A habitable (human could survive in its atmosphere without suits) desert planet which used to have large oceans is quite common in ScFi. I was thinking about the processes which could have been responsible for the loss of water.

I want to leave biological processes—like sandworms from Dune, out. Likewise, I'm aware of technological solutions, like orbital rings, but I want something natural. Also, water swallowing holes in the ground don't work out either and I want real water loss, not just deeper oceans and more land.

So I came up with with 3 ways how a planet might lose a significant portion of its surface water. For all discussions, I want to use an earth twin (0.02% by mass surface water) as reference, unless mass or orbital parameters need to be altered to make something happen. The definition for desert planet I'm gonna go by is less than 10% surface coverage by water.

  1. Atmospheric stripping and subsequent ocean loss, like it happened on Mars. But this destroys the atmosphere and further outgassing would bring up water again.
  2. Stripping via impacts. To achieve this, we need to hit the planet with a rock (which has at least the mass of the oceans) at escape velocity. Due to several rather obvious issues the true figure is going to be several orders of magnitude higher. Also, I aim to create a habitable desert world and not a lava planet/asteroid belt.
  3. Break up the water bonds with strong radiation (from the star or an nearby event). But aren't the odds of survival on that world terrible during such an event? Furthermore I am aware that unicellular organisms deep within the crust could survive all of those events, but I want my survivors to be macroscopic and multicellular.

So how does such a transition happen in the least destructive way?

  • $\begingroup$ Welcome to Worldbuilding! Could you please clarify your technological + orbital rings statement? I think I'm missing something. $\endgroup$
    – Dan
    Commented Feb 19, 2019 at 20:20
  • 1
    $\begingroup$ Secondly, is there any time constraint on these processes? Or does the planet simply need to be discovered in these conditions without much concern to how long they took to develop? $\endgroup$
    – Dan
    Commented Feb 19, 2019 at 20:22
  • $\begingroup$ 1. I don't see why there should be a strict time constraint. But lets go with 7 billion years, since this is a timeframe that sounds plausable considering the metallicity the universe required for earthlike worlds. The planet sould by now be a habitable desertworld if it formed 7 billion years a go. 2. The technologie statment referred to any way of achiving the transition via technologie. I know of plausable ways to do it that way. Things like using electrolysis plants to split up the water to have it escape or transport it of the planet with space elevators or giant orbial rings. $\endgroup$ Commented Feb 19, 2019 at 20:45
  • $\begingroup$ @user58321 if the planet has billions of years to lose its water, then any mechanism is good. The only problem is to have breathable atmosphere at the end. $\endgroup$
    – Alexander
    Commented Feb 19, 2019 at 20:49
  • $\begingroup$ I wish I had the scientific knowledge to put this in an answer, but what about 2 separate processes? One to remove the water, another to reconstruct the atmosphere? Many theories suggest that Earth's water was actually delivered (via comets/asteroids) after it's formation, so why not do that in reverse? Remove the water (and significant portions of the atmosphere), and then redeliver the atmosphere? Or even do both at the same time, impacts eject water in to the upper atmosphere, where solar wind/etc blows it away, while at the same time delivering atmospheric components $\endgroup$
    – Harthag
    Commented Feb 20, 2019 at 19:19

4 Answers 4


First, welcome to Worldbuild SE!

Per the reality-check tag, I feel the desired conditions posed in the original question are not possible within the confines of present-day known science.

The main problem, which you correctly identify, is not the loss of liquid surface water but the subsequent retention of a breathable atmosphere.

First, these caveats you describe are indeed correct:

  • Irradiating the surface to break down the water would make the planet unlivable and compromise the habitability of the atmosphere as well.

  • Any collision strong enough to wipe away 90% of a planet's water would obliterate the surface in the process.

  • A scenario similar to Mars by definition would result in insufficient atmosphere. However, I did consider whether it might be possible for there to exist a sort of in-between period where the surface water was already gone, yet enough of a breathable atmosphere still remained for a time. Sadly, it doesn't seem like that could happen, and even if it were possible, it would almost certainly be outside the temperature range for humans.

Another major issue arises with the atmospheric oxygen itself: If the surface of an Earth-like world were suddenly made barren, there would be no new sources of oxygen. And as it happens, oxygen likes to react with just about everything. On geological time scales, it would be a very short time before the air lacked the O2 necessary for humans. (In fact, those reactions would probably create a lot of new surface water in the process, funny enough.)

Logan's idea is interesting, but I'm having difficulty seeing how circumventing the cold trap in some way wouldn't simply lead to Venus-like greenhouse conditions or result in an otherwise inhospitable atmospheric composition. I'd be curious to see this idea developed more fully.

In any case, the least-catastrophic way to rid a planet of 90% of its water (without removing its atmosphere altogether) would be to boil it off, but that will still result in a runaway greenhouse situation. In short, I don't see how any natural processes could cause a planet to literally lose 90% of it's water and still leave it with a breathable, hospitable atmosphere.

That said, I would very much look forward to editing my response in light of any comments or additional information to the original question. Sadly, as the question stands currently, I don't see a way to reconcile the desired environment with the reality of atmospheric science.

  • $\begingroup$ You mention that oxygen would be lost on a barren desert world due to a lack of refreshment. Yet barren was never the goal. What I consider a desert world should for sure have large, sterile deserts, but I see no reason why regions towards the poles shouldn't still have ecosystems (especially since there still may be 10 % of thes surface covered in water). A few dry adepted plants and animal in the moderate zones and even scrubland in the Arctic are on the table in my opinion. Maybe there would be less oxygen, but it might still be produced. $\endgroup$ Commented Feb 22, 2019 at 23:41
  • $\begingroup$ @TheDyingOfLight I agree, but it doesn't address the issue of losing water without compromising the atmosphere. If the main goal is to have a desert planet, then this is a much easier question to answer. But if the main goal is for there to have once been an abundance of water which subsequently was lost, then we're still looking at the same difficulties as before. $\endgroup$
    – Dan
    Commented Feb 23, 2019 at 0:10
  • $\begingroup$ @TheDyingOfLight Also, would you mind elaborating story-wise on the reason behind wanting the water loss? Maybe it'll trigger some additional ideas in people. $\endgroup$
    – Dan
    Commented Feb 23, 2019 at 14:51

Eliminate the tropospheric cold trap.

Earth retains water because the atmosphere gets cold enough for water vapor to mostly freeze out before it gets thin enough for Jeans escape to be a significant factor. If water vapor could freely mix into the ionosphere, it would both be directly lost to thermal escape, and broken down by unfiltered solar radiation, allowing the hydrogen to escape.

Getting rid of the atmospheric cold trap without making the planet uninhabitable is a bit tricky, though. You will want to make the stratosphere colder and the top of the troposphere warmer to ensure that the temperature gradient is both monotonic, and doesn't ever get too cold to trap water. If it just needs to be habitable to some kind of not-terribly alien life, using normal Earthling biochemistry, the problem isn't too bad; there are plenty of ways to fiddle with atmospheric composition in order to smooth out the temperature gradient. E.g., increasing the fraction of CO2 in the atmosphere will both cool the ionosphere by increasing infrared emissivity, and heat the troposphere; put the planet far enough from its sun that it doesn't trigger a runaway greenhouse, and you're set. If you want it to remain habitable for unmodified humans, however, the constraints are much stricter; figuring out how to manage that is left as an exercise for the reader.

  • $\begingroup$ Earth retains water because the atmosphere gets cold enough for water vapor to mostly freeze out before it gets thin enough for Jeans escape to be a significant factor. Jeans? Could you please explain. Thanks $\endgroup$ Commented Feb 19, 2019 at 21:44
  • $\begingroup$ How does the process stop before all the water is gone though? That would sterilise the place in the end, wouldn't it? The very slow desertification is a cool aspect, but would end up killing everything in the end. $\endgroup$ Commented Feb 19, 2019 at 21:49
  • $\begingroup$ Whould other stuff escape too using the same mechanism? $\endgroup$ Commented Feb 19, 2019 at 21:50
  • 3
    $\begingroup$ @user58321 "A number of different mechanisms can be responsible for atmospheric escape, operating at different time scales; the most prominent is Jeans Escape, named after British astronomer Sir James Jeans, who described the process of atmospheric loss to the molecular kinetic energy." - Wikipedia: en.wikipedia.org/wiki/Atmospheric_escape $\endgroup$ Commented Feb 19, 2019 at 21:54
  • $\begingroup$ @TheDyingOfLight Other light gases would also escape, but there aren't many to be worried about it if you've got an oxygen-rich atmosphere and water oceans--hydrogen, helium, ammonia, and methane would be the most prominent. Technically, oxygen and nitrogen would also continue escaping, but probably less than they already do from Earth--i.e., at rates that are not worth worrying about. After all, all atmospheres evaporate--what matters is just how long they take to do so. $\endgroup$ Commented Feb 20, 2019 at 0:05

Water could also be lost because of geological processes (I know it sounds a lot like "water swallowing holes in the ground", but I think it is different enough to be worth mentioning).

It is a well-known (and slightly disturbing) phenomenon even on the Earth: in the subduction zones under the oceans, the water is absorbed by rocks and brought deep down in the mantle.
This phenomenon should be contrasted by emission of water vapor into the atmosphere by other geological processes (eructions), but it is not known if there is a balance or Earth surface is steadily losing water toward the inner layer of the planet.
Per this article:

The finding has major implications for our current understanding of the global water cycle, according to Wiens. If three times more water is reaching Earth’s interior than we thought, he says, then the amount going in is much more than what current estimates suggest is coming back out at the surface, for example, through volcanic emissions and degassing. “If that [imbalance] persisted for many millions of years … then the ocean would go away.”

So your planet could undergo some anomalous geological processes that accelerate this phenomenon (but I think it would require anyway a timespan in the order oh the milions years), causing the loss of the surface water. The water is still on the planet, but at depths where it couldn't be reached by an artesian well or partecipate to the water cycle of the surface.

  • $\begingroup$ Interesting article. It mentions a global average, but this, too, is a tenuous educated guess. We simply don't know how much is really being absorbed into the crust (or how quickly it is released again). But on a fictional planet, it might be feasible to devise conditions where, at least for the time being, its surface is inhabitable. I imagine it would require a somewhat different geological picture that we have currently on Earth, but similar enough that if the OP wants to relax their requirements, this might be a good approach. $\endgroup$
    – Dan
    Commented Feb 23, 2019 at 2:29
  • $\begingroup$ Subduction doesn't just trap water underground. It also involves chemical reactions of the form: oceanic crust rock + water → continental crust rock. (The product of the reaction is less dense than the oceanic crust rock, so it tends to pile up into continents.) $\endgroup$
    – Jasper
    Commented Jul 1, 2019 at 0:37

Chemical binding - a large swarm of meteors crashed into one of your oceans, and cracked open. The interior of the meteors contained vast deposits of Alkali metals (Lithium, Sodium, etc), which began reacting with the water to produce Salts and Hydrogen. The Hydrogen drifted into the upper atmosphere, and escaped into space.

  • $\begingroup$ Would there be such asteroids though? Since those elements are highly reaktive, they should react in the proto planetary disk. Furthermore this process sounds quite violent energy wise and will poison the environment. So not so great for habitability afterwards. Or am I missing something? $\endgroup$ Commented Feb 22, 2019 at 23:50
  • $\begingroup$ @TheDyingOfLight You are right that the chances of such an asteroid forming are slim (although not quite impossible). Yes, the reaction would be energetic - you would probably have a bright light, immense heat, and large plumes of steam given off. However, given the reactants, the only "poisoning" of the environment will be that your ocean gets saltier. $\endgroup$ Commented Feb 24, 2019 at 22:59

You must log in to answer this question.

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