≡ Menu

Habitable Moons and Kepler

While we’ve looked several times in these pages at David Kipping’s work on exomoons, the investigation of moons much closer to home reminds us that finding a habitable satellite of another planet may not be out of our reach. After all, we’re gaining insights into possible habitats for at least microbial life on (or in) places like Europa and Enceladus, and speculations about similar biospheres within some Kuiper Belt objects also keep them in contention.

So what about a habitable moon around a distant gas giant? Kipping (University College London) has now gone to work on the question in relation to the Kepler space telescope. His findings are striking: A Saturn-sized planet in the habitable zone of an M-dwarf star would allow the detection of an exomoon down to 0.2 Earth masses.


Image: A habitable exomoon would offer an exotic vista, a view that may be more common in the galaxy than we have previously imagined. Credit: Dan Durda.

Now that sounds unusual, given that Kepler can’t find planets of such small size. How, then, does Kipping hope to find exomoons at this scale? The answer is that an exomoon detection depends on two measurements, neither of which demands observing the dip in starlight caused by the moon itself. What Kipping’s team is looking for is the effect the moon has on the planet, and the transit of that Saturn-class world around an M-dwarf is something Kepler can work with.

The method relies on two sets of observations, the first being transit timing — variations in the amount of time it takes a transiting planet to complete its orbit can be the signal of a moon. Add transit timing variation to the second measurement — transit duration — and you can nail down the presence of that moon. Transit duration measures the speed at which the planet actually passes in front of the star. Detecting Life-Friendly Moons, on the Astrobiology Magazine site, explains that the two signals occur separately when a moon is involved, screening out other possible causes.

The Kepler researchers have their hands full looking for exoplanets, but an exomoon survey using Kepler data is certainly in the running for future investigation. If a moon as small as a third of Earth mass can hold onto a magnetic field, life could develop despite the presence of nearby planetary radiation belts. Such a moon, Kipping believes, would be detectable in the right transit situation as much as 500 light years from the Sun.

“There may be just as many habitable moons as habitable planets in our galaxy,” Kipping tells Astrobiology Magazine, an audacious thought with all kinds of ramifications for the Fermi paradox. Remarkably, combining the two measurements his team works with would allow Kipping to estimate both the moon’s orbital period and its mass. Run that to its limit and you get this: Knowing the orbital period would allow prediction of a lunar eclipse whose spectroscopic study could then reveal the signature of atmospheric gases on the moon’s surface. That’s pushing current technology to the max, but it’s increasingly feasible as we tune up our resources.

The paper is Kipping et al., “On the detectability of habitable exomoons with Kepler-class photometry,” Monthly Notices of the Royal Astronomical Society, published online 24 September, 2009 (abstract).


Comments on this entry are closed.

  • andy October 27, 2009, 13:58

    The suggestion is that moon masses cannot exceed a small fraction of the mass of the parent planet, but various methods for forming objects of a few times Jupiter’s mass, including ones that are more starlike. This would presumably have some impact on the circumplanetary environment, so perhaps satellite systems could be used to distinguish between giant planets and low-mass brown dwarfs?

  • Mike October 27, 2009, 16:00

    To Andy, regarding your previous comment about exomoons could you please describe it in more detail. I didn’t really understand the first sentence.
    Thank you.

  • spaceman October 27, 2009, 17:13

    Very interesting. Before the discovery of hot-jupiters and other close-in jovian worlds, this idea was not in vogue. But now we know of jupiter-like exoplanets in the habitable-zones of their parent stars. I wouldn’t be surprised if habitable moons are found to exist at some point. I would guess that they are probably not quite as common as habitable planets, but imagine the views from the surfaces of such orbs!

    Also, speaking of transit photometry more generally, it was announced that the CoRoT mission has been extended to the year 2013. Good news!

  • Allan R. Schmitt October 27, 2009, 17:27

    Is anyone willing to go out on a limb and predict when the first exomoon for a planet in the HZ of a star will be detected? My optimistic guess is the latter half of 2010. By then Kepler should have identified all of the Saturn class exoplanets in the HZ of M class stars. One of the follow-up ground based observatories will take the inititive and perform an intensive TTV and TDV study of the nearest candidate exoplanet(s). It would be a truely stunning achievement if sussessful only slightly less important than the discovery of an earth sized exoplanet in the HZ of a sun-like star.

    I’m very excited about the possibility. What exciting times we live in!

  • Athena Andreadis October 27, 2009, 17:59

    I’m also very excited. Given the situation in our solar system, extrapolations argue that exomoons are numerous, varied and inhabitable… inhabited?

  • andy October 27, 2009, 18:25

    Mike: apparently I dropped a few words while editing it. Point is the theoretical models of satellite formation around giant planets predict a ratio of satellite system mass to planet mass of a few times 1/10000. However if the planet forms by methods such as gravitational collapse, the circumplanetary environment may well be very different.

  • Dave Weeden October 27, 2009, 18:55

    The obvious good thing about exomoons in the ‘habitable zone’ round a red dwarf is that these can’t be tidally locked to the star. So the whole planet will get starlight at some point, which should attenuate the extreme conditions one would expect of a tidally locked earth-like world.

    But, while the abstract says, “The effects of stellar variability, instrument noise and photon noise are all accounted for in the analysis” the paper doesn’t seem to consider the possibility of there being so many moons (Jupiter has 60) that the observations can never be certain of anything other than that there are some moons and a rough guess at their total mass. I mean, if Saturn and all its satellites orbited a nearby red dwarf, could this method detect Europa unambiguously? I really suspect that the signal-noise ratio leans heavily toward noise here.

  • ljk October 28, 2009, 0:08


    Friday, October 23, 2009

    Naming the Exoplanets

    The International Astronomical Union is refusing to name the exoplanets. That seems unnecessarily curmudgeonly.

    Since 1995, astronomers have found more than 400 planets orbiting other stars. And yet not one of them has a formal name, other than their orginal scientific designation such as MOA-2008-BLG-310-L b, (a sub-Saturnian mass planet recently detected in the Galactic Bulge). How come?

    The official reason from the International Astronomical Union is that astronomers expect to find numerous exoplanets, so many in fact that it would be impractical to name them.

    Today Wladimir Lyra at the Max Planck Institute for Astronomy in Germany says this stance makes no sense. He points out that stars are even more common and it is indeed impractical to name them all but that hasn’t stopped us naming some of them.

    The IAU’s stance seems difficult to justify, not least because it is unnecessarily curmudgeonly. There are good reasons for thinking that some of these exoplanets, perhaps many of them, will turn out to play significant roles in the history of astronomy, perhaps by helping us understand the nature of planets and the possibility that life may exist on some of them.

    Lyra points out that some astronomers have begun using unofficial names for some planets but he wants to go further. He thinks they should all have names and has come up with one for everyone of the 403 exoplanets discovered to date.

    Lyra’s scheme uses names from Greek and Roman mythologies, the same system used to name other heavenly bodies, such as planets stars and asteroids. He’s even explained the mythological significance of the names he has chosen.

    So far example CoRoT-3 b becomes Cratos, WASP-6 b becomes Teucrus and HD 132406 b becomes Atlas. For the full list, see the paper.

    That’s a creditable effort but surely we could all have a little more fun with an exoplanet naming system, perhaps by opening it up to the next generation of astronomers. What better way to inspire children interested in astronomy than by allocating planets to high schools around the world and asking the kids themselves what these bodies should be called?

    Ref: http://arxiv.org/abs/0910.3989: Naming the extrasolar planets

  • keith October 28, 2009, 8:50

    I think objects should be actually IMAGED before anyone has the right to name them. If you are allowed to name things whose presence is only inferred or calculated, where does it end?

    Right now, I want to name the valley on Pluto, the one that I infer exists. I am staking my claim to name this valley, on Pluto, “Keith’s Valley”. OK?

  • Athena Andreadis October 28, 2009, 10:15

    It’s discouraging that Lyra’s names are all male and come from European mythology (even if it’s Greek). As for Keith’s point — many things have been named that have never been seen. All deities, to give one example.

  • philw1776 October 28, 2009, 12:05

    Alan writes, “Is anyone willing to go out on a limb and predict when the first exomoon for a planet in the HZ of a star will be detected? My optimistic guess is the latter half of 2010. By then Kepler should have identified all of the Saturn class exoplanets in the HZ of M class stars.”

    I believe that the longest HZ ‘year’ of the brightest M stars M0 is around 2 months. Kepler started science in late May, therefore it will have had the required 3 congruent observations of any HZ planets and satellites by Thanksgiving 2009. The delay in ground confirmation (fighting for BIG telescope time) and the delay in publishing derived papers will gate any announcements. I’m hopefull that the Astronomy convocations in January 2010 will be the vehicle for the first Kepler planet announcements. M star exoplanet moons in HZ, if any, could well be published in papers and announced, likely in the 2nd half of 2010. Most K stars have HZ years 4 months or less. 2010 will be fun!

  • Ronald October 28, 2009, 13:46

    I am pretty confident that once we find an earthlike planet in a solartype star’s HZ, it will be named (and even much more so when it shows a biosignature!).

    Right now, and even more so in the near future, trying to name all found (inferred) planets seems like ‘boire la mer’ (drinking the sea), or trying to name grains of sand.
    In the course of this century, unless a global disaster throws back human civilization, we will probably discover many millions of new planets, increasingly by fully automated ways of detection.

  • Ronald October 28, 2009, 14:20

    I don’t want to be the party pooper here, but I think a few words of caution may be due.

    As andy is also implying, I understood from earlier threads on this topic here, that there is an upper mass limit on moons (of gas giants) of around 2*10^-4 or about 0.02 % (Ward, Nature 2006). Furthermore the largest Jovian moons are not or hardly larger than the largest Saturnian (Titan), it just has more of them.

    According to this, to get a moon around 1/3 of earth mass would require a gas giant of approx. 3 – 5 Jupiter masses.

    Correct me if I am mistaken, but so far, gas giants around M dwarfs seem to be relatively rare anyway. I wonder how many M dwarfs possess super gas giants of required mass.

    This in turn would imply some nasty radiation belts. Then again, the moon’s atmosphere and magnetic field may be sufficient to deal with that.

  • Ronald October 28, 2009, 14:22

    Now, a habitable exomoon around a (early) K or G star may be a more promising prospect. But will Kepler be able to detect those as well and how long will that take?

  • keith October 28, 2009, 14:46

    Athena Andreadis
    It’s not the same ! Here we are talking about science and astronomy and naming real concrete objects, not fantasy.

    What things have been named by their discoverer that have not been seen or imaged in some way?

  • andy October 28, 2009, 16:13

    Regarding Athena Andreadis’s comments on the matter of naming: there are quite a few female names in there (though I haven’t checked through the relative proportion). They don’t seem to have been “ghettoised” to one constellation in a similar way to the way that the nomenclature of planetary geography seems to tend to do in relegating female names to Venus either. I agree that the choice of European culture for the nomenclature is disheartening though: the justification offered in the text basically boils down to “we haven’t acknowledged other cultures in our planetary names before, so why start now”. Not convinced – rings very hollow to me.

    In fact I disagree that the exoplanets should be given names at the present time: the vast majority of them are very poorly characterised. A radial velocity variation gives only a minimum mass: we cannot be sure that these objects are necessarily exoplanets at all. The example I’m thinking of here is HD 33636b, for which the radial velocity method gave a minimum mass of 10 times Jupiter. Astrometric measurements then revealed that the orbit is near face-on, and that the “planet” is actually a low mass red dwarf star.

    Another notable incidence of things being not quite as they seem was the assignment of the name “Goldilocks” to the object orbiting 70 Virginis. Refined parallax measurements showed the star was further away and more luminous than had been thought, so instead of being in the habitable zone, the planet was far too hot for habitability. Besides, the planet has a minimum mass large enough to imply it is, at the least, a gas giant. So much for a “Goldilocks planet”.

    Lyra’s naming system has a few oddities where the name doesn’t seem to reflect very well the name of the object, for example he assigns the name Phlegethon (the river that flows with fire which burns and does not consume) to HD 28185b, an object which (according to current information) orbits within its parent star’s habitable zone. Seems like an odd choice to me.

    If the current naming convention is difficult, it is surely because of the designations for the parent stars. There are fairly unwieldy ones like “MOA-2008-BLG-310L”, numerical designations which have the potential for confusion (both HD 73256 and HD 73526 are exoplanet host stars). It might make things somewhat more accessible to name the exoplanet host stars, rather than the exoplanets themselves. We are already used to such names as “Sirius B” for companion stars, and the planet orbiting Fomalhaut is usually referred to as “Fomalhaut b”. Certainly it would lead to less of a mess in the nomenclature if, say, something called “Huitzilopochtli b” turned out to be a star: just change the designation to a capital letter.

  • Athena Andreadis October 28, 2009, 16:44

    Ok, Keith, let’s stick to real examples: most elementary particles were named when they were hypothesized, which for some of them was/is way before direct or indirect observation (still AWOL: Higgs boson, graviton, gravitino… you get the gist).

    One reason to name planets as long as they become a strong probability is that names like HD 73256 don’t roll off the tongue and as a result never stick in people’s memories. I also agree with Andy that if the names are evocative, they need to fit the target.

  • Athena Andreadis October 28, 2009, 16:44

    Oops, typo! I mean “as soon as they become a strong probability”

  • andy October 28, 2009, 19:12

    Regarding M-dwarfs, it is interesting that the vast majority of known giant planets around such stars are located where pre-51 Pegasi theories predicted gas giants would be: in long-period orbits beyond the snowline. The first known planetary system around an M-dwarf, Gliese 876, appears to be a rather rare case of an M-dwarf which has its giant planets located further in.

  • ROCA October 29, 2009, 8:11

    The existence of an exomoon habitable orbiting an exoplanet in a M-red-dwarf system is exciting because this exomoon would have days and nights cycles. The exoplanets in M-dwarf habitable zones are tidally locked and exhibit the same face to their star. That should make these extrasolar worlds truly inhospitable for complex organism development. But the exomoons wouldn’t have this problem…

  • ljk October 29, 2009, 10:09

    Keep in mind that these planets might already be named by the natives
    and/or the local interstellar empire.

  • Athena Andreadis October 29, 2009, 14:50

    Hehe! Larry is absolutely right. True for earth as well, from Aotearoa to Massachusetts.

  • philw1776 October 31, 2009, 9:03

    Really BAD news for KEPLER…

    Kepler, NASA’s mission to search for planets around other stars, will not be able to spot an Earth-sized planet until 2011, according to the mission’s team. The delays are caused by noisy amplifiers in the telescope’s electronics.

    “We’re not going to be able to find Earth-size planets in the habitable zone — or it’s going to be very difficult — until that work gets done,” says Kepler principal investigator William Borucki.

    Borucki says that the noise will hinder searches for a rarer scenario: Earth-sized planets that orbit more quickly around dimmer, cooler stars — where the habitable zone is closer in. These planets could transit every few months.


    Apparently the software fix for this known before ;aunch issue will take until 2011 or thereabouts.

  • Ronald November 1, 2009, 4:52

    Ref. philw1776;
    I understand from the article, that it wil mainly effect the detection of earthsized planets in close orbit around M dwarfs and the like, no so much that of the earthlike planets in more earthlike orbits around solar type stars (because their orbits are so much longer and solid confirmation takes at least 3 orbits).

    Quote: “Greg Laughlin, an astronomer based at the University of California at Santa Cruz, says that the delay for Kepler makes it “more likely that the first Earth-mass planet is going to go to the radial-velocity observers”.”

  • ljk November 2, 2009, 10:59

    For what this is worth, James Cameron’s new SF film Avatar, coming to the big screen, as they say, next month, apparently takes place on an Earth type world that appears to be a moon of a gas giant planet, complete with a blue version of Jupiter’s Great Red Spot – judging by this nice artwork from the film:


    Perhaps Avatar will help make the general public more aware of the possibility for exomoons to be places of life as well as Earth type worlds. I know Star Wars had Yavin 4 and the “forest moon” of Endor, but the first one’s nature was not terribly emphasized and the latter apparently had its gas giant parent somehow explode before the events in the film. In other words, it is definitely time for a new SF flick to focus on the concept.

    Now we need a film that focuses on the idea of aliens whose origins do not rely on a planet of any sort to break loose from that paradigm as well.

  • ljk December 1, 2009, 22:07


    Tuesday, December 01, 2009

    The Hunt for Habitable Exomoons

    Forget exoplanets, the first habitable body to be discovered beyond our solar system could turn out to be an exomoon in a habitable zone.

    The discovery of habitable planets around other stars is one of the great goals of modern astronomy. But it’s not just planets that can host life. Astronomers have long believed that moons orbiting Jupiter-like planets in the habitable zone could have Earth-like qualities. The problem is how to detect them.

    One of the best ways to spot exoplanets is to look for changes in a parent star’s brightness as the they pass in front of its disc. So why not look for exomoons in the same way? A planet-moon system orbits a common centre of mass which itself orbits the star. Consequently, the planet’s position can be shifted by a small amount as it moves in front of the disc, leading to a small change in the time that a transit begins and ends. This transit timing variation (TTV) is a signal that could be used to spot an exomoon.

    In theory, at least. The trouble is that that up to 98 per cent of promising exoplanet signals turn out to be false alarms caused by other effects. It turns out there is no way of teasing the exomoon signal out of this mess.

    And there the field of exomoon hunting would have remained were it no for the work of David Kipping at the Harvard-Smithsonian Center for Astrophysics in Cambridge and a few buddies. Earlier this year, these guys showed that exomoons produce another signal. They pointed out that not only does an exomoon change the start time of a transit, it should also change the duration of the transit and that together, these signals can uniquely identify an exomoon.

    Today, Kippling and co say this method is possible now. Earlier this year, NASA launched the Keppler Space Telescope to stare continuously at a fixed region of the sky and measure the changes in brightness of some 100,000 stars.

    After crunching a few numbers, Kippling and co say that Keppler should be able to see exomoons smaller than Earth around Saturn-like exoplanets if they are orbiting any of 25,000 stars in the Keppler’s field of view.

    That’s exciting news. Keppler is already sending back data. If habitable exomoons are out there, we’ll see them soon.

    Ref: http://arxiv.org/abs/0911.5170: Pathways Towards Habitable Moons