sign in sign up
<<
Home , Gliese 581, Gliese 581d

is the first discovered terrestrial- in the habitable zone

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Wordsworth, Robin D., François Forget, Franck Selsis, Ehouarn Millour, Benjamin Charnay, and Jean-Baptiste Madeleine. 2011. “Gliese 581d is the first discovered terrestrial-mass exoplanet in the habitable zone.” The Astrophysical Journal Letters 733 (2) (May 12): L48. doi:10.1088/2041-8205/733/2/l48. http:// dx.doi.org/10.1088/2041-8205/733/2/L48.

Published Version doi:10.1088/2041-8205/733/2/L48

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:27846843

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Open Access Policy Articles, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#OAP Gliese 581d is the first discovered terrestrial-mass exoplanet in the habitable zone

Robin D. Wordsworth,1∗ Fran¸coisForget,1 Franck Selsis,2,3 Ehouarn Millour,1 Benjamin Charnay,1 Jean-Baptiste Madeleine1 1Laboratoire de M´et´eorologieDynamique, Institut Pierre Simon Laplace, Paris, France 2CNRS, UMR 5804, Laboratoire d’Astrophysique de Bordeaux, 2 rue de l’Observatoire, BP 89, F-33271 Floirac Cedex, France 3Universit´ede Bordeaux, Observatoire Aquitain des Sciences de l’Univers, 2 rue de l’Observatoire, BP 89, F-33271 Floirac Cedex, France

ABSTRACT It has been suggested that the recently discovered exoplanet GJ581d might be able to support liquid due to its relatively low mass and orbital distance. However, GJ581d receives 35% less stellar energy than and is probably locked in tidal resonance, with extremely low insolation at the poles and possibly a permanent night side. Under such conditions, it is unknown whether any habitable climate on the would be able to withstand global glaciation and / or atmospheric collapse. Here we present three-dimensional climate simulations that demonstrate GJ581d will have a stable and surface liquid water for a wide range of plausible cases, making it the first confirmed super- (exoplanet of 2-10 Earth ) in the habitable zone. We find that with over 10 bar CO2 and varying amounts of background gas (e.g., ◦ N2) yield global mean above 0 C for both land and ocean-covered surfaces. Based on the emitted IR radiation calculated by the model, we propose observational tests that will allow these cases to be distinguished from other possible scenarios in the future. Subject headings: —planet- interactions— and satellites: atmospheres— techniques: spectroscopic

1. Introduction the team responsible for finding the other four planets in the system (Kerr 2010; Tuomi 2011). The local (20.3 ly from For the moment, therefore, GJ581g remains un- the , M = 0.31MSun, L = 0.0135LSun, spec- confirmed. tral type M3V) (Hawley et al. 1997) has received arXiv:1105.1031v1 [astro-ph.EP] 5 May 2011 GJ581d, in constrast, which was first discov- intense interest over the last decade due to the low ered in 2007 and has a minimum mass between mass discovered around it. As of early 5.6 and 7.1 M , has now been robustly con- 2011 it has been reported to host up to six planets Earth firmed by (RV) observations (Udry (Udry et al. 2007; Mayor et al. 2009; Vogt et al. et al. 2007; Mayor et al. 2009; Vogt et al. 2010). 2010). One of these, GJ581g, was announced in Due to its greater distance from the host star, September 2010 and estimated to be in the hab- GJ581d was initially regarded as unlikely to have itable zone (the orbital range in which a planet’s surface liquid water unless strong warming mech- atmosphere can warm the surface sufficiently to al- anisms due to e.g., CO clouds (Forget and Pier- low surface liquid water) (Kasting et al. 1993; Pier- 2 rehumbert 1997; Selsis et al. 2007) were present in rehumbert 2011). However, its discovery has been its atmosphere. Recently, simple one-dimensional strongly disputed by other researchers, including

1 radiative-convective studies (Wordsworth et al. Restricting the composition of the atmosphere to 2010b; von Paris et al. 2010; Kaltenegger et al. two species in this way allows us to determine con- 2011) have suggested that a dense atmosphere servative conditions for habitability, as it neglects could provide a significant greenhouse effect on the warming due to other greenhouse gases like GJ581d. However, the planet’s tidal evolution CH4 or buffer gases like N2 or Ar (von Paris et al. poses a key problem for its habitability. 2010; Goldblatt et al. 2009; Li et al. 2009). As it is likely either in a pseudo-synchronous The simulations were performed using a new state with a rotation period that is a function of type of GCM that we developed specifically for ex- the eccentricity, or in spin- resonance like oplanet and paleoclimate studies. It uses radiative in our (Leconte et al. 2010; transfer data generated directly from high resolu- Heller et al. 2011), GJ581d should have extremely tion spectra, which allows the accurate simulation low insolation at its poles and possibly a perma- of climates for essentially any atmospheric cock- nent night side. Regions of low or zero insolation tail of gases, aerosols and clouds for which optical on a planet can act as cold traps where volatiles data exists. The dynamical core of the model was such as H2O and CO2 freeze out on the surface. adapted from the LMD Mars GCM, which uses A few previous studies (Joshi et al. 1997; Joshi an enstrophy and total angular momentum con- 2003) have examined atmospheric collapse in 3D serving finite difference scheme (Sadourny 1975; with simplified radiative transfer, but only for Forget et al. 1999). Scale-selective hyperdiffusion Earth-like atmospheric pressures or lower (0.1 to was used in the horizontal plane for stability. The 1.5 bar). For low values of stellar insolation and planetary boundary layer was parameterised us- large planetary radii, even dense CO2 atmospheres ing the method of Mellor and Yamada (1982) to will be prone to collapse, which could rule out a calculate turbulent mixing, with the latent heat of stable water cycle altogether for a super-Earth H2O also taken into account in the surface tem- like GJ581d. To conclusively evaluate whether perature calculations when necessary. A standard −2 GJ581d is in the habitable zone, therefore, three- roughness coefficient of z0 = 1 × 10 m was used dimensional simulations using accurate radiative for both rocky and ocean surfaces for simplicity, transfer are necessary. although we verified that our results were insensi- Here we present global climate model (GCM) tive to variations in this parameter. Spatial reso- simulations we performed to assess this issue. In lution of 32×32×20 in longitude, latitude and al- Section 2 we describe the model used. In Section titude was used for all simulations; we performed 3 we describe our results, and in Section 4 we dis- one comparison test at the highest rotation rate cuss implications and propose future observational with 64 × 64 × 20 resolution and found that the tests for the simulated habitable scenarios. differences were small. Our radiative transfer scheme was similar to 2. Method that we developed previously for one-dimensional simulations (Wordsworth et al. 2010a,b). For a In our simulations we made the initial hypoth- given mixture of atmospheric gases, we computed esis that GJ581d has a climate dominated by the high resolution spectra over a range of tempera- greenhouse effects of CO2 and / or H2O, as is tures, pressures and gas mixing ratios. For this the case for all rocky planets with atmospheres study we used a 6 × 9 × 7 , pressure in the Solar System (, Earth and Mars). To and H2O volume mixing ratio grid with values T = assess the influence of water on the climate in- {100, 150,..., 350} K, p = {10−3, 10−2,..., 105} dependently, we considered two classes of initial −7 −6 −1 mbar and qH2O = {10 , 10 ,..., 10 }, respec- condition: a rocky planet with no water, and an tively. The correlated-k method was used to pro- ocean planet, where the surface is treated as an duce a smaller database of coefficients suitable infinite water source. CO2 was taken as the pri- for fast calculation in a GCM. The model used mary constituent of the atmosphere and H2O was 38 spectral bands in the longwave and 36 in the allowed to vary freely, with surface ice / liquid shortwave, and sixteen points for the g-space in- and cloud formation (including radiative effects) tegration, where g is the cumulated distribution taken into account for either gas when necessary. function of the absorption data for each band.

2 In most simulations CO2 was assumed to be the Manabe and Wetherald (1967). Precipitation of main constituent of the atmosphere, except the H2O due to coagulation was also included using a locally habitable ice planet experiments (see Sec- simple threshold parameterisation (Emanuel and tion 3), where N2 was used. CO2 collision-induced Ivkovi-Rothman 1999). absorption was included using a parameterisation On the surface, the local varied accord- based on the most recent theoretical and experi- ing to the composition (rocky, ocean, CO2 or H2O mental studies (Wordsworth et al. 2010a; Gruszka ice; see Table 1). In the wet simulations, ice for- and Borysow 1998; Baranov et al. 2004). For the mation (melting) was assumed to occur when the stellar spectrum, we used the Virtual Planet Lab- surface temperature was lower (higher) than 273 oratory AD Leo data (Segura et al. 2005). AD K, and temperature changes due to the latent heat Leo is a M-class star with effective temperature of fusion were taken into account. In all cases, Teff = 3400 K, which is acceptably close to the the simulations were run until collapse/glaciation most recent estimates of Teff = 3498 ± 56 K occurred or steady states of thermal equilibrium (von Braun et al. 2011) for the purposes of cli- were reached. The time taken to reach ther- mate modelling. A two-stream scheme (Toon et al. mal equilibrium can be estimated from the atmo- 1989) was used to account for the radiative ef- spheric radiative relaxation timescale (Goody and fects of both clouds and Rayleigh scattering, as in Yung 1989) Wordsworth et al. (2010b). cpps τr = 3 , (1) In the water cycle and cloud modelling, care σgTe

was taken to ensure that the parameterisations where cp, ps, g, Te and σ are the specific heat ca- used were based on physical principles and not pacity of the atmosphere, the mean surface pres- tuned to Earth-specific conditions. When this sure, the surface , the atmospheric emis- was not possible (as for e.g., the density of con- sion temperature and the Stefan-Boltzmann con- densable cloud nuclei in the atmosphere Nc), we stant, respectively. Taking Te = 200 K, cp = 850 −1 −1 −2 tested the sensitivity of our results to variations in JK kg , ps = 30 bar and g = 16.6 m s those parameters. Three tracer species were used yields τr ∼ 4000 Earth days (60 GJ581d ). in our simulations: CO2 ice, H2O ice and H2O This was close to the timescales we observed in vapour. Tracers were freely advected in the at- the model by plotting time series of mean surface mosphere, subject to changes due to sublimation temperature. / evaporation and condensation and interaction We used the minimum mass for GJ581d given with the surface. For both gases, condensation was by Mayor et al. (2009) instead of the smaller value assumed to occur when the atmospheric temper- (Mmin = 5.6MEarth) proposed by Vogt et al. ature dropped below the saturation temperature. (2010). We took the actual mass of GJ581d to Local mean CO2 and H2O cloud particle sizes were ◦ be M = Mmin/sin 60 = 8.2MEarth, given that determined from the amount of condensed mate- the statistically most probable value for the in- rial and the density of condensable nuclei Nc. This clination angle is 60◦. Radius and gravity for 5 −1 parameter was set to 10 kg in most of our sim- rocky and ocean / ice cases (Table 1) were then ulations; we tested the effect of varying it over the determined from theoretical models (Sotin et al. 4 6 −1 range 10 to 10 kg and found that the maxi- 2007). In the latter case, the assumed bulk com- mum difference in mean surface temperature after position of the planet was 50% H2O. Model tests 60 orbits was less than 5 K. As a further test of using M = Mmin and M = 1.6Mmin (the latter the robustness of our results, we also performed value comes from dynamical stability considera- some tests with cloud radiative effects removed al- tions (Mayor et al. 2009)) did not reveal significant together (see Section 3). differences from our main results. Ice particles of both species were sedimented To produce emission spectra, we recorded top- according to Stokes law (Forget et al. 1999). Be- of-atmosphere longitude-latitude maps of outgoing low the stratosphere, adjustment was used to relax fluxes computed by the GCM over one orbit. In- temperatures due to convection and / or conden- clination angle of the orbit relative to the observer sation of CO2 and H2O. For H2O, moist and large- was assumed to be 60◦. An isotropic (Lambertian) scale convection were taken into account following distribution of specific intensities at the top of

3 the atmosphere was assumed; comparison with a sumptions, due to the scattering of radi- line-by-line radiative transfer code at wavelengths ation from the ground (Forget and Pierrehumbert where the limb-darkening was most pronounced 1997). To check the robustness of our results, we revealed that the disk-integrated flux error due to also performed tests with the (unrealistic) assump- this effect was below 5%. tion of no CO2 cloud radiative effects, and found that the atmospheres were colder, but still stable 3. Results above around 10 bar. In the ocean planet simulations, we neglected We performed simulations with 5, 10, 20 and all oceanic horizontal heat transport. This as- 30 bar atmospheric pressure and 1:1, 1:2 and 1:10 sumption led us to overestimate global tempera- orbit-rotation resonances for both rocky and ocean ture differences and hence the probability of atmo- planets (see Table 1). Eccentricity was set to zero spheric collapse, in keeping with the aim of a con- in the simulations we present here, with the stellar servative habitability estimate. We found that at- flux set to the true value at a = 0.22 and not in- mospheric H O vapour greatly increased warming, creased to account for orbital averaging as in our 2 while H O cloud formation low in the atmosphere 1D study (Wordsworth et al. 2010b). However, we 2 tended to cool the planet by increasing the plan- also performed tests with e = 0.38 and found that etary albedo (Fig. 3). This led to a transition in eccentricity was not critical to the results. Simi- the climate as total pressure increased. At 10 bar larly, we assumed zero obliquity to assess the like- and below, cooling effects dominated and runaway lihood of atmospheric collapse in the most severe glaciation occurred, followed by atmospheric col- cases, but sensistivity tests showed that its influ- lapse. From 20 bar, the positive feedback of water ence on the climate was second order at the high vapour on greenhouse warming raised mean tem- pressures where the atmosphere remained stable. peratures significantly compared with the rocky CO2 gas had a powerful greenhouse warming case (Fig. 2b). In the intermediate region, climate effect in our simulations because GJ581d orbits a stability was difficult to assess as surface temper- red dwarf star. In the Solar System, Rayleigh scat- ature trends were extremely small (as low as a few tering reflects the bluer incoming starlight much K per 100 orbits in some cases). There, our re- more effectively, and hence a planet receiving the sults were somewhat sensitive to our microphysical same flux as GJ581d would be uninhabitable for assumptions; for example, lowering the precipita- any CO pressure. In our rocky simulations, for 2 tion threshold l0 (see Table 1) resulted in optically pressures below ∼10 bar the atmosphere was in- thinner H2O clouds and stable, cooler climates deed unstable, and began to condense on the dark at lower total pressures. Nonetheless, at higher side and / or poles of the planet (Fig. 2a). How- pressures we found stable, hot climates with little ever, for denser atmospheres, we found that hori- global temperature variations even in the tidally zontal heat transport and greenhouse warming be- locked cases, regardless of the choice of microphys- came effective enough to remove the threat of col- ical parameters. lapse and allow surface temperatures above the We investigated the effect of starting with a melting point of water (Fig. 1). Rotation rate surface covered by H O ice. For the dense atmo- affected the atmospheric stability through both 2 spheres where the climate was stable, this only the insolation (the planet has permanent and weakly changed the planetary albedo, and the ice night sides in the most extreme 1:1 resonance case) melted for pressures of 20 bar or more. Hence and the horizontal heat transport (faster rotat- the runaway (H O) glaciation that is possible on ing atmospheres were less efficient at transporting 2 Earth (Budyko 1969) would be unlikely to occur heat polewards). on an ocean-planet GJ581d. Our results here con- Even in the stable simulations, some CO2 usu- trast with those of Spiegel et al. (2009), who used ally condensed in the middle atmosphere (Fig. 3), a 1D energy balance model to investigate the ef- leading to CO2 ice cloud formation as occurs on fects of obliquity on , be- present-day Mars (Montmessin et al. 2007). CO2 cause they neglected the dependence of planetary clouds typically had a net warming effect (up to albedo on the total pressure in dense atmospheres. 12 K), depending on the cloud microphysical as- We also investigated the possibility that an ice-

4 covered, tidally locked GJ581d with permanent means that it can retain an atmosphere more eas- day and night sides could be locally habitable on ily than Earth or Venus, but the increased XUV the day side due to partial melting. However in and ion fluxes from M-class early in their his- this scenario, the day side only stayed warm when tory means that removal processes can be much the atmosphere was thin and hence inefficient at more efficient than in the Solar System (Tanigawa transporting heat. As a result, the planet’s dark and Ikoma 2007; Tian 2009). The escape of at- side became cold enough for the collapse of even mospheric hydrogen depends sensitively on both + an N2 atmosphere. An ice planet with only a thin XUV flux levels and radiative cooling by H3 ions, sublimation-driven H2O atmosphere could have both of which are poorly constrained for GJ581d dayside temperatures above 273 K and contin- (Selsis et al. 2007; Koskinen et al. 2007). More- ual transport of H2O to the dark side, but the over, stripping by the stellar wind, which does not low atmospheric pressure would preclude liquid discriminate between light and heavy atoms, could water except in extremely limited sub-surface re- have removed large amounts of any gas (including gions. Clearly, dense, stable atmospheres offer CO2) (Lammer et al. 2007). Detailed modelling better prospects for habitability. of atmospheric escape may be able to constrain the effects of these processes on the atmosphere 4. Discussion further. Unlike the majority of the Kepler planetary Are CO partial pressures of over 10 bar a real- 2 candidates, Gliese 581d is relatively close to Earth, istic possibility for GJ581d? Given the planet’s as- so in the future it will be possible to establish sumed gravity, this corresponds to a CO column 2 which scenario applies to it through direct spectro- of only 4 to 6 bars on Earth. Even though Venus scopic observations. To determine the sensitivity and Earth are smaller, their total CO invento- 2 required for this, we used the outgoing longwave ries scaled by planetary mass are around 10-100 radiation (OLR) from our simulations to produce times greater than this. On Venus, it is thought synthetic emission spectra for the habitable cli- that a large fraction of this inventory is in the at- mates just discussed, along with those for a planet mosphere (Bullock and Grinspoon 1996), while on with no atmosphere (Fig. 4). We did not calculate Earth, the atmospheric partial pressure is regu- emission spectra for the H -He dominated case, lated over geological timescales by the carbonate- 2 but equilibrium chemistry calculations1 allowed us silicate cycle (Walker et al. 1981) via plate tecton- to establish that for an atmosphere where H is ics. The geophysics of GJ581d is unknown, but if 2 the dominant species by mass, the abundance of a similar mechanism were present there, its atmo- CO should be below the limit of detectability. spheric CO would stabilize above the level needed 2 2 For a rocky / icy planet where volatiles like CO to maintain a liquid water cycle by negative feed- 2 have collapsed on the surface and the atmosphere back. is thin or non-existent, the flux variations over one CO2-H2O-(buffer gas) atmospheres are clearly orbit will be large (grey region in Fig. 4). Obser- only a subset of all the possible scenarios for vations of CO2 and H2O absorption bands with GJ581d. If small quantities of additional green- low phase variations (red / blue regions in Fig. house gases such as CH4 or SO2 were present, 4) would hence be a strong indicator of the kind they would increase warming further. However, if of stable, dense habitable atmospheres discussed these gases were chemically unstable, they would in this paper. From Fig. 4, the flux sensitivity need to be continually emitted by e.g., volcanic necessary for this in the infrared will be of order or biological sources to have a long-term effect on 10−21 W m−2 µm−1, corresponding to a planet the climate. More drastically, the planet could / star ratio lower than 10−6 for wavelengths < have a thick H2-He envelope like Uranus or Nep- 20 µm. This is beyond the capabilities of current tune, or its atmosphere could have been removed space and ground-based observatories, but it will entirely by increased stellar activity early in the of the host star. Both scenarios would ex- 1These were performed using a code developed by R. clude surface liquid water (the former through an Bounaceur (LRGP, CNRS, Nancy, France), based on an excess of surface pressure). GJ581d’s large mass algorithm to minimize the free energy of a given mixture of gases (Reynolds 1986).

5 become possible with future improvements in in- Physics, 93:1007–1016, 1998. doi: 10.1080/ strumental technology. 002689798168709. S. L. Hawley, J. E. Gizis, and N. I. Reid. The F.S. acknowledges support from the Euro- Palomar/MSU Nearby Star Spectroscopic Sur- pean Research Council (Starting Grant 209622: vey.II.The Southern M Dwarfs and Investiga- E3ARTHs). tion of Magnetic Activity. Astronomical Jour- nal, 113:1458–+, April 1997. doi: 10.1086/ REFERENCES 118363. Y. I. Baranov, W. J. Lafferty, and G. T. Fraser. Infrared spectrum of the continuum and dimer R. Heller, J. Leconte, and R. Barnes. Tidal obliq- absorption in the vicinity of the O2 vibrational uity evolution of potentially habitable planets. fundamental in O2/CO2 mixtures. J. Mol. Astronomy and Astrophysics, 527:accepted for Spectrosc., 228:432–440, December 2004. doi: publication, 2011. doi: 10.1051/0004-6361/ 10.1016/j.jms.2004.04.010. 201015809. M. I. Budyko. The effect of solar radiation vari- M. Joshi. Climate Model Studies of Synchronously ations on the climate of the earth. Tellus, 21: Rotating Planets. Astrobiology, 3:415–427, 611–619, 1969. 2003. doi: 10.1089/153110703769016488. M. A. Bullock and D. H. Grinspoon. The stability M. M. Joshi, R. M. Haberle, and R. T. of climate on Venus. Journal of Geophysical Reynolds. Simulations of the Atmospheres of Research, 101:7521–7530, 1996. doi: 10.1029/ Synchronously Rotating Terrestrial Planets Or- 95JE03862. biting M Dwarfs: Conditions for Atmospheric Collapse and the Implications for Habitability. K. A. Emanuel and M. Ivkovi-Rothman. Develop- Icarus, 129:450–465, 1997. doi: 10.1006/icar. ment and Evaluation of a Convection Scheme 1997.5793. for Use in Climate Models. Journal of Atmo- spheric Sciences, 56:1766–1782, 1999. L. Kaltenegger, A. Segura Peralta, and S. Mo- hanty. Model Spectra of the First Potentially F. Forget and R. T. Pierrehumbert. Warming Habitable Super-Earth - Gl581d. Accepted Early Mars with Clouds That for publication in The Astrophysical Journal, Scatter Infrared Radiation. Science, 278:1273– March 2011. 1276, November 1997. J. F. Kasting, D. P. Whitmire, and R. T. F. Forget, F. Hourdin, R. Fournier, C. Hourdin, Reynolds. Habitable Zones around Main Se- O. Talagrand, M. Collins, S. R. Lewis, P. L. quence Stars. Icarus, 101:108–128, January Read, and J.-P. Huot. Improved general circu- 1993. doi: 10.1006/icar.1993.1010. lation models of the Martian atmosphere from the surface to above 80 km. Journal of Geo- Richard A. Kerr. First goldilocks exoplanet may physical Research, 104:24155–24176, 1999. not exist. Science, 330(6003):433, 2010. doi: 10. 1126/science.330.6003.433. URL http://www. C. Goldblatt, M. W. Claire, T. M. Lenton, sciencemag.org. A. J. Matthews, A. J. Watson, and K. J. Zahnle. Nitrogen-enhanced greenhouse warm- T. T. Koskinen, A. D. Aylward, and S. Miller. A ing on early Earth. Nature Geoscience, 2:891– stability limit for the atmospheres of giant ex- 896, December 2009. doi: 10.1038/ngeo692. trasolar planets. Nature, 450:845–848, Decem- ber 2007. doi: 10.1038/nature06378. R. M. Goody and Y. L. Yung. Atmospheric radi- ation : theoretical basis. 1989. H. Lammer, H. I. M. Lichtenegger, Y. N. Kulikov, J.-M. Grießmeier, N. Terada, N. V. Erkaev, M. Gruszka and A. Borysow. Computer simula- H. K. Biernat, M. L. Khodachenko, I. Ribas, tion of the far infrared collision induced ab- T. Penz, and F. Selsis. Coronal Mass Ejection sorption spectra of gaseous CO2. Molecular (CME) Activity of Low Mass M Stars as An

6 Important Factor for The Habitability of Ter- W. C. Reynolds. The Element Potential Method restrial Exoplanets. II. CME-Induced Ion Pick for Chemical Equilibrium Analysis: Implemen- Up of Earth-like Exoplanets in Close-In Habit- tation in the Interactive Program STANJAN. able Zones. Astrobiology, 7:185–207, February Department of Mechanical Engineering, Stan- 2007. doi: 10.1089/ast.2006.0128. ford University, 1986. J. Leconte, G. Chabrier, I. Baraffe, and B. Lev- R. Sadourny. The Dynamics of Finite-Difference rard. Is sufficient to explain Models of the Shallow-Water Equations. Jour- bloated exoplanets? Consistent calculations ac- nal of Atmospheric Sciences, 32:680–689, 1975. counting for finite initial eccentricity. Astron- omy and Astrophysics, 516:A64+, June 2010. A. Segura, J. F. Kasting, V. Meadows, M. Co- doi: 10.1051/0004-6361/201014337. hen, J. Scalo, D. Crisp, R. A. H. Butler, and G. Tinetti. from Earth-Like King-Fai Li, Kaveh Pahlevan, Joseph L. Planets Around M Dwarfs. Astrobiology, 5:706– Kirschvink, and Yuk L. Yung. Atmo- 725, December 2005. doi: 10.1089/ast.2005.5. spheric pressure as a natural climate regulator 706. for a with a biosphere. Proceedings of the National Academy of Sci- F. Selsis, J. F. Kasting, B. Levrard, J. Paillet, ences, 106(24):9576–9579, June 2009. doi: I. Ribas, and X. Delfosse. Habitable planets 10.1073/pnas.0809436106. around the star Gliese 581? Astron. Astro- phys., 476:1373–1387, December 2007. S. Manabe and R. T. Wetherald. Thermal Equi- librium of the Atmosphere with a Given Dis- C. Sotin, O. Grasset, and A. Mocquet. Mass ra- tribution of Relative Humidity. J. Atmos. Sci., dius curve for extrasolar Earth-like planets and 24:241–259, May 1967. ocean planets. Icarus, 191:337–351, November 2007. M. Mayor, X. Bonfils, T. Forveille, X. Delfosse, S. Udry, J.-L. Bertaux, H. Beust, F. Bouchy, D. S. Spiegel, K. Menou, and C. A. Scharf. Hab- C. Lovis, F. Pepe, C. Perrier, D. Queloz, and itable Climates: The Influence of Obliquity. N. C. Santos. The HARPS search for south- The Astrophysical Journal, 691:596–610, Jan- ern extra-solar planets. XVIII. An Earth-mass uary 2009. doi: 10.1088/0004-637X/691/1/596. planet in the GJ 581 . Astron- T. Tanigawa and M. Ikoma. A Systematic Study omy and Astrophysics, 507:487–494, November of the Final Masses of Gas Giant Planets. The 2009. Astrophysical Journal, 667:557–570, September G. L. Mellor and T. Yamada. Development of 2007. doi: 10.1086/520499. a Turbulence Closure Model for Geophysical F. Tian. Thermal Escape from Super Earth Atmo- Fluid Problems. Reviews of Geophysics, 20: spheres in the Habitable Zones of M Stars. The 851–875, 1982. Astrophysical Journal, 703:905–909, September F. Montmessin, B. Gondet, J.-P. Bibring, 2009. doi: 10.1088/0004-637X/703/1/905. Y. Langevin, P. Drossart, F. Forget, and O. B. Toon, C. P. McKay, T. P. Ackerman, and T. Fouchet. Hyperspectral imaging of convec- K. Santhanam. Rapid calculation of radiative tive CO2 ice clouds in the equatorial meso- heating rates and photodissociation rates in in- sphere of Mars. JGR (Planets), 112:11–+, homogeneous multiple scattering atmospheres. November 2007. JGR, 94:16287–16301, November 1989. R. T. Pierrehumbert. A Palette of Climates for M. Tuomi. Bayesian re-analysis of the radial ve- . The Astrophysical Journal Let- locities of Gliese 581. Evidence in favour of ters, 726:L8+, January 2011. doi: 10.1088/ only four planetary companions. Astronomy 2041-8205/726/1/L8. and Astrophysics, 528:L5+, April 2011. doi: 10.1051/0004-6361/201015995.

7 S. Udry, X. Bonfils, X. Delfosse, T. Forveille, Table 1: Standard parameters used in the climate M. Mayor, C. Perrier, F. Bouchy, C. Lovis, simulations. F. Pepe, D. Queloz, and J.-L. Bertaux. The Stellar L [L ] 0.0135 HARPS search for southern extra-solar planets. Sun Orbital semi-major axis a [AU] 0.22 XI. Super- (5 and 8 M+) in a 3-planet e 0.0 system. Astron. Astrophys., 469:L43–L47, July Obliquity φ 0.0 2007. doi: 10.1051/0004-6361:20077612. Tidal resonance n 1, 2, 10 S. S. Vogt, R. P. Butler, E. J. Rivera, N. Haghigh- Initial atmospheric pressure ps [bars] 5-30 ipour, G. W. Henry, and M. H. Williamson. Radius (rocky) r [rEarth] 1.8 The Lick-Carnegie Exoplanet Survey: A 3.1 Radius (ocean) r [rEarth] 2.3 MEarth Planet in the Habitable Zone of the (rocky) g [m s−2] 25.0 Nearby M3V Star Gliese 581. The Astrophysi- Surface gravity (ocean) g [m s−2] 16.6 cal Journal, 723:954–965, 2010. Surface albedo (rocky) As 0.2 Surface albedo (ocean) As 0.07 K. von Braun, T. S. Boyajian, S. R. Kane, G. T. Surface albedo (ice) As 0.6 −2 van Belle, D. R. Ciardi, M. L´opez-Morales, Surface roughness coefficient z0 [m] 1 × 10 −1 H. A. McAlister, T. J. Henry, W.-C. Jao, A. R. Precipitation threshold l0 [kg kg ] 0.001 −1 5 Riedel, J. P. Subasavage, G. Schaefer, T. A. ten Number of cloud condensation nuclei Nc [kg ] 1 × 10 Brummelaar, S. Ridgway, L. Sturmann, J. Stur- mann, J. Mazingue, N. H. Turner, C. Farring- ton, P. J. Goldfinger, and A. F. Boden. As- trophysical Parameters and Habitable Zone of the Exoplanet Hosting Star GJ 581. The As- trophysical Journal Letters, 729:L26+, March 2011. doi: 10.1088/2041-8205/729/2/L26. P. von Paris, S. Gebauer, M. Godolt, J. L. Grenfell, P. Hedelt, D. Kitzmann, A. B. C. Patzer, H. Rauer, and B. Stracke. The ex- trasolar planet Gliese 581d: a potentially hab- itable planet? Astronomy and Astrophysics, 522:A23+, November 2010. doi: 10.1051/ 0004-6361/201015329. J. C. G. Walker, P. B. Hays, and J. F. Kasting. A negative feedback mechanism for the long-term stabilization of the earth’s surface temperature. Journal of Geophysical Research, 86:9776–9782, October 1981. R. Wordsworth, F. Forget, and V. Eymet. In- frared collision-induced and far-line absorption in dense CO2 atmospheres. Icarus, 210:992– 997, December 2010a. doi: 10.1016/j.icarus. 2010.06.010. R. Wordsworth, F. Forget, F. Selsis, J.-B. Madeleine, E. Millour, and V. Eymet. Is Gliese 581d habitable? Some constraints from radiative-convective climate modeling. Astron- omy and Astrophysics, 522:A22+, 2010b. doi: This 2-column preprint was prepared with the AAS LATEX 10.1051/0004-6361/201015053. macros v5.2.

8 Fig. 3.— Latitude-altitude map of H2O (blue) and CO2 (red) cloud coverage for a 20-bar ocean sim- ulation with 1:2 resonance, averaged in longitude and over one orbit. While the cloud deck altitudes Fig. 1.— Surface temperature snapshots after 60 were similar for all simulations, the latitudinal dis- orbits integration time for rocky planet simula- tribution depended strongly on the atmospheric tions with a) 1:1, b) 1:2 and c) 1:10 tidal resonance dynamics and hence on the rotation rate and to- and a 20-bar CO2 atmosphere. tal pressure.

Fig. 4.— Infrared emission spectra for an ocean / rocky GJ581d with a 20-bar atmosphere (blue / red) and a rocky GJ581d with no atmosphere (grey). Thickness of the lines corresponds to the Fig. 2.— Surface temperature vs. pressure for maximum / minimum difference in flux over one a) rocky and b) ocean planet simulations. Red, orbit, for the most probable observation angle of green and blue markers represent tidal resonances 60◦. The maximum (minimum) surface temper- of 1:1, 1:2 and 1:10. Points are separated for ature in the airless case is 271 K (37 K). For clarity; simulations were performed at 5, 10, 20 the cases with atmospheres, emission of the ocean and 30 bar for each all cases. Crosses and error planet is higher at most wavelengths due to the bars show mean and maximum / minimum tem- increased planetary radius. The major absorption peratures (sampled over one orbit and across the features are labelled by molecule / process (CO2- planet’s surface) in stable cases, while circles in- CO2 corresponds to CO2 collision-induced absorp- dicate simulations where the atmosphere began to tion). collapse or runaway glaciation occurred.

9