Sunday, January 13, 2013

Earth 2

Figure 1. Terrians, those mysterious inhabitants of the planet G889 in the 1990s TV series Earth2.

I just started a new job, and yesterday was my first day off since the first of the year. A few days ago I noticed some headlines about the discovery of an “Earth twin” – e.g., Most Earth-Like Alien Planet Possibly Found. Since similar headlines have been pretty frequent recently, I figured the subject matter would be, as usual, some familiar phantom planet, such as GJ 581 g or Tau Ceti e. So with all my work-related chaos, I didn’t pay much attention. I surmised that if this were a truly new and remarkable discovery, there’d be plenty of follow-up forthcoming in higher-impact press.

As it turns out, there’s been very little. Yet this really is novel stuff.

The object in question is a Kepler Object of Interest or KOI, meaning it’s a candidate transiting planet, not a confirmed detection. Its current designation is KOI-172.02. The fullest account I could find through a casual search was the NASA press release of January 7 announcing 461 new planet candidates. The authors, alas, were stingy with their details. They write that “four of the potential new planets are less than twice the size of Earth and orbit in their sun’s habitable zone,” but with a single exception they never mention which four KOIs those might be. The only member of the Goldilocks quartet to be identified by number was 172.02, which is described as “approximately 1.5 times the radius of Earth” and orbiting in a period of 242 days. Not much to go on, and certainly no hint as to how far away this putative system might be.

Statements by astronomers in various news stories offered a tad more detail. Natalie Batalha noted that the host star has spectral type G, and Mario Livio speculated, “Maybe there's no land life, but perhaps very clever dolphins.”

Yesterday I finally had time to dig around. I consulted the nifty NASA Exoplanet Archive, which maintains a list of all KOIs with the most recent data. From that archive I constructed my own short list of potentially Earthlike KOIs. Given the currency of the parameters, it’s more robust than similar lists that have appeared over the past two years. Table 1 presents my back alley distillation of the latest numbers:

Table 1. Current Kepler Objects of Interest (KOI) with potentially Earthlike equilibrium temperatures

Column 1 gives the KOI designation; column 2 the planet period in days; column 3 the semimajor axis in astronomical units (AU); column 4 the planet radius in Earth units (Rea); column 5 the planet equilibrium temperature (Teq); column 6 the stellar effective temperature (Teff), and column 7 the stellar radius in Solar units (Rsol). Temperatures are in Kelvin.

I used an arbitrary lower limit of 4000 K for stellar Teff to screen out the squirmiest M dwarfs, and my cut-offs for planet Teq were guided by published values for Venus, Earth, and Mars (not all necessarily reliable). Let me go on the record as a moaner and complainer about the scarcity of consistent Teq data for the Solar planets. I spent a lot of time Googling but turned up no reliable sources other than Wikipedia, which tells me the following: the mean surface temperature of Mars might be 210 K; the temperature at Venus’ cloud tops might be an oddly chilly 230 K (although as everybody agrees, the Venusian surface temperature is about 735 K); the Teq of Earth might be 255 K; and the mean surface temperature of Earth might be 288 K. Those numbers inspired my upper and lower limits of 210 and 300 K for Table 1.

To select planetary radii, I used the rule of thumb that any object larger than 2 Rea most likely has a hydrogen atmosphere, especially objects in the Super Earth mass range (which today I’m calling 2 Mea-9 Mea). Fortunately, Li Zeng and Dimitar Sasselov just circulated a preprint of their new article on the mass-radius relationship for solid planets, which updates the scales published last year by various Kepler scientists (Zeng & Sasselov 2013). According to their new chart, a planet of 1.4 Rea (e.g., KOI-3010.01) is likely to have a mass between 2 and 5 Mea, with the most lightweight possibility involving a layered composition of 25% rock/metal and 75% water, and the most heavyweight involving a collisionally stripped core composed mainly of metal. But since nobody knows if such severe collisional stripping ever happens outside numerical simulations (Marcus et al. 2010), a more likely upper mass limit would be 3.5 Mea, corresponding to the same rock/metal structure as Earth. Thus, 1.4 Rea is a suitable radius for anybody in search of singing cetaceans or calculating lizards.

By the time we get to 1.9 Rea (e.g., KOI-2762.01), outcomes become much fuzzier. That girth might represent an all-water world with a mass of 3 Rea, a 50-50 water plus rock/metal planet with a mass of 4.5 Mea, or a truly terrestrial planet with a rock/metal composition like Earth and a chunky mass of 10-12 Mea. Still worse, for planets on temperate orbits (i.e., all the KOIs in Table 1), it could also include an ice/rock or rock/metal planet with a hydrogen atmosphere and a mass between 3 and 8 Mea. As a result, I refuse to get excited by any candidate bigger than 1.7 Rea.

Now I’m more curious than ever to know which three KOIs apart from 172.02 were nominated by the Kepler folks for the title of “potentially possibly habitable.” Several of the objects in Table 1 have been noted in passing in previous mission publications (Kaltenegger & Sasselov 2011, Batalha et al. 2012), but without any fanfare whatsoever.

The current poster child, KOI-172.02, has the longest orbital period and the widest semimajor axis of any object on my list, and not coincidentally it’s the one that’s least likely to be tidally locked. That seems to be the main reason why it was singled out. 

According to Franck Selsis and colleagues, tidal locking can be predicted for planets on circular orbits within a radius of 0.6 AU around a star of 0.9 Msol; within 0.5 AU around a star of 0.8-0.6 Msol; and within 0.45 AU around a star of 0.5 Msol (Selsis et al. 2007). A moderate orbital eccentricity would be the only protection against locking into a permanent dayside and a permanent nightside, since an elliptical orbit like Mercury’s, with an eccentricity of 0.2, is more likely to result in spin-orbit resonance than in a 1:1 lock of rotation and revolution. Since most of the stars in Table 1 have Teq between 4000 and 5000 K, and since the main sequence mass range corresponding to that temperature range is about 0.6-0.85 Msol, all the associated KOIs will have experienced some form of spin-orbit synchronization. Not that 1:1 tidal locking or a higher-order spin-orbit resonance (e.g., 2:1, like Mercury) necessarily poses any major threat to habitability – it just wouldn’t be Earthlike.

Other data on 172.02 make it less attractive as a candidate Earth twin. The host star's Teff is actually 100 K higher than the temperature of our Sun, while the planet's semimajor axis is only a little larger than that of Venus. The result is a planet substantially hotter than Earth. It would take a very special mix of physical conditions to create a habitable environment out of such an object, even in the best-case scenarios: a water world of 2 Mea or a rocky planet of 5 Mea, either of which is consistent with a radius of 1.54 Rea. My favorites in Table 1 are 701.03 and 3010.01. Both are cooler than 172.02, and as we saw above, the radius of 3010.01 is consistent with a biophilic composition.  Both planets would be tidally synchronized with their host stars, but that's not a deal breaker.

The bottom line is that we shouldn’t break out the good champagne for any of these KOIs. None has been confirmed even as a planetary candidate, and all associated data are subject to revision. Some of these interesting objects may turn out to be false positives; others may have their parameters redefined in ways that would exclude them from consideration of habitability. Regardless, we can still find solace, however small, in the fact that astronomers have finally identified a handful of intriguing and so far viable contenders for the title of Earth 2. Confirmation of one of them, or of some still unpublicized KOI or candidate, might come sometime in 2013.

Figure 2. Another Earth. Has anyone on this Earth actually seen that movie?

And yet – in the current Zeitgeist, is it already too late to celebrate another Earth? Science headlines (as well as some publicity-seeking astronomers) have cried “Earth twin!” for so long now that their audience has gotten jaded. When the NASA news on KOI-172.02 perfused cyberspace this past Thursday, a blogger with the Washington Post, Alexandra Petri, responded tartly with a post entitled No More New Planets, Please. Most of the 61 accumulated comments shared Ms. Petri’s lack of enthusiasm, as the online discussion quickly devolved into jokes and bickering about politics, religion, and the end of the world. One early commenter identified as “felpin” made me laugh out loud: “I turned off the science news on my Google news because [of] the endless parade of planets .... almost as boring as the Kardashian news.”

So what’s the antidote for Kardashian exoplanets?

Less hype and more context in astronomical journalism. More cautious choice of sound bites in press releases and media conferences. Richer presentation of detail in all media. More education of the curious about the mysteries and wonders of our planet-haunted universe.

In short, this blog.

The morning after I uploaded this post I saw a brand-new preprint by Eric Gaidos on the very same topic - “Candidate Planets in the Habitable Zones of Kepler Stars.” I was perplexed by the fact that for candidates around 2 Rea or smaller, his list and my list show minimal overlap. Even for overlapping systems, the parameters he reported sometimes differed from the ones shown on the NASA site. I gather that Gaidos used the brand-new data release but conducted his own analyses to recalibrate many values. For the record, the systems that we identified in common are KOI-701.03, KOI-2418.01, and KOI-3010.01, but Gaidos presents a much larger radius for the first of these (2.04 Rea), which would have made me disqualify it. He also reports an age that is only one-eighth of what NASA says, and a stellar effective temperature that is much hotter. Notably, Gaidos doesn't include NASA's poster child, KOI-172.02, and generally his list seems to have few KOIs with only 3 digits preceding the decimal point. Puzzling.

My takeaway from this exercise is that when you're trying to understand potential Earth twins located hundreds of parsecs away, you can be pretty darn sure that most or all of your data are unreliable. Gaidos notes that different analyses of the Teff of Kepler stars may disagree by 200 K, which is more than enough to push planetary Teq out of the habitable range. So, for now, most everybody seems to be engaging in elaborate hand-waving.


Batalha N, Rowe JF, Bryson ST, Barclay T, Burke CJ, Caldwell DA, and 66 others. (2012) Planetary candidates observed by Kepler III: Analysis of the first 16 months of data. In press. Abstract:
Gaidos E. (2013) Candidate planets in the habitable zones of Kepler stars. Abstract: 
Harrington JD, Johnson M. (2013) NASA’s Kepler mission discovers 461 new planet candidates. Release 13-008. Monday, January 7, 2013.
Kaltenegger L, Sasselov D. (2011) Exploring the habitable zone for Kepler planetary candidates. Astrophysical Journal Letters 736, L25
Marcus RA, Sasselov D, Lars Hernquist, and Sarah T. Stewart. (2010) Minimum radii of Super-Earths: Constraints from giant impacts. Astrophysical Journal Letters 712, L73–L76.
Selsis F, J. F.Kasting, B. Levrard, J. Paillet, I. Ribas, and X. Delfosse. (2007) Habitable planets around the star Gliese 581? Astronomy & Astrophysics 476, 1373-1387.
Zeng L, Sasselov D. (2013) A detailed model grid for solid planets from 0.1 through 100 Earth masses. In press. Abstract:

Tuesday, January 1, 2013

Year of the Signal

Figure 1. “Raw” (top) versus “non-noisy” (bottom) periodograms for radial velocity data from Alpha Centauri B. Analyses by Xavier Dumusque and colleagues led to the announcement of Alpha Centauri Bb, a hot, rocky exoplanet just 4 light years away. (Based on Figure S10 from Dumusque et al. 2012.)

Compared to the excitement of 2011, the offerings of the exoplanetary year just ended seem a bit sedate. Granted, the Kepler Mission continues to report new worlds, and its vast and growing dataset will surely feed astronomical research for years to come. But the pace of breakthroughs has slackened. Kepler candidates in 2012 were typically more of the same (a few more circumbinaries, many more hot, low-mass planets), while the mission itself lost a bit of luster through a combination of hardware failure, higher-than-anticipated rate of false positives (Cameron 2012), and serious scarcity of Earth-size planets on Earth-like orbits. Ground-based programs continued to yield notable results, but overall, exoplanets bagged fewer headlines than phenomena in adjacent disciplines.

In particle physics, the Higgs boson underwent a media apotheosis as the God Particle, with articles about its discovery leading most science news for several weeks over the summer. In planetary astronomy, space robots found ice at Mercury’s north pole (Paige et al. 2012) and a dry riverbed on Mars, making both places feel ever so slightly more homey. The best that the exoplanetary headlines could muster were more of those deeply dubious “habitable Super Earths” (which generally turn out to be Mini Neptunes, when they turn out at all) and a few days of sci-fi speculation on rogue planets hurtling sunless across the Galaxy.

Figure 2. Artist’s view of Alpha Centauri Bb. Credit: University of Puerto Rico Arecibo

Nevertheless, this past year saw provocative and likely historic developments in exoplanetary science:
  • A team of astronomers using the HARPS spectrograph announced an Earth-mass planet orbiting Alpha Centauri B (the K-type component of the very nearest star system) in a period of just 3.24 days (Dumusque et al. 2012). Detecting such a faint radial velocity signal pushed instrumental sensitivity to the limit.
  • The Kepler team reported the first system of multiple planets in a circumbinary configuration – Kepler-47, where a Sun-like star and an M dwarf share a circular orbit of 7.5 days, and two Uranus-like planets (gas dwarfs) orbit them on periods of 50 and 303 days (Orosz et al. 2012).
  • An international team using the Herschel Space Observatory reported an extensive debris disk – much larger than our own Kuiper Belt – around the nearby M dwarf GJ 581, better known as the host of a multiplanet system with a Super Earth or Mini Neptune orbiting near the habitable zone (Lestrade et al. 2012). This remarkable discovery, which has gone largely unnoticed by the science journalists, adds to our exiguous knowledge of debris disks around M dwarf stars. Only one other red dwarf system (AU Microscopii) supports a dust disk that has been spatially resolved. Just as important, the newly revealed parameters of the GJ 581 disk bring us closer to understanding the overall architecture of a complex and perennially fascinating exoplanetary system. 
  • Several candidate low-mass planets, some orbiting in their systems’ habitable zones, were proposed by astronomers who applied sophisticated statistical methods to extract faint signals from publicly available radial velocity data on nearby stars, all but one of which are already known to host planets. The list of revisited systems includes GJ 667C, GJ 676A, HD 10180, HD 40307, and – most notably – Tau Ceti, the nearest Solar analog (i.e., single G-type star). Most of these candidate exoplanets remain controversial.
  • The orbiting CoRoT spacecraft, launched in late 2006 to search for transiting planets, suffered an apparently fatal computer error. I haven’t yet seen an official post mortem, but from my back alley perspective, this five-year mission has not met expectations. Despite much pre-launch optimism (Leger et al. 2004, Garrido & Deeg 2006), the CoRoT team so far has reported a single transiting Super Earth (CoRoT-7b, a prototypical Hellworld with a “year” that lasts 20 hours), a single transiting Warm Jupiter (CoRoT-9b, with a period of 95 days), and dozens of transiting Hot Jupiters. Early CoRoT predictions, by contrast, envisioned the detection of “between 10 and 40 terrestrial planets . . . some of them in the ‘habitable zone’ ” (Boisnard & Auvergne 2006).
  • The homepage of the indispensable Extrasolar Planets Encyclopaedia underwent its first major redesign since its launch in 1995. Despite persistent bugs, the site now looks at home in the 21st century.
I’ve already blogged about the two top discoveries, and I plan to discuss the GJ 581 system in an upcoming post about the immediate Solar neighborhood. So for now, I’ll just take a look at the fourth phenomenon – planet candidates at the very limit of detection, and the potential reality thereof – and then conclude with a few remarks about the notion of confirmed vs. controversial discoveries.

planetary conflict in Ursa Major and Libra

Controversy is an inevitable outcome of scientific endeavor. A classic example in exoplanetary astronomy involves 47 Ursae Majoris (47 UMa), a nearby G-type star that became one of the first confirmed exoplanetary hosts. In 1996, Butler and Marcy announced a single gas giant, 47 UMa b, orbiting at a semimajor axis of 2 AU (Butler & Marcy 1996). The planet was included in the census of the Extrasolar Planets Encyclopaedia (EPE), and within a few years its orbital parameters were independently confirmed by other studies (Naef et al. 2004, Wittenmyer et al. 2007). In 2002 the original discovery team, with three new collaborators, analyzed a longer data series that indicated the presence of a second gas giant on a wider orbit (Fischer et al. 2002). The new planet, christened 47 UMa c, was estimated to have a period of at least 7 years, corresponding to the largest semimajor axis then known (3.73 AU).

Unfortunately, other search programs could not detect planet c (Naef et al. 2004, Rivera & Haghighipour 2007, Wittenmyer et al. 2007, 2009). EPE continued to list it as a confirmed exoplanet, despite the controversy, and some years later, a single member of the 2002 team collaborated with Philip Gregory on a new study (Gregory & Fischer 2010). By then the radial velocity observations of 47 UMa accumulated by the Lick and McDonald Observatories covered 22 years. After conducting “a Bayesian adaptive hybrid Markov chain Monte Carlo analysis” of the data series, Gregory & Fischer felt confident in confirming planet c, now with a slightly shorter period of 6.55 years and a slightly smaller semimajor axis of 3.6 AU. Their analysis also detected an uncertain signal potentially corresponding to a third gas giant on a still wider orbit, with a semimajor axis of at least 11.6 AU (compare Saturn at 9.5 AU, with a period just under 30 years), They cautioned, however, that “the longest period orbital parameters are still not well defined.” Nevertheless, all three planets continue to be listed by EPE, and no further challenges have emerged. For the time being, then, 47 UMa has three gas giants.

Another controversy has not subsided so quietly. In 2010, the same year in which Gregory & Fischer published their results on 47 UMa, a team led by Steven Vogt announced the discovery of a new Super Earth in the habitable zone of GJ 581 (Vogt et al. 2010). This nearby M dwarf was already known to host four low-mass planets – i.e., Mini Neptunes or Super Earths – which were originally reported by the HARPS team. The new planet candidate, GJ 581 g, was nicknamed “Zarmina’s World” after the lead astronomer’s wife, and received instant media acclaim as the first habitable exoplanet. Alas, even before the official publication of the discovery paper, astronomers with the HARPS survey sharply challenged the reality of Zarmina’s World. As Francesco Pepe argued, “The signal amplitude of this potential fifth planet is very low and basically at the level of the measurement noise” (Mullen 2010). He considered it likely that the purported “signal” was “just produced ‘by chance’ out of the noise.”

This signal vs. noise controversy continues into the present. Because it involves a “habitable Super Earth” instead of a potential Saturn analog (without the rings), it has attracted much more attention than the challenges to 47 UMa c. Although Guillem Anglada-Escude quickly defended the findings of Vogt and colleagues in an unpublished study (Anglada-Escude 2010), prevailing opinion has gone against them (Gregory 2011, Tuomi 2011, Tuomi & Jenkins 2012). In fact, a recent study challenges the existence not only of GJ 581 g but also of GJ 581 d (Baluev 2012), which has otherwise retained its crown as the most likely “habitable Super Earth.” In the case of Zarmina, I’ll go on the record in support of the skeptics. For me the most persuasive evidence is an as-yet unpublished study by the HARPS team featuring new radial velocity data that, in their analysis, contain no evidence for more than four planets around GJ 581 (Forveille et al. 2011). Yet despite so many naysayers, the ghost of planet g has found no rest, as evidenced by its prominent appearance in a poster just published by the University of Puerto Rico:

Figure 3. Recent poster by the University of Puerto Rico Arecibo

phantom planets in the reality continuum

The chronicles of 47 UMa and GJ 581 set the stage for the still more remarkable detections proposed in 2012. We have already met the astronomers – Philip Gregory, Guillem Anglada-Escude, Mikko Tuomi, and their colleagues – as well as their methods, among which Bayesian inference looms large. A handful of systems are involved, all but one quite nearby. I’ll consider the most representative case first.

GJ 667C  A red dwarf with spectral type M1.5 and mass 0.31 Solar, GJ 667C is the third member of a triple star system located just 7 parsecs (23 light years) away (Anglada-Escude et al. 2012 say 6.8 pc, Gregory 2012 says 7.23 pc). Members of the HARPS team have announced two low-mass planets: GJ 667C b with a period of 7 days and a minimum mass of 5.46 Mea, and GJ 667C c with a period of 28 days and a minimum mass of 4.25 Mea (Bonfils et al. 2011, Delfosse et al. 2012). Those numbers make GJ 667C c the best candidate yet for the media status of “habitable Super Earth,” although in my judgment the minimum mass is probably high enough to exclude it from consideration.

Two additional studies by astronomers outside the HARPS team followed soon afterward, one by Gregory and the other led by Anglada-Escude and including 16 additional collaborators. Gregory reanalyzed the data used by Delfosse et al.; Anglada-Escude reanalyzed the data of Bonfils et al., adding observations by members of his own group. Both studies agreed on the reality of the 7-day and 28-day planets, with Gregory suggesting a mass of 4.8 Mea for planet c and Anglada-Escude suggesting 4.54 Mea. Thereafter their conclusions diverged quite dramatically. Anglada-Escude’ group noted a potential signal with a period of 75 days, corresponding to an object similar in mass to planets b and c; nevertheless, they recommended “due caution” in considering a planetary interpretation. Gregory, on the other hand, found evidence for three more planets, two of them also in the habitable zone, and all in the mass range of Super Earths. Gregory’s five-planet solution includes planets b (7.2 days, 5.4 Mea), c (28.1 days, 4.8 Mea), d (30.8 days, 3.1 Mea), e (38.9 days, 2.4 Mea), and f (91.3 days, 5.4 Mea). This is a remarkably crowded system of Super Earths, but nowhere do we see an object with a period near 75 days.

It’s impossible for anyone without the requisite education and experience to assess the likelihood of either reanalysis. If Gregory’s findings are accurate, we’d best launch that interstellar probe to GJ 667 C without delay. Yet EPE currently lists only two planets in this system; candidate d is tagged “controversial” and candidates e and f are not acknowledged. To me that suggests a skeptical position. Even Gregory does not claim that his numerical model corresponds to a working dynamical architecture. As he puts it, “N-body simulations are required to determine which of these signals are consistent with a stable planetary system.” That leaves me wondering whether we are contemplating pretty planets or pretty numbers.

GJ 676A  Although classified M0, this star is as robust as many K dwarfs (0.71 Solar masses). It has a smaller companion star of type M3 at a separation of at least 800 AU. In 2011, Thierry Forveille and colleagues reported a surprisingly massive gas giant (4.9 Mjup) orbiting star A in a period of almost three years (Forveille et al. 2011). While noting that the object might actually have a longer period and a higher mass, making it a brown dwarf or a very dim red dwarf, they were more confident of a planetary origin for the radial velocity signals. EPE duly added this object to the catalog of confirmed exoplanets, where it appeared as recently as May 27, 2012.

Over the next few months, Anglada-Escude & Tuomi collaborated on successive drafts of a study in which they applied Bayesian methods to the published HARPS radial velocities of GJ 676A. Not only did they confirm the giant orbiting at about 1.8 AU; they also detected two low-mass planets on short-period orbits – a potential Super Earth with a minimum mass of 4.4 Mea orbiting in 3.6 days, and a Warm Uranus with a minimum mass of 11.5 Mea orbiting in 35.4 days (Anglada-Escude & Tuomi 2012). This architecture, if real, would make GJ 676A the only known system beside HD 10180 and our own to host at least two low-mass planets inside the ice line and at least one gas giant outside it.

But how likely is it to be real? The Exoplanet Orbit Database, produced by Jason Wright and Geoff Marcy, lists only planet b for GJ 676A, while EPE, for reasons unstated, has completely expunged this star from its catalogs! Evidently the purge happened in June or July, and it was as total as anything envisioned (or unenvisioned) by George Orwell. Not even the freak tent of “unconfirmed, controversial, or retracted planets” has room for this oversized dwarf. GJ 676 Ab is, at least for now, an unplanet. [UPDATE: As of January 2, 2013, planet b is back in the freak tent, tagged by EPE as a "controversial" object. The other candidates remain unplanets.] 

HD 10180  This yellow star, about 39 parsecs (128 light years) away, was identified in 2010 as the host of at least five low-mass planets orbiting within 1.42 AU, with a small gas giant at about 3.4 AU and a seventh suspected planet with a mass similar to Earth’s and an orbital period shorter than two days – i.e., a Hellworld somewhat like CoRoT-7b (Lovis et al. 2011). Until the announcement of the Kepler multiplanet systems, this was the most packed system architecture known. Yet even amid such an abundance of companions, not a single one is both cool enough and lightweight enough to qualify as a potentially habitable rocky planet.

Enter Mikko Tuomi, who as before used Bayesian methods to reanalyze the published HARPS radial velocities for HD 10180. His reanalysis confirms all seven planets suggested by Lovis and colleagues and adds two more low-mass objects: planet i, which orbits between planets c and d in a period of about 10 days, and planet j, which orbits between planets d and e in a period of about 68 days. Both candidates have minimum masses below 5.5 Mea. Neither has even a remote chance of habitability.

As before, Tuomi’s candidates remain candidates; EPE continues to list only planets c through g in the main catalog, while planets b, i, and j are “unconfirmed, controversial.”

Figure 4. Highly imaginative artist’s view of the hypothetical planet HD 40307 g, along with two better constrained siblings. Credit: J.T. Pinfield

HD 40307 Back in 2008, Michel Mayor and colleagues reported a system of three Super Earths orbiting this amber star of spectral type K2, which is located 12.8 parsecs (42 light years) away. Confined within a semimajor axis of just 0.13 AU, all three must support environments the likes of which even Dante never imagined. In 2012, Tuomi and Anglada-Escude led a team of eight astronomers (including such noted planet hunters as R. Paul Butler, Eugenio Rivera, and Steven Vogt) in yet another Bayesian reanalysis of published HARPS data. They recovered three additional signals: planet e, with a period of 35 days and a minimum mass of 3.5 Mea; planet f, with a period of 52 days and a minimum mass of 5.2 Mea; and planet g, with a period of 198 days and a minimum mass of 7.1 Mea. Since HD 40307 is substantially less luminous than our Sun, the orbit of planet g occupies the system habitable zone, a feature that guaranteed a bumper crop of headlines in the weeks before Thanksgiving.

Remarkably, EPE accepted these results and added all three Bayesian planets to the official catalog. We can only wonder why this reanalysis seemed more persuasive than others by members of the same team. HD 40307 now officially qualifies as a system of six planets, all with minimum masses under 10 Mea and orbital periods shorter than a Venusian year. And as we saw in Figure 3 above, HD 40307 g has joined the happy club of possibly potentially hypothetically habitable Super Earths.

Ah, fluffy white clouds, deep blue seas, rugged green continents! Figure 4, an artist’s view of this new world, gives us all those delightfully familiar options. Nevertheless, as I’ve taken pains to establish in earlier posts, a cool planet with a minimum mass as high as 7 Mea is extremely unlikely to resemble Earth. Such a massive object is likely to retain a deep hydrogen atmosphere, and it is doubtful that it could sustain either plate tectonics or oceans (which according to some theorists are mutually constitutive phenomena – you can’t have one without the other).

But I say let the dreamers go on dreaming if they won’t wake up.

Figure 5. Artist's view of the proposed Tau Ceti system. Credit: J.T. Pinfield

Tau Ceti  Now we reach the biggest dream on our end-of-the-year agenda. Just in time for Christmas, Tuomi and crew (including many of his collaborators from the study of HD 40307) reported results on a star with no previously suspected planets: Tau Ceti, the nearest Solar analog. Located just 3.65 parsecs (12 light years) away in the constellation of Cetus the Whale, this star was always high on those lists of “nearby stars that might have solar systems” that circulated during the remote era before radial velocity and transit searches (i.e., my childhood).

Along with Alpha Centauri and Epsilon Eridani, Tau Ceti was targeted by all the earliest search programs. Null results continued even as other nearby systems – Gamma Cephei, Epsilon Eridani, 82 Eridani, 61 Virginis – began to yield up their secrets. Eventually, infrared observations revealed the presence of a two debris belts around Tau Ceti: one cool and massive and the other warmer and more tenuous (Greaves et al. 2004, Di Folco et al. 2007). Since debris is widely regarded as a signpost of planet formation, it seemed unlikely that Tau Ceti could harbor asteroids and comets but no planets. Nevertheless, the Great Whale refused to sing.

By the time Tuomi and colleagues began their Bayesian reanalysis, they were able to use extensive, high-quality datasets from three different search programs, including HARPS, with solid coverage over 13 years. Using sophisticated and highly complex modeling, they recovered signals corresponding to a system of five planets (b-f) that range in minimum mass from 2 to 6.6 Mea, and in period from 14 to 642 days.

Most notable is candidate planet e, with a mass of 4.3 Mea and a period of 168 days. Given the host star’s spectral type and effective temperature (G8, 5344 K), this orbit occupies the circumstellar habitable zone. Thus, provided it exists, Tau Ceti e meets most definitions of a potentially habitable Super Earth. If the results of Tuomi and colleagues had been firm enough to count as a detection, Tau Ceti would have topped my list of Notable Exoplanet News of 2012, hands down.

I’m hardly alone in my eagerness to find planets around Tau Ceti. Astronomer Greg Laughlin, never one to suffer fools, commented favorably on Tuomi and colleagues’ study in a New Year’s Eve blog post, while EPE included their findings as one of only two notable “News” items for all of 2012 (the other one was neither Alpha Centauri Bb nor Kepler-47, but “Hint for a transiting extended atmosphere on 55 Cnc b”).

Yet no one, not even Tuomi and his collaborators, is pressing this claim too strongly. EPE lists the “Tau Ceti Five” as “unconfirmed, controversial,” and Tuomi’s study concludes, “these issues remain merely speculative until the planetary origin of the signals can be verified by an independent detection.” Once again we contemplate pretty numbers instead of pretty planets.

where is the referee?

The International Astronomical Union (IAU) is the only organization officially vested with the authority to decide whether an object is a planet, and what to call it if it is (or isn't). In 2006, the IAU memorably intervened in the debate over the nature of Pluto. As soon as it classified that distant icebox as a dwarf planet and decreed that our Solar System has only eight full-size companions, textbooks were revised, museum exhibits were redesigned, and a surprising number of ordinary people complained.  Nevertheless, the decision stands, and virtually all astronomers are satisfied.

So far the IAU has made no decisions about exoplanets. That's part of the reason why we have no universally accepted definition of a Super Earth and no consensus on which planetary candidates are real and which are imaginary: there's no exoplanetary referee in deep space. EPE has therefore become the de facto decider, given its frequent citation in exoplanet studies and the wide acceptance of its published extrasolar census.

As we have seen, however, EPE isn't consistent in its decisions, and still worse, the site typically provides no alerts when it deletes or reclassifies an object, nor any explanation for such changes. In the ensuing vacuum, so to speak, we get publications like the poster shown in Figure 3. The fact that the authors have added asterisks to two of the names (GJ 581 g and HD 40307 g) inspires lots of lulz at my coordinates. On the one hand, study after study has rejected the existence of GJ 581 g, while on the other, HD 40307 g appears as a bona fide planet in EPE. So why do both get the same asterisk?

Still worse, we see this funny object designated Gliese 163 c, with no asterisk or scare quotes to tip us off to the fact that it has yet to be published, even in a preprint. Oh wait, it's listed in EPE anyway -- but as a small gas giant with a minimum mass of 0.226 Mjup (72 Mea). That's not much less than the mass of Saturn. One can only shrug.

Meanwhile, I'm waiting for that extrasolar umpire to turn up with an unequivocal call. I wonder how many more light years it/they/she/he still has to traverse.


Anglada-Escudé G. (2010) Aliases of the first eccentric harmonic: Is GJ 581g a genuine planet candidate? (First version, unpublished; a second unpublished draft with the collaboration of Rebekah Dawson is available at
Anglada-Escudé G, Arriagada P, Vogt SS, Rivera EJ, Butler RP, Crane JD, and 11 others. (2012) A planetary system around the nearby M dwarf GJ 667 C with at least one super-Earth in its habitable zone. Astrophysical Journal Letters 751, L16. Abstract:
Anglada-Escudé G, Tuomi M. (2012) A planetary system with gas giants and super-Earths around the nearby M dwarf GJ 676 A. Optimizing data analysis techniques for the detection of multi-planetary systems. Astronomy & Astrophysics 548, A58. Abstract:
Baluev R. (2012) The impact of red noise in radial velocity planet searches: Only three planets orbiting GJ 581? Monthly Notices of the Royal Astronomical Society, in press.
Boisnard L, Auvergne M. (2006) COROT in brief. Proceedings of The CoRoT Mission Pre-Launch Status - Stellar Seismology and Planet Finding (ESA SP-1306). Editors: M. Fridlund, A. Baglin, J. Lochard, L. Conroy. ISBN 92-9092-465-9
Cameron AC. (2012) Astrophysical false positives. Nature 6 December 2012 (492:48-50). 
Di Folco E, Absil O, Augereau JC, Merand A, Coude du Foresto V, Thevenin F, Defrere D, Kervella P, ten Brummelaar TA, McAlister HA, Ridgway ST, Sturmann J, Sturmann L, Turner NH. (2007) A near-infrared interferometric survey of debris disk stars I. Probing the hot dust content around Epsilon Eridani and Tau Ceti with CHARA/FLUOR. Astronomy & Astrophysics 475, 243-250.
Dumusque X, Pepe F, Lovis C, Segransan D, Sahlmann J, Benz W, Bouchy F, Mayor M, Queloz D, Santos N, Udry S. (2012) An Earth-mass planet orbiting Alpha Centauri B. With supplemental material. Nature 17 October 2012.
Delfosse X, Bonfils X, Forveille T, Udry S, Mayor M, F. Bouchy F, et al. (2012) The HARPS search for southern extra-solar planets. XXXV. Super-Earths around the M-dwarf neighbors Gl 433 and Gl 667C. Astronomy & Astrophysics in press.
Forveille T, Bonfils X, Lo Curto G, Delfosse X, Udry S, Bouchy F, et al. (2011) The HARPS search for southern extra-solar planets. XXVI. Two giant planets around M0 dwarfs. Astronomy & Astrophysics 526, A141. Abstract:
Forveille T, Bonfils X, Delfosse X, et al. (2011) The HARPS search for southern extra-solar planets XXII. Only 4 planets in the Gl 581 system. Astronomy & Astrophysics in press. Abstract:
Garrido R, Deeg HJ. (2006) Exoplanet detection and the CoRoT Mission. Real Sociedad Española de Física. 
Greaves JS, Wyatt MC, Holland WS, Dent WRF. (2004) The debris disc around Tau Ceti: a massive analogue to the Kuiper Belt. Monthly Notices of the Royal Astronomical Society 351, L54-L58. Abstract.
Gregory PC, Fischer DA. (2010) A Bayesian periodogram finds evidence for three planets in 47 Ursae Majoris. Monthly Notices of the Royal Astronomical Society 403, 731-747.
Gregory PC. (2011) Bayesian re-analysis of the Gliese 581 exoplanetary system. Monthly Notices of the Royal Astronomical Society 415, 2523-2545. Abstract:
Gregory PC. (2012) Additional Keplerian signals in the HARPS data for Gliese 667C from a Bayesian re-analysis. Monthly Notices of the Royal Astronomical Society in press. Abstract:
Leger A, Baglin A, Barge P, Borde P, Defay C, Deleuil M, Rouan D, Schneider J, Vuillemin A. (2004) Detecting Earth-Uranus Class Planets with the Space Mission COROT. In Penny A, Artymowicz P, Lagrange A-M, Russell S. Planetary Systems in the Universe: Proceedings of the IAU Symposium No. 202.
Lestrade J-F, Matthews BC, Sibthorpe B, Kennedy GM, Wyatt MC, Bryden G, Greaves JS, Thilliez E, Moro-Martın A, and 11 others. (2012) A DEBRIS disk around the planet hosting M-star GJ 581 spatially resolved with Herschel. Astronomy & Astrophysics 548, A86.
Lovis C, Segransan D, Mayor M, Udry S, Benz W, Bertaux J-L, et al. The HARPS search for southern extra-solar planets XXVIII. Up to seven planets orbiting HD 10180: probing the architecture of low-mass planetary systems (2011) Astronomy & Astrophysics 528, A112.
Mullen L. (2010) Doubt Cast on Existence of Habitable Alien World. Astrobiology Magazine, 12 October 2010. Linked article.
Naef D, Mayor M, Beuzit JL, Perrier C, Queloz C, Sivan JP, Udry S. (2004) The ELODIE survey for northern extra-solar planets III. Three planetary candidates detected with ELODIE. Astronomy & Astrophysics 414, 351-359.
Orosz JA, Welsh WF, Carter JA, Fabrycky DC, Cochran WD, Endl M, and 33 others (2012) Kepler-47: A Transiting Circumbinary Multiplanet System. Science 28 August 2012.
Paige DA, Siegler MA, Harmon JK, et al. (2012) Thermal stability of volatiles in the north polar region of Mercury. Science 29 November 2012.
Rivera E, Haghighipour N. (2007) On the stability of test particles in extrasolar multiple planet systems. Monthly Notices of the Royal Astronomical Society 374,: 599–613.
Tuomi M. (2011) Bayesian re-analysis of the radial velocities of Gliese 581. Evidence in favour of only four planetary companions. Astronomy & Astrophysics 528, L5. Abstract:
Tuomi M. (2012) Evidence for nine planets in the HD 10180 system. Astronomy & Astrophysics 543, A52. Abstract:
Tuomi M, Anglada-Escudé G, Gerlach E, Jones HRA, Reiners A, Rivera EJ, Vogt SS, Butler RP (2012) Habitable-zone super-Earth candidate in a six-planet system around the K2.5V star HD 40307. Astronomy & Astrophysics in press. Abstract:
Tuomi M, Jenkins JS. (2012) Counting the number of planets around GJ 581. False positive rate of Bayesian signal detection methods. Astronomy & Astrophysics in press. Abstract:  
Tuomi M, Jones HRA, Jenkins JS, Tinney CG, Butler RP, Vogt SS, Barnes JR, Wittenmyer RA, and 7 others. (2012) Signals embedded in the radial velocity noise. Periodic variations in the Tau Ceti velocities. Astronomy & Astrophysics in press. Abstract:
Vogt SS, Butler RP, Rivera EJ, Haghighipour N, Henry GW, Williamson MH. (2010) The Lick-Carnegie exoplanet survey: A 3.1 MEA planet in the habitable zone of the nearby M3V star Gliese 581. Astrophysical Journal 723, 954-965.
Wittenmyer RA, Endl M, Cochran WD. (2007) Long-period objects in the extrasolar planetary systems 47 UMa and 14 Her. Astrophysical Journal 654, 625-632.
Wittenmyer RA, Endl M, Cochran WD, Levison HF, Henry GW. (2009) A search for multi-planet systems using the Hobby–Eberly telescope. Astrophysical Journal Supplement Series 182, 97-119.