Additional Planets at Gliese 581

Geneva Team's envy? Who knows

Well I've done circular fits for the HARPS 2009 data using Systemic and I still don't see planets f or g. Guess the next step would be to construct synthetic datasets of various system configurations (all-circular 4-planet, eccentric 4-planet, all-circular 6-planet) to see how likely it is that the datasets would be able to detect them.

Right I actually decided to try the synthetic data trick. Starting from the model given in Vogt et al. (2010) using P, K and M, and perturbing it by the quoted error values, with a jitter value of 1.4 m/s, I generated 20 fake datasets with observations at the time and with the errors of the HARPS 2009 dataset. I then used Systemic to fit circular-orbits solutions to the fake data.

In 12 of the datasets (60%) I could detect all six planets, though for 5 of these the ability to do this depended on the correct choice of period vs alias.

In 6 of the datasets (30%) I could detect the four previously-known planets and one only of the new planets. 5 of these I could detect g but not f, and the other time I could detect f but not g.

In 1 dataset (5%) I could only detect the four previously-known planets, and it still preferred the 82-day alias period for planet d, despite using the 67-day period to generate the data. If you set the period for planet d to the 67-day value (though there's no particular indication that you should do this: the 67-day period is not noticeable in the periodogram) you can still find planet g with this dataset.

In 1 dataset (5%) I could only detect planets b, c and e (!)

Probably need a larger sample of synthetic data to get anything meaningful out (and I may have made several incorrect assumptions when generating the datasets), but at least it looks plausible that the published HARPS dataset would likely have indicated the presence of f and/or g.

Well, further observation can only improve parameters.

So let me get this straight.

The Lick-Carnegie Exoplanet Survey dataset, by itself, does not show the two new putative planets, due to a lack of sufficient phase coverage of all orbits.

The original Geneva dataset, by itself, also does not show additional planets for the same reason.

Vogt's team states that the two datasets combined do show the two new plants, mainly because the two datasets are complementary in phase coverage.

The Geneva team states that their data, including new measurements, still does not show additional planetary signals. But what do their new data add in terms of phase coverage to their original data?

Has anybody combined the expanded Geneva dataset with the Lick-Carnegie dataset to see if the two new putative planets still appear in the data? Surely that is the obvious next step.

I don't think the phase coverage is the issue. The HARPS 2009 data has quite even phase coverage for the period of planet g. The HARPS phase coverage for f is a bit more patchy, I guess that's why I found planet g in my trial data somewhat more often than I found planet f.

Unfortunately until the Geneva group publish the updated HARPS dataset, I'm stuck with working on the 2009 version. But as far as I can see, it isn't looking particularly good for f or g.

Aliases of the first eccentric harmonic : Is GJ 581g a genuine planet candidate?
http://arxiv.org/abs/1011.0186

Quote :

The radial velocity (RV) method to detect extrasolar planets has been the most successful to the date. The RV signal imprinted by a few Earth-mass planet around a cool star is at the limit of the typical single measurement uncertainty on state-of-the-art spectrographs. As a consequence, one has to rely on statistics to unearth signals buried below the noise. Observationally related issues such as unquantified systematic errors or artifacts introduced by the observing cadence can produce spurious signals. Also, they can mask genuine signals that should be easily detected otherwise. This is the case for the planetary system around the low mass star GJ 581 for which there has been recently announced a 3.1 Earth mass planet in its habitable zone. Even though it is a very stable star, the combination of the observing cadence and the presence of multiple planets has already caused a number of period misinterpretations in the past. We discuss here a particularly devious statistical degeneracy that derives from the aliasing of the first eccentric harmonic of an already detected planet with the characteristic one year sampling frequency. Such a degeneracy can prevent the detection of the true signal and correlate the eccentricities of known planets with the mass determinations of additional low amplitude companions. By performing a number of statistical tests, we conclude that even though the statistical degeneracy is clearly present, the existence of GJ 581g remains well supported by the available data.

Looks like that assumes as given that the merging of the HARPS+HIRES data is correct. It would be far more useful to have an analysis of whether the HARPS or HIRES data individually should have been able to see the data. HARPS people have claimed that when they insert the planet into the dataset they can recover it, and I get similar results from constructing fake datasets.

I think it is going to re-enter in the exoplanets' list soon

Edasich wrote:

I think it is going to re-enter in the exoplanets' list soon

Would be nice, but we shall see...

Lazarus wrote:

Edasich wrote:

I think it is going to re-enter in the exoplanets' list soon

Would be nice, but we shall see...

I am crossing fingers.

Gliese 581g: now you see it, now you don’t

Quote :

“Even in the [Americans’] published data, after taking out the first four planets, what’s left looks like noise,” said Joseph Catanzarite, an astrophysicist at NASA’s Space Interferometry Mission, where he hunts for habitable planets.

That was exactly the conclusion of the Swiss team, which couldn’t find Goldilocks using a new data set either.

If Goldilocks exists, she appears to move her central star by one meter per second. Problem is, instruments measuring the shifting of stars can only register a bump of about one meter per second. The planet’s signal is so close to the instrument’s threshold that it may or may not be noise.

Quote :

“The question isn’t quite settled but it’s close—it looks like the planet isn’t there,"

Still not looking good I see. Then again I suspect that more planets than the currently-known set of 4 are there since the temperature of the detected dust disc suggests it lies substantially further out than the orbit of planet d, so there could well be additional planets between planet d and the disc.

From here

Quote :

Even with Mayor's team and Vogt's team trying to confirm Gliese 581g, "it may take another season or two to really know, so I expect we won't see any possible confirmation (or refutation) for another year or two," Vogt said. "In the meantime, we are hard at work not only gathering more data on this system, but on others that may yield similar potentially habitable planets."

Bayesian Re-analysis of the Gliese 581 Exoplanet System
http://arxiv.org/abs/1101.0800

Quote :

A re-analysis of Gliese 581 HARPS and HIRES precision radial velocity data was carried out with a Bayesian multi-planet Kepler periodogram (from 1 to 6 planets) based on a fusion Markov chain Monte Carlo algorithm. In all cases the analysis included an unknown parameterized stellar jitter noise term. For the HARPS data set the most probable number of planetary signals detected is 5 with a Bayesian false alarm probability of 0.01. These include the $3.1498\pm0.0005$, $5.3687\pm0.0002$, $12.927_{-0.004}^{+0.006}$, and $66.9\pm0.2$d periods reported previously plus a $399_{-16}^{+14}$d period. The orbital eccentricities are $0.0_{-0.0}^{+0.2}$, $0.00_{-0.00}^{+0.02}$, $0.10_{-0.10}^{+0.06}$, $0.33_{-0.10}^{+0.09}$, and $0.02_{-0.02}^{+0.30}$, respectively. The semi-major axis and $M sin i$ of the 5 planets are ($0.0285\pm0.0006$ au, $1.9\pm0.3$M$_{\earth}$), ($0.0406\pm0.0009$ au, $15.7\pm0.7$M$_{\earth}$), ($0.073\pm0.002$ au, $5.3\pm0.4$M$_{\earth}$), ($0.218\pm0.005$ au, $6.7\pm0.8$M$_{\earth}$), and ($0.7\pm0.2$ au, $6.6_{-2.7}^{+2.0}$M$_{\earth}$), respectively. The analysis of the HIRES data set yielded a reliable detection of only the strongest 5.37 and 12.9 day periods. The analysis of the combined HIRES/HARPS data again only reliably detected the 5.37 and 12.9d periods. Detection of 4 planetary signals with periods of 3.15, 5.37, 12.9, and 66.9d was only achieved by including an additional unknown but parameterized Gaussian error term added in quadrature to the HIRES quoted errors. The marginal distribution for the sigma of this additional error term has a well defined peak at $1.8\pm0.4$m s$^{-1}$. It is possible that this additional error arises from unidentified systematic effects. We did not find clear evidence for a fifth planetary signal in the combined HIRES/HARPS data set.

Nice to see that analysis, especially as it takes the datasets on their own and in combination. With the HIRES data it does seem to validate my own investigations that only planets b and c can be found in there. Interesting to see planet f coming out of the HARPS dataset, I did not find this when I investigated the system. (On looking at it again I can get a period at ~200d to show up, but this results in an extreme eccentricity >0.99 which is unlikely to be real. Some of the plots in the paper do indicate some solutions at ~200d for the fifth planet though.)

Unlikely to be the last word on all this though...

Here's a paper about the use of reduced χ², and the pitfalls involved with using this statistic for nonlinear models.

One of the examples used is Gliese 581: using the parameters given in Vogt et al. (2010) and the same data, they point out that use of the Kolmogorov-Smirnov test suggests the four planet configuration is the most likely one.

Astrophysical Parameters and Habitable Zone of the Exoplanet Hosting Star GJ 581
http://arxiv.org/abs/1102.0237

Quote :

GJ 581 is an M dwarf host of a multiplanet system. We use long-baseline interferometric measurements from the CHARA Array, coupled with trigonometric parallax information, to directly determine its physical radius to be $0.299 \pm 0.010 R_{\odot}$. Literature photometry data are used to perform spectral energy distribution fitting in order to determine GJ 581's effective surface temperature $T_{\rm EFF}=3498 \pm 56$ K and its luminosity $L=0.01205 \pm 0.00024 L_{\odot}$. From these measurements, we recompute the location and extent of the system's habitable zone and conclude that two of the planets orbiting GJ 581, planets d and g, spend all or part of their orbit within or just on the edge of the habitable zone.

Quote :

Thus, we confirm the results in Boyajian et al. (2010) that theoretically calculated radii of M dwarfs are smaller than the directly measured counterparts.

Definitely it is looking like theoretical models of red dwarf star mass/radius relationships are in serious need of revision.

So now with Kepler showing us yet more examples of the cooler low mass planets being mini-Neptunes, I am now far more sceptical than I was that any of the Gliese 581 planets are habitable.

I guess we need some more models of what happens to the habitable zone if you start adding significant quantities of ice and hydrogen. Models based on the assumption that these worlds are mainly rocky probably don't cut it.

Agreed, I am coming to the conclusion that most Earth-mass planets (other than the really close-in ones) end up as mini-Neptunes, with only those objects that experience a giant impact after the protoplanetary disc has dissipated, resulting in their dense primordial atmosphere (and most of their water?) being stripped away, left to resemble an Earth/Venus style terrestrial planet. Mars might be too small to have retained its primordial atmosphere, or even gained one in the first place. Mercury might have been large enough, but it evidently experienced a giant impact as well, leaving only the nickel-iron core and a relatively thin mantle behind. Three giant impacts for four planets seems unlikely, perhaps Jupiter had something to do with this, or it could just be an unlikely coincidence.

It certainly would explain why Earth, Venus and Mercury appear to have undergone a giant impact at some point, maybe if they had not, they would still be like mini-Neptunes in composition.

It also could help explain the Fermi paradox, with a lot fewer terrestrial planets in the Galaxy than otherwise expected.

We really need more observations, to verify or disprove this idea.

Bayesian re-analysis of the radial velocities of Gliese 581. Evidence in favour of only four planetary companions
http://arxiv.org/abs/1102.3314

Quote :

The Gliese 581 planetary system has received attention because it has been proposed to host a low-mass planet in its habitable zone. We re-analyse the radial velocity measurements reported to contain six planetary signals to see whether these conclusions remain valid when the analyses are made using Bayesian tools instead of the common periodogram analyses. We analyse the combined radial velocity data set obtained using the HARPS and HIRES spectrographs using posterior sampling techniques and computation of the posterior probabilities of models with differing numbers of Keplerian signals. We do not fix the orbital eccentricities and stellar jitter to certain values but treat these as free parameters of our statistical models. Hence, we can take the uncertainties of these parameters into account when assessing the number of planetary signals present in the data, the point estimates of all of the model parameters, and the uncertainties of these parameters. We conclude that based on the Bayesian model probabilities and the nature of the posterior densities of the different models, there is evidence in favour of four planets orbiting GJ 581. The HARPS and HIRES data do not imply the conclusion that there are two additional companions orbiting GJ 581. We also revise the orbital parameters of the four companions in the system. Especially, according to our results, the eccentricities of all the companions in the system are consistent with zero.

Wonder when the HARPS people are going to release the extended dataset. Will be very interesting to see how these analyses change in light of the new data.

From today's arXiv, here's a study of the habitability of Gliese 581g using geodynamic modelling. Looks like it would be close to the optimum part of the habitable zone.

Pity that the existence is so uncertain, and that if it does exist there appears to be a good possibility of it being a mini-Neptune anyway.

Lazarus wrote:

Wonder when the HARPS people are going to release the extended dataset. Will be very interesting to see how these analyses change in light of the new data.

Seems they did not see any signs of planets f or g.