Atmospheric Circulation and Cloud Evolution on the Highly Eccentric Extrasolar Planet HD 80606b
N. K. Lewis, V. Parmentier, T. Kataria, J. de Wit, A. P. Showman, J. J. Fortney, M. S. Marley
11 Space Telescope Science Institute, Baltimore, MD 21218, USA;
Draft version May 8, 2018
Preprint typeset using L A TEX style emulateapj v. 01/23/15
ATMOSPHERIC CIRCULATION AND CLOUD EVOLUTION ON THE HIGHLY ECCENTRIC EXTRASOLARPLANET HD 80606B
Nikole K. Lewis , Vivien Parmentier , Tiffany Kataria , Julien de Wit , Adam P. Showman , Jonathan J.Fortney , Mark S. Marley Draft version May 8, 2018
ABSTRACTObservations of the highly-eccentric (e ∼ Spitzer have provided someof best probes of the physics at work in exoplanet atmospheres. By observing HD 80606b dur-ing its periapse passage, atmospheric radiative, advective, and chemical timescales can be directlymeasured and used to constrain fundamental planetary properties such as rotation period, tidal dis-sipation rate, and atmospheric composition (including aerosols). Here we present three-dimensionalgeneral circulation models for HD 80606b that aim to further explore the atmospheric physics shap-ing HD 80606b’s observed Spitzer phase curves. We find that our models that assume a planetaryrotation period twice that of the pseudo-synchronous rotation period best reproduce the phase vari-ations observed for HD 80606b near periapse passage with
Spitzer . Additionally, we find that therapid formation/dissipation and vertical transport of clouds in HD 80606b’s atmosphere near peri-apse passage likely shapes its observed phase variations. We predict that observations near periapsepassage at visible wavelengths could constrain the composition and formation/advection timescalesof the dominant cloud species in HD 80606b’s atmosphere. The time-variable forcing experienced byexoplanets on eccentric orbits provides a unique and important window on radiative, dynamical, andchemical processes in planetary atmospheres and an important link between exoplanet observationsand theory.
Keywords: atmospheric effects - methods: numerical - planets and satellites: general - planets andsatellites: individual (HD 80606b) INTRODUCTION
HD 80606b is a ‘hot Jupiter’ ( M p =3.94 M J , R p =0.98 R J ) on an extremely eccentric orbit ( e =0.9332) (Pont et al.2009). First discovered by Naef et al. (2001) using radial velocity observations, HD 80606b was later determined to beeclipsed by (Laughlin et al. 2009) and transit (Moutou et al. 2009) its host star. The extreme eccentricity of the orbitof HD 80606b is the result of “Kozai Migration” in the presence of a binary star system (Wu & Murray 2003). Duringits ∼
111 day orbit, HD 80606b experiences extreme ( ∼ × ) shifts in the amount of incident flux it receives from itshost star from apoapse to the periapse of its orbit. These variations in stellar insolation are likely to cause not onlydramatic changes in the thermochemical structure of the planet, but also in global scale wind and cloud patterns.Observations with the Spitzer Space Telescope (Werner et al. 2004) presented in de Wit et al. (2016) and Laughlinet al. (2009) have shown that HD 80606b exhibits large variations in its flux as a function of orbital phase nearperiapse passage. Such flux variations were predicted by atmospheric models for exoplanets on eccentric orbits, whichexhibit significant variations in atmospheric temperature and wind speeds that directly inform atmospheric radiativeand dynamical timescales (Langton & Laughlin 2008; Lewis et al. 2010; Cowan & Agol 2011; Kataria et al. 2013;Lewis et al. 2014). Laughlin et al. (2009) were the first to directly measure the atmospheric radiative timescale for anexoplanet with their
Spitzer µ m observations of HD 80606b. Through their 30 hour observation of HD 80606b nearsecondary eclipse, when the planet passes behind its host star, and periapse passage, Laughlin et al. (2009) estimateda radiative timescale of 4.5 hours near the 8 µ m photosphere of the planet.Building on the original analysis of HD 80606b presented in Laughlin et al. (2009), de Wit et al. (2016) revisitedthe 8 µ m Spitzer observations given the recent evolution in data reduction techniques (e.g. Lewis et al. 2013) andperformed a combined analysis with HD 80606b observations taken in the 4.5 µ m channel of Spitzer in 2009. TheHD 80606b observations taken at 4.5 µ m provided a significantly longer temporal baseline around the periapse ofHD 80606b’s orbit, with 80 hours worth of data in total. This increased baseline allowed de Wit et al. (2016) to notonly estimate the radiative timescales in HD 80606b’s atmosphere near its 4.5 and 8 µ m photospheres ( ∼ ∼
40 hours.The interpretation of the HD 80606b phase variations observed by
Spitzer presented in Laughlin et al. (2009) andde Wit et al. (2016) relied on semi-analytic (e.g. Cowan & Agol 2011; de Wit et al. 2016) and two-dimensional Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, MD 21218, USA Department of Planetary Sciences and Lunar and Planetary Laboratory, The University of Arizona, Tucson, AZ 85721, USA Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Astronomy & Astrophysics, University of California, Santa Cruz, CA 95064, USA NASA Ames Research Center 245-3, Moffett Field, CA 94035, USA Sagan Postdoctoral Fellow a r X i v : . [ a s t r o - ph . E P ] J un Lewis et al. hydrodynamical models (Langton & Laughlin 2008). Such models allow for rapid exploration of how specific changesin model parameters might provide a better match to the observed atmospheric flux variations. However, such modelsdo not fully capture the interplay of radiative, dynamical, and chemical processes in a planet’s atmosphere. Here wepresent new atmospheric models for HD 80606b that more fully explore the physical processes at work in HD 80606b’satmosphere. Our models are three-dimensional (latitude, longitude, and pressure) and include full radiative transfercalculations that consider equilibrium chemistry processes in HD 80606b’s atmosphere throughout its highly eccentricorbit. We specifically explore how assumed rotation period plays a role in shaping global circulation patterns and howthe formation and evolution of clouds might play a role in shaping the observed flux variation of HD 80606b. Thefollowing sections outline the specifics of our modelling approach, explore the complex atmospheric physics at work inHD 80606b, and provide observational predictions that give further insights into the current
Spitzer observations andguide future atmospheric characterization observations of the HD 80606 system. MODELS
The results presented here for our atmospheric modeling effort rely on a combination of simulation tools that linkthe physical processes potentially at work in HD 80606b’s atmosphere and what can be observed. In the followingsections we present an overview of the global circulation, cloud, and observational phase curve models employed inthis study.
Global circulation model
In the simulations presented here we employ the Substellar and Planetary Atmospheric Radiation and Circulation(SPARC) model first presented in Showman et al. (2009) as applied to HD 189733b and HD 209458b and subsequentlyused in a number of other exoplanet studies (e.g. Lewis et al. 2010; Kataria et al. 2013; Lewis et al. 2014; Katariaet al. 2014, 2015; Showman et al. 2015; Kataria et al. 2016; Parmentier et al. 2016). At the core of the SPARC modelis the MITgcm (Adcroft et al. 2004), which solves the primitive equations with pressure as the vertical coordinate.We determine the amount of heating/cooling at each grid point by the divergence of radiative fluxes in each modellayer. Radiative fluxes are calculated using the two-stream non-gray radiative transfer model of Marley & McKay(1999). We use the correlated-k method with 8 k-coefficients inside each of our 11 wavelength bins. These simulationsof HD 80606b utilize the cubed-sphere grid (Adcroft et al. 2004) with a horizontal resolution of C16 ( ∼ ×
64 inlatitude and longitude) and a vertical pressure range from 200 bar to 0.20 mbar broken down into 39 layers with evenlog-pressure spacing with a top layer that extends to zero pressure.In the simulations presented here we assume an atmospheric composition of 1 × solar values and have excluded TiOand VO from the opacity tables since it is likely that these species are ‘cold trapped’ deep in the atmosphere (Fortneyet al. 2008). Clouds are neglected in the global circulation model. The effect of clouds is, however, considered inthe post-processing of the simulation. We define our planetary and stellar parameters using the values from Pontet al. (2009). The nominal rotation period of HD 80606b is determined to be 40.4761 hours assuming the pseudo-synchronous rotation relationship presented in Hut (1981). Because it is possible that HD 80606b’s rotation periodmight significantly deviate from this nominal value, we construct additional atmospheric models where the rotationperiod is assumed to be half (20.2380 hours) and twice (80.9522 hours) the nominal pseudo-synchronous rotation period.We initialize the model with wind speeds set to zero and each column of the grid assigned a pressure-temperature profilederived from one-dimensional radiative equilibrium calculations assuming that the planet is at periapse. The planetloses ‘memory’ of this initial temperature distribution fairly rapidly and we have found no significant difference betweensimulations started near periapse and those initialized near apoapse with the corresponding pressure-temperatureprofile.We find that in order to maintain numeric stability near periapse that the timestep used to solve the relevantmomentum and energy equations must be very short ( ∼ ∼
300 simulated days). Integrating our simulations beyond this point results in only small changes in planetarywind and thermal patterns that are confined to pressures well below photospheric pressures ( ∼
300 mbar).
Cloud modeling
As shown in Section 3, our global circulation model predicts that the dayside temperature of the planet can vary bymore than 700 K between periapse and apoapse in the observable portion of the atmosphere. As such, the photosphericlayers (1 bar to 1 mbar) can cross the condensation curves of several potential cloud species (e.g. Morley et al. 2012;Marley et al. 2013). To simulate the formation, dissipation and general evolution of clouds in HD 80606b’s atmosphere,we use the cloud model presented in Parmentier et al. (2016). In a given atmospheric column, a material is condensedif the temperature is cooler than its condensation temperature at a given pressure level in the atmospheric general tmospheric Circulation and Clouds of HD 80606b −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r )
600 600
800 800 (K) 60080010001200140016001800 −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r )
150 150 (m s −1 )5001000150020002500300035004000 −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r )
800 800 (K) 60080010001200140016001800 −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r ) (m s −1 )50010001500200025003000350040004500 −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r )
800 800 (K) 60080010001200140016001800 −1 −0.5 0 0.5 1 1.5 2 2.5 3 3.510 −3 −2 −1 time from periapse (Earth days) p r ess u r e ( b a r )
150 150
200 200
600 600
800 800 8001000 (m s −1 )50010001500200025003000350040004500 Figure 1.
Average temperature (left) and RMS horizontal velocity (right) as a function of time relative to periapse passage for ourhalf-nominal (top), nominal (middle), and twice-nominal (bottom) rotation period models. The temperatures and RMS velocities representaverages over latitude and longitude as a function of pressure. circulation model. Clouds are assumed to condensed with a uniform particle size of a , which is a free parameterranging from 0.1 µ m to 10 µ m. In this study of HD 80606b, we consider the following possible cloud species: MgSiO (enstatite), MnS (manganese sulfide), and Na S (sodium sulfide). Cloud species ZnS and KCl are not considered inthis study as they are fully evaporated from the planet’s dayside during periastron passage and are expected to form
Lewis et al. significantly less massive, and less opaque, clouds compared with MgSiO , MnS, and Na S (e.g. Fortney 2005).
Theoretical Phase Curve Calculation
To calculate theoretical phase curves, planetary flux as a function of time, we solve the two-stream radiative transferequations along the line of sight for each atmospheric column and for each planetary orbital position. In our radiativetransfer calculations, we consider absorption, emission and scattering. This method, similar to the calculations ofFortney et al. (2006), naturally takes into account geometrical effects such as limb darkening and the variation of thestellar and planetary orientations with respect to each other and a earth-based observer. The stellar flux is assumed tobe a collimated flux propagating in each atmospheric column with an angle equal to the angle between the local verticaland the direction of the star. We use 196 frequency bins ranging from 0.26 to 300 µ m and integrate the resultingoutgoing flux over a variety of observational bandpasses spanning the visible to infrared wavelengths (0.3-28 µ m).The gaseous opacities used to calculate the values of the k-coefficients are the same as the ones used for the globalcirculation model calculations. Rayleigh scattering by the gas and Mie scattering by the clouds are combined with thegaseous opacities to produce our theoretical spectra as a function of time for HD 80606b.In this study, cloud opacities are included in post-processing and not self-consistently included within the frameworkof the global circulation model. This is done to maintain the computational feasibility of our atmospheric models. Asingle SPARC model can require weeks to months of computational time in a cluster computing environment to reacha stable solution, largely limited by computational requirements of the fully non-gray radiative transfer calculations.Incorporation of cloud formation, evolution, and radiative feedback is not impossible within the framework of theSPARC model, but would currently require prohibitively long computation times. As shown by (Parmentier et al.2016), the cloud distribution in hot Juptiers is primarily determined by the thermal structure of the planet, withradiative feedback and the dynamical mixing of the clouds being secondary effects. Therefore, the cloud distributionswe derive here from our models should be a good approximation for the spatial and wavelength dependent opacitieswe could expect to shape HD 80606b’s phase curve. We are currently working on further optimization to the SPARCand cloud formation models to more robustly treat clouds without requiring highly tuned or parameterized schemes. RESULTS
The following sections overview the key results from our atmospheric modelling effort for HD 80606b. We firstpresent changes in the global scale winds and temperatures of HD 80606b as a function of both pressure and timefor our half-nominal, nominal, and twice-nominal rotation period models. We then compare the thermal patterns andwinds predicted for HD 80606b near the periapse and apoapse of its orbit from our half-nominal, nominal, and twice-nominal rotation period atmospheric models. We further focus on the thermal structure and vertical wind profilesas a function of longitude that develop in our models near periapse. Using the post-processing method described inSection 2.2, we predict the evolution of cloud coverage in HD 80606b. Finally, we present theoretical phase curvesderived from our simulations for each rotational period case assuming a range of cloud properties and compare themwith the
Spitzer observations at 4.5 and 8 µ m presented in de Wit et al. (2016). Global Scale Winds and Temperatures
Since our simulations are performed in three-dimensions, we can investigate changes in the overall temperaturesand wind speeds of HD 80606b as a function of depth in the atmosphere as well as time. Figure 1 presents globallyaveraged temperatures and root mean square (RMS) horizontal velocities as a function of pressure and time relativeto the periapse of HD 80606b’s orbit for each of our rotation period cases. The peak global average atmospherictemperatures near the 340 mbar level in our simulations occurs on average 11 hours after periapse passage. Nearthe 10 mbar level of our simulation where radiative timescales are shorter, peak temperatures are reached on average4.5 hours after periapse passage. In all cases peak wind speeds are achieved ∼ ∼ ∼
10 mbar. In the half-nominal rotationperiod case, the high latitude eastward jets are maintained through out the orbit, strengthening and weakening withthe amount of incident flux on the planet. tmospheric Circulation and Clouds of HD 80606b −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r ) −400 −400 −
800 800 (m s −1 )−400−2000200400600800 −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r )
200 200 200 (m s −1 )050100150200250300 −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r ) −800 −800−600 −600−400 −400−200 −
200 200200
400 400 (m s −1 )−1400−1200−1000−800−600−400−2000200400 −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r ) (m s −1 )050010001500 −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r ) −1200−1000 − −800 −800−600 −600 −600−400 −400 −400−200 − −2000 (m s −1 )−1200−1000−800−600−400−2000200400 −80 −60 −40 −20 0 20 40 60 8010 −3 −2 −1 latitude ( ° ) p r ess u r e ( b a r ) − (m s −1 )−2000200400600800100012001400 Figure 2.
Zonal-mean zonal winds for HD 80606b near periapse (left) and apoapse (right) for our models that assume the half-nominal(top), nominal (middle), and twice-nominal (bottom) rotation period for the planet. Apoapse and periapse occur at true anomalies ( f ) of180 ◦ and 0 ◦ respectively. The colorbar shows the strength of the zonally averaged winds in m s − . Contours are spaced by 100 m s − .Positive wind speeds are eastward, while negative wind speeds are westward. Note the significant change in the jet structure as a functionof orbital position. Lewis et al. −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 5006007008009001000110012001300 −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 420430440450460470480490500510520 −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 5006007008009001000110012001300 −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 420430440450460470480490500510520 −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 5006007008009001000110012001300 −150 −100 −50 0 50 100 150−80−60−40−20020406080 longitude ( ° ) l a t i t ud e ( ° ) (K) 420430440450460470480490500510520 Figure 3.
Temperature (colorscale) and horizontal winds (arrows) at the 340 mbar level of our HD 80606b model near periapse (left) andapoapse (right) for our half-nominal (top), nominal (middle), and twice-nominal (bottom) rotation period cases. The length of the arrowsrepresent the strength of horizontal winds. The longitude of the substellar point is indicated by the solid vertical line. The Earth facinglongitude is indicated by the dashed vertical line.
The rapid heating that HD 80606b experiences as it passes through periapse results not only in changes in theplanet’s jet structure, but also produces significant changes in its horizontal temperature distribution. Figure 3 showsthe temperature distribution and wind vectors at the 340 mbar level of our simulations near the periapse and apoapseof HD 80606b’s orbit. Near apoapse, HD 80606b’s atmospheric temperature distribution at the 340 mbar level of oursimulations is fairly uniform with longitude but shows an equator to pole temperature difference that increases withdecreasing rotation period. The efficiency of equator to pole heat transport is strongly influenced by the assumedrotation period of the planet. The larger latitudinal gradient in temperature in our half-nominal rotation period caseresults from the fact that equator to pole heat transport becomes less efficient as the rotation period is reduced (e.g.,Showman et al. 2015).Near the periapse of its orbit, the temperature distribution of HD 80606b becomes dominated by a strong day-nighttemperature contrast (Figure 3). The strength of this day-night temperature contrast is determined in part by the tmospheric Circulation and Clouds of HD 80606b φ (cid:63) ( t ))becomes stationary near periapse and the heating becomes confined to the region near that longitude. For our half-nominal rotation period case, the time rate of change of φ (cid:63) slows, but continues in a westward motion heating otherlongitudes and muting the day/night temperature contrast. In our twice-nominal rotation period case, φ (cid:63) ( t ) changesin an eastward motion near periapse, revisiting previously heated regions of the atmosphere and increasing the widthof the hot spot on the dayside. This, in turn, affects the magnitude of the day-night temperature contrast and alsothe efficiency of day-night versus equatorial flow. In all of our rotation period cases, we see some convergence of theflow from the planet’s dayside in the region between the eastern terminator and the nightside of the planet (left sideof Figure 3).The orientation of HD 80606b’s orbit ( e = 0 .
93 and ω = 300 . ◦ , Pont et al. (2009)) is such that secondaryeclipse, when the planet passes behind the host star as seen from Earth, occurs just three hours before periastronpassage. This is an observationally advantageous orbital configuration for studies of planetary atmospheric responsenear periapse passage and has been leveraged in the phase-curve studies of Laughlin et al. (2009) and de Wit et al.(2016). Phase-curve observations largely probe longitudinal brightness variations (Cowan & Agol 2008), which in thecase of HD 80606b vary strongly as a function of time. Figure 4 shows the latitude-weighted average temperature asa function of longitude from substellar point and pressure that evolves in our models of HD 80606b near periapse andapoapse. Near apoapse, our atmospheric models show virtually no variation in temperature from the planet’s dayside tonightside, with temperatures ranging from ∼ ∼
500 K above 10 bar) temperature variations with longitude. The longitudinal locationthe peak temperatures and the depth to which the thermal structure of the planet is altered during periapse passage isa function of the assumed rotation period (Figure 4). The offset of peak temperatures away from the substellar pointincreases with decreasing rotation period of the planet as the strength of the day-night vs equatorial flow decreases.Similarly, the depth to which the planet’s thermal structure is significantly altered increases with increasing rotationperiod, which reflects the duration that particular longitudes are exposed to incoming stellar radiation during periapsepassage.The transport of cloud material in HD 80606b’s atmosphere near periapse passage will depend critically on bothhorizontal (see Figures 2 and 3) and vertical transport in the planet’s atmosphere. Figure 5 shows the latitude-weightedaverage vertical wind as function of longitude from the substellar point and pressure in our HD 80606b models nearperiapse and apoapse. Near apoapse, vertical velocities in our HD 80606b models are fairly weak ( ∼ − ),with patterns of updrafts and downdrafts that vary based on the assumed rotation period of the planet. Nearperiapse, strong updrafts develop on the dayside and strong downdrafts on the nightside of the planet in our models ofHD 80606b’s atmosphere. These regions of updrafts and downdrafts near periapse correspond to regions of horizontalwind divergence and convergence seen in Figure 3. The location and depth of the strongest regions of updrafts anddowndrafts vary with assumed rotation period, with the twice-nominal rotation period model manifesting the deepestupdraft that is mostly closely located near the substellar longitude. Under all rotation period assumptions, the presenceof strong dayside updrafts could allow for cloud material located at depth in HD 80606b’s atmosphere to be loftedinto the visible regions of the planet’s atmosphere. Cloud Formation and Evolution
Rapid formation/dissipation and transport of clouds is likely to occur in the atmosphere of HD 80606b duringperiapse passage. As shown in Figure 6 the evolution of the cloud coverage through periapse passage depends onthe cloud species’ condensation temperatures. When HD 80606b’s atmospheric temperatures do not exceed the cloudcondensation temperature, such as for MgSiO or MnS clouds, the planet is fully cloudy all the time. When theatmospheric temperatures get hotter than the condensation temperature, such as for Na S clouds, then a hole in theclouds forms when the planet gets close to periapse (see Figure 4), qualitatively affecting the flux maps shown inFigure 6 . By measuring the evolution of the cloud coverage in HD 80606b, one could possibly determine the mostprobable cloud composition. For all cloud compositions considered in this study, the nightside should remain cloudyduring periapse passage, leading to a dimming of the atmosphere at infrared wavelengths.Atmospheric transport of cloud particles, not self-consistently considered here, could affect the picture presented inFigure 6. At apoapse, all the clouds modeled here have cloud bases well below the photosphere near 300 mbar. Ifvertical mixing is small, then these species are trapped near the cloud base and should not be present in the observableportion of the atmosphere. During periapse, strong updrafts on the dayside can transport the cloud material from itsoriginal cloud base to the observable atmosphere (see Figure 5). Horizontal winds can also potentially play a role intransporting material lofted on the dayside to the nightside of the planet (see Figures 2 and 3).The depth of the cloud bases near apoapse strongly depends on the under-constrained internal temperature profile ofHD 80606b. Tidal dissipation (e.g. Bodenheimer et al. 2001) and other atmospheric processes (e.g. Ohmic dissipation,Batygin & Stevenson (2010)) could significantly increase the deep temperature profile of HD 80606b and raise thecloud deck closer to the photosphere. In our models we have assumed an internal temperature ( T int ) of 100 K, whichis consistent with the internal temperature of Jupiter (Fortney et al. 2011). If the assumed internal temperatureof HD 80606b was raised to 500 K or 1000 K the temperature at 200 bar level in our model atmospheres wouldbe increased to roughly 3000 K and 4000 K respectively. This increase in the internal temperature would cause an‘upward’ shift in the thermal structure of the models presented in Figure 4, which would result in the cloud basepressure for MgSiO to move from roughly 100 bars to 10 bars and 1 bar under the assumption that T int equals 500 K Lewis et al.
Figure 4.
Temperature (colorscale) averaged in latitude as a function of pressure and degrees from the substellar longitude near periapse(left) and apoapse (right) for our half-nominal (top), nominal (middle), and twice-nominal (bottom) rotation period cases near periapse.Temperatures represent average values weighted by cos φ , where φ is latitude. Contours in panels represent condensation points, in pressure-temperature space, for the cloud species considered in this study. Note the thermal inversion that develops at pressure less than 1 bar nearperiapse. tmospheric Circulation and Clouds of HD 80606b Figure 5.
Vertical velocity averaged in latitude as a function of pressure and degrees from the substellar longitude near periapse (left)and apoapse (right) for our half-nominal (top), nominal (middle), and twice-nominal (bottom) rotation period cases. Vertical velocitiesrepresent average values weighted by cos φ , where φ is latitude. Positive vertical velocity values represent updrafts while negative verticalvelocities represent downdrafts. Note the significant updrafts that develop on the dayside of HD 80606b in our models near periapsepassage. Lewis et al.
Temperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-11hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-6hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=-1hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+4hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+9hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+14hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+19hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+19hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mt=+19hTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S cloudsa=1 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S cloudsa=1 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S cloudsa=1 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO cloudsa=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO clouds Na S cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ mNo Clouds No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ mt=+24h No Clouds No CloudsTemperatureat 0.1bar Flux at 0.7 µ m Flux at 3.6 µ m Flux at 4.5 µ mNa S clouds MgSiO clouds Na S clouds MgSiO clouds Na S clouds MgSiO cloudsa=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ m a=1 µ m Figure 6.
Temperature at 0 . bar (left column) and flux emerging from the Earth-facing planet hemisphere at three different wavelengthsassuming either that the atmosphere is cloudless or that Na S or MgSiO clouds are present. The different rows are for different times, with t = 0 being the periapse passage. The temperature scale ranges linearly from 500K to 1200K. The brightness scale is linear and normalizedfor each wavelengths to the brightest point simulated, so brightness maps can be compared between different times and different cloudspecies but not between different wavelengths. At 0 . µm the planetary flux is dominated by reflected light and cloudy parts appear brightand cloudless parts appear dark. At longer wavelengths, dominated by thermal emission, the presence of clouds raises the photosphereto higher, cooler layers and the planet appears dimmer. Na S clouds disappear partially from the dayside at t = − h when the planetbecomes hotter than the condensation temperature. On the contrary, MgSiO clouds are always present. and 1000 K, respectively.During periapse passage, an updraft with a vertical velocity of on average 1 m s − develops on the dayside ofour HD 80606b atmospheric models that persists for ∼ ∼
100 km in atmospheric height, corresponding to two scale heights and thus one order of magnitude in pressure. As aconsequence, only particles from pressures shallower than ∼ ∼
10 bar can be present in the observable portionof HD 80606b’s atmosphere. In the case of a cool deep temperature profile (T int =100 K), only Na S clouds can belofted above the photosphere during periapse passage. If deep atmosphere is hotter ( T int =500 K), then MgSiO andMnS clouds could also be present. Observational determination of the composition of clouds present in HD 80606b’satmosphere during periapse passage could help constrain the deep pressure temperature profile of the planet. Theoretical Light Curves
The SPARC model is uniquely equipped to produce theoretical light curves and spectra for HD 80606b that accountfor spatial thermal variations, cloud coverage and dynamics within the atmosphere (see Fortney et al. 2006; Showmanet al. 2009). This capability allows us to make observational predictions based on a variety of assumptions aboutthe planet’s atmospheric composition (in this case clouds) and rotation period. In this section, we focus our model-observation comparisons on the
Spitzer µ m bandpasses used in de Wit et al. (2016) to study HD 80606b.We consider both cloud-free and cloudy models when exploring derived theoretical phase curves and comparing themwith the phase variations observed for HD 80606b near periapse. Model-Observation Comparisons: Cloud-Free
Given the potential for clouds to be sequestered well below the photosphere of HD 80606b or rapidly dissipated nearperiastron passage (see Figures 4), it is possible that clouds play a negligible role in shaping HD 80606b’s time-varyingflux. Figure 8 presents theoretical light curves at
Spitzer’s µ m assuming the half-nominal, nominal and tmospheric Circulation and Clouds of HD 80606b Figure 7.
Dayside average vertical velocity (m s − ) as a function of pressure and time from periapse passage for our half-nominal (top),nominal (middle), and twice-nominal (bottom) rotation period cases. Positive vertical velocity values represent updrafts while negativevertical velocities represent downdrafts. The strong updraft that develops in the dayside of our HD 80606b models near periapse persistsin the atmosphere for ∼ Lewis et al. -20 0 20 40Time from Periapse (h)020040060080010001200 P l ane t/ S t a r F l u x R a t i o ( pp m ) τ rot τ rot τ rot µ m P l ane t/ S t a r F l u x R a t i o ( pp m ) τ rot τ rot τ rot µ m −10 −5 0 5 10Time from Periapse (h) 40060080010001200 P l ane t/ S t a r F l u x R a t i o ( pp m ) P l ane t/ S t a r F l u x R a t i o ( pp m ) Figure 8.
Observed Planet/Star flux ratio (in parts per million, ppm) as a function of time from periapse passage in the
Spitzer µ m channels presented in de Wit et al. (2016). Dashed lines represent the 1 σ envelopes for the observed phase variations. Theoreticalphase curves derived from the half-nominal (red), nominal (blue), and twice-nominal (green) rotation period cloud-free models of HD 80606bpresented in this study are included for comparison. Right panel is a zoom-in of the left panel in the region near periapse passage. Verticallines in right panel represent location of the peak of the planetary flux. Note that all of our cloud-free models over-predict the time of fluxmaximum. The twice-nominal (green) model provide the best match of the observed phase amplitude, which is consistent with the rotationperiod derived by de Wit et al. (2016). twice-nominal rotation period for the planet. The amplitude and shape of the light curves near periapse passage is astrong function of the assumed rotation period. The orbital configuration of HD 80606b is such that an earth observersees the full dayside hemisphere three hours before the periapse passage. As HD 80606b continues to rotate, moreand more of the nightside hemisphere contributes to the observable flux from the planet. We find that peak of theobservable planetary flux from our HD 80606b models peaks very near periapse as the result of the combination ofplanetary rotation period, radiative timescales, and observing geometry of the system.The timing of the predicted peak of the planetary flux does vary with the assumed planetary rotation period and theobservational bandpass being considered. In the 8 µ m bandpass, our models predict the timing of the peak observableplanetary flux to occur 0.5 hours before, 0.75 hours after, and 1.3 hours after periapse passage for the half-nominal,nominal, and twice-nominal rotation periods respectively. In the 4.5 µ m bandpass, our models predict the timing ofthe peak observable planetary flux to occur 0.2, 1.1, and 1.6 hours after periapse passage for the half-nominal, nominal,and twice-nominal rotation periods respectively. In all cases, our models over-predict the time of peak observable fluxfor HD 80606b compared with the ∼ µ m channels.We also find that the magnitude of the peak of the planetary flux is a strong function of the assumed rotationperiod. This is not surprising since the rotation period of the planet determines both the rate at which the substellarpoint sweeps in longitude, but also the fraction of the dayside hemisphere viewed by an earth observer. We find peakvalues in the planet/star flux ratio at 8 µ m of 790 ppm, 1030 ppm, and 1180 ppm for the half-nominal, nominal, andtwice-nominal rotation period cases respectively compared with the observed peak value of 1130 ppm from de Witet al. (2016). We find peak values in the planet/star flux ratio at 4.5 µ m of 320 ppm, 500 ppm, and 620 ppm for thehalf-nominal, nominal, and twice-nominal rotation period cases respectively compared with the observed peak value of740 ppm from de Wit et al. (2016). The twice-nominal rotation period model for HD 80606b provides the best matchto the amplitude of the observed flux variation presented by de Wit et al. (2016).In addition to the peak in the planetary flux near periapse, we find that in each of our rotation cases secondarypeaks in planetary flux occur. For our half-nominal rotation case, the peaks occurs on an interval approximately equalto the assumed rotation period of 20.2380 hours. The second peak in the half-nominal rotation period case, occursslightly before 20.2380 hours from the first peak, which is a result of the sub-Earth longitude probing the remnantsof a compressional heating event. The secondary peaks in the nominal and twice-nominal rotation period cases occurbetween one and two days after periapse passages are also the result of the sub-Earth longitude crossing the longituderegion affected by a compressional heating event in each case. The observed phase variations of HD 80606b do noexhibit the predicted ‘ringing’ (Cowan & Agol 2011; Kataria et al. 2013), which was cited by de Wit et al. (2016) asfurther support for the exceptionally long planetary rotation period inferred from the observational data ( ∼
93 hours).
Model-Observation Comparisons: Cloudy tmospheric Circulation and Clouds of HD 80606b L o n g e r R o t a t i o n P e r i o d − → − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ mM n S . µ m N a S . µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O . µ m M n S . µ m N a S . µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ mM n S µ m N a S µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ m M n S µ m N a S µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ mM n S µ m N a S µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ m M n S µ m N a S µ m . µ m − T i m e f r o m P e r i ap s e ( h ) Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ mM n S µ m N a S µ m . µ m Planet/Star Flux Ratio (ppm) N o C l oud s M g S i O µ m M n S µ m N a S µ m . µ m F i g u r e . P l a n e t / S t a r flu x r a t i o ( i np a r t s p e r m illi o n , pp m ) a s a f un c t i o n o f t i m e f r o m p e r i a p s e p a ss ag e a s o b s e r v e d w i t h S p i t z e r ( s o li db l a c k li n e s w i t hd a s h e db l a c k li n e σ e n v e l o p e s ) a nd a s p r e d i c t e d f r o m o u r m o d e l s w i t hh a l f - n o m i n a l ( l e f t p a n e l s ) , n o m i n a l ( m i dd l e p a n e l s ) , a nd t w i ce - n o m i n a l ( r i g h t p a n e l s ) w i t h M g S i O ( g r ee n ) , M nS ( m ag e n t a ) , a nd N a S ( y e ll o w ) c l o ud s w i t hp a r t i c l e ss i ze s o f . µ m (t o pp a n e l s ) , . µ m ( m i dd l e p a n e l s ) , µ m ( b o tt o m p a n e l s ) . A cc o un t i n g f o r t h e p r e s e n ce o f c l o ud s d o e s i m p r o v e t h ec o n s i s t e n c y o f t h e m o d e l ph a s ec u r v e p r e d i c t i o n s w i t h t h e S p i t z e r o b s e r v a t i o n i n t h e r e g i o n s a w a y f r o m p e r i a p s e p a ss ag e . Lewis et al.
It is possible that clouds will form in or be lofted into the photospheric regions of HD 80606b’s atmosphere duringperiapse passage (see Section 3.2). Figure 9 compares phase variations predicted using our models assuming a rangeof condensate species and particle sizes (see discussion in Section 2.2). We find that in the 8 µ m bandpass assumedparticle size (0.1-10 µ m) does not significantly change the phase curve predictions for a given cloud species, but in allcases the presence of MgSiO clouds better suppresses the flux in region outside of periapse passage. A range of ofboth cloudy and cloud-free models produce theoretical phase curves that are consistent with the 8 µ m data given theuncertainties on the observed flux variations for HD 80606b at that wavelength. The 8 µ m data are best matched bymodels that assume either the nominal or twice-nominal rotation period for the planet, consistent with the rotationperiod of the planet estimated by de Wit et al. (2016) (93 ± hours).The half-nominal rotation period model, while inconsistent with observations, does highlight a potentially interestingatmospheric effect that is lacking from our other models. In the half-nominal rotation period model, the presence ofMnS or Na S clouds suppresses the flux near periapse passage while the presence of MgSiO clouds actually enhancesthe planetary flux near 8 µ m. This increase in the peak planetary flux when including MgSiO clouds for the half-nominal rotation period model, as well as the nominal rotation period model with 0.1 µ m size MgSiO cloud particles,occurs because the significant increase in atmospheric opacity causes the 8 µ m photosphere to be pushed to shallowerpressures that are dominated by the thermal inversion seen in the left panels of Figure 4. In general, the presenceof MgSiO clouds tends to cause the predicted peak in the planetary flux to occur earlier than the cloud-free andMnS or Na S cloud models. This behavior in the peak flux timing can also be attributed to enhanced opacity causingshallower pressures, with shorter radiative timescales, to be probed.Much like with the cloud-free cases (Figure 8), the model predictions for HD 80606b’s phase variations are not ableto reproduce the observed flux variations at 4.5 µ m. However, a few key trends can be noted that move us towardunderstanding the atmospheric processes shaping the Spitzer µ m observation. First, as postulated in de Wit et al.(2016), clouds are necessarily to suppress the flux from the planet in the regions outside of periapse passage. Second,the assumed average particle size of a particular cloud species does play a role in the predicted phase-curves behaviorsat 4.5 µ m, especially for MgSiO clouds. Overall, The twice-nominal rotation model including clouds, in particular1 µ m particle size Na S clouds, best reproduce the observed HD 80606b phase variations at 4.5 µ m. The complexitiesof cloud formation and evolution are not fully captured by our current models, but our results indicate that aerosolsplay a significant role in shaping the flux variations observed by Spitzer for HD 80606b. DISCUSSION
Our three-dimensional general circulation models of HD 80606b provide for self-consistent treatment of a number ofcomplex atmospheric processes without requiring model tuning and/or parametric schemes (e.g. Langton & Laughlin2008; Cowan & Agol 2011). However, the complexity of our models does require detailed analysis of the circulationpatterns, thermo-chemical structure, and cloud coverage that evolves, as we’ve performed in the previous sections.Here we discuss the key atmospheric physical processes that manifest in our models of HD 80606b that likely play asignificant role in shaping the observed flux variation of HD 80606b near the periapse of its orbit.Two scenarios currently exist to explain the lack of thermal emission from HD 80606b outside of periapse passagein the 4.5 µ m Spitzer observations presented in de Wit et al. (2016). The first scenario is that the visible part ofHD 80606b lacks clouds, because they could potentially be confined to depths well below the photosphere, and alsodoes not have a significant internal luminosity. The second scenario is that HD 80606b has a significant internalluminosity, potentially due to tidal heating, and optically thick clouds impeding the internal heat flux to escape atthe observed wavelengths. Our cloud-free models have a very small internal luminosity (T int =100 K), yet still predicta significant flux from HD 80606b post periapse passage inconsistent with the observations. Our cloudy models withoptically thick MgSiO clouds provide the best method for suppressing the planetary flux outside periapse. Althoughnone of our models can provide a ‘perfect’ fit to current observations, the most plausible scenario for HD 80606b isthat an optically thick cloud deck is formed via lofting material from depth during periapse passage.In de Wit et al. (2016), the lack of any visible atmospheric ‘ringing’ (Cowan & Agol 2011; Kataria et al. 2013)combined with the width of planetary flux ‘bump’ was used to constrain HD 80606b’s rotation period to be significantlylonger than the pseudo-synchronous rotation period. As we’ve described here, the lack of atmosphere ‘ringing’ couldin fact simply be due to the development of an optically thick cloud deck post periapse passage. However, we findthat our models that assume twice the pseudo-synchronous rotation period still provide the best explanation of thewidth and amplitude of the flux variations of HD 80606b near periapse. Our models both with and without cloudsover-predict the time of peak planetary flux relative to periapse. We see better alignment between the observed andpredicted timing of the peak in planetary flux when a significant opacity source, such as MgSiO clouds, results inshallower pressures, with shorter radiative timescales, being probed by a given observational wavelength. In our cloud-free models, the radiative timescale we measure near the 4.5 and 8.0 µ m photospheres ( ∼ ∼ ∼ clouds in our models for HD 80606b raise the 4.5 and 8.0 µ m photospheres to closerto 10 mbar, where the radiative timescales we measure in our models are more consistent with the observed values.It is likely that the pressures we’re probing in generating our synthetic phase curves are too deep in the atmosphere,which could be compensated for through enhanced opacity from aerosols, a more metal-rich atmosphere, and/or astrengthening of the thermal inversion that develops in the upper levels of our HD 80606b models (Figure 4). Observational Probes Beyond Spitzer tmospheric Circulation and Clouds of HD 80606b -20 0 20 40Time from Periapse (h)0500100015002000 P l ane t/ S t a r F l u x R a t i o ( pp m ) No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m µ m −20 0 20 40Time from Periapse (h)0500100015002000 P l ane t/ S t a r F l u x R a t i o ( pp m ) No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m µ m −20 0 20 40Time from Periapse (h)0500100015002000 P l ane t/ S t a r F l u x R a t i o ( pp m ) No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m µ m -20 0 20 40Time from Periapse (h)0500100015002000 P l ane t/ S t a r F l u x R a t i o ( pp m ) No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m µ m Figure 10.
Theoretical Planet/Star flux variations for HD 80606b derived from our twice-nominal rotation period model assuming arange of cloud species and particle sizes for a range of
Spitzer bandpasses. These wavelengths will also be accessible to JWST.
Observations at both longer and shorter wavelengths than those probed by
Spitzer in de Wit et al. (2016) willhelp to better constrain the physical processes shaping HD 80606b atmosphere during its periapase passage. As seenin Figure 9, predicted and observed phase variations for HD 80606b near the periapse of its orbit have a strongdependence on both wavelength and assumptions regarding the chemistry and bulk rotation period of the atmosphere.In particular, observations at wavelengths longward of 8 µ m might better constrain the thermal structure of the planetwhile observations at wavelengths shortward of 4.5 µ m might provide more information regarding the formation andevolution of clouds in HD 80606b’s atmosphere near periastron passage.The James Webb Space Telescope (JWST), scheduled for launch in October 2018, will have access to wavelengthslonger than what is currently available to
Spitzer during the warm phase of its mission. Figure 10 shows predictedphase variations using our twice-nominal rotation period model for HD 80606b in four
Spitzer bandpasses that will beaccessible to JWST. We focus here on the twice-nominal rotation period model because it provides the best match tocurrently available observational data and is in line with the rotation period estimate presented in de Wit et al. (2016).The predicted signal from HD 80606b at 24 µ m is a robust 1500 ppm signal with a peak planet/star flux ratio of nearly2000 ppm. At these long infrared wavelengths, our predicted phase curves do not depend strongly on the assumed cloudspecies or its average particle size, but with 100 ppm precision a determination of probable cloud species and rotationperiod of the planet could be made. Similar determinations could be made from 16 µ m observations with strongervariations based on composition and average particle size of any clouds present in HD 80606b’s atmosphere. At both16 and 24 µ m JWST observations will likely reveal this secondary ‘bump’, thereby more definitively determining therotation period of HD 80606b, key for understanding tidal dissipation and spin-synchronization for exoplanets.In the warm phase of its mission, Spitzer can provide exceptionally stable long temporal baseline photometery atboth 3.6 and 4.5 µ m. Although 3.6 µ m observations of HD 80606b were obtained by Spitzer , the systemmatics presentin the data have prevented a full reduction and analysis. Figure 10 shows that the expected signal from HD 80606bat 3.6 µ m will be on the order of a few hundred ppm, which is challenging for Spitzer but should be achievable withJWST. Both the 3.6 µ m and 5.8 µ m wavelength spectroscopic regions could provide valuable insights concerning boththe composition and average particle size of any clouds present in HD 80606b, although some ambiguity may still6 Lewis et al. -20 0 20 40Time from Periapse (h)020406080100120140 P l ane t/ S t a r F l u x R a t i o ( pp m ) WFC UVISF390W392 nm
No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m -20 0 20 40Time from Periapse (h)020406080100120140 P l ane t/ S t a r F l u x R a t i o ( pp m ) WFC3 UVISF350LP585 nm
No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m -20 0 20 40Time from Periapse (h)020406080100120140 P l ane t/ S t a r F l u x R a t i o ( pp m ) STISG750L1027 nm
No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m -20 0 20 40Time from Periapse (h)020406080100120140 P l ane t/ S t a r F l u x R a t i o ( pp m ) WFC3 IRG1411400 nm
No CloudsNo CloudsMgSiO , 0.1 µ mMgSiO , 0.1 µ mMgSiO , 1.0 µ mMgSiO , 1.0 µ mMgSiO , 10 µ mMgSiO , 10 µ mMnS, 0.1 µ mMnS, 0.1 µ mMnS, 1.0 µ mMnS, 1.0 µ mMnS, 10 µ mMnS, 10 µ mNa S, 0.1 µ mNa S, 0.1 µ mNa S, 1.0 µ mNa S, 1.0 µ mNa S, 10 µ mNa S, 10 µ m Figure 11.
Theoretical Planet/Star flux variations for HD 80606b derived from our twice-nominal rotation period model assuming a rangeof cloud species and particle sizes for a range of potential HST observations. The no cloud (black lines) theoretical phase curves representcontributions due to thermal emission/Rayleigh scattering. exist. JWST will also provide the potential for spectroscopic exploration of HD 80606b, which could provide furtherinsights into the chemistry, and its relevant timescales, at work in this intriguing atmosphere.There also exist the potential for observations of HD 80606b near its periapse passage with the
Hubble SpaceTelescope (HST). Because HST is in low Earth orbit, it is unable to provide the semi-continuous temporal coveragefor phase curve observations that
Spitzer provides. Despite this fact, a handful of successful exoplanet phase curvestudies have been executed with HST (e.g. Stevenson et al. 2014). HST provides for both photometric and spectroscopicobservations spanning the far-ultraviolet to the near-infrared. Figure 11 presents a few HST instrument and filter/grismcombinations that have the potential to provide high signal-to-noise observations of HD 80606b near secondary eclipseand periastron passage in the optical to near-infrared where scattered light from the planet should dominate the signal.These predictions for HST were derived using our twice-nominal rotation period HD 80606b model for consistencywith observations at longer wavelengths, but we see only small variations in our predictions based on assumed rotationperiod at HST wavelengths.The signal amplitudes predicted from our models at optical to near-infrared wavelengths are considerably smallercompared with observations at infrared wavelengths (Figure 10), but they have the potential to provide a wealth ofcritical information concerning the formation and evolution of clouds in HD 80606b’s atmosphere. The cloud-freemodel predictions presented in Figure 11 have almost vanishingly small amplitudes, but they represent contributionsfrom the planet’s thermal emission (near-infrared) and Rayleigh scattering (blue-visible) at those relevant wavelengths.The presence of clouds would likely dramatically increase the amplitude of any signal from HD 80606b that could beobserved with HST.In the HST bandpasses presented in Figure 11, differences in signals produced by particular cloud species and particlesizes are readily apparent. The amplitude of the predicted HD 80606b phase curve at wavelengths probed by HSTdepends on a combination of particle size and the particle composition, the case of silicate (MgSiO ) clouds with aparticle size of 0 . µ m presenting the most favorable signal at optical to near-infrared wavelengths. The shape of theobserved flux variations can give insights into the cloud composition. Clouds that would evaporate during periapse tmospheric Circulation and Clouds of HD 80606b S clouds, lead to a peculiar triangular-shaped phase curve whereas clouds species that do notevaporate near periapse, such as MgSiO , have a more symmetrical, bell-shaped phase curve (see also Figure 6).Observations with HST at one or more of the bandpasses presented in Figure 11 of HD 80606b near the periapse ofits orbit would provide for the most robust interpretation of cloud properties in its atmosphere.With a V-band magnitude of 9.0, observations of the HD 80606 system with WFC3-UVIS should yield a 55 ppmprecision per 51 sec frame. Therefore, a program targeting HD 80606b’s flux modulation around periapse passageat visible wavelengths with HST could allow to disentangle between MgSiO /MnS clouds, Na S clouds, and a cloud-free atmosphere. It is possible that aerosols not considered in this study (e.g. hazes), would produce a phase curvesignature distinct at visible wavelengths. An observation window of 22 hours centered around the periapse passagewould yield a sufficient precision on the phase curve shape to constrain the cloud composition. In particular, it shouldyield a detection of 1 µ m size MgSiO /MnS and Na S clouds at the ∼ σ and ∼ σ levels, respectively. CONCLUSIONS
Our atmospheric models of HD 80606b present a rich area for investigating atmospheric circulation and cloudformation under extreme time-variable forcing conditions. Here we’ve shown that bulk rotation period assumptionscan dramatically alter the atmospheric circulation of a planet like HD 80606b, especially near the periapse of its orbit.We have shown that a number of cloud species could play a significant role in shaping the flux variations of HD 80606bnear the periastron of its orbit. In comparing theoretical phase curves for HD 80606b derived from our atmosphericmodels, we find that our twice-nominal rotation period model with clouds included as an opacity source best matchthe
Spitzer observations presented in de Wit et al. (2016).The de Wit et al. (2016) study used a semi-analytic cloud-free model to constrain the rotation period, radiativetimescale, and baseline brightness temperature of HD 80606b from
Spitzer µ m observations. de Witet al. (2016) concluded that the lack of significant flux from the planet outside of the periapse passage region was anindication that HD 80606b’s possesses a small internal luminosity. We find that the lack of planetary flux outside ofperiapse, combined with the timing of the peak planetary flux relative to periapse passage, is more likely an indicationof the presence of an optically thick cloud deck composed of material lofted from deep within HD 80606b’s atmosphereduring periapse passage. This cloud formation and lofting mechanism would work most efficiently in scenarios whereHD 80606b has significant internal flux, which provides an important constraint on the tidal and internal evolution ofthe planet.In future work, we will include more self-consistent treatment of cloud formation and evolution in our models and anexploration of a larger range of values for the internal temperature and rotation period of HD 80606b. Observations atboth shorter and longer wavelengths than those currently available for HD 80606b with JWST and HST would providemore robust constraints on the physical processes at work in HD 80606b’s atmosphere. Such atmospheric constraintscan then be extended to our global understanding of exoplanet atmospheres and allow us to further refine our modelsas we begin to explore cooler and smaller worlds beyond our solar system.NKL thanks H. Wakeford, K. Stevenson, J. Fraine, J. Valenti, S. H¨orst, and M. Lewis for their support during thewriting of this manuscript. This work was performed in part under contract with the California Institute of Technology(Caltech) funded by NASA through the Sagan Fellowship Program executed by the NASA Exoplanet Science Institute.REFERENCESm observations. de Witet al. (2016) concluded that the lack of significant flux from the planet outside of the periapse passage region was anindication that HD 80606b’s possesses a small internal luminosity. We find that the lack of planetary flux outside ofperiapse, combined with the timing of the peak planetary flux relative to periapse passage, is more likely an indicationof the presence of an optically thick cloud deck composed of material lofted from deep within HD 80606b’s atmosphereduring periapse passage. This cloud formation and lofting mechanism would work most efficiently in scenarios whereHD 80606b has significant internal flux, which provides an important constraint on the tidal and internal evolution ofthe planet.In future work, we will include more self-consistent treatment of cloud formation and evolution in our models and anexploration of a larger range of values for the internal temperature and rotation period of HD 80606b. Observations atboth shorter and longer wavelengths than those currently available for HD 80606b with JWST and HST would providemore robust constraints on the physical processes at work in HD 80606b’s atmosphere. Such atmospheric constraintscan then be extended to our global understanding of exoplanet atmospheres and allow us to further refine our modelsas we begin to explore cooler and smaller worlds beyond our solar system.NKL thanks H. Wakeford, K. Stevenson, J. Fraine, J. Valenti, S. H¨orst, and M. Lewis for their support during thewriting of this manuscript. This work was performed in part under contract with the California Institute of Technology(Caltech) funded by NASA through the Sagan Fellowship Program executed by the NASA Exoplanet Science Institute.REFERENCES