25% of Sun-Like Stars Could Host Earth-Sized Worlds

Article Updated: 24 Dec , 2015


A five-year survey of nearby solar-mass stars has provided astronomers with an estimate of how many stars of this type could have Earth-size planets. Andrew Howard and Geoffrey Marcy from the University of California Berkeley studied 166 G and K stars within 80 light-years of Earth, determining the number, mass and orbital distance of any of the stars’ planets. Since Earth-sized worlds have not yet been found, they extrapolated the number of that size of planets, based on the fraction of stars that host Neptune to super-Earth sized planets. Their findings are encouraging, since it means planets the size of Neptune and smaller are probably much more common than gas-giant planets, like Jupiter. But what they found also conflict with current models of planet formation and migration.

“Of about 100 typical sun-like stars, one or two have planets the size of Jupiter, roughly six have a planet the size of Neptune, and about 12 have super-Earths between three and 10 Earth masses,” said Howard. “If we extrapolate down to Earth-size planets – between one-half and two times the mass of Earth – we predict that you’d find about 23 for every 100 stars.”

“This is the first estimate based on actual measurements of the fraction of stars that have Earth-size planets,” said Marcy. Previous studies have estimated the proportion of Jupiter and Saturn-size exoplanets, but never down to as small as this study, and the astronomers say this enabled them to estimate the Earth-size planets.

“What this means,” Howard added, “is that, as NASA develops new techniques over the next decade to find truly Earth-size planets, it won’t have to look too far.”

Using the 10-meter Keck telescopes in Hawaii, the astronomers measured the small wobble of each star from the tug of orbiting planets. For systems with multiple planets, teasing out the radial velocity signature of each planet is very complex, since each signature is extremely small. The more times a star is observed, the better the data. Current techniques allow detection of planets massive enough and near enough to their stars to cause a wobble of about 1 meter per second. That means they saw only massive, Jupiter-like gas giants up to three times the mass of Jupiter (1,000 times Earth’s mass) orbiting as far as one-quarter of an astronomical unit (AU) from the star, or smaller, closer super-Earths and Neptune-like planets (15-30 times the mass of the earth). An AU is 93 million miles, the average distance between the earth and the sun.

Histogram of stellar masses for Eta-Earth stars. Credit: Howard, et al.

Only 22 of the stars had detectable planets – 33 planets in all – within this range of masses and orbital distances. After accounting statistically for the fact that some stars were observed more often than others, the researchers estimated that about 1.6 percent of the sun-like stars in their sample had Jupiter-size planets and 12 percent had super-Earths (3-10 Earth masses). If the trend of increasing numbers of smaller planets continues, they concluded, 23 percent of the stars would have Earth-size planets.

Based on these statistics, Howard and Marcy, — who is also member of NASA’s Kepler mission to survey 156,000 faint stars in search of transiting planets — estimate that the telescope will detect 120-260 “plausibly terrestrial worlds” orbiting some 10,000 nearby G and K dwarf stars with orbital periods less than 50 days.

“One of astronomy’s goals is to find ‘eta-Earth,’ the fraction of sun-like stars that have an earth,” Howard said. “This is a first estimate, and the real number could be one in eight instead of one in four. But it’s not one in 100, which is glorious news.”

They were able to only detect close-in planets, so they say there could be even more Earth-size planets at greater distances, including within the habitable zone — or Goldilocks zone — located at a distance form the star where conditions are not too hot or too cold to allow the presence of liquid water.

But the researchers note that their results conflict with current models of planet formation and migration, where it is thought that nascent planets spiral inward towards the sun because of interactions with the gas in the disk. Such models predict a “planet desert” in the inner region of solar systems. But that’s where all the planets are being found.

“Just where we see the most planets, models predict we would find no cacti at all,” Marcy said. “These results will transform astronomers’ views of how planets form.”

Howard and Marcy report their results in the Oct. 29 issue of the journal Science.

Sources: UC Berkeley, Science

15 Responses

  1. tacitus says:

    I don’t believe adjusting upwards the likely number of planets in the Universe is going to do much to close the missing mass gap.

    Note: still not digging the link explosion on every article. We have three links for variations of “the mass of the Earth” in this one, three links for “the Sun” and three for “stars” (one of which points erroneously to “A-star”.

    If this is really just about helping the reader, then this is the wrong way to go about it.

    Firstly, there are just too many terms being picked out. I can’t imagine there are many readers who need to know what a planet or star is — certainly not enough to warrant highlighting the terms several times in every single article. The same goes for all the planet names and other very commonly used astronomical terms. One helpful link to one introductory article (or series) on the page would be more than sufficient.

    Secondly, you need to filter the terms your system is picking up for duplicates. The algorithm you are using is not working, since it is cluttering the page with duplicate links (e.g. sun, Solar, the Sun, etc).

    Thirdly, this type of link explosion apes the worst type of web sites that scatter dozens of links throughout all their text in order to either gain a better search ranking, or just as a way to place more ads. I have no reason to suspect that is what you are doing with UT, but I don’t think you want to be compared with these sites in any way or form.

    Fourthly, these links are indistinguishable from real links to articles and information pertaining to the actual story (instead of just background). This makes looking for these one or two real links an exercise in frustration as you have to hover over several to find the right one..

    So why not create a glossary widget in your sidebar instead of piling up on the links? The same (well, improved) code can dig out the terms in the article, and place them in the glossary widget, providing the same access to useful information without messing up the main article.

    I don’t like complaining, but as a long time reader of this site, I hate to see this happening, and I hope the feedback is useful.

  2. tacitus says:

    As for hypertextng, that is I think automatic. Hypertexting words like planets reflects the fact grade school kids may be visiting this site.

    Yes, I am aware it is automatic, but there is code behind the linking, and there is a coder behind the code πŸ™‚

    I agree that if grade school kids really are using this facility to learn more about astronomy then great, and leveraging the accumulated body of knowledge on this site is a great idea, but I think the way the system works at the moment is as much confusing as it is elucidating.

    I am a WordPress developer (written several plugins and have created several WP-based sites), so if there is anything I can do to help, just let me know.

  3. Aqua says:

    It appears that matter isn’t as dark as some models would have us believe? How much ‘missing matter’ is implied? What does this do for the Drake Equation?

  4. Aqua says:

    A refresher?

    N = N* fp ne fl fi fc fL

    N* represents the number of stars in the Milky Way Galaxy
    Current estimates are 100 billion.

    fp is the fraction of stars that have planets around them
    Current estimates range from 20% to 50%

    (The above states that the current estimate for this number is 23% = low end)
    ne is the number of planets per star that are capable of sustaining life
    Current estimates range from 1 to 5.

    fl is the fraction of planets in ne where life evolves

    fi is the fraction of fl where intelligent life evolves.

    fc is the fraction of fi that communicate
    What percentage have the desire to communicate?

    fL is fraction of the planet’s life during which the communicating civilizations live

    N, the number of communicating civilizations in the galaxy.

  5. tacitus says:

    I’ll see what I can do to remove it from the newest news, so it doesn’t clutter things up.

    Actually, that’s not a bad compromise, Fraser. Delaying the linking for a few days would probably mean that most people who follow the site regularly don’t have to wade through them, but you still get the longer term benefits.

    Thanks for listening.

    (I would still look at tweaking the algorithm though πŸ™‚ )

  6. Torbjorn Larsson OM says:

    I’ll be damned – Gliese 581 g exists after all! (Since its discovery place Earth analogs in 10-30 % of systems, commensurate with this result.)

  7. swampfox48 says:

    Why estimate? To estimate how many possible planets with intelligent life? Fools gold. Our sun and planet is 85% of the age of the universe. If biologists are correct with how evolution works only similar aged and older Earth-like planets have a possibility. Take into consideration how one gamma ray explosion kills all life within hundreds of parsecs and your at a pretty low figure.

  8. Lawrence B. Crowell says:

    These additional planets are a billionth the missing mass attributed to dark matter.

    As for hypertextng, that is I think automatic. Hypertexting words like planets reflects the fact grade school kids may be visiting this site.

    It has to be remembered that even if 25% of solar mass stars have Earth mass planets that a fraction of those are in a habitable zone and that many of these stellar systems are not gravitationally optimal to keep a small planet in a stable orbit. If Jupiter were located where Mars is then Earth would have been gravitationally pertrubed a whole lot more. It is questionable whether life could have evolved or even persisted this long.


  9. Sirius_Alpha says:

    ” This makes looking for these one or two real links an exercise in frustration as you have to hover over several to find the right one..”

    I was honstly unaware that there were links in here that weren’t automatically generated other than the links that state the source.

  10. Fraser Cain says:

    Hi Tacitus, you’re absolutely right, this is being done automatically by a WordPress plugin. I’ll see what I can do to remove it from the newest news, so it doesn’t clutter things up. But it’s essentially what Wikipedia is doing. There’s a huge Wikipedia-style encyclopedia of information underneath Universe Today, so a person arriving at the site can follow these links and get more information on almost any topic.

    It helps bring in more readers through people searching topics, and it provides a benefit to people doing research reports.

  11. Uncle Fred says:


    If I’m reading this article right, the researchers only extrapolated sub-Earth to super-Earth sized planets within the regions close to their parent star. These are the very regions that planetary models predict that planets should not be able to exist for long periods – or form at all.

    It could be possible that the chances for a habitable body within the Goldilocks zone of a sun-like star is actually greater than 25%. Yet at this time we have no data to extrapolate from for these circumstances.

    Can’t wait for those multi-year Keplar results…

  12. Kawarthajon says:

    This logic of this study simply does not convince me – because we can’t find the planets, but the ratio of the size to frequency of planets increases as the planets get smaller, therefore there must be 23% Earth like planets.

    Don’t get me wrong, I’m sure that there are Earth like planets out there with life on them, but to somehow pull 23% out of a hat when they weren’t able to find a single one and they surveyed a very small number of stars compared to how many stars there are in our general neighbourhood. I

    t seems like a pretty big leap of faith to come up with such a specific number when we don’t know the value of all of the variables and our planetary formation models don’t match what we have observed so far.

  13. gopher65 says:

    Fraser Cain: I’m an admin on 2 different Wikimedia sites, and you’re incorrect that this is how Wikipedia (and its sister sites) do things. Each Wikimedia site has its own ruleset (including each of the various language editions of Wikipedia), so things aren’t exactly the same between sites (IE, Wikipedia’s rules are different than Wikinews rules). But even with the different rules, things are fairly similar.

    For the most part not only are links not automatically generated (we do them *all* manually. And yes, it’s a huge amount of work), but on most Wikimedia sites there are specific rules saying “don’t like common words. People know what “plant” means, so you don’t need to link it. That just creates clutter. People hate that, so don’t do it.”

    So no, what you do isn’t what Wikipedia (or other Wikimedia Foundation sites) does.

  14. gopher65 says:

    *looks at all the typos in his previous post*

    The lack of an edit feature on WordPress makes me sad:(. But it’s still better than Disqus.

  15. gopher65 says:

    Ok. My previous post was just a blather of garbage, so let me try this again, as simply as possible:

    Links on Wikimedia sites are done manually. Each link is suppose to be carefully considered before placement. If the word is too common, like “plant” or “planet”, then it shouldn’t normally be linked. Only links that actively contribute to the article in question should be placed. The reason for this is that if you link common words, your article becomes cluttered with links. A solid wall of blue linky text is neither fun to read nor helpful to the readers.

    Because of this links are suppose to be created only if one of two conditions are met:

    1) The word is uncommon or has multiple meanings, and could therefore be reasonably misunderstood by an ordinary person not familiar with the topic in question, or

    2) The word or phrase is topical. This means that linking the word or phrase will allow the reader to quickly surf through the various different pages of the topic in question. For instance, on the English Wikipedia page “Canada”, the first sentence links to two different articles: “North America”, and to another article that has detailed information on Canada’s provinces and territories. Those are things that are integrally related to the topic of Canada, and giving the reader the chance to click those links increases the reader’s chance of locating the specific information that they are looking for.

    And that last sentence is key. Links should be either explanatory or they should exist to make locating additional information on the *specific topic of the page* easier.

    I notice that the page I’m currently typing this on is about sunlike stars having planets. Linking to “sun like stars” is fine. Linking to “gas giants” is fine. Linking to “Kepler mission” is perfect. That’s a great link; it’s very topical. But general links like “telescope”, “astronomy”, and “Saturn”? How do those help you find out more about the topic of “planets around sunlike stars”?

    That’s why links need to be done manually rather than automatically. We’ve discussed doing them automatically on some of Wikipedia’s sister wikis, but in the end we always ended up realizing that links need to have some intelligence behind them to be useful. Otherwise they’re just random garbage that the audience filters out while reading.

Comments are closed.