinder. So the loops of the wind currents will suffer a slight deformation as depicted below. Note the additional effect of the 3 mirrors in the center of the cylinder: the 3 mirrors will be heating directly the air in the central axis. So, adding this to the convection currents, will result in a region of motionless hot air in the center axis of the cylinder. enter image description here Finally, please remember (and here comes the disclaimer) no one has ever built anything like this (well, not in our Solar System). So we really can´t be completely sure about the winds or anything weather-related inside a colony like that (we can´t even predict completely our own weather here on Earth).


Ash and boxcartenant's answers both claim that there will be a constant wind on the habitat, perceived as anti-spinward by residents, due to "resistance" against the spinning. However, after some long conversations, I am reasonably sure that this is a misunderstanding, and will add my own answer:

Thermal equilibrium

Note that this ignores thermal effects of nearby stars, so you have a universe with uniform background radiation (to keep the fluid from freezing), and you spin up a ring or cylinder or torus with fluid inside it.

The torus is easiest, so I'll start with that:

At first, the fluid touching the surface will be dragged along with it (due to "no slip" condition), and the fluid in the center of the tube will not (due to inertia). It will have the same parabolic velocity profile as fluid laminar flow in a pipe:

Parabolic flow velocity profile

From the perspective of someone standing on the inside of the tube, this would appear as a "jet stream" of wind in the center altitude above their heads.

After a sufficient period of time, however, of the tube rotating at constant speed, no longer accelerating, the momentum from the surface will be transferred from the outer layers to the inner layers, and so on, until all of the fluid is spinning at the same constant rate as the tube. The fluid is now in hydrostatic equilibrium, the entire habitat in uniform solid body rotation, and there is no wind. Any turbulence or cycles will wear themselves out and dissipate into heat. This is the state it will stay in forever, without any outside interference.

Because of the rotation, the fluid near the outer wall will be more compressed than the fluid near the inner wall, so it will have a pressure gradient like Earth's atmosphere.

In a walled cylinder habitat, the same thing will happen, except there is no inner wall to drag the fluid.

In an open cylinder or ring, like a Banks orbital, the fluid in the center will just drift away and be lost, because there are no walls and it's not adhering to the surface. Only particles that collide with the surface will become carried along with it.

Initially, the particles are not rotating, but as long as they have non-zero drift velocity, they will eventually encounter the ring, which will impart momentum on the particle and then capture it.

So there is no inherent wind just because the habitat is rotating. This will only happen at spin-up or spin-down (as described in Rendezvous with Rama).

Not equilibrium

If you add a star nearby, of course, the thermal gradients from cycles of night and day will cause new flows of fluid, relative to the surface.

I'm much more hazy on what happens in this case. The Coriolis effect acts vertically instead of horizontally, so either there won't be cyclones, or they won't have a preferred rotation direction. This effect may also cause a prevailing wind, I don't know.

  • 1
    $\begingroup$ I agree with you that the equilibrium system will exhibit solid or rigid body motion in the absence of thermal effects. Thank you for that insight. That's a very cool physics simulation program. I believe the vertical Coriolis effect will cause a prevailing wind with cycles of spinward wind deflected down, downward wind deflected anti-spinward, anti-spinward wind deflected upwards, and upwards wind deflected spinwards. This should create a prevailing anti-spinward wind near the outer surface of the ring assuming the effect is strong enough. $\endgroup$ Commented Jul 9, 2021 at 16:25
  • $\begingroup$ @MikeNichols I don't know that the Coriolis effect is that strong, though. I think convection from day temperature and night temperature will be much stronger? And this will happen at both sunrise and sunset in opposite directions? Can we recruit a meteorologist to answer this? $\endgroup$
    – endolith
    Commented Jul 12, 2021 at 22:26
  • $\begingroup$ The magnitude of the Coriolis force is a function of how fast the world is spinning. An orbital ring spins much, much faster than the Earth so its Coriolis force is much stronger. As an additional consequence the faster rotation will lead to shorter days and nights (on the order of hours or even minutes) and so less temperature variance and less wind driven by convection between day and night. Hot air will still rise from the surface but more uniformly and be driven spinward by the Coriolis force. For those reasons I think Coriolis winds will be the prevailing ones. $\endgroup$ Commented Jul 17, 2021 at 1:13
  • $\begingroup$ @MikeNichols I'm thinking of a Banks orbital which spins at the same rate as the Earth and has the same day-night cycle $\endgroup$
    – endolith
    Commented Jul 17, 2021 at 18:16
  • 1
    $\begingroup$ Ah, then yes, for a ring on the order of millions of kilometers in diameter the Coriolis force would be much less important, but for my purposes, I'm trying to represent something more on the scale of tens of thousands of kilometers in radius. Still, instructive to recognize that Coriolis forces and differential heating are both a function of ring-size, thank you. $\endgroup$ Commented Jul 17, 2021 at 21:25

Assuming this is an open ring (i.e. no ceiling), in order for the atmosphere to be dense at the edge of the ring, the gasses in the atmosphere would have to be affected by the same centrifugal forces which are simulating the gravity, so they have to be spinning with the ring. Also, as long as there is any friction or obstruction at all on the surface of the ring, it will drag the gas along with it. So in general, it will feel like there is no wind at the surface

Now, the air closer to the center of the ring will be less affected by the friction and obstructions on the surface (in addition to having a slower linear speed). If the ring is open, then the upper atmosphere will not be as motivated to keep pace with the angular speed at the surface, so when people go up on a high mountain, they will feel some wind in the anti-spinward direction. (It will feel like wind to observers, but the air really just has a slower angular speed).

Like Carlos Zamora said, though, the more variables you add, the more realistic your model will be. If you have some other body in near orbit to the ring, its gravity will affect tides and wind; if you have a night/day cycle, the temperature will affect wind; if you have mountains, they will affect the shape of the wind by obstructing it, and their gravity will affect the barometric pressure there; etc..

  • 1
    $\begingroup$ I’m not following fully your argument for anti-spinward winds at high altitude. I agree with you that there will be very little force acting on the air at high altitude to have it keep pace with the surface of the ring. However, in order for it to have an anti-spinward velocity, there must be some force acting against the wind, slowing it down, to counteract the small amount of friction acting to speed the wind up. Otherwise, eventually, the system reaches an equilibrium where there won’t be net wind. What is that "drag" force? $\endgroup$ Commented Aug 7, 2018 at 14:34
  • $\begingroup$ The force acting against the wind is the fact that it wants to move in a straight line, but it has to keep changing direction to rotate with the ring. This means it can't keep its own momentum, and will always tend to slow down a little. $\endgroup$ Commented Aug 7, 2018 at 15:34
  • $\begingroup$ "then the upper atmosphere will not be as motivated to keep pace with the angular speed at the surface" Why wouldn't it be? It has friction rotating it, and nothing stopping it from rotating $\endgroup$
    – endolith
    Commented Oct 5, 2020 at 14:32
  • $\begingroup$ @endolith since it has to keep changing direction (rotating), it will have a tendency to slow down. $\endgroup$ Commented Oct 22, 2020 at 23:11
  • $\begingroup$ @boxcartenant That's the centrifugal force pulling it down towards the surface. It wouldn't change it's speed relative to the surface $\endgroup$
    – endolith
    Commented Oct 23, 2020 at 0:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .