In this article, scientists are investigating river channels on Earth, Mars, and Titan to see how rivers modify their landscapes. The following quote caught my eye:

Rivers are, in a way, topography bystanders that always flow downhill.

But is the above statement necessarily true? In this question the OP asks about terrestrial rivers temporarily flowing uphill on a map, but most of those answers involve "resolution" issues (where average gradient is uphill but local gradient is downhill), pressurization, or alternate physics. The answers to the linked question are all local geographic tricks. OP there is focused on earth and cartography. I'm looking for physics on alternate worlds, not necessarily talking about water. Titan has hydrocarbon rivers, Europa has cryovolcanoes, Venus has molten rivers.

Would it be possible that a certain combination of environment, chemistry and physics could create "uphill rivers" across significant distances and timescales?

For instance, a naturally occurring superfluid that flows upward between multiple reservoirs or something more exotic?

EDIT: To clarify the parameters of the question, a river is defined as a naturally occurring stream of water or other fluid flowing through a channel toward a reservoir, such as an ocean or lake. This means things like underground streams certainly classify as rivers, but something like a geyser does not (since it does not flow toward a reservoir).

  • 3
    $\begingroup$ So what's the difference between these two questions? For me it looks like got ate asking the same thing again not because its different but because you dislike answers there, and that's what bounties are for. $\endgroup$ – Mołot May 22 '17 at 15:31
  • 1
    $\begingroup$ Well, rivers are all about following the path of least resistance downwards, right? So either change the meaning of 'least resistance' (complicated), or change 'downward' (not so complicated). Eg, a local gravitational inconsistency would do it, no? $\endgroup$ – Konchog May 22 '17 at 16:04
  • 2
    $\begingroup$ I don't see how this question is different from the question you linked...in your question. At least one answer specifically talks about longer distances etc. $\endgroup$ – James May 22 '17 at 16:19
  • 1
    $\begingroup$ @Mołot The answers to the linked question are all local geographic tricks. OP there is focused on earth and cartography. I'm looking for physics on alternate worlds, not necessarily talking about water. Titan has hydrocarbon rivers, Europa has cryovolcanoes, Venus has molten rivers. $\endgroup$ – Isaac Kotlicky May 22 '17 at 20:36
  • 1
    $\begingroup$ The fact you dislike answers is meaningless when it comes to duplicate votes. What matters is how questions differ. $\endgroup$ – Mołot May 23 '17 at 14:22

There are a few possibilities that spring to mind:

  • Capillary action: This requires very thin spaces for the water to move through, it could be some absorbent rock with small regular spacing. You see this effect if you fill a bath and then hang a towel over the edge, the water will soak up the towel and drip from the end hanging outside the bath.

  • Very high surface tension (and viscosity): As long as the other side of the hill was lower and the surface tension of your liquid (we won't be using water here) is high enough then we can get the liquid to flow over the top of the hill and down the other side. This 'river' would be more likely something you could walk across than swim through though, something like pitch or treacle.

  • Magnets: Perhaps your liquid is a ferofluid (a liquid that becomes magnetic) and your hilltops are magnetised. Once the liquid gets up there maybe wind or just the potential energy are enough to push it down the other side.

I doubt the last two are likely to occur and the first is not a 'river' as such, but I'm not sure what it is you're after so perhaps those ideas will fit.

  • $\begingroup$ Interesting answers. Whether it's likely isn't the question, only whether it's a feasibility under the known laws of physics. I'm curious if there's a naturally occurring compound that would fit for your second and third answers... $\endgroup$ – Isaac Kotlicky May 24 '17 at 15:20
  • 1
    $\begingroup$ For a real-life example of the second option: youtu.be/bOSVX8zOPkc $\endgroup$ – Tyrannosaur May 25 '17 at 20:01
  • $\begingroup$ @Lio Elbammalf - Capillary action may help the water go against gravity, but in that case, the water cannot simply drip and flow unless the exit point is lower than the point of entry to the capillaries. This, should be done through the pressure of the water column above it or through evaporation. That is why the Boile's flask cannot work. rationalwiki.org/wiki/… $\endgroup$ – Christmas Snow Jan 25 '19 at 9:31
  • $\begingroup$ For option 3: if the magnetic hill forms a naturally occurring Halbach Array, then all you need is for the momentum of the flow to carry it past the array and into the "no field" zone $\endgroup$ – Chronocidal Jan 25 '19 at 12:54

Rivers do flow uphill. Twice a day, most of the time.

Severn Bore

This is the Severn bore in England. The water on the right is the normal, downhill flow of the river; the water on the left is the incoming tide, flowing uphill. Tides are a pretty complex phenomenon, but the important part (for our purposes) is that the gravity of the Moon is causing water to flow differently on Earth.

Let's define 'downhill' as 'towards sea level'. That's generally what it refers to. In that case, to get a river to flow 'uphill', or away from sea level, all one needs is a temporary or local effect that draws the water up along a slope in order to balance itself out. For instance, if a river encountered a gentle ridge that had a very, VERY, VERY dense object inside it, it might be pulled up to the top of the ridge, briefly flowing uphill, then drop away in a waterfall on the other side.

Another possibility is something called a Hydraulic Jump. This happens when fast-moving water reaches a slow-moving body, such as when a waterfall or rapids meet a slow-moving stream. The pressure of the fast-moving water forces the level of the slower-moving stream upwards, which could cause the river to briefly flow uphill.

  • $\begingroup$ With your "very, VERY, VERY dense object" do you mean something massive enough to pull the water up using gravity? I kinda get the feeling that something that dense would mean you get a point between it and the center of Earth where things would float because the pulls cancel...not really a sciency person but I would guess that would have something unrealistic about it? $\endgroup$ – FreeElk May 26 '17 at 8:29
  • $\begingroup$ @FreeElk Yes, something that massive would have to be either magical or technological in nature. You probably wouldn't find a point where the gravity between it and earth would cancel out, though, because it would be a part of the planet...just a really, really dense part. $\endgroup$ – Werrf May 26 '17 at 13:41

Self-siphoning non-Newtonian fluids

from https://www.teachersource.com/product/poly-ox-a-self-siphoning-gel/chemistry These viscous fluids can siphon themselves. The intramolecular adhesion between the long molecules are such that the fluid can be dragged uphill as long as the ultimate destination is lower than the start.

Polyox is the one you always see. The molecule has loads of intermolecular hydrogen bonds which is (I think) how it can pull itself along.

enter image description here

Supposedly PVA slime (the glue/borax slime) is also self siphoning & it should be. But I bet it is tricky to get the consistency right, and I found no images. Super long carbohydrates like dextran should also be able to self siphon.

It would be a gooey river for sure. If the waterfall part were slow and gooey enough I wonder if you could float up it against the current, Minecraft style?


How about a dense gas?

You have something hot, a geothermal vent or something, at the bottom of a hill. Your liquid passes over this and evaporates but, because it is dense, still stays low down. With the heat rising up the hillside the gas moves with the air current until it gets to the top of a hill, where it is cooler, and condenses again and runs down as a liquid.

So it isn't a river of water but more of a thick fog...but it might work.

  • $\begingroup$ This is somewhat similar to the superfluid I mentioned. It's a good start, but can you flush it out with some more science? $\endgroup$ – Isaac Kotlicky May 24 '17 at 15:17
  • $\begingroup$ If it's so dense, why is it moving up the hill? Wouldn't the hot air-current, being less dense, just flow straight up to the sky, leaving the dense gas to fall in below? $\endgroup$ – colmde May 25 '17 at 9:25
  • $\begingroup$ @colmde I was just thinking the wind could blow it along. It wouldn't go up so much because its heavier than the air but it isn't heavier than the ground on the hill. You get fog staying at ground level on hills and stuff. $\endgroup$ – FreeElk May 26 '17 at 8:32

What about an underground river? Of course it would have to go downhill eventually, but the water may be "sucked" uphill from a large reservoir, but exit the earth at some point far away that's lower than the reservoir.

As for how it would form in the first place I don't know - perhaps it was once a more traditional underground river, but it dried up for a time, and then a huge flood started the water flowing "up" the channel (at this point the reservoir level was higher than the highest point of the underground channel) and pushed it far enough to come out the other side. After the flood receded, the siphoning process continued.

Of course if the reservoir ever dried up enough that air could get into the tunnel, then the flow of the river would stop forever (or at least until the next superflood)

  • $\begingroup$ Side point - Looking into this I'm learning we're not 100% certain how siphons work... science is wild! $\endgroup$ – Isaac Kotlicky May 24 '17 at 15:23

If river, of any kind, would flow "uphill" it would actually flow upward. The only reason water river goes downhill is that there is something beneath it. If the surface ends water flow down and we call it waterfall.

So the easiest river of your kind to imagine is a river of hot air being obstructed by ceiling tilted as some degree.

  • $\begingroup$ Flowing directly upward doesn't fit the definition of a river, which is a flow in a channel. $\endgroup$ – Isaac Kotlicky May 24 '17 at 9:22
  • $\begingroup$ That' why you need some kind of ceiling that will stop the upward movement and force material to make and look for easiest way up thus making stream bed. $\endgroup$ – SZCZERZO KŁY May 24 '17 at 9:48
  • $\begingroup$ can you explain how such a system would operate and be naturally occurring? $\endgroup$ – Isaac Kotlicky May 24 '17 at 15:29
  • $\begingroup$ Imagine a reversed river section. So you have rock at the top, water and then air. If you switch water with some other fluidy thing and air with another substance you can have reversed river. Kind like with this toys s-media-cache-ak0.pinimg.com/736x/4f/10/a0/… $\endgroup$ – SZCZERZO KŁY May 25 '17 at 10:15

I can think of three possibilities that would make an upward-flowing river:

Bubble River

When we think of rivers, we generally think of water flowing over rock or earth, down towards the ocean. But what if the "river" were lighter than air? A light gas would float upward, along a ceiling. If the gas originated deep in the earth, it would follow a long, winding channel, not unlike a river, just upside down. For a full cycle, the gas undergoes some chemical change at the end point - a "lake" of gas, that turns to liquid and runs down into the earth, eventually returning to the original path.

Alternately, the river could be underwater; perhaps oil, floating on still water; it bubbles to the surface, following underwater channels. A river of oil would seem to run uphill, as the water it is displacing forces it up.

Outside Forces

The second way a river could flow uphill is external physical forces. Tides are one example, but a moon large enough to cause significant tides in rivers would be... dangerous. Instead, natural vents cause inverse waterfalls; as water, quickly speeding along a narrow river, flows over a thin vent, the force of the super-heated gas shoves it upward and forward, launching it into a pool a few feet above. The water continues to rush forward, possibly encountering other vents that jump it up again and again.

Alternately, a natural wind, shaped by a rock cavern, whirls around over a pool, whipping the water into the air, where it splashes against the walls of he cavern. The water runs down from the top and is caught by a lip, which funnels the water out, now much higher than it was. A series of caverns like this could transport water up a considerable distance, as long as the wind is blowing - and there are places on earth the wind almost always blows. For a beautiful image, the water may be transported high into the air, then splashes down as a "vanishing waterfall" - as the water plummets, the wind simply blows it away. It's likely the water eventually makes its way down to a lake below, where the process repeats.

Illusion of Height

It could be that highly-pressurized water from deep in the earth, or water that's come from a great height, is forced through a natural nozzle; the water would arch upwards, appearing to travel up-stream. However, the enormous erosion forces would keep this from lasting very long, and technically, the water would always be flowing down - but for a brief moment, the water would jet upwards.


Maybe an underground pipe-like tunnel filled with water would work. The river would have a brief uphill journey, then a longer, steeper downhill journey so the downhill journey causes a 'vacuum' that sucks the water to go uphill a little. The tunnel the water is in would have to be airtight otherwise the water wouldn't get pulled up by the downhill stretch of river - air would be pulled in.

This system would act like a hose where one end is in a body of water and the other end is much lower than the body of water. If the hose if filled with water, it keeps pulling water out of the body of water because of gravity.


A world with different physics might have that. The master of this, Greg Egan, has exactly that in his Dichronauts universe.

He works out the consequences of a “double plus, double minus” space-time metric. More generally, you can invent reasons why some directions are different from others, either due to the nature of space itself or some external effect like the magnetic field of Forward’s Dragon’s Egg.

His web page includes an explaination of the effect and a physics sandbox where you can stack blocks and see it slide “up” the ramp in suitable conditions!

Since the force of gravity is directed vertically downwards, we will define gravitational potential energy to be a positive multiple of the height of the object above the ground. This is consistent with the kinetic energy associated with vertical motion always being positive. So far, this is the same as the situation on the surface of the Earth.

But because we have chosen the horizontal direction of the ramp to be of the opposite kind to the vertical direction, the kinetic energy associated with horizontal motion will have the opposite sign to that for vertical motion. If the ramp is at a slope of less than forty-five degrees, an object sliding along the ramp will be moving more rapidly in the horizontal direction than the vertical, so its overall kinetic energy will be negative. That enables it to ascend up the ramp, because its gain in gravitational potential energy is balanced by its negative kinetic energy.

I must say though, seeing it in a simulation of the new physics is a lot “harder” SF than any amount of handwaving!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.