Gliese 892 / HD 219134 - six planets (at least one planet in transit)

The mass value seems to have been revised somewhat, the Motalebi et al. paper gave 2.67 Earth masses, while Vogt et al. gave 0.11 Jupiter masses (=3.5 Earth masses)

I think you mean 0.011. I noticed also much higher mass for c than previously announced (maybe revised stellar mass and radius) and I wonder whether mass of b is revised upward. Nevertheless chance for transits of outer planets increased slightly

Two other slides :

2 planets only?

tommi59 wrote:

2 planets only?

2 certain ones for sure.

Though it's hard to read the connotations from the slide alone, that is far too pessimistic.

First, the periods used to refer to "e" and "h" (which, in the discovery papers, refer to the same signal) are different to the previous values. The period of "e" is now close to the proper period of "h", while "h" has moved to the actual period of the magnetic cycle. This doesn't actually change the planet count because they were the same object in the first place.

Unlike rotation, it is physically not possible for harmonics (integer divisions) of magnetic cycles to cause an RV signal because they do not represent surface variability, so "e" cannot be activity. The same conclusion is reached by Johnson et al., and I cannot imagine why the opposite is insinuated by this slide.

The rotational period of the star appears to be about the period of d (~42 d, which Motalebi et al. detect), so the first harmonic is about the period of f (~23 d, which is the signal Johnson et al. detect). However, Motalebi et al. find the RV signal to be more coherent than the activity signal and, more convincingly, the two variables are decorrelated (figure 8 ). Though their model for d is actually a combination of d and f, it remains that there would be a correlation with the activity indicators if the RV signal was activity, which is not observed.

Conclusions for f are more difficult because Motalebi et al. do not parse it, but Vogt et al. find that the signal is coherent:

Quote :

In each case, the P =22.8-day periodicity is the location of the strongest peak in the power spectrum of the velocity residuals (between
P = 10 days and P = 30 days).  This suggests that the P = 22.8-day signal is not the  product  of  spot  modulation,  and  that  it  has been present and stable throughout the full time span of our observations.

Though the conclusion is weaker, the evidence favours a Keplerian origin for the signal than activity.

As for g, I don't even know. I guess because it's about twice the rotational period? But multiples of rotational periods don't mean anything - it is not possible for spots or plages to work like that, because then fluid dynamics would break.


Stellar activity, be it rotational or basal convection, possess three important properties. They are not coherent over long timescales, they are not perfectly periodic, and they leave signals in certain spectral lines (which is how activity indicators are derived). It is very unwise to dismiss planets because they happen to have periods sort of like activity without actually testing whether this is supported by activity indicators, especially when simple tools like correlations are enough to determine whether an RV signal is activity or not.

But then again, all of the papers I'm citing (and this slide cites…) reach the same conclusions as what I just wrote, so I don't actually know who I should be directing this to. And there may be some missing context.

Johnson et al. (2016) favour the interpretation that the 46.7-day and 2200-day signals are planetary, in which case the system consists of at least 4 planets (b, c, d, h). They don't seem to recover the signal for g from their data but they note (§4.1) that the uneven time sampling in the Keck data is problematic, they don't seem to detect c either (but transits give a fairly good indication that it does exist).

From the conclusion:

Quote :

Conversely, however, our work provides evidence that the 46.7 day HD 219134 d and the 2200 day HD 219134 h are likely to be actual planets; M15 and V15, respectively, had expressed some concerns about whether these RV signals could be related to stellar rotation or activity.

It will be interesting to see what further data reveals about the system. The revised masses in those slides do suggest that there is a new analysis of the RVs.

They don't seem to detect c either  (but transits give a fairly good indication that it does exist).

So if transit would not happen then planet would go to unconfirmed ? The 22.8 days signal is very strong and looks like genuine planet what a pity inclination will not allow it to transit (just only 2.4 % chance for transit).We can not rule out planet existence only because rotational period of the host is close to period of the planet

http://www.nature.com/articles/s41550-017-0056?utm_content=buffer91f35&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

Nature :

Two massive rocky planets transiting a K-dwarf 6.5 parsecs away

The transiting nature of both HD 219134 b and c increases the
probability that planets d and f also transit. Using the formalism of
previous work4, we compute posterior transit probabilities of 13.1%
and 8.1% for planets f and d
(…)
Although such a transit search is probably out of reach of ground-based
telescopes, it could be performed again by Spitzer, whose operations have
been extended to end-2018, or by the space missions TESS and
CHEOPS, which are both due to launch in 2018.

So I hope for transit of 22.7 days planet although.At what inclination such transit would occur? 89.3 degrees?

From what I can see they do not seem to discuss whether or not the 22.7-day signal is activity-related, as previously suggested. Though Johnson et al. (2016) did suggest that detection in the HARPS-N data would favour the stellar origin interpretation:

Quote :

The activity level of HD 219134 is currently increasing, with the next maximum expected around late 2018. If the 22.8-day RV signal begins to appear in continuing HARPS-N observations as the activity level increases, this would be strong evidence for the stellar origin of this signal. In order to test for this possibility V15 split their dataset into three portions, and did recover the 22.8-day periodicity in all three subsets; however, they did not quote the significance level of the recovery or if the other parameters are consistent between the different subsets, preventing us from making a more detailed analysis of this issue. Additionally, high-cadence observations near the cycle maximum could potentially probe whether the 22.8 day periodicity is a harmonic of a longer rotation period.

The model used in the Nature paper takes account of R_HK but not sure how this relates to the S_HK analysed in the Johnson et al. paper.

On arXiv :

https://arxiv.org/abs/1703.01430

Two massive rocky planets transiting a K-dwarf 6.5 parsecs away

Dorn & Heng "Secondary atmospheres on HD 219134 b and c"
https://arxiv.org/abs/1711.07745

The atmospheres of these planets are likely to be secondary rather than primordial. HD 219134 b likely contains a significant amount of volatiles if its rocky core has the composition inferred from stellar abundances, while HD 219134 c may be rocky.

One of the papers in this Russian astrophysics journal, if I've understood it correctly through Google translate, seems to imply that the transit of HD 219134b is deeper in the UV than in the Spitzer IR.

Abstract wrote:

Приведены и анализируются результаты наземных наблюдений, подтверждающих открытие транзитной суперземли в системе звезды HD 219134. Признаки существования этой планеты были обнаружены ранее при анализе лучевых скоростей звезды и по наблюдению транзитов космическим телескопом «Спитцер» в инфракрасном диапазоне. В расчетное время мы зарегистрировали транзит в ближнем ультрафиолетовом диапазоне несколько раз. Измеренная глубина транзита в фильтре U системы Джонсона составляет 0.13% ± 0.027%, что глубже известной по измерениям «Спитцера.» Обсуждаются вероятные причины этой разницы.

Google Translate wrote:

The results of ground-based observations confirming the discovery of transit super-earth in the star system HD 219134 are presented and analyzed. Signs of the existence of this planet were discovered earlier when analyzing the radial velocities of the star and observing the transit of the space telescope "Spitzer" in the infrared range. At the estimated time, we registered transit in the near ultraviolet range several times. The measured transit depth in filter Johnson's U filter is 0.13% ± 0.027%, which is deeper than that known from Spitzer's measurements. The probable causes of this difference are discussed.

Presumably this could indicate the presence of an extended (escaping?) atmosphere.

Additional planets in the system?

An Integrative Analysis of the HD 219134 Planetary System and the Inner Solar System: Extending DYNAMITE with Enhanced Orbital Dynamical Stability Criteria

Quote :

Planetary architectures remain unexplored for the vast majority of exoplanetary systems, even among the closest ones, with potentially hundreds of planets still ``hidden" from our knowledge. DYNAMITE is a powerful software package that can predict the presence and properties of these yet undiscovered planets. We have significantly expanded the integrative capabilities of DYNAMITE, which now allows for (i) planets of unknown inclinations alongside planets of known inclinations, (ii) population statistics and model distributions for the eccentricity of planetary orbits, and (iii) three different dynamical stability criteria. We demonstrate the new capabilities with a study of the HD 219134 exoplanet system consisting of four confirmed planets and two likely candidates, where five of the likely planets are Neptune-size or below with orbital periods less than 100 days. By integrating the known data for the HD 219134 planetary system with contextual and statistical exoplanet population information, we tested different system architecture hypotheses to determine their likely dynamical stability. Our results provide support for the planet candidates, and we predict at least two additional planets in this system. We also deploy DYNAMITE on analogs of the inner Solar System by excluding Venus or Earth from the input parameters to test DYNAMITE's predictive power. Our analysis finds the system remains stable while also recovering the excluded planets, demonstrating the increasing capability of DYNAMITE to accurately and precisely model the parameters of additional planets in multi-planet systems.