Where did Earth’s (and the asteroid belt’s) water come from?

Imagine a dry Earth. No waterfalls. No oceans. No beer (or people to drink it).  A sad place.

Water is pretty important stuff.  In this post we will go deep into how Earth got its water (much deeper than in this post from a while back).  And since it’s connected with how Earth formed, our story will span the Solar System.  We’ll find out that Earth and the asteroid belt got their water from the same place, and that Jupiter’s growth was the key.

Part 1. Water in the inner Solar System

Oceans cover three quarters of the planet.  But within our planet as a whole water is pretty scarce.

We know how much water is on Earth’s surface. Let’s call that one “ocean” of water. So, one “ocean” includes the Pacific, Atlantic, Indian and Arctic oceans plus all the lakes, rivers and such. It feels like a lot because we humans are tiny compared with a planet. But one ocean of water is only one 4000-th the mass of the Earth as a whole.  Not a lot.

The relative size of the rocky Earth and a ball containing all of Earth’s surface water (one “ocean” of water).  Credit: Jack Cook/WHOI/USGS

We don’t know how much water is inside the Earth. Some could be trapped as hydrated rocks in the mantle, or as hydrogen bonds in the core. Some scientists think that there is very little extra water inside the Earth.  Others argue that there is more than ten times as much as on Earth’s surface. Still, water makes up no more than a few parts in a thousand of Earth.

Where is the rest of the water inside Jupiter’s orbit?  Mars and Venus are each pretty dry, although we think they had more water in the past.  Mercury and the Moon have a dusting of ice.  But the main reservoir is the asteroid belt.

Asteroids come in different flavors.  Most of them are found between about 2 and 3.2 Astronomical Units (AU).  That is the main belt. The inner part of the main belt is dominated by S-type asteroids and the outer part by C-types.  There are plenty of other types but they don’t contain as much mass and are not important for this story.

The distribution of different types of asteroids.  The main asteroid belt is located between about 2 and 3.2 AU (1 AU is the Earth-Sun distance).  Most of the actual mass in the belt is in C-types (which are wet) and S-types (which are dry).  Credit: astrobites.org, Gradie & Tedesco (1982)

We can use meteorites to figure out what asteroids are actually made of.  Meteorites are pieces of asteroids that land on Earth.  By determining what different types of meteorites would look like if they were floating around in space, we can figure out which meteorites go with which types of asteroids.

It turns out that the inner part of the asteroid belt is drier and the outer part is wetter. S-type asteroids (associated with ordinary chondrite meteorites) are dry.  C-type asteroids (associated with carbonaceous chondrite meteorites) are wet, with about 10% water by mass.

So now we know where the water is in the inner Solar System. Let’s gather one more piece of information. Let’s take the chemical fingerprint of these sources of water.

Isotopes are when an element has a different number of neutrons. Isotopic ratios are just the relative amount of different isotopes. These are the “fingerprints” that will help us figure out how Earth got its water.  The most important element in water (H2O) is hydrogen, and the isotope we can most easily observe is called deuterium.  It’s just a hydrogen atom with an extra neutron.

Here is what the D/H (Deuterium-to-Hydrogen) ratio for different water-rich bodies in the Solar System look like:

D/H ratios for different Solar System bodies. Here, Asteroids basically represent C-type asteroids (from carbonaceous chondrite meteorites). Comet 67P/Churyumov–Gerasimenko was visited by the Rosetta spacecraft and has a high measured D/H.  Credit: K. Altwegg.

This figure is pretty complicated but it tells us a lot. First, the Sun (bottom left point) and the gas giants have a very low D/H ratio compared with Earth. This is thought to represent the fingerprint of the gas in the Sun’s planet-forming disk. So Earth’s water probably did not come directly from the disk (although some may have).

Second, C-type asteroids have the same chemical fingerprint as Earth’s water.  This puts them high on the list of possible suspects.

Third, comets are confusing.  The measured Oort cloud comets (in purple) have D/H about twice as high as Earth. Only three Jupiter-family comets have been measured: two have D/H similar to Earth and one has D/H 3.5 times higher.  The simplest interpretation is that comets are not all the same, but rather sample a wide range of conditions.  Some comets share the same fingerprints as asteroids and some don’t.  The ones with high D/H ratios are bad candidates for Earth’s water, but maybe the comets with D/H like Earth’s contributed.

OK, we’re in good shape.  We’ve got the information we need to jump into….

Part 2: How exactly did Earth get its water?

The story of where Earth’s water came from starts back during the early stages of Solar System formation.  It’s like ice cream — it all starts back with the cows…

Planets form in giant frisbee-shaped disks around young stars.  Close to the star it’s too hot for ice so planets can only grow from rock and iron.  Farther away it’s colder and planets can form from ice (as well as rock and iron).

The “snow line” in a planet-forming disk is the distance beyond which it is cold enough for water to condense as ice. Closer to the star water exists as vapor. Credit: Univ. of Hawaii ‘s Institute for Astronomy.

Earth formed in the rocky part of the disk.  It didn’t just grab onto its water from the gas in the disk (because the D/H fingerprint doesn’t match).  So how did Earth get its water?

The simple answer is: water-rich objects bashed into the growing Earth.  The next question is: well, where did these water-rich objects come from?

Today, there is a rough boundary between dry S-type asteroids and wet C-types.  It’s simple to think that the C-type asteroids were always where they are now.  This is actually not a great assumption for a few reasons.  One reason is that ordinary and carbonaceous chondrite meteorites have very different compositions, and it’s hard to imagine how they could have formed right next to each other. Another reason is that the “snow line”, the boundary beyond which ice can freeze, was constantly moving.  Finally,  Jupiter is awfully massive, and its growth must have shaken things up.  To get a handle on all this we really need to understand what happened pretty early on.

Let’s rewind the clock to when Jupiter was forming. We think that Jupiter and Saturn formed in two steps. First they grew large solid cores that were a few times more massive than Earth.  Then, they gravitationally captured gas from the disk.

Cartoon of the growth of a gas giant, starting from a solid core that captures gas from the disk.  Credit: oklo.org

This process starts off slow but then has a runaway period during which the planet grows super fast.  The planet’s growth slows as it carves a gap in the disk.

The gap carved in a gaseous protoplanetary disk by a Jupiter-like planet.  Credit: Phil Armitage.

Growing giant planets are surrounded by small bodies that feed them. What happens to those small bodies (called “planetesimals”) when the giant planet goes runaway and grows big fast?

When a giant planet grows fast, nearby small bodies are gravitationally kicked onto stretched-out orbits that cross the inner Solar System.  They get thrown in all directions. Often they are eventually kicked out of the Solar System entirely. But there is still gas in the disk and this causes their orbits to become more circular, effectively moving them inward.  Here is what it looks like (in this case Jupiter and Saturn both grow and scatter planetesimals colored by their starting orbital positions):


As Jupiter grows, planetesimals are kicked up onto stretched-out “eccentric” orbits and then rain down into the asteroid belt! This happens again after Saturn grows. This is how the C-type asteroids were implanted into the belt. The C-types were implanted from the Jupiter-Saturn region, although when the ice giants are included, some C-types can have originated as far out as Neptune’s current orbit.

Some planetesimals were gravitationally kicked too hard to be implanted into the asteroid belt. These bodies were kicked onto orbits that are too stretched-out to be stable.  They are up to the left in the movie.  Those scattered bodies cross the orbits of the growing terrestrial planets. They delivered water to the Earth. 

The planetesimals that delivered water to Earth came from the same source as those that ended up as C-type asteroids.  So they must have the same chemical fingerprints. As we saw above, Earth’s water and the C-types share the same D/H ratio (as well as other isotopic fingerprints I didn’t discuss).  Boom!

This story all fits. Jupiter and Saturn’s growth naturally pollutes the inner Solar System with water-rich planetesimals.  That is where the C-type asteroids and the terrestrial planets’ water came from.

What happened next?  In the classical model of Solar System formation (animation here), the gaseous planet-forming disk evaporated and the planets finished growing (although the classical model has some issues).  In the Grand Tack model, Jupiter and Saturn migrated across the inner Solar System to keep Mars small.

The pieces in the puzzle of how the Solar System formed are up for debate.  What we are certain of is that Jupiter and Saturn grew.  The movie above shows that the giant planets’ growth disturbs nearby planetesimals, implanting some into the asteroid belt and  scattering some toward the growing Earth.  This is a simple answer to a long-standing problem.  It also fills a hole in some models of Solar System formation, and improves upon all of the current models.

Punchline: The inner Solar System’s water — both on the rocky planets and in the asteroid belt — is a simple byproduct of Jupiter and Saturn’s growth.

Our paper on the subject is going to be published in Icarus. It explains all of this in much much more detail. If you are interested you can download it here.

For those of you who are wondering how this story would change if Jupiter and Saturn’s migration was taken into account, the answer is: not much. I put a couple of animations of simulations with migration on YouTube.  See here and hereHere is another simulation without migration but that runs longer so that the population of water-delivering planetesimals is easier to see.


Questions, comments, words of wisdom?


The history of this project.  Back in the Spring of 2008, when I was living in Boulder, Morby, Dave O’Brien  (both visiting) and I met to talk about the small Mars problem.  At that meeting, spurred along by our discussions, I came up with the idea that eventually became the Grand Tack model. (It would take Morby a year or so to become a believer and then hire Kevin Walsh to wrap things up). The Grand Tack was my baby, but I saw that it was missing a piece: Jupiter’s growth must have itself shaped the distribution of water before it migrated.  That was the start of this project.  I planned to publish this paper before the Grand Tack model, as a prequel that built things up piece by piece.  However, my code has a persistent annoying bug. Plus, the Grand Tack paper was ready before this paper was. So, I put this paper to the side.  Flash forward to 2015: I talked with Andre Izidoro, postdoc extraordinaire, about this project and showed him my code.  He solved the bug in less than an hour, and upgraded the gas drag calculation in the process.  A part of the project was initially included in a paper on chaotic excitation of the asteroid belt, but was later removed.  The paper was accepted to Icarus in June 2017, nine years after I started the project.


48 thoughts on “Where did Earth’s (and the asteroid belt’s) water come from?

  1. Very cool! This theory makes good sense. When the initial inner solar system planetesimals formed was the sun shining and the frost line beyond them, or could they have formed with some native water? If that is the case when the frost line moved out, would they have lost that water or retained some of it?
    Many thanks for the post!

  2. Good question! As soon as the Sun had a stable disk (within about a hundred thousand years of turning on), it would have had a “frost line”. But the frost line actually moves around in time and does move inward. Some models actually put the frost line around Earth’s orbit later in the disk’s lifetime. But we don’t think that’s where Earth’s water came from because the “primitive” asteroids in the inner belt are pretty dry. It seems like there is a sign that everything was dry out to maybe 2-2.5 AU. Still, this issue is not completely solved and is something astronomers are thinking about right now.

  3. I see objects scatttered out too, could this explain an irregular satellite like Himalia being a C-type?

    1. Yes, things are scattered out too (we discuss this in the paper). I am not all that familiar with Himalia but with a quick Google I would say, sure, this could potentially explain that!

  4. Hi Sean. Very useful article – many thanks.

    Do you have a way of signing up to get an email notification when you post a new blog post? I only saw this by chance when I made one of my infrequent visits to Twitter. (I am highly resistant to social media so don’t suggest more Twitter time!)

    Other word press blogs that I’m interested in have email alerts, so it must be possible.

    1. HI Margarita — yes, you just need to “follow” the blog and you will get an email every time there is a new post (which is not super often at the moment since I’m busy with other stuff…)

      1. Odd – I clicked to follow your blog some months back and, tho your blog appears on my WordPress Reader, I didn’t get an email notification for this post. I’ll go and check it.

      2. Got it! I had specifically to switch on receiving email notifications from your blog.
        I wonder if I’ll ever catch up with information technology??

  5. Does this produce a correlation between inclination and eccentricity, with the highest inclination objects retaining larger eccentricities?

  6. Is there a good explanation for why deuterium gets concentrated by a factor of 10 or more in asteroids, & comets?

    1. Ahh, well, it’s actually the other way around. Stars and planets form from clouds of cold gas with very high D/H ratios. Only close enough to the star does it get hot enough to process the gas and dust and change the D/H. And yes, there are models for this, but they don’t agree in terms of what the initial distribution of D/H ratios in the primitive solar system looked like…

  7. The Grand Tack theory might be wrong, it is one of the best empirical models, but what it is says here it is only one possibility, do not believe too much to this article. Many scientists push to their theories for they ego, not for science.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s