47 Ursae Majoris d

Well since Early Access articles have trackable DOIs so they can be cited in advance of print, I guess this is kosher to post here. Odd that it hasn't been picked up by EPE yet though.

Gregory and Fischer (2010) A Bayesian periodogram finds evidence for three planets in 47 Ursae Majoris

38⁺¹¹_-14 years period for outermost planet, obviously with that kind of orbit it has not yet completed one revolution since sufficiently-high precision measurements began, so some pretty large uncertainties on the parameters.

Wow! Great to see some new results on some of the first known planetary systems, 47 Ursae Majoris and Upsilon Andromedae.

Quote :

The three-planet model is found to be 10⁵ times more probable than the next most probable model which is a two-planet model.

Lol. That's pretty encouraging.

Well, well, well... Gregory is the same of 2 planets at HD 208487 and 3 planets at HD 11964. Both systems have turned out not hosting additional planets. I hope this time he's right

Well to be fair, he did note with the HD 11964 result that the period was suspiciously similar to 1-year: i.e. the data reduction process that had been used had problems.

And why focus so much on the failures, and ignore the result he got for HD 73526, where he found the evidence pointed towards the 2nd planet. (Although at the time, it was two different 1-planet models being compared)

Listed at EPE between confirmed candidate. Hope it may remain there without retraction

http://exoplanet.eu/star.php?st=47+Uma

With a three-planet model being 10⁵ times more probable than the two-planet model, the planet's going to have to be destroyed to be unconfirmed.

One further point about all this is that it is nice to see 47 UMa c in there (though somewhat different parameters to what's on EPE). Previously its existence had been questioned in several papers.

The question is what that 10⁵ means. Probably it refers to whether this period is present in the dataset. Of course there are other reasons such a period could exist than a planet, e.g. stellar variation or artefacts of the data reduction process (as turned out to be the case with HD 11964).

They said in the abstract that,

Quote :

radial velocity data confirms and refines the properties of two previously reported planets with periods of 1079 and 2325 d.

So I am taking this to mean that the c was confirmed more-or-less as advertised.

It's interesting to compare the parameters found by Wittenmyer et al. in their analyses, with the parameters for planet d in the three-planet solution.

In Wittenmyer et al. (2006), the parameters given for "47 UMa c" are 7586±727 days, mass of 1.34±0.22 times Jupiter. The 1σ error bars do not quite overlap. Wittenmyer et al. (2009) describes the best-fit period for the outer planet in their 2-planet solution as being ~9660 days, within the error bars for planet "d" from Gregory and Fischer (2010).

So maybe, considering that for the 3 planet solution in Gregory and Fischer (2010), planet c would have the lowest RV semi-amplitude, Wittenmyer et al. were detecting planet d but not c?

I can see how the RV curve of d would make c's period seem longer, but I would think that they truly did detect c, instead of d, as I doubt the baseline for the observations were enough to really get c and d confused. I'm not sure how a 14,000 day periodicity would appear to represent a ~9,000 day period.

I vaguely remember a revision to the fit for 47 UMa that placed c at ~2,500 days. I don't remember where the paper was though. I modeled it in Celestia, noted that the orbits of b and c were very close, then you reminded me that the parameters for b were different, I re-did it and it worked out better.

I don't seem to remember where that thread is. Or the paper.

Take a look at the error bars on the period for planet d

14002⁺⁴⁰¹⁸_-5095 days

Range inside the error bars is 8907–18020 days, I guess this is the 1σ range. Remember we're dealing with an incomplete orbit, there's going to be substantial period-eccentricity degeneracy.

I always forget about the error bars.

You're right. It will be interesting to see if the discovery paper mentions anything about this. Would Hipparcos data compared to modern data be sufficient to allow an estimate of the inclination? I recall the Hipparcos astrometry estimates are generally not trusted though.

http://arxiv.org/PS_cache/arxiv/pdf/1003/1003.5549v1.pdf

Quote :

Mon. Not. R. Astron. Soc. 403, 731, 2010, Accepted 20 Dec. 2009, submitted 27 Aug. 2009
The definitive version is available at www.blackwell-synergy.com

A Bayesian analysis of 47 Ursae Majoris (47 UMa) radial velocity data confirms and
refines the properties of two previously reported planets with periods of 1079 and
2325 days. The analysis also provides orbital constraints on an additional long period
planet with a period  10000 days. The three planet model is found to be 105 times
more probable than the next most probable model which is a two planet model. The
nonlinear model fitting is accomplished with a new hybrid Markov chain Monte Carlo
(HMCMC) algorithm which incorporates parallel tempering, simulated annealing and
genetic crossover operations. Each of these features facilitate the detection of a global
minimum in 2. By combining all three, the HMCMC greatly increases the probability
of realizing this goal. When applied to the Kepler problem it acts as a powerful multi-
planet Kepler periodogram.

Can't see any changes from what's already up at EPE though

(moved to the thread already discussing this result)