Habitable Planets: Working the Odds

by Paul Gilster on March 10, 2010

Want to play around with some numbers? The process is irresistible, and we do it all the time when plugging values into the Drake equation, trying to find ways to estimate how many other civilizations might be out there. But a question that is a bit less complicated is how many terrestrial planets exist in the habitable zones of their stars? It’s a question recently addressed by Jianpo Guo (National Astronomical Observatories, Kunming, China) and colleagues via simulations. By ‘terrestrial’ world, the researchers refer to planets between one and ten Earth masses, although they note that some scientists would take this figure lower, to perhaps 0.3 Earth masses, which may be enough to retain an atmosphere over long geological timescales and to sustain tectonic activity.

Guo’s team is interested in the distribution of terrestrial planets in our galaxy, and the simulations that grew out of this study create a probability distribution of such planets in habitable zones. The paper is laced with the specifics, but let’s cut to the chase. Guo’s figures show 45.5 billion terrestrial planets in the habitable zones of host stars in our galaxy. The team also worked out the probability for planets in the habitable zones of different types of stars, concluding that M-class dwarfs host 11.5 billion such terrestrial worlds, while K-class stars are the most fecund, with 12.9 billion. G-class stars like our Sun weigh in with 7.6 billion, while F-class stars show 5.5 billion.

Image: M81. Our estimates of the habitable worlds in galaxies like these are widely variable, but they all imply countless chances for life to get its start. Credit: Jonathan Irwin, DSS2.

It’s interesting to weigh these numbers against the year-old estimates of exomoon hunter David Kipping (University College, London). Kipping starts with the galactic distribution of stellar types. He’s assuming about 300 billion stars in the Milky Way (increasingly cited as the best estimate) and noting that 90 percent of these are main sequence and thus stable for long periods of time. He goes on to whittle the number down, eliminating M-dwarfs because of tidal lock and also cutting out short-lived stars higher than F-class. 22.7 percent of main sequence stars in classes F, G and K thus remain.

Citing Michel Mayor’s Geneva team, which found that roughly 30 percent (give or take 10%) of F, G and K-class stars have super-Earth or Neptune-mass planets, Kipping narrows the field yet again:

Using 30% as a fixed value and assuming that very roughly half of this sample correspond to rocky planets and half to Neptune-like gas giants then we may write down that 15% of all F, G and K-type stars have rocky planets around them. It should be noted that this value is very likely an underestimate due to fact planets of Earth mass are currently below the detection threshold.

But how many of these planets would exist in the habitable zone? Kipping was working with 330 exoplanets then discovered, with about thirty in the habitable zone of their host star, and so he suggested a fraction of 10 percent would be a safe estimate based on current knowledge. He then factors in a galactic habitable zone, assuming that one may exist and that any value he obtains will therefore be an underestimate if it does not. This takes the number of stars with habitable planets down to 5 percent, but still leaves him with 50 million habitable-zone exoplanets in the Milky Way. We can contrast that with Alan Boss’s prediction of ten billion habitable exoplanets in our galaxy and, of course, with Guo’s team, whose whopping 45.5 billion is the largest estimate I’ve ever seen.

The weird thing, as Kipping confirmed this morning, is that his 50 million estimate was actually rounded up from 45.5 million, a figure exactly 1000 times less than Guo and team’s number. Our numbers, then, seem to be all over the map, and Kipping also notes Bond and Martin’s 1978 estimate of 10 million habitable exoplanets. But Kipping is the only one I know who takes a shot at the intriguing question of habitable moons. He is, after all, a specialist in detection methods for moons around exoplanets, studying methods that may help us detect large satellites during exoplanet transits. Noting that a large moon could be found around anything from an Earth-class planet to a gas giant, he boosts Mayor’s 30 percent figure to 50 percent, for any kind of planet. And this is interesting:

Let us also assume that the habitable zone for exomoons is extended by around 50% due to the possibility of tidal heating maintaining temperate conditions in traditionally cold-zones. This means that 15% of all planets can host a habitable exomoon.

How many planets have large moons? Kipping notes how little information we have, but using the Solar System as an example, he finds two planets out of eight where a moon has been formed through a capture/impact process, which he believes to be a requirement for a large moon. Assume, then, that 10 percent of planets host a large moon and you wind up with a figure of 25 million habitable exomoons in the Milky Way. But we have to keep these figures in context. Kipping again:

So our calculation suggests that there are roughly half the number of exomoons than exoplanets. One important thing to realize is that these calculations are based on many guesses but many of the assumptions underlying each calculation are the same. Whether the ratio is 0.5, 1, 2 or 5 is not very reliable right now, but what does seem perhaps more persuasive is that if we talk about ‘order of magnitude’ kind of figures, the number of habitable exoplanets and exomoons is ball-park equal.

Kipping’s figures are truly mind-blowing when he turns to the larger universe. A figure of roughly 100 million habitable environments per galaxy can now be turned around for an estimate of habitable worlds in the visible universe. The number works out to 1018, or 10 million trillion. Even allowing the vast play in the numbers between our low-ball and high-end estimates of habitable planets, the universe is likely to be filled with environments conducive to life. My guess is that it’s out there in fantastic abundance. But how much of it has gone on to sentience and, perhaps, technology? That’s the question SETI continues to poke at, and it’s one that’s emphatically still in play.

The Guo paper is “Probability Distribution of Terrestrial Planets in Habitable Zones around Host Stars,” Astrophysics & Space Science Vol. 323, No. 4 (October, 2009). Preprint available. David Kipping’s Web pages are packed with good information. Start with these articles.


kurt9 March 15, 2010 at 14:20

Excellent analysis, Ronald.

I think your estimates of “habitable” planets is likely correct. I would also comment about the emergence of the Eukaryote and the inclusion of mitochondria in cells as a result of the Hydrogen hypothesis of Endosymbiosis is sufficiently unlikely that we are likely the only advanced life in the galaxy. Also, consider that Eukaryotic-based photo-synthesis is likely necessary to generate the Oxygen atmosphere that is needed for animal life. Earth’s prokaryote atmosphere had, at most, 10% the partial pressure of Oxygen of our current atmosphere.

For those so inclined, I recommend Nick Lane’s “Power, Sex, and Suicide, the Story of Mitochondria”.

Eniac March 15, 2010 at 22:00


Considering how rare advanced intelligence is on earth, how long it took to arise, and how totally unnecessary high intelligence is for life in general to survive, the question then becomes which small fraction of these will possess an (advanced, techno-capable) ETI. Most probably a minute fraction.

I don’t see this at all. It took not long to arrive (a mere 400-500 million years or so after the first animals left the sea), and it is not rare, the Earth is covered in it. We are the first, yes, but does that make us rare? Someone has to be first, right? Probably there will be no seconds, but that is because of us, not because of any intrinsic unlikeliness of the event. There were, after all, runner-ups who perished, most likely at our hands.

Eniac March 15, 2010 at 22:25


On earth this seems to have required the oxygenation of the atmosphere – both for energy generation and creation of structural support (oxygen is a critical component of both collagen and lignin for example) needed to build big complicated bodies.

Agree on energy, but not on structural support. Collagen and lignin are both proteins, which are an integral part of life as far back as we can imagine. The oxygen in them comes from H2O and CO2, there was plenty of it, always. We are talking atmospheric oxygen, and it is for energy, exclusively.

4. It is actually very probable if certain other enabling preconditions are also met (such as oxygen atmosphere or something like that), but these preconditions depend on random events, and it just so happened that on earth these preconditions were not met in its first 3 billion years or so.

Not random events, really. Both the buildup of oxygen and the increase in solar power were far from random. Inexorable would be a better description. Slow, but inexorable.

5. It actually didn’t take 3.5 billion years on earth, because we are not the first technological species. We were actually preceded, perhaps by billions of years, by an earlier technological species that left the planet/went extinct, and we have just not yet found evidence of them yet because of the long time and sparcity of the fossil record.

That seems rather unlikely, given how well we can account for the entire tree of life and the fossil record. I am no biologist or archeologist, but I think neither might take too kindly to this suggestion of yours… :-)

So one question would be whether or not oxygen specifically is necessary or if other chemical process can substitute, and if so, how common/likely to arise are these chemical processes?

Now, this is an interesting suggestion. Let’s talk about carbon-based life, first. The primary source of carbon is in CO2, at least on this carbon-poor, oxygen-rich Earth of ours. Thus, the liberation of oxygen is an unavoidable consequence of fixing carbon into biomass.

If life was based on silicon, things would be no different. The primary source of silicon is silicon oxide (rock), and its fixation would inevitably produce free oxygen.

So, either way, the oxygen would be there, and it would be foolish of animal life not to evolve to take advantage of it.

Could there be completely different chemistries? There probably could, but with oxygen being the most common element on Earth, and the third most common in the universe, it would be really hard to avoid, especially since it is also so reactive.

So, I think it is safe to say that oxygen is the key to animal life (with its high energy density), and animal life is key to mobility, which in turn is key to intelligence. In my view, this relationship is also driving, i.e. one leads to the other, inexorably, but that is certainly debatable.

Eniac March 15, 2010 at 22:46

Does anyone else find it eerie that the Earth (and the rest of the universe) is much better endowed with raw materials for machines (silicon and metal oxides, aka rock) than with that for carbon-based life (carbon dioxide)? Doesn’t it feel like we are sitting on a Petri dish made for something else? Iron, the most stable element. Coincidence? I think not. It all can only mean one thing: Carbon based life is too frail and short-lived to travel amongst the stars. The age of the Machines is drawing near. We humans will soon fall into the dustbin of history, mere temporary ushers for the true masters of the universe.

All hail the Machines!


Ronald March 16, 2010 at 6:08



Yes, although, of course still extremely premature, it seems as if the more reasonable estimates of number of (potentially) habitable planets are hovering around the same figures, or at least same order of magnitude.

However, I find it a bit too pessimistic and premature to state that “we are likely the only advanced life in the galaxy”. Though I agree that there is nothing simple about single-celled Eukatyotes.

One might also argue that, apparently, there are certain driving, ‘creative’ if you will, biochemical mechanisms inherent in nature working toward higher degrees of organization.
But I admit that there may be a degree of wishful thinking here.

Anyway, always look at the bright side of (biological) life: if life in general, or at least any higher life, appears to be very rare, even on habitable planets, it just means that there is an enormous amount of free real estate out there.

Brings us back then to the topic of seeding life.

To me the issue is really that simple: either there is already biological life on habitable planets, or we will, eventually, bring it there.

Only remains the ethical issue of how to (not) deal with present ‘simple’ microbial life. Honestly, I wouldn’t worry too much about that, we would most probably have a very hard time even trying to exterminate it.

Ronald March 16, 2010 at 6:10

Eukatyotes = Eukaryotes, of course.

Eniac March 16, 2010 at 11:16


One might also argue that, apparently, there are certain driving, ‘creative’ if you will, biochemical mechanisms inherent in nature working toward higher degrees of organization.

One might, and one would be right. These creative mechanisms are known as “evolution” and “natural selection”, and they are very driving.

Eniac March 16, 2010 at 11:26

Ooops, Lignin, of course, is not a protein. Still, it requires no free oxygen to make, being found in plants which do not require oxygen.

kurt9 March 16, 2010 at 13:54

“Anyway, always look at the bright side of (biological) life: if life in general, or at least any higher life, appears to be very rare, even on habitable planets, it just means that there is an enormous amount of free real estate out there.”

You got it, dude!

Ronald March 17, 2010 at 5:48

Encouraged by kurt9 ;-) , I elaborated my above ballpark figures a little, to calculate the average distance between aby two (potentially) habitable planets in the Galactic Habitable Zone (GHZ).

The result is the following:

– We know from surveys, such as NStars, Recons, Hipparcos, etc., that the average density of stars (excluding brown dwarfs) in the galactic disk is about 0.002 per cubic lightyear, i.e. 1 per 500 cubic lightyear.

– In my above calculations I indicated that about 1.5 – 3% of all the stars in the GHZ can be called sufficiently solar type (i.e. non-variable, main-sequence, luminosity 0.2 – 1.5 solar, corresponding with spectral type from F7 – K3, singular or wide binary, sufficient metallicity).
Let’s optimistically take the 3% (some estimates go a bit higher than this, so 3% seems reasonable).

– Based on various planetary survey sources and in line with Kipping, we assume that 10% of those solar type stars actually possess an earthlike planet in the Habitable Zone.

– Even if we take the higher end of the above-mentioned estimated fraction of solar type stars, 3%, this means that the average density of ‘stars with a (potentially) habitable planet’ in our GHZ is only 0.000006 per cubic ly, or about 1 per 167,000 cubic ly.

– From this we can derive that the average distance between any two potentially habitable planets in our GHZ is then 68 ly. Note: the calculated radius of this volume (= 4/3*pi*r^3) of space around each habitable planet is 34 ly, however, for the distance between two I think we should take twice this radius. Correct me if I am wrong here.

Mind, we are not yet talking about advanced (complex) life here, let alone intelligence, not even mentioning a techno civ. Just any potentially habitable planet.
Purely statistically speaking, we would be exceedingly lucky if Alpha Centauri A or B possess a habitable planet.
But maybe I am a bit too pessimistic here, after all it’s just statistics, based on few data and several assumptions.

William rothamel May 23, 2010 at 20:45

I suspect that the uncertainty of these estimates is very large. Also. there are many more criteria for the existence of a habitable planet than those discussed. I venture to say that if all of the necessary conditions were written down, and the correct probabilities of occurrence were written down: we would not even find the likelihood of even one such planet in the Milky Way–This goes to say that I believe we are most likely an extremely rare confluence of life essential conditions.

Time and new instruments will tell the true story.

Ivan Grosz December 5, 2010 at 22:38

I concur to a substantial extent with William Rothamel’s comments but with the following note. I believe that in a practically infinite number of planets and stars any singularity is highly unlikely. Still I believe that the habitability requirements are extremely severe and therefore the places which have supported intelligent life in the past or possibly at present are extremely limited and outside of the communication range. This does not have to stop us from trying.

Charles January 13, 2011 at 8:35

I have a question. Do any of these calculations take into consideration the probability that in denser regions of the galaxy, near the center where most stars are, planets are much more likely to have been cleansed of any life by radiation from nearby nova? I would think this would reduce the number of habitable planets by a significant percentage.

Charles January 13, 2011 at 8:44

(this should be included in the previous comment)

Another consideration for life is, as I understand it, latest theories are that a sizable moon is required to produce tidal pools for life to begin. I would suspect that this requirement would reduce the number of worlds where life could originate by at least 50% from the number of rocky planets with an athmosphere inside the temperate zone of a star. Comments?

Comments on this entry are closed.

{ 2 trackbacks }