In the book Dune by Frank Herbert. Arrakis is a desolate planet composed entirely of deserts. Water is a precious commodity due to its rarity. The Fremen have found ways to live on its surface, despite its seeming lack of habitability. My question concerns the following:

Could a planet with an atmosphere that is similar in proportion to Earth's form without any major bodies of water on its surface?

The atmosphere of Earth has two major components:

  1. Oxygen- 20%
  2. Nitrogen- 78%

Is the presence of water necessary to get such a balance in the first place so that Earth-like organisms could survive on its surface?

  • $\begingroup$ Are we assuming that it has to be habitable for organisms with biochemistries similar to those of life on Earth? Cool, thanks. $\endgroup$
    – HDE 226868
    Commented Dec 13, 2015 at 16:39
  • $\begingroup$ @HDE 226868 Yes we are assuming this, I'll update the question. $\endgroup$
    – Quiquȅ
    Commented Dec 13, 2015 at 16:39
  • $\begingroup$ Rivers, lakes and oceans beneath the solid surface would be okay? What about a multitude of island-sized surface pieces (made of ice or something else) floating atop a global ocean which was only visible at some boundaries? I think there are moons in our solar system which are thought to be something like that. $\endgroup$
    – Crissov
    Commented Dec 13, 2015 at 19:29

3 Answers 3


I think it's unlikely but possible. The thing about water is that it's a very light molecule. Mars started with water like Earth, but lower gravity and missing magnetic field meant that it rapidly (geologically speaking) lost that water into space.

So if this world starts with little water how soon until it has none? Not long enough for photosynthesis to evolve and run for long enough to create an oxygen atmosphere. (That took the best part of three billion years on Earth).

So to the unlikely possibility. This is a dying world. It started a lot like Earth but held its initially abundant water less well. Now it has just a few million years left before life can no longer adapt to the dessicating biosphere. But that's plenty of time for a story on a human scale. Arrakis, with deep time stealing its water rather than sandworms.

Note: I'm assuming that Earth like planets will always start with a reducing atmosphere (mostly methane) like Earth did. Oxygen is highly reactive and unstable in the presence of methane or iron-2+, so I think an oxygen atmosphere not created by life or created rapidly can be ruled out.

  • $\begingroup$ (I know this is an old answer, but) an oxygen atmosphere could be created by the very same process that led to the water loss. Sunlight splits water into H and OH in the upper atmosphere, and H, being the lightest molecule of all, is lost more rapidly than OH or H2O. So an initially water-rich but reducing planet could easily end up oxygen-rich by the time all the water is gone, especially if it lacks plate tectonics (which would otherwise expose new reducing rocks, depleting the O2 supply). Mars' surface is quite oxidising for exactly this reason, I believe. $\endgroup$
    – N. Virgo
    Commented Nov 28, 2022 at 8:26

I do not think it is likely for life to evolve without water in reasonable time scales. For a hypothetical metabolism we need a solvent which:

  • is polar
  • is a good solvent for ions, molecules, protein/DNA equivalents
  • is very abundant
  • is liquid under given temperatures and pressure
  • has sufficiently low viscosity
  • is not too aggressive against the cell
  • for ecosystems a density anomaly is also desirable

If you think about these points it is unlikely, that there are many other alternatives. Methane is e.g. liquid and abundant on Titan, but the low temperatures will not allow reaction in sufficiently small time scales for life to evolve fast enough. Ammonia might be the closest bet, because at -33°C it is liquid. However according to the RGT rule the chemical reaction speed would be 1/2^4 to 1/2^5 in comparison to a water system. So it may take already four times longer for life to evolve and in case of an event it may be too slow to adapt.


Not only is it possible, but that's how it happens. Earth formed with no water--the impact events during it's formation were too energetic for it to retain any primordial water. Earth's water supply came from cometary impacts.

Thus to get a planet with basically no water you need a cataclysmic impact event after most of the debris has been swept up. Something perturbs the orbit of a planet or dwarf planet enough to cause an impact event. The water boils off. Note that this takes out the atmosphere, also.

If you want life on the world you'll need some more impacts to bring in some water but you can decide how many (by how late in the formation of the planetary system the late impact event happened) and thus how much water and atmosphere.

If you'll accept another approach to the world you're after, simply take a warm planet. Earth is not big enough to retain hydrogen--water vapor that gets high enough in the atmosphere to be disassociated results in hydrogen leaking away. This is a very slow process on Earth because the stratosphere is cold enough that only a trickle of water can make it through.

If the atmosphere were warmer (look at the projections of Earth's very long term future) this gate gets opened wider and the ocean slowly bleeds away to space. With a warm atmosphere this process would be basically complete in far less time than the Earth's age.


You must log in to answer this question.

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