The objective of this meeting is to bring together Earth, Solar System and Exoplanet atmosphere specialists to discuss recent results and the way ahead. The general emphasis is on comparative planetology, with each specialty bringing a different angle to the study of planetary atmospheres.

For more information visit: exoclimes.org

An outline of the paper: HST Spectral Mapping of L/T Transition Brown Dwarfs Reveals Cloud Thickness Variations. by Apai et al.

2M2139 and SIMP0136 are two brown dwarfs at the transition between L and T spectral types, that fluctuate in brightness on timescales of a few hours. Since the L-T transition is characterised by the disappearance of silicate/iron clouds in brown dwarfs, the variability is thought to be caused by patches of clouds rotating in and out of view as the brown dwarfs spin. Their rotation periods are 8 hours and 2 hours, compared to 10 hours for Jupiter.

The authors have measured the changes in the spectrum of the two objects in the near infrared (1-2 microns) with the WFC3 camera on the Hubble Space Telescope. Measuring the variations in the spectrum over several hours can show whether the brightness fluctuations are due to holes in the clouds, temperature changes in the cloud deck, or the presence of several distinct populations of clouds.

This is similar to the study of Buenzli et al. earlier this year (see our post), except that the two objects have variations of much higher amplitude (about 5% for SIMP0136 and up to 25% for 2M2139), so that they afford more constraining scenarios.

Apai et al. find that the best way to reproduce the Hubble observations is with a mixture of two cloud layers, high clouds and low clouds.

Sketch of the clouds on the two brown dwarfs [Figure 5 of the paper].

In the best fit models, one of the cloud layers is distributed in at least three large patches, on a uniform background provided by the other layer.

Two possible cloud distributions accounting for the observations of 2M2139: three dark spots of high clouds on a brighter low cloud deck, or three bright gaps of low clouds in a higher cloud deck [Figure 9 of the paper].

A look at Jupiter shows how simplistic a cloud distribution in terms of two layers and circular patches might be. Indeed, the Hubble Space Telescope has recently acquired a long sequence of Jupiter images that will allow us to consider it “as an exoplanet” and test our capacity to recover cloud features from integrated lightcurves.

The Apai et al. study is sobering in the context of exoplanet atmospheres: the interpretation remains difficult and ambiguous even with vastly more precise and abundant data that we could dream of for exoplanets.

An outline of the paper: 3D mixing in hot Jupiter atmospheres I:
application to the day/night cold trap in HD 209458b, by Parmentier, Showman and Lian

This paper examines the vertical mixing in a 3-D model of the atmospheric circulation of hot Jupiters – using the output of one of the latest models by the group of Adam Showman.

The results are important for two pressing issues concerning hot Jupiter atmospheres:

(1) strong vertical mixing is required to lift titanium oxide vapors up in the atmosphere of hot Jupiters.

Titanium oxide is expected to be the dominant source of opacity in the visible and responsible for the stratospheric temperature inversion observed in hot Jupiters. However, it condensates into grain and will rain out of the atmosphere if it is not brought back at a sufficient rate by vigorous vertical mixing.

(2) vertical transport is also crucial to the formation of silicate and iron clouds.

Titanium oxide, silicon and iron grains are known to form clouds in brown dwarfs with temperatures comparable to hot Jupiters. In brown dwarfs however, the vertical mixing is done by thermal convection, because the dominant energy transfer is the leakage of internal heat towards space. In hot Jupiters however, the dominant transport is the day-side to night-side redistribution of the energy from the irradiation of the host star. Thermal convection is suppressed, and horizonal motions are 100 to 1000 faster than vertical motions.

The paper by Vivien Parmentier and collaborators examines the amplitude and distribution of the vertical motions in a 3-D circulation model -tuned to the parameters of HD209458b- and compare the mixing timescale to the settling timescale of condensate grains. The result in a nutshell is that vertical mixing from the global circulation is rather vigorous, sufficient to keep small grains afloat (grains smaller than a micron). The mixing is sufficient to keep titanium oxide in the atmosphere, and to lift a haze of silicate grains that formed on the cold night side  - what would be required to explain the haze observed in HD 189733b if it is indeed made of silicate grains.

Interestingly, the strongest vertical mixing occurs at two specific points in the planet: near the Equator on the morning side, a strong updraft forms and lifts air over many atmospheric scale heights, with a corresponding downwards draft on the evening side.

Figure: Map of temperature (left) and vertical velocity (right) in the simulation (the arrows show the dominant wind directions). Notice the fast downdraft and updraft near the Equator at 40^oW and 160^oE, with velocities up to 300 m/s (1000 km/h). Grains that form on the cold night side can be lifted in the high atmosphere by the “morning” updraft. If the grains are small enough, they will not have time to rain down before reaching the night side again.

There is a strong caveat to interpreting the results of this study in terms of what is actually happening in hot Jupiter atmospheres: in the simulations, the condensates are injected only as tracers of the flow, like invisible dust. Actual clouds are not transparent, so their formation triggers a strong feedback effect because they change the transfer of radiation, and therefore modify the temperature profile and the circulation. The next step for this kind of simulations is thus to include the effect of clouds on the radiative transfer.

An outline of the paper: Detecting bio-markers in habitable-zone earths transiting white dwarfs, by Loeb & Maoz

Detecting bio-markers in the atmospheres of planets that orbit main-sequence stars like the Sun is extremely challenging, Abraham Loeb and Dan Maoz, from the Universities of Harvard and Tel-Aviv, suggest that White Dwarfs are far better suited to atmospheric studies and could allow us to detect molecular oxygen.

White Dwarfs have previously been proposed as ideal targets for Earth-sized planet searches since they are much smaller than main-sequence stars and would provide a bigger relative dip in flux when planets transit them. In principle, the same idea can be applied to atmospheric observations: a larger percentage of star light would pass through the atmosphere of the planet, making molecular signals much easier to extract.

White Dwarfs have long-lived habitable zones by virtue of their long cooling timescale, so a planet orbiting a White Dwarf has plenty of time to develop life. Molecular oxygen only began to accumulate in the Earth’s atmosphere after life began: detecting it in the atmosphere of another planet could be a sign that there is life there too.

Loeb and Maoz calculate that the James Webb Space Telescope (JWST), due for launch in 2018, will be capable of detecting molecular oxygen in the atmospheres of planets orbiting White Dwarfs after just over five hours of exposure time.

A drawback to this proposal is the current lack of White Dwarf planet discoveries. Assuming that there is an Earth-mass planet orbiting one in every three White Dwarfs, the odds of us actually seeing a transit is just one in five-hundred. These odds look long, however we have plenty of time to get searching – JWST won’t launch until 2018, and five years is a long time in the field of exoplanets. Plus, an up-coming parallax mission, Gaia, will significantly improve our chances.

Detecting molecular Oxygen in the atmosphere of any extra-solar planet would be a huge forward leap for astronomy, so Loeb and Maoz’s idea may be well worth pursuing despite the long-sounding odds.

Feature image: NASA, ESA and G. Bacon (STScI)

In 2011, Nicolas Cowan and Eric Agol published a paper  trying to make some sense of the available observational data. For each of the 24 hot Jupiters in their sample, they attempted to put constraints on two basic quantities that give a first order characterisation of the global atmospheric energy budget: the Bond albedo and the circulation efficiency. The Bond albedo is the fraction of incident stellar radiation that gets scattered back to space by the planetary atmosphere, integrated across all wavelengths. The circulation efficiency then quantifies the fraction of absorbed energy that gets redistributed to the nightside of the planet by winds.

For the majority of cases, the authors only had secondary eclipse depths at infrared wavelengths to work with, mostly from Spitzer/IRAC datasets. These measurements probe the planetary dayside emission, but the constraints they provide are degenerate between the Bond albedo and circulation efficiency. This degeneracy is quite intuitive, but see Figure 1 of the paper for further details if you’re interested. It can be broken if the planet’s nightside emission is also known, but this was only the case for 3 of the planets in the sample because it’s more expensive to measure, requiring full-phase observations.

The likelihood distribution of Bond albedos and circulation efficiencies derived by Cowan & Agol for their sample of 24 hot Jupiters. Black to white indicates increasing likelihood. Taken from Cowan & Agol (2011).

The resulting likelihood map for the distribution of Bond albedo versus circulation efficiency is shown in the figure above. Despite the degeneracy between the two parameters, there’s clearly a strong preference for Bond albedos less than about 0.5. This is perhaps reassuring, because it’s broadly consistent with what’s been predicted from models (see for instance Sudarsky et al 2000). On the other hand, high Bond albedos of ~0.5 have been predicted for the hottest (>1700K) hot Jupiters, which may form reflective layers of silicate haze in their upper atmospheres. Given that over half the planets in the sample have estimated dayside temperatures in the >1700K range, the fact that they’re consistent with Bond albedos below 0.5 could be a useful constraint for such models.

In terms of reducing the degeneracy seen in the plot above, one line of attack will be to measure more secondary eclipses at wavelengths closer to the peak of the host star spectra. This is because the continuum emission from a planet will have two peaks: one corresponding to reflected starlight and another corresponding to intrinsic thermal emission. Provided the wavelength separation between these two peaks is wide enough, the contribution from the reflected starlight can be isolated, giving a measure of the atmosphere’s reflective properties independent of the circulation efficiency. Unfortunately, this isn’t always straightforward for hot Jupiters, simply because they’re so hot, and their thermal emission profiles are shifted to shorter wavelengths. As a result, the planetary thermal emission can seep into the passband you’re trying to measure the reflected starlight in, and disentangling the relative contributions can be difficult. This is a particular problem for secondary eclipse signals that have been detected in Kepler lightcurves, as Kepler has a broad passband that extends to 850nm (e.g. Christiansen et al 2010, Demory et al 2011, Kipping & Spiegel 2011, Désert et al 2011).

Despite challenges such as these, Frédéric is currently leading an HST program to try measure the reflected light signal of HD189733 b with the STIS camera. Those data were obtained on 21 December 2012 for a single eclipse and the analysis is ongoing. Combined with the infrared full-phase observations already published by Agol et al (2010) and Knutson et al (2012), we’ll hopefully be able to get a handle on the Bond albedo of this well-studied hot Jupiter. This will be an important piece of the puzzle when it comes to understanding the reflective properties of the haze material that must be responsible for the scattering signature seen so prominently in the optical transmission spectrum.

An outline of the paper: Re-Evaluating WASP-12b: Strong Emission at 2.315 micron, Deeper Occultations, and an Isothermal Atmosphere, by Crossfield et al.

WASP-12b is an unusually hot planet even for a hot Jupiter. It is one of the first exoplanets to be studied with large wavelength coverage in transmission, with preliminary results used by us to predict what the sunset would look like on such a planet. WASP-12b also stands out because previous published observations have indicated that it may have an unusually high C/O ratio (Madhusudhan et al. 2011). Recently, the picture became more confusing when Cowan et al. (2012) measured two full orbits with Spitzer, one at 3.6 microns and one at 4.5 microns and found that the resulting secondary eclipse depths were too different to reconcile with either a solar or a high C/O ratio. This means that the characteristics of WASP-12b pose quite a challenge to our current understanding. A possible explanation was revealed recently: it was announced by Crossfield et al. at the Heidelberg conference that WASP-12 actually has an M dwarf companion star (initially reported by Bergfors et al. 2011,2012), which causes the transit to be diluted and the inferred radius to be smaller than the real radius. It also affects the planetary spectrum because the spectral type of the companion is different to WASP-12.

Firstly, the authors present new narrowband photometry with Subaru at 2.315 microns (FWHM of 27 nm) and find a deeper than expected eclipse depth. The measurement rules out at the 3 σ level the strong absorption feature predicted at this wavelength by high C/O ratio (> 1) models.

Top: narrowband photometry of WASP-12b at 2.315 microns. The model is shown with a red solid line with dotted lines showing the 1 sigma range. Open points are excluded in the analysis due to sky background problems.

It was during these observations that the authors noticed the companion star, which they call Bergfors-6. They followed up the initial discovery with a series of K band observations using IRTF/SpeX to estimate the flux ratio of WASP-12 to Bergfors-6.

Left: Subaru image at 2.315 microns. Right: follow-up image from IRTF/SpeX K band.

The authors also use their i,z and K band photometric measurements to find that Bergfors-6 has a temperature of 3840±70 K with spectroscopic constraints giving a slightly cooler temperature.  Spectral lines are also used to measure the radial velocities, which are 16.5±2.6 km/s and 19.7±1.3 km/s for WASP-12 and Bergfors-6 respectively. This could mean that the two stars are gravitationally bound, but could also be a coincidence. Combining the photometry and spectroscopy implies that Bergfors-6 lies 50 % closer to the Earth than WASP-12. There is still a possibility of Bergfors-6 being bound to WASP-12 if Bergfors-6 is in fact a binary system observed near conjunction, where the fluxes would be greater than the flux from one star.

Finally, all existing data for the planetary atmosphere are corrected for Bergfors-6, with the severity of contamination depending on the aperture size initially used for analysis. The model that now best matches all the current data is a 3000 K black body (an isothermal atmosphere), which is quite unexpected since it is so different from the previously assumed models. The narrowband Subaru measurement fit the blackbody to within 2σ. Alterations to the model (changing the temperature-pressure profile and thermal inversion, changing abundances) to match this point lead to overall worse fits.

Bottom axis is wavelength in microns. Solid points are dilution-corrected measurements. Top: the dark spectrum is a model with a thermal inversion. The grey spectrum is this at 100 times higher resolution. Bottom panel: 3000 K blackbody spectrum overplotted onto the data.

An outline of the paper: IRSF SIRIUS JHKs Simultaneous Transit Photometry of GJ1214b, by Narita et al.

The transmission spectrum of GJ 1214b has in previous papers by Berta et al. (2012) and Bean et al. (2011) amongst others, shown itself to be a flat, featureless spectrum. This has lead to theories that GJ 1214b either has an atmosphere with a large scale height obscured by high altitude haze or clouds, or a water dominated atmosphere, undetectable due a large mean molecular weight. For more background info see our previous post on the atmospheric chemistry of GJ 1214 b.

The only sign of a significant spectral absorption feature was presented in a paper by Croll et al. (2011) who found a larger effective radius in the K_s band (~2.15 microns). For this prominent spectroscopic feature to exist, the atmosphere would have to be H/He dominated with a massive scale height making the super-Earth more of a mini-Neptune. It was suggested in this paper and subsequent papers that this detection be remeasured, and remeasured it was.

Narita et al. remeasured the effective radius of GJ 1214 b in the K_s band and found it to be much shallower than what Croll et al. had measured, resulting in an inconsistency at the 4σ level. As mentioned in their paper, they find no good explanation for this discrepancy. It might be tempting to think that the measurement by Croll et al. is a statistical outlier, but as three separate transits events yielded a deeper transit in the Ks band compared to the J band, it is not that easy.

The GJ 1214b transmission spectrum in the region 1-2.5μm. The grey solid line represents the water-dominated atmosphere whilst the grey dotted line represents the hydrogen-dominated (Solar composition) atmosphere. The results by Narita et al. is shown in red, whilst the results from Croll et al. are shown in purple.

de Mooij et al. got a similar result to Croll et al. in the Ks band, but warn that their uncertainties are large due to systematic features in the light curves.

It is clear from the studies mentioned above that more observations in the K_s band is warranted, but as Narita et al. suggest, observations in the blue optical bands would be very useful. This is because the observations could help discern between the water-dominated atmosphere and a H/He + haze atmosphere, as the latter would show an increasing absorption towards the blue due to Rayleigh scattering. The main problem doing this, however, is that the M-star GJ1214 is emitting fewer photons towards bluer wavelengths, requiring longer observing times at expensive facilities.

For now, GJ1214 remains the ideal candidate for super-Earth atmospheric studies, being the only super-Earth within the parameter space were an atmospheric detection is possible. This is likely to be the case until surveys like the MEarth survey find another super-Earth or until the upcoming  Next Generation Transit Survey comes online, estimated to find many of them.

Featured image found here.

• “On the probability of habitable planets” by Forget
• “3D gas dynamic simulation of the interaction between the ‘hot Jupiter’ planet and its host star” by Bisikalo et al
• “Spectral Fingerprints of Earth-like Planets Around FGK Stars” by Rugheimer et al
• “On the potential of the EChO mission to characterise gas giant atmospheres” by Barstow et al
• “XUV exposed non-hydrostatic hydrogen-rich upper atmospheres of terrestrial planets. Part I: Atmospheric expansion and thermal escape” by Erkaev et al
• “XUV exposed, non-hydrostatic hydrogen-rich upper atmospheres of terrestrial planets II: Hydrogen coronae and ion escape” by Kislyakova et al
• “The Ultraviolet Radiation Environment Around M dwarf Exoplanet Host Stars” by France et al
• “The most common habitable planets – atmospheric characterization of the subgroup of fast rotators” by Pinotti
• “A Photometric Study of the Hot Exoplanet WASP-19 b” by Lendl et al
• “The 0.81 – 4.2 micron ground-based transmission spectra of the hot jupiter HD189733 b” by Danielski et al

Feature image taken from Rugheimer et al

About the paper “A 3D general circulation model for Pluto and Triton with fixed volatile abundance and simplified surface forcing” by Zalucha & Michaels

The authors have run a general circulation model on the cases of the atmospheres of Pluto and Triton.

The two have similar atmospheres. Pluto is the prototype of the  “dwarf planets” according to the new IAU nomenclature whilst Triton is a satellite of Neptune, thought to be a captured Pluto-like body. Their atmospheres are made primarily of nitrogen, with ground pressure around 10 microbars.

That’s right: microbars. Collapsed to liquid form, the atmospheres would amount to less than one millimetre of liquid nitrogen over the surface of Pluto and Triton. Ten microbars in a lab on Earth qualifies as a very good vacuum.

Amazingly, this minute amount of air suffices to exhibit most of the features of a full-blown atmosphere: the atmospheres of Pluto and Triton extends over dozens of kilometres vertically, they have eastwards jets, photochemical hazes, and condensation clouds.

The main data comes from occultations of background stars by the planets. During ingress and egress the atmosphere shows are a progressive dimming, an occultation lightcurve.

Unlike Earth, temperatures get colder near the ground. The reason, in the words of to ESO press release 0908 (about an occultation measurement with CRIRES on the VLT), is the following:

The reason why Pluto’s surface is so cold is linked to the existence of Pluto’s atmosphere, and is due to the sublimation of the surface ice; much like sweat cools the body as it evaporates from the surface of the skin, this sublimation has a cooling effect on the surface of Pluto. In this respect, Pluto shares some properties with comets, whose coma and tails arise from sublimating ice as they approach the Sun.

Artist’s impression of the surface of Pluto. Credit: ESO/L. Calçada

In the simulations, the day-night temperature contrast is 5 K, and the maximum wind speed 10 m/s.

How small can an atmosphere be before it ceases to be an atmosphere? Io, down to nanobars, has no circulation to speak of, and the structure of its atmosphere depends on the latest volcanic events, so Pluto and Triton are probably near the lower limit of applicability for the usual tools of comparative planetology.

An outline of the paper: Vertical atmospheric structure in a variable brown dwarf: pressure-dependent phase shifts in simultaneous HST-Spitzer light curves, by Buenzli et al.

It is thought that the transition from L-type to T-type brown dwarfs, at temperatures around 1400 K,  is characterised by the disappearance of silicate clouds that dominate the late L-style atmospheres. Two brown dwarfs near the L-T transition show strong variability on the timescale of their rotation (a few hours), and this has led to one of the most exciting hypothesis of recent years in the field, the “patchy cloud” scenario. According to this scenario, what we see is a cloud coverage that breaks into patches near the L-T transition, causing flux variations as the breaks in the cloud cover rotate in and out of view.

Buenzli et al. have used HST and Spitzer to monitor the flux of the late-T brown dwarfs 2MASSJ22282889-4310262 (and we sometimes complain about unwieldy exoplanet names!) over 12 hours. They find a clear 2% variability of the flux along the rotation period of 1.4 hours.

Lightcurve of 2MASS2228 with Hubble and Spitzer space telescopes, showing out-of-phase variations in flux of about 1% at different infrared wavelengths.

Most interestingly, they found that while the amplitude of the variation in different wavelength ranges remain similar, the phase of the variation is strongly offset. The phase shift seems to evolve smoothly as wavelengths probe deeper layers in the atmosphere (the wavelength coverage is 1.2-1.7 microns for HST and 4-5 microns for Spitzer, which according to models should span the 1-10 bar range in the atmosphere of the brown dwarf).

Phase shift in the light curve of the brown dwarf 2MASSJ2228, as a function of the pressure probed by the observations according to atmosphere models.

Thanks to the different wavelengths, the authors can test whether the variations result from changes in opacity (patchy clouds) or in temperature (atmospheric circulation).

Their conclusion is that the patchy-cloud explanation doesn’t work on its own, since in that case one would expect the amplitude of the variation to change with wavelength rather than the phase. What works best is a mixture of the two effect – in other words a full “weather” situations with hot and cold currents as well as forming and dissipating cloud structures.

We may need this kind of data for many more objects before clear patterns start emerging. This is slightly disturbing when we think that in hot-Jupiter studies we can only dream of having the kind of extensive constraints available for these brown dwarfs.

Feature Image: Jon Lomberg