≡ Menu

Sharp Early Returns from Kepler

Unlike the Cassini Saturn orbiter, which we looked at yesterday in the context of cryovolcanism on Titan, the Kepler spacecraft has but a single scientific instrument. It’s a photometer based on a Schmidt telescope design with a 95 cm aperture and a field of view larger than 100 square degrees. Measuring brightness variations for over 100,000 stars, Kepler is the first mission that should be able to detect Earth-size planets in the habitable zones of their stars.


That made yesterday’s news conference an eagerly anticipated event, but we have to remember that it’s going to be a while before we start talking about terrestrial planet detections. It takes multiple transits and much data analysis to make that possible, and a transiting world at roughly Earth-like distance from its star will demand several years of work. Kepler’s baseline mission is three and a half years, more than enough to make such detections, and the good news is that the instrument works.

Image: Magnified Kepler measurements of the planet HAT-P-7b showing transits and occultations. Credit: NASA.

The lightcurve of the planet HAT-P-7b shown at the news conference yesterday was dramatic proof. It was based on a mere ten days of test data collected during Kepler’s commissioning period, before science operations officially began. And even before the instrument has been fully calibrated and its data analysis software fine-tuned, it was able to detect HAT-P-7b’s atmosphere. The level of exactitude in these measurements has everyone talking. Here’s William Borucki, Kepler principal science investigator:

“When the light curves from tens of thousands of stars were shown to the Kepler science team, everyone was awed; no one had ever seen such exquisitely detailed measurements of the light variations of so many different types of stars.”

The paper on HAT-P-7b is being published in Science today, describing work on a planet that is roughly a thousand light years from Earth. It was a useful early target for calibration given that this gas giant orbits in a mere 2.2 days, a ‘hot Jupiter’ some 26 times closer to its star than Earth is to the Sun. That makes for numerous transits in short order, and in this case offers observations of a planet that is hot enough to be glowing like the burner of a stove (more in this news release).

Both initial transit and occultation were clearly visible as the planet first passed in front of, then behind the star as seen from Kepler’s vantage point. What we learn is that HAT-P-7b’s atmosphere has a dayside temperature of more than 2350 degrees Celsius (4310 degrees Fahrenheit). And here’s the key: The observed brightness variation is a mere one and a half times what would be expected from a terrestrial planet transit.


The prospects for terrestrial planet detections, in other words, have never looked better. The NASA image above gives us an idea how the story will play out. Here are planets graphed by mass and orbital distance, with plentiful representation at the high end and no planets of Earth mass or lower yet detected in the habitable zones of their stars. We should be able to offer a significantly different chart within just a few years.

The paper is Borucki et al., “Kepler’s Optical Phase Curve of the Exoplanet HAT-P-7b,” Science Vol. 325. no. 5941 (7 August 2009), p. 709 (abstract).


Comments on this entry are closed.

  • tacitus August 7, 2009, 13:02

    I wonder if they’ll also detect:

    a) planets that don’t quite transit, but are aligned well enough to generate the “phased light curve” as exists in HAT-P-7b.

    b) non-transiting planets though variations in the timing of an observable planet’s transits.

    Something tells me we will be squeezing many more planets out of the raw data as the years go by — possible for many years after the mission has ended.

    Fantastic stuff!

  • Ron S August 7, 2009, 14:35

    That last chart, which is really nice, is not quite right. The shown radius for the habitable zone is only true for stars with luminosities equal to the Sun’s. Is there a similar chart that shows orbital radii relative to each system’s habitable zone?

    tacitus (a): Non-transiting planets that are of Earth’s size and orbital radius will likely be below the detection threshold (someone know the exact values?) since the ripple of the light curve will be of much, much lower amplitude. The same may be true of Jovians in a Jupiter-sized orbit.

    In the likely case where there are multiple planets in the system, the ripples would also have to be separated (Fourier analysis?). It will also be true that where one planet is in a true transiting orientation, the other planets would (mostly) not transit, even though they are roughly in the same plane. There is going to be a lot of detailed analysis before we see results.

    Regardless, this is great stuff: data, data, data!! Bring it on!

  • Administrator August 7, 2009, 16:11

    Ron S writes:

    Is there a similar chart that shows orbital radii relative to each system’s habitable zone?

    Would like to see such myself, and if I locate one, I’ll see if I can post it here.

  • Adam August 7, 2009, 18:01

    A chart which defined the AU for a system as the 1 Earth insolation level might be useful. Wonder how hard it would be to write a program to that effect?

  • NS August 7, 2009, 18:32

    In the press conference they said the slope in the light measurements as the planet revolved around its star showed the variation in the light the planet was reflecting toward us. But they also said this indicated a large difference between its day- and night-side temperatures. I wasn’t clear on how they could determine that.

  • andy August 8, 2009, 4:28

    A chart which defined the AU for a system as the 1 Earth insolation level might be useful. Wonder how hard it would be to write a program to that effect

    Well the mathematics isn’t too difficult, it’s a matter of getting reliable data on the bolometric luminosities of the stars in question. You’d also have to scale for orbital eccentricity: planets with the same semimajor axis but with more eccentric orbits receive more energy from their star over the course of an orbit.

    The question of where in the habitable zone (if at all) such an insolation level corresponds to is another matter… there is a dependence of albedo on the incident spectrum, so Earth insolation does not necessarily put you in the HZ…

  • Dave Weeden August 8, 2009, 7:49

    Why are we looking for terrestrial planets exactly? I realise that the confirmed discovery of one would make the news, and that might affect funding and graduate applications down the line, but I can’t see the utility myself. We can’t even go to Mars, realistically, so whether or not we could live on a exoplanet is really just academic.

    Whether a planet could support life – well, I really don’t see why very large planets couldn’t, and the ‘habitable zone’ applies to planets around a sun-like star, which have an ozone layer to protect the surface from UV rays, have a magnetic field, a relatively large moon etc etc. Life, that is self-replicating chemicals, may be very common, and we’re nowhere near confirming the rest of the solar system as sterile. At the moment, the interesting questions we’re some way to gathering hard data on concern planetary systems origins, development, and likely forms. Gas giants seem to be very common, but many are in very close orbits around their stars. Why is this? and does this mean that they ‘swept up’ rocky (‘habitable’) planets when their orbits spiralled inward (if they did)?

    Even if we do spot a terrestrial planet in the right orbit, I don’t see any way we can make testable biochemical hypotheses about life. Finally, there’s only been life on earth for 2-3 billion years out of 4.5 billion. That says to me that there’s a 1 in 2 or 1 in 3 chance that any world we find of the right mass in the right orbit will be barren anyway. Looking back at the graph, there seem to a lot of Jovian planets closer in to their primaries than Jupiter is (though I can’t tell from the graph how bright those primaries are), so there may be plenty of moons like Enceladus and Titan but a little warmer. If we’re really looking for life, we need to keep very open minds …

  • Allan R. Schmitt August 8, 2009, 9:59

    What about the detection of earth-size planets in the habitable zone of M class stars? Based on a lower mass star, the orbital period would be a few months rather than a year. So wouldn’t Kepler be able to report results much sooner for those planets? Or am I wrong in assuming Kepler can detect planets around low mass (luminosity) stars?

  • Administrator August 8, 2009, 10:36

    Allan, you’re on target. Orbital periods will be correspondingly shorter for planets in the habitable zone around an M-dwarf, and Kepler will be examining such targets as well as G-class stars like the Sun and hot A-class stars.

  • philw1776 August 8, 2009, 12:37

    Word up is to expect Kepler discovery announcements at the Jan 2010 AAS meeting. Planets with orbital periods under 5 weeks or thereabout. Discoveries will be verified by large ground telescope radial velocity measurements where possible. A slow process. I believe that Kepler can only detect planets around the brighter M stars. The ‘Dynamics of Cats’ blog by a Penn State astrophysicist is a nice source of Kepler advanced informed speculation.

  • Administrator August 8, 2009, 13:29

    Dynamics of Cats is Steinn Sigurðsson’s site, always worth reading for those of you who don’t already know it:


  • yeti101 August 9, 2009, 8:39

    re Dave Weedens post.

    Why are we looking for terrestrial planets? Ever since mankind has known the stars in the sky are other “suns” we have asked the questions : Are there other planets like earth around those stars? Are there other beings like us on those planets? Kepler is the first step towards answering those questions.

    Kepler is designed to tell us how common terrestrial planets in the HZ are for the whole galaxy. The sample is big enough for us to extrapolate it galaxy wide. That in itself will be very interesting to know.

    Kepler wont be able to determine the chemical composition of the planets atmospheres that it finds. Future space telescopes such as Darwin/TPF will. We run the light through a spectrometer and If we find a planet with a similar chemical signature & age as earth it looks very good for complex life existing there.

    its going to take these discoveries to fire peoples imgaination about interstellar travel. If we found an earth like planet 10ly away you can bet more attention would be paid to solving how we get there, or send a probe there.

    I agree it may take finding alot of these planets before we find one similar to earth. Right now about 30% of systems studied have a hot jupiter

  • philw1776 August 9, 2009, 11:50

    The 100,000+ Kepler watched stars are mostly hundreds to a thousand or more LY away. Darwin/TPF are designed to analyze planets within less than 50 LY, assuming they meet current proposed specs. As previously stated, what Kepler will give us for the first time is hard, significant statistical data on the percentages of terrestrial planets in the HZ and data on solar systems in general. Are solar systems like ours common, or a statistical freak? We’ll likely know in 4 years or so.

  • Athena Andreadis August 10, 2009, 10:45

    I want to second all of Yeti101’s points (great handle, by the way, and fitting in this venue: SETI, METI, YETI…). Biology is still lagging badly in astrobiology for lack of a second sample. And if you have a specific destination with a possibility of extant complex life, the motivation will be much stronger to fund and send an expedition (robotic or crewed) rather than the general fishing expeditions we have to mount now.

  • george scaglione August 12, 2009, 9:18

    athena,you are certainly correct it would be a huge “shot in the arm” for space if we only had that”second example”.could not fail to help.if only we could get some information in this regard it would indeed make history.thank you very much,your friend george

  • ljk August 13, 2009, 10:42

    HAT-P-7: A Retrograde or Polar Orbit, and a Second Planet

    Authors: Joshua N. Winn, John Asher Johnson, Simon Albrecht, Andrew W. Howard, Geoffrey W. Marcy, Ian J. Crossfield, Matthew J. Holman

    (Submitted on 12 Aug 2009)

    Abstract: We show that the exoplanet HAT-P-7b has an extremely tilted orbit, with a true angle of at least 86 degrees with respect to its parent star’s equatorial plane, and a strong possibility of retrograde motion. We also report evidence for a second planet in a more distant orbit.

    The evidence for the unparalleled orbit and the additional planet is based on precise observations of the star’s apparent radial velocity. The anomalous radial velocity due to rotation (the Rossiter-McLaughlin effect) was found to be a blueshift during the first half of the transit and a redshift during the second half, an inversion of the usual effect, implying that the angle between the sky-projected orbital and stellar angular momentum vectors is 182.5 +/- 9.4 deg.

    The second planet is implicated by excess radial-velocity variation of the host star over 2 yr. Possibly, the second planet tilted the orbit of the inner planet through a close encounter or the Kozai effect.

    Comments: Submitted to ApJ Letters [6 pages]

    Subjects: Earth and Planetary Astrophysics (astro-ph.EP); Solar and Stellar Astrophysics (astro-ph.SR)

    Cite as: arXiv:0908.1672v1 [astro-ph.EP]

    Submission history

    From: Joshua N. Winn [view email]

    [v1] Wed, 12 Aug 2009 10:44:57 GMT (67kb)


  • ljk August 13, 2009, 10:44

    First Evidence of a Retrograde Orbit of Transiting Exoplanet HAT-P-7b

    Authors: Norio Narita, Bun’ei Sato, Teruyuki Hirano, Motohide Tamura

    (Submitted on 12 Aug 2009)

    Abstract: We present the first evidence of a retrograde orbit of the transiting exoplanet HAT-P-7b. The discovery is based on a measurement of the Rossiter-McLaughlin effect with the Subaru HDS during a transit of HAT-P-7b, occured on UT 2008 May 30.

    Our model shows that the spin-orbit alignment angle of this planet is $\lambda = -132.6^{\circ} (+12.6^{\circ}, -21.5^{\circ})$. The existence of such a retrograde planet had been predicted by recent planetary migration models considering planet-planet scattering processes or the Kozai migration. Our finding provides an important milestone that supports such dynamic migration theories.

    Comments: 12 pages, 3 figures, 3 tables, PASJ submitted (on August 5)

    Subjects: Earth and Planetary Astrophysics (astro-ph.EP); Solar and Stellar Astrophysics (astro-ph.SR)

    Cite as: arXiv:0908.1673v1 [astro-ph.EP]

    Submission history

    From: Norio Narita [view email]

    [v1] Wed, 12 Aug 2009 10:45:15 GMT (284kb)


  • spaceman August 18, 2009, 4:32

    One concern I have is that based on graphs I’ve seen in published scientific papers it looks as though Kepler will need the full 4 years in order to securely detect an earth-sized planet at 1 A.U. around a solar type star. How can this be done if the mission is only slated for 3.5 years? And this line of thought leads me to another question: has anybody heard anything about the possibility of the Kepler mission being extended beyond 3.5 years so that it can bag smaller planets as well as planets in wider orbits?