Gliese 581g
Gliese 581g was a candidate exoplanet postulated to orbit within the Gliese 581 system, twenty light-years from Earth. It was discovered by the Lick–Carnegie Exoplanet Survey, and was the sixth planet claimed to orbit the star; however, its existence could not be confirmed by the European Southern Observatory / High Accuracy Radial Velocity Planet Searcher survey team, and was ultimately refuted. It was thought to be near the middle of the habitable zone of its star, meaning it could sustain liquid water—a necessity for all known life—on its surface, if there are favorable atmospheric conditions on the planet.
Gliese 581g was claimed to be detected by astronomers of the Lick–Carnegie Exoplanet Survey. The authors stated that data sets from both the High Resolution Echelle Spectrometer and HARPS were needed to sense the planet; however, the ESO/HARPS survey team could not confirm its existence. The planet remained unconfirmed as consensus for its existence could not be reached. Additional reanalysis only found evidence for four planets, but the discoverer, Steven S. Vogt, did not agree with those conclusions. In 2012, a reanalysis by Vogt supported its existence. A new study in 2014 concluded that it was a false positive, a conclusion which has been further confirmed by subsequent studies. The planet was thought to be tidally locked to its star. If the planet has a dense atmosphere, it may be able to circulate heat. The actual habitability of the planet depends on the composition of its surface and the atmosphere. It was thought to have temperatures around. By comparison, Earth has an average surface temperature of —while Mars has an average surface temperature of about. The planet was said by Vogt to have a "100%" chance of supporting life. The supposed detection of Gliese 581g was said to foreshadow what Vogt called "a second Age of Discovery".
History
Discovery
The planet's discovery was claimed in September 2010, to have been detected by astronomers in the Lick–Carnegie Exoplanet Survey, led by principal investigator Steven Vogt, professor of astronomy and astrophysics at the University of California, Santa Cruz, and co-investigator R. Paul Butler of the Carnegie Institution of Washington. The discovery was made using radial velocity measurements, combining 122 observations obtained over 11 years from the HIRES instrument of the W. M. Keck Observatory with 119 measurements obtained over 4.3 years from the HARPS instrument of the ESO 3.6 m Telescope at La Silla Observatory. In addition, brightness measurements of the star were confirmed with a robotic telescope from Tennessee State University.After subtracting the signals of the previously known Gliese 581 planets, b, c, d and e, the signals of two additional planets were apparent: a 445-day signal from a newly recognized outermost planet designated f, and the 37-day signal from Gliese 581g. The probability that the detection of the latter was spurious was estimated at only 2.7 in a million. The authors stated that while the 37-day signal is "clearly visible in the HIRES data set alone", "the HARPS data set alone is not able to reliably sense this planet" and concluded, "It is really necessary to combine both data sets to sense all these planets reliably". The Lick–Carnegie team explained the results of their research in a paper published in the Astrophysical Journal, which were also made available in preprint version on arXiv. Although not sanctioned by the IAU's naming conventions, Vogt's team informally referred to the planet as "Zarmina's World" after his wife, and in some cases simply as Zarmina.
During a press release announcing the discovery, Vogt et al. acknowledged that the "Gliese 581 system has a somewhat checkered history of habitable planet claims," as two previously discovered planets in the same system, Gliese 581c and d, were also regarded as potentially habitable, but later evaluated as being outside the conservatively defined habitable zone.
Nondetection in new HARPS data analysis
Two weeks after the announcement of the discovery of Gliese 581g, another team—led by Michael Mayor of the Geneva Observatory—reported that in a new analysis of 179 measurements taken by the HARPS spectrograph over 6.5 years, neither planet g nor planet f was detectable. An astronomer who works on HARPS data at the Geneva Observatory, Francesco Pepe, said in an email for an Astrobiology Magazine article republished on Space.com, "The reason for that is that, despite the extreme accuracy of the instrument and the many data points, the signal amplitude of this potential fifth planet is very low and basically at the level of the measurement noise". The Geneva team had also published their paper on arXiv, but it appeared to not have been accepted for publication.Vogt responded to the latest concerns by saying, "I am not overly surprised by this as these are very weak signals, and adding 60 points onto 119 does not necessarily translate to big gains in sensitivity." More recently, Vogt added, "I feel confident that we have accurately and honestly reported our uncertainties and done a thorough and responsible job extracting what information this data set has to offer. I feel confident that anyone independently analyzing this data set will come to the same conclusions."
Differences in the two groups' results may involve the planetary orbital characteristics assumed in calculations. According to Massachusetts Institute of Technology astronomer Sara Seager, Vogt postulated the planets around Gliese 581 had perfectly circular orbits whereas the Swiss group thought the orbits were more eccentric. This difference in approach may be the reason for the disagreement, according to Alan Boss. Butler remarked that with additional observations, "I would expect that on the time scale of a year or two this should be settled." Other astronomers also supported a deliberate evaluation: Seager stated, "We will have consensus at some point; I don't think we need to vote right now." Ray Jayawardhana noted, "Given the extremely interesting implications of such a discovery, it's important to have independent confirmation." Gliese 581g is listed as "retracted" in the Extrasolar Planets Encyclopaedia.
Further analyses of HIRES/HARPS data
In December 2010, a claimed methodological error was reported—by a group led by Rene Andrae of the Max Planck Institute for Astronomy—in the data analysis that led to the discovery of Gliese 581f and g.In 2011, another reanalysis—performed by a group led by Philip Gregory of the University of British Columbia—found no clear evidence for a fifth planetary signal in the combined HIRES/HARPS data set. The claim was made that the HARPS data provided only some evidence for 5 planet signals, while incorporation of both data sets actually degraded the evidence for more than four planets. Mikko Tuomi of the University of Hertfordshire performed a Bayesian reanalysis of the HARPS and HIRES data with the result that they "do not imply the conclusion that there are two additional companions orbiting GJ 581".
"I have studied in detail and do not agree with his conclusions," Steven Vogt said in reply, concerned that Gregory has considered the HIRES data as more uncertain. "The question of Gliese 581g's existence won't be settled definitively until researchers gather more high-precision radial velocity data", Vogt said. However, Vogt expects further analysis to strengthen the case for the planet.
By performing a number of statistical tests, Guillem Anglada-Escudé of the Carnegie Institute of Washington concluded that the existence of Gl 581g was well supported by the available data, despite the presence of a statistical degeneracy that derives from an alias of the first eccentric harmonic of another planet in the system. In a preprint posted to arXiv, Anglada-Escudé and Rebekah Dawson claimed that, "with the data we have, the most likely explanation is that this planet is still there."
2012 reanalysis of HARPS data
In July 2012, Vogt reanalyzed the 2011 data proposed by Forveille et al., noting that there were five objects. Planet g was orbiting around 0.13 AU with an orbital period of thirty-two days, placing it inside the habitable zone. Vogt concluded that the object had a minimum mass of 2.2 M and had a false positive probability of less than 4%. Vogt also said that they couldn't come to same conclusion as the Geneva team, without removing data points, "I don't know whether this omission was intentional or a mistake," he said, "I can only say that, if it was a mistake, they've been making that same mistake more than once now, not only in this paper, but in other papers as well." Vogt then said that the planet was there as long as all of the planets had circular orbits, and that the circular orbits work because “of dynamic stability, goodness-of-fit, and principle of parsimony."Further studies and refutation
Two studies in 2013 did not find evidence of Gliese 581g, only finding evidence for four—or three—planets in the system.A study in 2014—published in Science— led by postdoctoral researcher Paul Robertson concluded that Gliese 581d is "an artifact of stellar activity which, when incompletely corrected, causes the false detection of planet g." "They were very high value targets if they were real," Robertson said, "But unfortunately we found out that they weren't." It was pointed out—during a press release by Penn State University—that sunspots could sometimes masquerade as planetary signals. An additional study concluded that Gliese 581g's existence depends on Gliese 581d's eccentricity. The planet was later delisted from the Habitable Exoplanets Catalog, which is run by the University of Puerto Rico at Arecibo. Later, in October that year, Abel Mendez wrote—in a blog post characterizing "false starts" in exoplanet habitability— that the planet does not exist.
In 2015, a pair of researchers led by Guillem Anglada-Escudé of the University of London questioned the methodology of the 2014 study and suggested planet Gliese 581d really could exist, despite stellar variability, and the 2014 refutation of the existence of Gliese 581d and g was triggered by poor and inadequate analysis of the data, saying that the statistical method used by Robertson's team was "simply inadequate for identifying small planets like Gliese 581d", urging that the data be reanalyzed using a "more accurate model." However, this response did not make any claim for the existence of Gliese 581g, and was published along with a rebuttal by the team that published the 2014 refutation. Most further studies have confirmed the stellar, rather than planetary, origin of the signal corresponding to Gliese 581d, and consequently Gliese 581g, with one such study explicitly refuting g.