Earth is the only planet in our solar system that has copious amounts of water on it. Where did this water come from and why is there so much water on Earth compared to every other planet in the solar system?

  • 10
    $\begingroup$ Isn't there more water on one of Jupiter's moons? Yeah, Europa. $\endgroup$
    – naught101
    Apr 16, 2014 at 8:37
  • 1
    $\begingroup$ @naught101, I didn't refer to moons in my question I referred to planets. $\endgroup$
    – Kenshin
    Apr 16, 2014 at 22:55
  • 2
    $\begingroup$ I don't see why the process would be any different. Also, there are likely to be planets in other solar systems with surface water oceans, so a more general answer could be quite useful. $\endgroup$
    – naught101
    Apr 17, 2014 at 1:16
  • $\begingroup$ +1 for correctly referring to the name of the planet as Earth, not the Earth. Just like we don't say the Mars. :) $\endgroup$
    – kwknowles
    Apr 19, 2014 at 21:44
  • 1
    $\begingroup$ keep in mind all four the gas giants have a lot of water. $\endgroup$
    – John
    Oct 26, 2017 at 3:28

5 Answers 5


Your assumption that there is not a lot of water elsewhere in the solar system is incorrect. According the this article on NASA's website;

Missions in recent years have overturned our view of a dry solar system, returning mounting evidence of ample water from a vast array of locations.

Comets from the remote corners of our solar system are made of water and other ices. Orbiters, landers, and rovers reveal Mars as a watery world in the distant past -- a world that today may contain entire underground oceans of frozen water.

It also points out that

Jupiter's moon Europa has a frozen crust of water, covering a thick global ocean. By current estimates, it has twice as much water as all of Earth's oceans and rivers!

The perception that the rest of the solar system does not have a lot of water is probably due to the fact that the Earth is in the 'Goldilocks zone' where water can exist in all three phases at the surface of the planet. The temperatures on the rest of the planets in the solar system are either so hot that it can only exist in a vapor phase or so cold that it can only exist in ice at the surface or liquid beneath.


The water was already present when the Earth assembled itself out of the accretionary disk. Continued outgassing of volcanoes transferred the water into the atmosphere which was saturated with water. And rain transferred the water onto the surface.

Compared to other planets and smaller solar system object Earth has a big advantage. It is large enough to prevent water molecules to leave the gravitational field, and it has a magnetic field which prevents atmospheric erosion (Wikipedia). This is due to the Earth's outer core being liquid (Moving charged liquid = magnetic field). Mars probably had oceans until its outer core solidified so much, that the convection was stopped. After the magnetic field disappeared a few million years of solar radiation removed all of the atmosphere and the oceans.


An important issue to realize is that water must have been very abundant in the protosolar disc, as tobias already stated. To expand on that I just want briefly touch upon the atomic abundances that we measure in the corona of the sun, as presented by a wikimedia-commons graph:

enter image description here (Sidenote: Those abundances compare well to the revised Asplund2009-abundances)
We think that those numbers are representative of the bulk composition of the Sun, as it is only burning Helium since 4.567 Gyrs. Thus the composition is generally assumed to be primordial, or what the solar system started out with.

Now let us imagine this atomic mixture accumulating around the young terrestrial planets and focus on the most abundant elements H, He, C, N and O. In a thick planetary envelope, shielded by UV, equilibrium chemistry will set in. Then a lot of $H_2$ will form, He will stay inert and C, N, O will try to react with Hydrogen, simply because the number of encounters is much higher with that, than inside of the CNO group. Some CO will form, as this is a very stable molecule. But as C becomes depleted, and our protosolar nebula has $C/O \approx 0.5$, there is still a lot of Oxygen left. Thus it will inevitably combine to $H_2O$.

The result is, that we expect there to be really a lot of water around in Planet-forming discs.

The escape or destruction of water after this period however, is apparently also very efficient, and other answers have touched upon the retention of water on Earth. So in fact Astronomers at the moment rather wonder "where did all the water go?"


Oceans did form on other rocky planets - at least Venus and Mars, and moreover, many moons of Jupiter and Saturn as well.

The problem is that of the two other terrestrial "uberplanets" who had oceans - i.e. the aforementioned Venus and Mars - they lost them, but in rather different ways.

On Venus, what happened appears to have been that, synthesizing the best theories and evidence so far, as the Sun steadily grew hotter as a result of the natural main-sequence stellar evolution, the planet essentially left the habitable zone. The oceans evaporated and were lost to space. The way that works is that as the planet heats up, the water vapor content in the atmosphere increases, and some of that vapor rises to the top of the atmosphere. At the top, the UV flux begins to break apart (photodissociation) the water molecules, releasing the hydrogen, which then easily escapes as its gravitoionization energy is very low (for Venus, with an escape velocity of 10.36 km/s, it is similar to Earth - stripping one $\mathrm{H}_2$ molecule, a mass of ~3.4 yg, costs 0.18 aJ or 1.1 eV, which can easily be supplied by photons, stellar wind, or other factors. Stripping nascent [monoatomic] hydrogen, if such could happen without molecularization, takes half this.). The end result is an oxygen-rich atmosphere (nearly pure oxygen) at very high pressure, though chemical reactions with the rock below may strip some of this.

But as you may note, the current atmosphere is mostly carbon dioxide, with a very high surface temperature, and thus while the above is fairly plausible, one must wonder how this atmosphere obtained. Well, here's the thing. Much like Earth, Venus likely had plate tectonics early in its history when it had oceans. Ocean water seems to be a necessary component of plate tectonic activity on planets at least around the size of Earth, acting as a sort of "lubricant", and without it, plate tectonics cannot operate. Hence, once the water evaporated from Venus, any tectonics it had would have stopped.

The cessation of plate tectonics, while it may not seem like much to the uninitiate (so what? The continents stop moving, and you stop getting earthquakes?), is actually catastrophic for this kind of planet. You see, plate tectonics does more than just move continents, mountains, and earthquakes. What it most crucially does is to serve as a sort of "release valve" for the heat that is being generated inside the planet as a result of the decay of primordial radionuclides mixed with the mantle rocks (mostly $^{238}\mathrm{U}$). This is because that, at spreading zones, hot mantle material oozes out (Iceland is a good Earthbound example of where this process happens to just come above water and can be seen with the naked eye - youtube it if you're too poor to travel), and at subduction zones, volcanoes are created, like in the Pacific Ring of Fire on Earth. The released material transfers the heat to the atmosphere and thus ultimately it gets radiated into space, hence shedding the internal heat. Without this process, that heat is shed far less efficiently - essentially only by conduction through the ground, which is terrible.

As a result, once you lose tectonics, the radioactive heat begins to build within the interior, and the mantle temperature starts to rise considerably. And when that starts, eventually, something has to give. And what "gives" is not quite well-understood but it seems to be either that tectonics restarts for a "short" time "with a vengeance", or that there commence widespread basaltic flood eruptions, or some combination of the two. In any case, the result is a rapid - on a geological scale of course, actually about 100 million years - "global resurfacing event" resulting in either or a combination of total subduction/reemplacement of the old crust (hence that is a considerably faster rate of tectonic motion than on Earth - to recycle all its crust over 100 Myr would take plate motion on the order of 20 cm/yr, where the fastest plate movement [India] in recent time was only around 4-6 cm/yr), or its progressive burial in flows similar to those from the Siberian Traps, likely from multiple vent sources, extending over such a long enough period as to cover all low-lying areas at least. In both cases there occurs the widespread release of volcanic gases which chiefly includes carbon dioxide, and that quickly converts the atmosphere to a "runaway greenhouse" state like we see on the planet today. The last GRE on Venus was finished at about 500 Ma, and there may have been more before that until the time when it last had oceans, and it is possible that the current thickness was the result of such a chain of episodes instead of just one.

The oceans on Mars, on the other hand, seem to have suffered a rather quieter demise. In this case, it seems the chief problem was the planet was simply too small, and the cause was not the result of external changes in the form of those in the Sun, but rather internal ones. As a result of its lack of size and hence of mass, there was both less primordial heat within the core and a smaller radionuclide load to maintain the internal temperature. The interior would have cooled to the point its iron inner core froze completely, and that shut down the planetary magnetic dynamo, which we know existed in the past thanks to there still being today small areas of "fossilized" magnetization in the crust due to rich concentrations of magnetizeable iron minerals.

And this is a problem because the magnetic field on a terrestrial planet serves the purpose of deflecting the solar wind away from the atmosphere. Without it, the wind directly strikes the atmosphere and begins to "sputter", or blow, it away. Mars thus lost its atmosphere in a comet tail-like effect, and with that also went the oceans thanks to the resulting pressure drop. What was left was the small amount of heavy (and hence hard to remove) carbon dioxide. No runaway greenhouse was produced because the thermal situation in the interior was just the opposite - it wasn't overheating, but rather, it was cooling down.

Generally, as long as the material from which you form planets - i.e. the star's proplyd - has sizeable quantities of water, as typically is the case, the planets will form oceans thanks to the chemical differentiation process leading to accumulation of a surface reservoir. Indeed, in many cases, they can form much more ocean as we've seen by observation of extrasolar planetary systems with planets that could have "ocean" layers hundreds of km thick (technically, this would only be liquid ocean proper down to about 100 km or so - at that point, the pressure becomes high enough [though the precise depth depends on the local gravity] to freeze the water by compression [pressure approx 1 GPa, for comparison, the Marianas on Earth, at 10 km depth is around 100 MPa], and the rest of the layer below that point is various high-pressure phases of ice) hence no bare land. It all depends on the water content of the material that forms it, and that is quite variable.

(One could, actually, consider certain moons of Jupiter and Saturn as actually Solar System examples of this kind of planetary composition, just at a much-reduced size as the extrasolar examples are north of one Earth mass ($6000\ \mathrm{Yg}$).)

  • $\begingroup$ This is a nice, detailed answer, but take into account that the "magnetic field protects atmosphere"-story is not supported by current data. See a recent paper adsabs.harvard.edu/doi/10.1093/mnras/stz1819 and refs in the introduction. Also question: Are you aware of any coupled geological+atmospheric evolution models that can quantitatively reproduce the stories as presented by you about Mars/Venus? Without quantitative underpinning, they're after all just that...stories. $\endgroup$ Jul 22, 2019 at 12:01
  • $\begingroup$ @AtmosphericPrisonEscape : Your linked article doesn't seem to really destroy the general thrust of the idea, but it does suggest it's a little bit simplistic and adds more detail, more interesting. Reading the link, a strong enough magnetic field will deflect the stellar wind indeed, and as it weakens this protection falters, just as I've been saying. But the wrinkle is that apparently, per the paper, once it drops past a certain strength, but before it has hit zero, it actually acquires an antagonistic effect where it assists the stellar wind in stripping the atmosphere. $\endgroup$ Jul 22, 2019 at 12:27
  • $\begingroup$ Thus if you think about a curve that has "magnetic field strength" on the x-axis and "protective capability" on the y-axis, it would look like a "bowl curve" - starts zero, goes down into negative capability (i.e. is harmful), reaches a minimum, then turns around and starts to climb again to and then past zero to positive protective (actually protecting) capability. $\endgroup$ Jul 22, 2019 at 12:28
  • $\begingroup$ Moreover, the precise when and where of this happening, i.e. the shape of the bowl, depends on the strength of the wind. Of course, it could also be that perhaps Mars started with field in the antagonistic range, and I believe the Solar wind in the primeval Solar System was stronger, too - I don't have the full paper accessible atm. $\endgroup$ Jul 22, 2019 at 12:30
  • $\begingroup$ @AtmosphericPrisonEscape : Also found this stuff: lpi.usra.edu/vexag/meetings/archive/vexag_4th/nov_2007/… talking about Venus, gives an outline at least similar to here though with only one resurfacing event. But not sure of any formation-to-present simulation of any model, if you have references on any that have been done that you know of, perhaps with a different scenario, I'd like to see it and then I could perhaps revise the answer with something better supported by stronger evidence. $\endgroup$ Jul 22, 2019 at 12:36

Water actually was never present on Earth before its creation and that is for sure.

Astronomers realized that there are two ready-made sources: comets and asteroids, the solar system’s gravel strewn among planetary boulders. The primary difference between the two is that comets typically have a greater concentration of ingredients that vaporize when heated, accounting for their iconic gaseous tails. Both comets and asteroids can contain ice. And if, by colliding with Earth, they added the amount of material some scientists suspect, such bodies could easily have delivered oceans’ worth of water. Accordingly, each has been fingered as a suspect in the mystery.

Adjudicating between the two is a challenge, and over the years scientific judgment has swung from one to the other. Nevertheless, recent observations of their chemical makeups are tipping the scale toward asteroids. Researchers reported last year, for example, that the ratios of different forms of hydrogen in asteroids appear to better match what we find here on Earth. But the analyses are based on limited samples, meaning there’s a good chance we’ve not yet heard the final word

  • 2
    $\begingroup$ References/ Citations would improve this answer $\endgroup$
    – Fred
    May 8, 2016 at 14:03
  • 3
    $\begingroup$ "Water actually was never present on Earth before its creation and that is for sure." Well that's a statement that I can't dispute... :-) $\endgroup$ May 9, 2016 at 11:49
  • 1
    $\begingroup$ Downvote: "Water actually was never present on Earth before its creation and that is for sure." Water was and is everywhere in the universe, as O and H are very abundant and H2O forms readily. $\endgroup$ May 10, 2016 at 21:54
  • $\begingroup$ earth was not present before the creation,but all the buildingblocks including water was present in the dust cloud that became the solar system. $\endgroup$ Mar 29, 2017 at 16:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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