The discovery of planet Gliese 581 g, an exoplanet just 3 times the mass of our Earth and located in its host star’s Habitable Zone, was one of the biggest science headlines of the year. The news broke, typically, somewhere between my observing proposal deadline and box number 15: “all the crap that didn’t fit into boxes 1-14″, and by the time I’d read the story the internets was awash with all the details already.
As it turns out, I’m glad I haven’t written about it yet, as the story just got a little bit more interesting. With the headline still hot on the media’s most-read lists, astronomers from the Geneva exoplanet group have this week cast doubt on the discovery. Using data of GJ581 from the HARPS spectrograph at the ESO 3.6-m telescope at La Silla, the instrument par excellence for planet hunting by radial velocities, they haven’t been able to confirm or reject the Californians’ findings, obtained with the HIRES spectrograph on Keck.
Francesco Pepe from Geneva reported this non-finding at the IAU Symposium on The Astrophysics of Planetary Systems: Formation, Structure, and Dynamical Evolution in Turin, from where it was broadcast to much of the astronomy world via Ray Jayawardhana‘s facebook status – like so:
I spotted the status update myself, and was reluctant to write about a bit of (unpublished) conference gossip without knowing the context or background – and exoplanet detection via radial velocities isn’t exactly my thang to begin with. But there’s been more discussion, and the story has been picked up by, amongst others, New Scientist, who publish some quotes from the scientists involved.
The original detection used a combination of data spanning many years of observations from HARPS and HIRES to get a 3-sigma detection of the 6th planet around Gliese 581, a 3.1 Earth-mass planet in the star’s HZ. The Geneva team tried to detect the planet using the same set of HARPS data as the Californians, and additional data from the same instrument. They were unable to detect the faint signal from Gliese 581 g.
It’s tempting to make a big deal out of this (cf. Daily Mail: “devastating claim”) or present it as some kind of “embarrassment” – but this stuff is the bread and butter of science. And a healthy level of competition between different groups does keep things exciting. A barely-3-sigma detection is a very weak signal indeed, so it’s not at all surprising that there is some discussion over the possible sources of error that may have been neglected and the validity of the result.
The Vogt et al paper interestingly chronicles the ongoing discovery ping pong between the groups over the GJ 581 system. Just last year Vogt and his collaborators retracted an earlier version of their paper in peer review after the referee, a scientist from the HARPS team, expressed concerns about residual systematics. They decided to go back to the telescope, gather more data, and think things through. The 2010 paper that appeared on Arxiv a few weeks ago was the result of that “year of introspection”.
They also add in their thoughtful summary of the paper:
“And to be completely fair, the HIRES data set could also have undiscovered systematic errors lurking within. This is very diﬃcult work and there is no shame or dishonor in uncovering residual systematic errors at these levels of precision. Collegial and unabashed inter-team comparisons on stars like GJ 581 and GJ 876 will be crucial to quantifying the true precision limits of any team’s data sets.”
I hope the Geneva team put pen to paper about their non-detection of the 7th planet so the rest of the community can compare the findings like for like. It is, in any case, a matter of time before the so-called Goldilocks planet gets found – whether it’s around GJ 581 or another one of the many exoplanet hosts we now know to exist.
Steven S. Vogt, R. Paul Butler, Eugenio J. Rivera, Nader Haghighipour, Gregory W. Henry, & Michael H. Williamson (2010). The Lick-Carnegie Exoplanet Survey: A 3.1 M_Earth Planet in the
Habitable Zone of the Nearby M3V Star Gliese 581 ApJ accepted arXiv: 1009.5733v1