Proceedings of the International Symposium on Planetary Science 2011,Edited by S. Okano, Y. Kasaba and H. Misawa, pp. 13–31. doi:10.5047/pisps.013c© TERRAPUB, Tokyo, 2019.

Origin and solar activity driven evolution of Mars’ atmosphere

H. Lammer1, P. Odert1,2, M. Leitzinger2, H. Groller1, M. Gudel3,K. G. Kislyakova1,2, M. L. Khodachenko1, and A. Hanslmeier2

1Space Research Institute, Austrian Academy of Sciences,Schmiedlstr. 6, A-8042 Graz, AustriaE-mail: helmut.lammer@oeaw.ac.at

2Institute for Physics/IGAM, University of Graz,Universitatsplatz 5, A-8010, Graz, Austria

3Institute for Astronomy, University of Vienna,Turkenschanzstr. 17, A-1180, Austria

The evolution of the martian atmosphere and its H2O inventory can be divided intoearly and late evolutionary epochs. Volatiles such as H2O and CO2 could have builtup a dense water vapour dominated H2O/CO2 steam atmosphere during the early im-pact and magma ocean solidification period. Due to the high activity of the youngSun, Mars’ atmosphere experienced soft X-ray and EUV-driven hydrodynamic blow-off during the early Noachian, which could have powered the loss of an outgassedinitial steam atmosphere with a surface pressure of ≥ 100 bar. After this initial at-mosphere was lost, impacts and mantle outgassing during later evolutionary epochsmay have built up around ∼4±0.2 Gyr ago, a CO2 atmosphere which could havebeen denser than the present ∼7 mbar. The CO2 surface pressure of such a secondaryatmosphere and its evolution followed a complex interplay of various nonthermal es-cape processes and carbonate precipitation. Finally we discuss how the discovery oftransiting terrestrial exoplanets will open the possibility to test atmospheric evolu-tion hypotheses as discussed in our study by observations and advanced numericalmodels.Key words: Young Sun, early Mars, magma ocean, atmosphere origin, atmosphereevolution, atmosphere escape, ENAs

1 IntroductionThe current martian atmosphere and its isotopic ratios (e.g., Becker et al., 2003)

have resulted from the action of impact erosion/delivery, SXR and EUV (XUV) andplasma driven thermal and nonthermal loss processes (e.g., Lundin et al., 2007; Lam-mer et al., 2008; Pham San et al., 2009; Tian et al., 2009) of volatiles as well as CO2

surface weathering processes (e.g., Zent and Quinn, 1995; Bandfield et al., 2003;Phillips et al., 2010; Lammer, 2013; Lammer et al., 2013, 2018).

The aim of this work is to discuss the origin and evolution of Mars’ atmosphereduring the early and late evolutionary epochs. In Section 2 we discuss the origin ofan impact induced initial dense water vapour dominated H2O/CO2 steam atmosphere

13

14 H. Lammer et al.

during/after the magma ocean solidification phase. The loss of this initial atmosphereduring the early Noachian and a possible growth of a secondary outgassed CO2 atmo-sphere ≤4±0.2 Gyr ago is discussed in Section 3. In Section 4 we focus on the escapeof Mars’ CO2 atmosphere during the past 3.5 Gyr, which resulted in the present ∼7mbar surface pressure. Because Mars’ upper atmosphere during the first ≤500 Myrmost likely under solar XUV-induced non-hydrostatic conditions we discuss in Sec-tion 5 how the UV observation of hydrogen- and energetic neutral atom (ENA) cloudsaround transiting terrestrial exoplanets within orbits of dwarf stars could confirm thedescribed atmospheric evolution hypothesis.

2 Origin of the Martian Atmosphere and Water InventoryThere are many pieces of evidence from the isotope records and dynamical rea-

sons that Earth’s initial water inventory originated from planetary embryos, whichhad their main origin in the asteroid belt (e.g., Morbidelli et al., 2000; Raymond etal., 2004; Lunine et al., 2011; Walsh et al., 2011; Marty, 2012; Brasser et al., 2017;Lammer et al., 2018). According to terrestrial planet formation models planets orig-inate not at the locations where they are located today (e.g., Raymond et al., 2004;Walsh et al., 2011). According to the “Grand-Tack” model scenarios it is found thatgiant planet migration could restrict the formation of the terrestrial planets to loca-tions between 0.7 and 1 AU (Walsh et al., 2011; Lammer et al., 2018).

Although, it is not clear where Mars initially formed, all formation theories agreethat the small planet remained significantly farther from the Sun than Earth duringthe most time of its history and its growth was stunted early so that its mass remainedrelatively low. According to a recent study by Brasser et al. (2017), which is alsobased on the “Grand Tack” model, it is expected that Mars’ formation required aspecial dynamical pathway, which was different compared to the Earth and Venus. Inthis studied formation scenario, Mars’ volatile budget is most likely different fromVenus and Earth.

Therefore, in the case of Mars the original D/H ratio in H2O, which was obtainedfrom planetesimals is not well known. During accretion, three possibilities or a mix-ture of them for the delivery of water to the growing planet are possible i.) planetaryembryos from beyond 2.5 AU, ii.) small asteroids from beyond 2.5 AU, and iii.)comets from the Jupiter region and beyond (Brasser, 2013).

Lunine et al. (2003) modeled how much H2O Mars could have acquired fromasteroids and comets by estimating the cumulative collision probability between as-teroids and comets with Mars. These authors assumed that comets consist of ∼50%of H2O-ice with a D/H ratio about 2 times the terrestrial value and that asteroids have∼10% H2O by mass with D/H values comparable to Earth’s ocean water. By usingsuch impactors it was found that Mars could acquire an amount of H2O equivalent to∼0.06–0.27 times the amount of an Earth ocean (EO) with a D/H ratio of ∼1.6 and∼1.2 times of Earth’s ocean water D/H ratio.

On Earth the mass of one evaporated EO, mEO ∼ 1.4 × 1024 g corresponds to asurface pressure of ∼270 bar (e.g. Kasting and Pollack, 1983; Zahnle et al., 1988).

Solar activity and Mars’ atmosphere evolution 15

This is the averaged surface pressure executed by the weight of an EO, i.e.

p = mEOg

4π R2pl

= G MplmEO

4π R4pl

, (1)

with the surface gravitational acceleration g, planetary radius Rpl, planetary mass Mpl,and Newton’s gravitational constant G. For Mars, one evaporated EO therefore cor-responds to a surface pressure of ∼360 bar, while the hydrogen content correspondsto 1.56 × 1023 g or EO/9, and hence a pressure of ∼40 bar. Thus, it follows fromthe results of Lunine et al. (2003) that early Mars could have produced an equivalentamount of H2O of ∼20–100 bar. This amount correspond to ∼600–2700 m worth ofH2O equivalent on the martian crustal regolith and surface and is in agreement withprevious studies (e.g., Baker, 2001) and recent arguments given in Brasser (2013).

Models for martian magma oceans of varying depths predict that Mars’ earli-est crust was most likely formed by decompression mantle melting which could oc-cur during gravitationally driven solid-state overturn of cumulates following magmaocean solidification (Elkins-Tanton et al., 2005; Elkins-Tanton, 2008; Erkaev et al.,2014). The mantle solidification is found to be 98% complete in less than 3 Myrfor various magma ocean depths and less than ∼105 years for low-volatile magmaoceans (Elkins-Tanton, 2008). Magma oceans with depths of 500 km or 2000 km,even with small initial volatile contents of ∼0.05 weight % (wt.%) H2O and ∼0.01wt.% CO2 outgassed an early atmosphere of ∼30–70 bars H2O and ∼5–12 bars CO2

(Elkins-Tanton, 2008; Lammer, 2013; Erkaev et al., 2014; Lammer et al., 2018).Model simulations with magma ocean depths larger than ∼1550 km produce twoseparate magmatic source regions during overturn that create compositionally dis-tinct magmas, which are consistent with major element and trace element data formartian SNC meteorites as well as with volume estimates of crust from the northernand southern sides of the planet’s dichotomy (Elkins-Tanton, 2005). A higher initialvolatile content of ∼0.5 weight % (wt.%) H2O and ∼0.1 wt.% CO2 results in densewater vapour dominated H2O/CO2 steam atmospheres on early Mars of ∼330–800bars. From these studies we can expect that the higher initial H2O content casesof Elkins-Tanton (2008) with ∼0.5 wt.%, which yield several hundreds of bar densewater vapour dominated H2O/CO2 atmospheres are not in agreement with the dynam-ical model results Lunine et al. (2003) and Mars’ formation studies (Brasser, 2013;Brasser et al., 2017).

In agreement with the formation and dynamical water delivery scenarios we as-sume that early Mars originated most likely with initial volatile contents in the orderof ∼0.05 wt.% and outgassed after its magma ocean solidification a water vapourdominated dense H2O/CO2 steam atmosphere within a range of ∼20–100 bar. Onecan expect that such an outgassed dense water dominated H2O/CO2 atmosphere re-mained more or less in its gas phase during a few tens of Myr as a steam atmosphere.Matsui and Abe (1986) studied the early stages of accretion and early impacts. Theseauthors suggested that during this early time period thermal blanketing and also fre-quent impacts which can rise temperatures up to ∼1500 K determined the environ-mental conditions. At the end of that epoch of a complex interplay between the

16 H. Lammer et al.

greenhouse effect and impacts, the planet’s surface and atmosphere cools, so that aremaining H2O vapour can condense and form oceans or ice in case if the climate iscold ice (Chassefiere, 1996; Lammer et al., 2013).

3 XUV-powered Atmospheric Escape during the Early NoachianThe Sun’s activity always played an important role in the escape and evolution of

Mars’ atmosphere. It is known from observations of young solar-type G stars sincedecades that opposite to a weaker total luminosity, the young stars are a much strongersource of X-rays, soft X-rays (SXR) and EUV (XUV) electromagnetic radiation (e.g.,Newkirk, 1980; Skumanich and Eddy, 1981; Zahnle and Walker, 1982). Due to thelack of detailed astrophysical observations of solar proxies with different ages, pre-vious studies on XUV-driven hydrodynamic escape of primitive atmospheres appliedonly rough XUV enhancement scaling factors (e.g., Sekiya et al., 1980; Watson etal., 1981; Kasting and Pollack, 1983; Zahnle and Kasting, 1986; Zahnle et al., 1988;Chassefiere, 1996).

Dorren and Guinan (1994) have studied the evolution of UV line fluxes in detailby using spectral measurements from the IUE satellite. Gudel et al. (1997) and Ribaset al. (2005) extended this research to X-rays and EUV which are the relevant wave-lengths (λ ≤ 1000 A) for ionization, dissociation and thermospheric heating (e.g.,Hunten, 1973, 1993; Hunten et al., 1987). Multi-wavelength XUV observations bythe ASCA, ROSAT, EUVE, FUSE, IUE, HST satellites of solar proxies with ages <

4.6 Gyr revealed a saturation of the XUV flux during the first 100 Myr (Gudel et al.,1997) at a value ∼100 times compared to that of today’s Sun and a decrease after thisearly active period according to an XUV enhancement factor power law (Gudel et al.,1997; Ribas et al., 2005; Gudel, 2007) SXUV

SXUV =(

tGyr

t0

)−1.23

, (2)

with t0 the age of the present Sun and tGyr the age of the Sun in time in units of Gyr.One should note that the power law above represents one that is valid for a typical

moderate rotating young G-type star. Evolutionary decay laws for XUV radiation,therefore, is a function of the stellar rotation periods (e.g., Tu et al., 2015; Lammer etal., 2018). Tu et al. (2015) found that, depending on whether a solar analog starts outas a slow or fast rotating young star after the disk phase, the XUV evolutionary tracksdiverge first and then converge again after several hundred Myr when the rotation pe-riods converge. The distribution of the XUV radiation is broadest in the range of a fewtens to a few hundreds of Myr. However, as shown recently by Odert et al. (2018), forlow mass bodies such as Mars the effect to mass loss of early magma ocean relatedcatastrophically outgassed steam atmospheres between a slow, moderate and fast ro-tating young G-type star is not very different because the main loss occurs duringthe early saturation phase of these stars. The effect in atmospheric escape related tothe spinning-down rotation-activity relation of the host star is more important dur-ing later evolutionary stages and the possible build up of secondary atmospheres (see

Solar activity and Mars’ atmosphere evolution 17

also: Lammer et al., 2018). Here, we report on the escape by considering the powerlaw above that corresponds to a moderate rotating young Sun-type star.

Because of the high XUV flux of the young Sun the H2O molecules in the ther-mosphere are dissociated and H2 and H atoms have dominated the upper atmosphereuntil they escaped to space. XUV photons that can be absorbed by H or H2 areuniquely suited to drive escape. Photons with longer wavelengths carry more energybut will in general be absorbed by good infrared radiating molecules, such as H2Oand CO2, so that this energy cannot be so easily transformed into escape. The heatingefficiency η which can be described as the net heating rate to the rate of solar energyabsorption is in the order of ∼15% (Chassefiere, 1996; Lammer et al., 2009; Shema-tovich et al., 2014; Odert et al., 2018). By applying a formula which is based on theenergy-limited equation where η = 100% for hydrodynamic escape (e.g., Hunten,1973, 1993; Hunten et al., 1987; Zahnle et al., 1988)

d M

dt= 3ηSXUV FXUV

4Gρpl, (3)

but modified for a lower more realistic η value of 15% (Chassefıere, 1996; Lammeret al., 2009), with Newton’s gravitational constant G, the mean density of the planetρpl and the present time XUV flux FXUV in Mars’ orbit we obtain a loss of hydrogenwhich results in a decrease of surface pressure during time. An example with threedifferent initial surface pressure values is shown in the left panel of Fig. 1. One cansee that an XUV-powered hydrodynamic blow-off removes a hydrogen content of 11bar which corresponds to an outgassed 100 bar steam atmosphere in ≤15 Myr. Even

Fig. 1. Left panel: Calculated amount of H atom loss resulting in a decrease of hydrogen partial surfacepressure in units of bar for a heating efficiency η = 15% over time. The initial partial surface pressurescorrespond to 88 bar (solid line), 50 bar (dashed line) and 11 bar (dashed-dotted line). Right panel:Critical temperature Tc leading to blow-off conditions on Mars as a function of exobase temperatureTexo and exobase altitude h = zexo. The symbol “+” marks the present day exobase temperature andaltitude, while the other symbols correspond to the modeled conditions of a CO2 atmosphere which isexposed to 3, 10 and 20 times higher XUV fluxes as that of the modern Sun (Tian et al., 2009).

18 H. Lammer et al.

the H-fraction of 88 bar which corresponds to a 800 bar H2O atmosphere is lost in≤150 Myr.

By comparing the flux of catastrophically outgassed volatiles during the magmaocean solidification phase with the corresponding escape flux to space we find thatthe source flux into the atmosphere is up to 2 magnitudes larger compared to the cor-responding escape flux. This means that a water vapour dominated dense H2O/CO2

steam atmosphere with several bars of CO2 and several tens of bars H2O could origi-nate, before XUV-powered hydrodynamic escape started to act efficiently.

One may argue that, even if the hydrogen content from a 100 bar H2O steamatmosphere is lost, the remaining part of ∼90 bar oxygen and several bar CO2 werekept. In a recent study by Tian et al. (2009) it is shown that even a several barCO2 atmosphere was most likely not stable on early Mars during the first 500 Myrafter the planet’s origin. These authors applied a 1D multi-component hydrodynamicplanetary thermosphere-ionosphere model and a coupled electron transport-energydeposition model to Mars and discovered that for XUV fluxes >10 times that ofthe present Sun CO2 molecules can efficiently dissociate so that less thermosphericIR-cooling CO2 molecules are available. Less IR-coolers result in higher heatingefficiency η for the remaining CO2 atmosphere and exobase temperatures comparedto previous model results (Kulikov et al., 2007). According to the study of Tian et al.(2009), the flux of the produced C and O atoms is >1011 cm s−1 which was higher asthe outgassing fluxes from volcanic activity, which depended on the remaining CO2

inventory ∼4–4.4 Gyr ago.That C and O atoms could escape from Mars to a great amount is also supported

by the XUV-induced non-hydrostatic hydrodynamic expansion of the exobase level.The results of Tian et al. (2009) indicate that the exobase level expanded from presenttime ∼200 km altitude up to ∼900 km and even to ∼10000 km (∼3Rpl) for XUVflux values which were ∼10 and ∼20 times larger than the present solar value. Thecorresponding exobase temperatures Texo moved from ∼800 K and ≥2500 K. Theconsequences of high Texo for the expansion of the exobase and thermal escape fromplanetary atmospheres is shown in the right panel of Fig. 1. The escape flux reachesblow-off when Texo≥Tcrit (e.g., Opik, 1963; Chamberlain, 1963; Erkaev et al., 2007)

(3

2

)kTcrit ≥ G Mplm

R, (4)

the thermal energy of the gas kinetic motion (3kTexo)/2 overcomes the gravitationalenergy corresponding to the thermal escape parameter X = (v∞/v0)

2 ≈ 1.5. v∞ isthe escape velocity and v0 is the most probable velocity of the Maxwellian velocitydistribution of temperature Tcrit, G is Newton’s Gravitational constant, Mpl planetarymass, m particle mass, radius R = Rpl + h, k is the Boltzmann constant and his the distance above the planetary surface. One can see from Fig. 1 (right panel)that because of the high XUV fluxes of the young Sun, due to its low mass, Mars’early atmosphere could not keep hydrostatic conditions and expanded dynamically toseveral Mars radii so that blow-off most likely occurred.

Solar activity and Mars’ atmosphere evolution 19

The hydrodynamic escape of oxygen from primitive atmospheres of Venus andMars was studied by Chassefiere (1996) who applied a SXUV of ∼25 times that oftodays Sun. This study is based on terrestrial planet formation models where the ac-cretion for Mars occurred quite late at ∼100 Myr after the origin of the Solar System(Wetherill, 1986). At that time the saturated XUV phase of the young Sun was overand the XUV flux began to decrease to lower values. However, there is much evidenceof faster protoplanet formation (e.g., Quingzhu et al., 2002; Lunine et al., 2011) andshort nebular evaporation time scales (e.g., Lunine et al., 2011) which indicate thatMars was most likely exposed during a time period of ∼50–70 Myr to an XUV fluxwhich was much more than 10 times higher compared to the modern Sun. Becauseof the different and most likely outdated assumptions of Chassefiere (1996) we carryout a rough investigation of the expected consequences of an earlier outgassed waterdominated H2O/CO2 steam atmosphere if embryo-size Mars accreted earlier.

In an atmosphere which is dominated by H2O vapour the water molecules will bemore or less dissociated at about 100 km above the base of the expanding atmospherewhich corresponds to the lower thermosphere (Chassefiere, 1996). Therefore one canexpect that O atoms are the major form of escaping oxygen. The escape flux FO ofthe O atoms which can be dragged by the outward flowing hydrogen flux FH (e.g.,Hunten et al., 1987; Chassefiere, 1996)

FO = XO

XHFH

(mc − mO

mc − mH

), (5)

with the mole mixing ratios XH and XO, the masses of the light and heavy speciesmH and mO, the so called cross over mass mc is

mc = mH + kT FH

bgXH, (6)

which depends on FH, the molecular diffusion parameter b (Zahnle and Kasting,1986; Chassefiere, 1996), gravity acceleration g, Boltzmann constant k and an av-erage upper atmosphere temperature T which can be assumed for hydrogen undersuch conditions in the order of ∼ 500 K (Zahnle and Kasting, 1986; Chassefiere,1996).

The left panel of Fig. 2 shows the loss related decrease of surface pressure byusing a heating efficiency η=15% during the saturated phase of the young Sun’sXUV flux. The initial surface pressure values are assumed to be 70 bar H2O and 12bar CO2. It is also assumed that Mars finished its accretion at ∼10 Myr after theorigin of the Sun (Hansen, 2009). One can see that such an atmosphere can loseits oxygen in ∼10 Myr. One should also note that under these extreme conditionsseveral bar of the outgassed CO2 would also be lost, either in dissociated form as Cand O atoms or as molecules. The maximum amount of an initially outgassed 340bar dense steam atmosphere that could be lost completely from early Mars duringthe saturated XUV flux of the young Sun is shown in the right panel of Fig. 2. Wepoint out that the results shown in this work are in agreement with results presented

20 H. Lammer et al.

Fig. 2. Left panel: Partial surface pressures in bar of an initially (t0 = 10 Myr) of a hydrodynamicallyescaping 70 bar dense water vapour dominated steam atmosphere with a CO2 content of 12 bar as afunction of time. Solid line corresponds to loss of the hydrogen content. The dash-dotted line corre-sponds to dragged carbon atoms and the thin dashed line is total sum of oxygen atoms which originatefrom dissociation of H2O and CO2, while the thick dashed line corresponds to the O content of the 70bar H2O vapour. If one assumes that 12 bar CO2 is not dissociated in C and O atoms it is shown inthe dashed-dotted line that it would also escape in molecular form. Right panel: Decreasing surfacepressure as a function of time for a maximum possible amount of H2O-vapor that can be lost within the∼100 Myr lasting XUV saturation phase of the young Sun under the assumption the RXUV ∼ Rpl . Thesolid line corresponds to the decreasing surface pressure caused by the escaping hydrogen. The dashedline corresponds to the partial surface pressure of oxygen which is dragged with the escaping hydrogenatoms.

in a recent review by Lammer et al. (2013) which assume that the altitude layer RXUV

of the upper atmosphere where the XUV radiation of the young Sun is absorbed andtransferred into heating of the thermosphere is close to Rpl. However, in a more recentstudy by Erkaev et al. (2014) it is shown that in cases of a hot steam atmosphere anda planet with a low gravity, the atmospheric layer where the solar XUV radiation isabsorbed and deposits its energy is more extended and RXUV is shifted to distances ofseveral Rpl. In such cases the fast escape of even a steam atmosphere with an amountof ∼70 % of an EO and ∼50 bar CO2 can be lost within several 104 years (Erkaev etal., 2014).

More recently, Odert et al. (2018) studied the escape and fractionation of volatilesand noble gases from Mars-sized planetary embryos at various orbital locations andpossible XUV flux evolution tracks. The results of these authors agree with the exam-ples discussed in this study and in Erkaev et al. (2014). Therefore, one can concludethat typical magma ocean related catastrophically outgassed steam atmospheres arelost in timescales of ten to a few tens of Myr from Mars-like bodies, depending onthe amount of degassed atmosphere, the orbital distance, and stellar activity evolution(Lammer et al., 2018; Odert et al., 2018).

Zahnle et al. (2008) applied a photochemical model for a dense early martian CO2

atmosphere and found that the CO2 could be photochemically converted into CO overtimescales of ∼0.1–1 Gyr. These results are in agreement with a previous study of

Solar activity and Mars’ atmosphere evolution 21

McElroy and Donahue (1972). The fast transformation of CO2 into CO could beprevented by catalytic recombination of CO with oxidants due to surface reactionsor warmer climate caused by additional greenhouse gases. Depending on the abso-lute values of the catastrophically outgassed volatiles, the planets impact history (e.g.Segura et al., 2012; Rodriguez et al., 2015; Iijima et al., 2017), possible additionalgreenhouse gases (i.e. CH4, H2, etc.) (e.g., Ramirez et al., 2014; Kerber et al., 2015;Fairen, 2017), large lakes consisting of liquid water may have been only episodicallypresent on the martian surface during the Noachian (e.g. Dohm et al., 2008; Rosen-bauer and Head, 2015; Lu et al., 2017). However, that Mars could have kept a denseCO2 atmosphere resulting in a warmer and wetter environment during the first billionyears seems unlikely due to the non-detection of large carbonate deposits (Bibring etal., 2005).

Results which indicate that Mars should have lost a part of its outgassed CO2

dragged by the hydrogen blow-off within 20 Myr are not in agreement with the mod-eled CO2 loss rates of Manning et al. (2006) who studied thick and thin models ofthe evolution of CO2 on Mars but considered only the not well understood process ofimpact erosion and delivery during the early Noachian. Because these authors did notconsider that Mars could have lost most of its initially outgassed CO2-inventory byXUV-powered escape, in their study the martian CO2 content decreased more slowly.

In case there are no undiscovered hidden large subsurface carbonate deposits(Lammer et al., 2013; and references therein), from dynamical and escape modelsone can conclude that Mars evolved most likely “cold and dry” during the Noachian.About 4±0.2 Gyr ago, when the XUV flux decreased to more moderate values of≤10 times that of the modern Sun, its atmosphere escaped at a lower rate. When thisoccurred, CO2, which was released into the atmosphere from volcanos may have builtup a secondary CO2 atmosphere which was denser than today’s 7 mbar, but has beenlost during the Hesperian and Amazonian by various nonthermal escape processes.

4 Nonthermal Escape since the End of the NoachianThe geological and mineralogical evidence of surface erosion features, such as

outflow channels, suggest at least sporadic periods of liquid surface water at the endof the Noachian (e.g., Baker, 2001). This indicates that Mars should have had a denserCO2 atmosphere compared to the present ∼7 mbar ∼4±0.2 Gyr ago. In the followingsubsections we discuss various atmospheric escape processes for CO2 from the lateNoachian until today.4.1 Solar wind induced ion erosion of the martian CO2 atmosphere

Barabash et al. (2007) used Mars Express ASPERA-3 ion escape data for theinvestigation of the solar wind erosion of the martian CO2 atmosphere related to theplanet’s history. This is also the time period where the martian magnetic dynamostopped to work so that atmospheric escape of ionized particles could also work (e.g.,Chassefiere et al., 2007). Barabash et al. (2007) estimated the total CO+

2 solar winderosion rates and found that Mars loses only a tiny amount at present in the order of∼8 × 1022 s−1. By estimating this loss rate backward to the end of the Noachian oneobtains a total loss equivalent to the surface pressure of about ∼0.2–4 mbar (Barabash

22 H. Lammer et al.

et al., 2007). Such low values for the molecular ion loss are within the minimum andmaximum values of theoretical studies which yield model depended amounts in theorder of ∼0.8–100 mbar (e.g., Ma et al., 2004; Modolo et al., 2005; Chassefiere etal., 2007; Lammer et al., 2008; Manning et al., 2010).4.2 Sputtering of carbon bearing species during the Hesperian and Amazonian

Atmospheric sputtering has been identified as an important escape source of heavyatoms from the upper atmosphere in the case of Mars, but of less importance formore massive planets like Venus (Luhmann and Kozyra, 1991). Leblanc and Johnson(2002) studied sputter rates of molecular species, such as CO2 and CO, at the mar-tian atmosphere during the past 3.5 Gyr. These authors coupled a test particle MonteCarlo model with a molecular dynamic model which described collisions betweenphotochemically produced “hot” atoms and molecules in the martian thermospherefor XUV fluxes which are 3 times and 6 times larger than that of the present Sun.According to Ribas et al. (2005) these XUV flux enhancements correspond to planetages of ∼2 Gyr and ∼3.5 Gyr ago. Leblanc and Johnson (2002) obtained from theirmodel a total loss of CO2 molecules during the past 3.5 Gyr of ∼6 mbar, for CO ∼7mbar and ∼15 mbar for C atoms. This yields a total sputter loss of CO2 in the orderof ∼60 mbar. One can see that, similar to the case of ion erosion, not much of themain CO2 atmosphere escaped via this nonthermal loss process since the end of theNoachian. We agree with Chassefiere et al. (2007) that it is important to know whenthe martian magnetic dynamo vanished. Because the sputtering efficiency is highlynonlinear with the solar XUV flux, a long period of intense escape by sputtering couldhave been initiated in case the martian magnetic dynamo stopped working before 4Gyr ago. But one should also note that it was shown by Terada et al. (2009) that, dueto the extreme solar wind-atmosphere interaction caused by the young Sun before ∼4Gyro ago, a stronger magnetic field of several hundreds of nT could have been in-duced on the entire dayside ionosphere, which should have decreased sputtering lossduring the transition period after the intrinsic dynamo stopped working.4.3 Escape of “hot” C and O atoms during the past 4 Gyr

Suprathermal O, C, and N atoms are found near and above the exobase as a resultof exothermal photochemical reactions, such as dissociative recombination (DR) ofO+

2 , N+2 or CO+ ions (e.g., Ip, 1988; Nagy et al., 1990; Kim et al., 1998; Lammer et

al., 2000; Fox and Hac, 2009; Krestyanikova and Shematovich, 2006; Valeille et al.,2010). The production of these hot atoms in the upper atmosphere significantly de-pends on the solar XUV radiation. For preliminary estimates of the efficiency of theCO2 loss related to the escape of “hot” C and O atoms since the past 4 Gyr we applythe “hot atom” Monte Carlo model of Groller et al. (2010) to the CO2 atmospheresmodeled by Tian et al. (2009). The calculation of the production rates of “hot” O andC atoms originating from DR of O+

2 and CO+ molecular ions, which yield O(3P, 1D),O(3P, 1S) and C(3P, 1D), O(3P, 1D, 1S) atoms, is based on an electron temperaturedependent rate coefficient, where the total energy of these newly produced particlesis a sum of their released energies (�E) according to a DR reaction channel of ki-netic and internal energy, the latter being stored in case of molecules as vibrationaland rotational energy. Excited C atoms are also produced by photo-dissociation of

Solar activity and Mars’ atmosphere evolution 23

CO molecules from CO + hν → C(3P) + O(3P) + �E, where �E is obtained as thedifference of the photon energy and the energy which is needed to dissociate themolecule and excite the newly produced atoms.

The magnitude of the initial velocity of a “hot” particle is randomly chosen fromthe calculated velocity distribution while the direction of the initial velocity is as-sumed to be isotropically distributed. Calculation of the new velocity vector of the“hot” atom after its collision is based on the energy dependent total and differentialcross sections for elastic, inelastic and quenching collisions (Groller et al., 2010). Aprimary “hot” atom can generate numerous cascaded “hot” particles. As a result, atany altitude, the velocity vectors of the suprathermal particles are known and the fluxfraction of “hot” C and O atoms with velocities larger than the escape velocity are cal-culated. For present Mars we obtain total escape rates for “hot” O atoms of ∼9×1025

s−1 and for “hot” C atoms ∼2.7 × 1025 s−1 during high solar activity conditions and∼3 × 1025 s−1 for “hot” O atoms and ∼3 × 1024 s−1 for “hot” C atoms during lowsolar wind conditions. Monte Carlo calculations based on the thermosphere profilesmodeled by Tian et al. (2009) during the earlier periods yield an integrated total CO2

loss from today to 4 Gyr ago of ∼200–400 mbar (Amerstorfer et al., 2017).By considering the integrated partial pressure values from the discussed nonther-

mal atmospheric escape processes above, the obtained total losses are in agreementwith Kurokawa et al. (2017) who estimated from N2 and Ar isotopic compositionmeasurements that Mars’ lost at least 500 mbar during the past 4 Gyr.4.4 Surface sinks for CO2

Besides escape of CO2 to space by the various nonthermal loss processes dis-cussed before, CO2 could also be trapped or stored in the regolith or in the form ofice. The SHARAD instrument on board of the Mars Reconnaissance Orbiter (MRO)spacecraft performed systematic radar soundings of the layered deposits at the mar-tian polar areas. From these sub-surface sounding experiments Phillips et al. (2010)found 4 regional reflection-free zones (RFZ) which distinguished qualitatively bytheir radar characteristics. From the reflection analysis the SHARAD-team discov-ered that at least one of these zones (RFZ3) occurs beneath the South polar residualcap and is composed of ∼5 m of solid CO2 underlain by an apparently thin layerof water ice. The equivalent atmospheric pressure of the extrapolated RFZ3 volumeis ∼4–5 mbar. If similar contents are stored in RFZ1 and RFZ2 one would obtainCO2 in ice of the order of ∼15 mbar. This is about the half amount of CO2 which isexpected to be desorbed in the regolith (Zent and Quinn, 1995). We note that someamount of CO2 which is trapped in the ice, was and will be released into the atmo-sphere during periods of climate change so that the CO2 content in the atmosphereand the surface pressure can vary.

Bandfield et al. (2003) measured ∼2 % carbonates in the martian dust, whichis also in agreement with findings of NASA’s Phoenix lander. It is not clear if thecarbonates in the dust can be considered as representative for the top 1–2 km of themartian crust or whether the carbonates formed preferentially in the dust/atmosphereand are artificially enriched in the dust. If the dust is representative of the first sugges-tion, it may be possible that a global layer with a thickness of ∼1–3 km can account

24 H. Lammer et al.

for ∼1–3 bar surface weathered CO2. However, as discussed in Section 3 it seemsvery unlikely that early Mars accumulated a large amount of oxygen which remainedfrom its outgassed water vapour dominated H2O/CO2 steam atmosphere, so that adense photochemically stable CO2 atmosphere (Zahnle et al., 2008) could be kept.

Thus, the amount of presently detected CO2 in the ice or regolith inferred from ob-servations corresponds to ≤30–50 mbar. The total model dependent sum of the CO2

atmosphere at the end of the Noachian ∼4 Gyr ago which escaped to space togetherwith the known CO2 in the surface yield ≤0.1–0.3 bar. This values are much lowerthan the ∼1–5 bar of CO2 required for an effective greenhouse effect responsible forthe “wet and warm” early Mars hypotheses (Forget and Pierrehumbert, 1997).

If early Mars could produce a CO2 atmosphere of several 100 mbar at the endof the Noachian, due to the lack of the efficiency of escape to space the majority of

Fig. 3. Illustration of the evolution of the Mars atmosphere and H2O inventory. As discussed in themain text a complex interplay between the high XUV activity of the young Sun and atmospheric escapeduring the first several hundret Myr kept early Mars most of its time cool and dry. After the solar XUVflux decreased at ∼4 ± 0.2 Gyr ago, outgassing from volcanos most likely resulted in the build-up of aslightly denser, warmer and wetter CO2 atmosphere which may have accumulated a surface pressure ofa few hundred mbars. This atmosphere was lost until present by various nonthermal atmospheric escapeprocesses (Lammer et al., 2013) such as dissociative recombination of carbon bearing molecular ions,solar wind induced sputtering and ion erosion, carbonate precipitation, and serpentinization during theHesperian and Amazonian epochs (Chassefiere and Leblanc, 2011).

Solar activity and Mars’ atmosphere evolution 25

this atmosphere should have been removed to the crust by sequestration in carbonaterocks and partially recycled to the atmosphere under reduced and/or oxidized form(Chassefiere and Leblanc, 2011; Lammer et al., 2013). If there are not enough CO2

deposits hidden under the planet surface, then alternative greenhouse gases are nec-essary for the explanation of standing bodies of liquid water on the planet’s surface∼3.5–4.0 Gyr ago. Figure 3 illustrates the evolution of the martian CO2 atmosphereaccording to our study.

5 Testing Atmosphere Evolution Hypotheses by UV Transit Observations of Ter-restrial Exoplanets within Orbits of Dwarf StarsAs discussed in Section 3, depending on the atmospheric species in the thermo-

sphere, non-hydrostatic dynamically expanded upper atmospheres should develop ifthe XUV flux of a star is several times larger compared to the present solar value.If light atoms such as hydrogen populate the thermosphere the upper atmosphere ex-pands beyond possible magnetosphere boundaries and energetic neutral atoms (ENAs)can be produced via charge exchange collisions with the stellar wind protons (Lam-

Fig. 4. Preliminary model simulation of an expanded hydrogen-rich thermosphere with an exobase level atabout 17REarth and related hydrogen corona which contains outward flowing planetary hydrogen atoms(yellow dots) and stellar wind (green dots) (nsw = 250 cm−3, vsw = 330 km/s, Tsw = 1 MK) producedhydrogen ENAs around an Earth-like planet within an orbit of an M-star habitable zone at 0.23 AU(after Lammer et al., 2011b; Kislyakova et al., 2013). The red and blue dots correspond to ENAs whichare moving towards (red dots) and away (blue dots) from the host star. The dark circle correspond tothe size of the planet. The dashed-line illustrates a planetary obstacle which can be a ionopause ormagnetopause.

26 H. Lammer et al.

mer et al., 2011a, b). If the atmospheric evolution hypothesis which is discussedin this work is correct, then every terrestrial planet with a hydrogen-rich upper atmo-sphere, and exposed to higher XUV fluxes compared to the present solar value shouldbe surrounded by a huge planetary hydrogen corona. This corona will interact withthe stellar wind plasma flow so that a huge ENA cloud can be produced. By applyinga direct simulation Monte Carlo code which is coupled with a stellar wind plasmaflow model one obtains results as shown in Fig. 4. The model approach of thesestudies which is described in detail in Lammer et al. (2011a, b) and Kislyakova et al.(2013) indicates that observations in UV of such hydrogen clouds around terrestrialexoplanets in orbits of low mass M-stars should be possible during transits. One cansee from these preliminary results that a huge fraction of the host star should be cov-ered with the outward flowing planetary neutral hydrogen atoms and produced ENAsduring a transit. By modeling virtual transits and investigation of the absorption of thehost stars’ Lyman-α emission (Holmstrom et al., 2008; Ekenback et al., 2010; Lam-mer et al., 2011a, b) one can study the parameter space which is necessary for testingthe discussed atmospheric evolution theories. By observing the size of XUV-heatedand extended upper atmospheres and related hydrogen and ENA-clouds and by de-termining the velocities of the surrounding hydrogen atoms, the upper atmospherestellar wind environment can be characterized. Due to the large number of M stars,their long periods of stellar activity and small size in comparison to G stars (Gudel,2007), M stars should represent promising candidates for the detection of extendedhydrogen clouds (Lammer et al., 2011b) via transit observations in the UV-range withnear-future space observatories such as the WSO-UV (Shustov et al., 2009) and willshed more light on the atmospheric evolution scenarios which are discussed before.

6 ConclusionsAt the present state of knowledge we conclude that due to Mars’ small size and

mass, together with the high XUV activity of the young Sun the planet’s initiallyproduced water vapour dominated H2O/CO2 steam atmosphere was lost within thefirst 100 Myr after its origin. The solar activity of the young Sun was most likelyhigh enough that a denser CO2 atmosphere could not be built up during most timeof the Noachian which is also in agreement with findings of the OMEGA instrumenton board of ESA’s Mars Express spacecraft of an early escape of most of Mars’ CO2

atmosphere (Bibring et al., 2005). At the end of the Noachian when the solar XUVflux decreased a secondary outgassed CO2 atmosphere may have build up, but waslost again during the Hesperian and Amazonian. The CO2 amount lost from theatmosphere at the end of the Noachian until today was most likely not much higherthan ∼0.3 bar. Until no large hidden subsurface carbonate deposits are discovered,the present picture for Mars’ evolution reveals a planet which was most of his life acold, frozen and dry body with sporadically warmer periods around ∼3.6–4 Gyr ago,but most likely no liquid water oceans on its surface.

Acknowledgments. H. Lammer, P. Odert, H. Groller, K. G. Kislykova, M. L. Khodachenko,acknowledge support from the Helmholtz Alliance project “Planetary Evolution and Life“. M.

Solar activity and Mars’ atmosphere evolution 27

Gudel, K. G. Kislyakova, M. L. Khodachenko, and H. Lammer acknowledge the support bythe FWF NFN project S116 “Pathways to Habitability: From Disks to Active Stars, Planetsand Life”, and the related FWF NFN subprojects, S116 604-N16 “Radiation & Wind Evo-lution from T Tauri Phase to ZAMS and Beyond”, S116 606-N16 “Magnetospheric Electro-dynamics of Exoplanets”, S116607-N16 “Particle/Radiative Interactions with Upper Atmo-spheres of Planetary Bodies Under Extreme Stellar Conditions”. M. Leitzinger acknowledgethe FWF project P22950-N16. A. Hanslmeier, P. Odert and M. Leitzinger acknowledge theFWF projects P22950-N16 and P30949-N36. The authors also thank the EU FP7 project IM-PEx (No.262863) and the EUROPLANET-RI projects, JRA3/EMDAF and the Na2 scienceWG4 and WG5 for support. H. Lammer also thanks his colleague N. Terada and the local or-ganizing committee of the International Symposium on Planetary Science 2011, held in Sendai,Japan during the week where the great earthquake and tsunami occurred on March 11 for theirhelp, support and hospitality during these hard days.

ReferencesAmerstorfer, U. V., H. Groller, H. Lammer, F. Tian, L. Noack, M. Sherf, C. Johnstone, L. Tu, and M.

Gudel, Escape and evolution of Mars’ CO2 atmosphere: influence of suprathermal atom, J. Geophys.Res., 122, 1321–1337, 2017.

Baker, V. R., Water and the martian landscape, Nature, 412, 228–236, 2001.

Bandfield, J. L., D. Glotch, and P. R. Christensen, Spectroscopic identification of carbonate minerals in theMartian dust, Science, 301, 1084–1087, 2003.

Barabash, S., A. Fedorov, R. Lundin, and J.-A. Sauvaud, Martian atmospheric erosion rates, Science, 315,501–503, 2007.

Becker, R. H., R. N. Clayton, E. M. Galimov, H. Lammer, B. Marty, R. O. Pepin, and R. Wieler, Isotopicsignatures of volatiles in terrestrial planets, Space Sci. Rev., 106, 377–410, 2003.

Bibring, J.-P. and the OMEGA team, Mars surface diversity as reveald by the OMEGA/Mars Expressobservations, Science, 307, 1576–1581, 2005.

Brasser, R., The formation of Mars: building blocks and accretion time scale, Space Sci. Rev., 174, 11–25,2013.

Brasser, R., S. J. Mojzsis, S. Matsumura, and S. Ida, The cool and distant formation of Mars, Earth Planet.Sci. Lett., 468, 85–93, 2017.

Chamberlain, J. W., Planetary coronae and atmospheric evaporation, Planet. Space. Sci., 11, 901–996,1963.

Chassefiere, E., Hydrodynamic escape of oxygen from primitive atmospheres: applications to the cases ofVenus and Mars, Icarus, 124, 537–552, 1996.

Chassefiere, E. and F. Leblanc, Methane release and the carbon cycle on Mars, Planet. Space Sci., 59,207–217, 2011.

Chassefiere, E., F. Leblanc, and B. Langlais, The combined effects of escape and magnetic field historiesat Mars, Planet. Space Sci., 55, 343–357, 2007.

Dohm, J. M., V. R. Baker, W. V. Boynton, A. G. Fairen, J. C. Ferris, M. Finch, R. Furfaro, T. M. Hare, D.M. Janes, J. S. Kargel, S. Karunatillake, J. Keller, K. Kerry, K. J. Kimi, G. Komatsu, W. C. Mahaneyk,D. Schulze-Makuch, L. Marinangeli, G. G. Ori, J. Ruiz, and S. J. Wheelock, GRS evidence and thepossibility of paleooceans on Mars, Planet. Space. Sci., 57, 664–684, 2008.

Dorren, J. D. and E. F. Guinan, The Sun in time: Detecting and modelling magnetic inhomogenities onSolar-type stars, in The Sun as a Variable Star: Solar and Stellar Irradiance Variations, (eds. Pap, J. M.,C. Frohlich, H. S. Hudson, and S. K. Solanki), proceedings of IAU Colloquium 143, held in Boulder,Colorado, USA, June 20–25, 1993, Cambridge University Press, Cambridge, U.K., New York, U.S.A.,pp. 206–216 pp., 1994.

Ekenback, A., M. Holmstrom, P. Wurz, J.-M. Grießmeier, H. Lammer, F. Selsis, and T. Penz, Energeticneutral atoms around HD 209458b: Estimations of magnetospheric properties, Astrophys. J., 709, 670–679, 2010.

28 H. Lammer et al.

Elkins-Tanton, L. T., Linked magma ocean solidification and atmospheric growth for Earth and Mars,Earth Planet. Sci. Lett., 271, 181–191, 2008.

Elkins-Tanton, L. T., E. M. Parmentier, and P. C. Hess, The formation of ancient crust on Mars throughmagma ocean processes, J. Geophys. Res., 110, E12S01, 2005.

Erkaev, N. V., Yu. N. Kulikov, H. Lammer, F. Selsis, D. Langmayr, G. F. Jaritz, and H. K. Biernat, Rochelobe effects on the atmospheric loss from “Hot Jupiters”, Astron. Astrophys., 472, 329–334, 2007.

Erkaev, N. V., H. Lammer, L. Elkins-Tanton, P. Odert, K. G. Kislyakova, Yu. N. Kulikov, M. Leitzinger,and M. Gudel, Escape of the martian protoatmosphere and initial water inventory, Planet. Space Sci.,98, 106–119, 2014.

Fairen, A. G., Icy Mars lakes warmed by methane, Nat. Geosci., 10, 717–718, 2017.

Forget, F. and R. T. Pierrehumbert, Warming early Mars with carbon dioxide clouds that scatter infraredradiation, Science, 278, 1273–1276, 1997.

Fox, J. L. and A. B. Hac, Photochemical escape of oxygen from Mars: A comparison of the exobaseapproximation to a Monte Carlo method, Icarus, 204, 527–544, 2009.

Groller, H., V. I. Shematovich, H. I. M. Lichtenegger, H. Lammer, M. Pfleger, Yu. N. Kulikov, W. Macher,U. V. Amerstorfer, and H. K. Biernat, Venus’ atomic hot oxygen environment, J. Geophys. Res., 115,E12017, 2010.

Gudel, M., The Sun in Time: Activity and environment, Living Rev. Sol. Phys., 4(3), 2007.

Gudel, M., E. F. Guinan, and S. L. Skinner, The X-ray Sun in Time: A study of the long-term evolution ofcoronae of solar-type stars, ApJ, 483, 947–960, 1997.

Hansen, B. M. S., Formation of the terrestrial planets from a narrow annulus, Astrophys. J., 703, 1131–1140, 2009.

Holmstrom, M., A. Ekenback, F. Selsis, T. Penz, H. Lammer, and P. Wurz, Energetic neutral atoms as theexplanation for the high-velocity hydrogen aroun d HD 209458b, Nature, 451, 970–972, 2008.

Hunten, D. M., The escape of light gases from planetary atmospheres, J. Atmos. Sci., 30, 1481–1494,1973.

Hunten, D. M., Atmospheric evolution of the terrestrial planets, Science, 259, 915–920, 1993.

Hunten, D. M., R. O. Pepin, and J. C. G. Walker, Mass fractionation in hydrodynamic escape, Icarus, 69,532–549, 1987.

Iijima, Y., K. Goto, K. Minoura, G. Komatsu, and F. Imamura, Hydrodynamics of impact-induced tsunamiover the Martian ocean, Planet. Space Sci., 95, 33–44, 2017.

Ip, W.-H., On a hot oxygen corona of Mars, Icarus, 76, 135–145, 1988.

Kasting, J. F. and J. B. Pollack, Loss of water from Venus. I. Hydrodynamic escape of hydrogen, Icarus,53, 479–508, 1983.

Kerber, L., F. Forget, and R. Wordsworth, Sulfur in the early Martian atmosphere revisited: experimentswith a 3-D global climate model, Icarus, 261, 133–148, 2015.

Kim, J., A. F. Nagy, J. L. Fox, and T. E. Cravens, Solar cycle variability of hot oxygen atoms at Mars, J.Geophys. Res., 103, 29339–29342 1998.

Kislyakova, G. K., H. Lammer, M. Holmstrom, M. Panchenko, M. L. Khodachenko, N. V. Erkaev, P.Odert, Yu. N. Kulikov, M. Leitzinger, M. Gudel, and A. Hanslmeier, XUV exposed, non-hydrostatichydrogen-rich upper atmospheres of Earth-size planets. II: Hydrogen coronae and ion escape, Astrobi-ology, submitted, http://arxiv.org/abs/1212.4710, 2013.

Krestyanikova, M. A. and V. I. Shematovich, Stochastic models of hot planetary and satellite coronas: Ahot oxygen corona of Mars, Solar System Res., 40, 384–392, 2006.

Kulikov, Yu. N., H. Lammer, H. I. M. Lichtenegger, T. Penz, D. Breuer, T. Spohn, R. Lundin, and H. K.Biernat, A comparative study of the influence of the active young Sun on the early atmospheres of Earth,Venus and Mars, Space Sci. Rev., 129, 207–243, 2007.

Kurokawa, H., K. Kurosawa, and T. Usui, A lower limit of atmospheric pressure on early Mars inferredfrom nitrogen and argon isotopic compositions, Icarus, 299, 443–459, 2017.

Lammer, H., Origin and evolution of planetary atmospheres: Implications for habitability, Springer Briefsin Astronomy, Springer Publishing House, Heidelberg / New York, 2013.

Solar activity and Mars’ atmosphere evolution 29

Lammer, H., W. Stumptner, and S. J. Bauer, Upper limits for the Martian exospheric number density duringthe planet B/Nozomi mission, Planet. Space Sci., 48, 1473–1478, 2000.

Lammer, H., J. F. Kasting, E. Chassefiere, R. E. Johnson, Yu. N. Kulikov, and F. Tian, Atmospheric escapeand evolution of terrestrial planets and satellites, Space Sci. Rev., 139, 399–436, 2008.

Lammer, H., P. Odert, M. Leitzinger, M. L. Khodachenko, M. Panchenko, Yu. N. Kulikov, T. L. Zhang,H. I. M. Lichtenegger, N. V. Erkaev, G. Wuchterl, G. Micela, T. Penz, H. K. Biernat, J. Weingrill,M. Steller, O. Ottacher, J. Hasiba, and A. Hanslmeier, Determining the mass loss limit for close-inexoplanets: What can we learn from transit observations?, Astron. Astrophys., 506, 399–410, 2009.

Lammer, H., V. Eybl, K. G. Kislyakova, J. Weingrill, M. Holmstrom, M. L. Khodchenko, Yu. N. Kulikov,A. Reiners, M. Leitzinger, P. Odert, M. Xian Gruß, B. Dorner, M. Gudel, and A. Hanslmeier, UV transitobservations of EUV-heated expanded thermospheres of Earth-like exoplanets around M-stars: Testingatmosphere evolution scenarios, Astrophys. Space Sci., 335, 39–50, 2011a.

Lammer, H., K. G. Kislyakova, M. Holmstrom, M. L. Khodachenko, and J.-M. Grießmeier, HydrogenENA-cloud observation and modeling as a tool to study star-exoplanet interaction, Astrophys. SpaceSci., 335, 9–23, 2011b.

Lammer, H., E. Chassefiere, O. Karatekin, A. Morschhauser, P. B. Niles, O. Mousis, P. Odert, U. V. Mostl,D. Breuer, V. Dehant, M. Grott, H. Groller, E. Hauber, and L. B. S. Pham, Outgassing history and escapeof the martian atmosphere and water inventory, Space Sci. Rev., 174, 113–154, 2013.

Lammer, H., A. L. Zerkle, S. Gebauer, N. Tosi, L. Noak, M. Scherf, E. Pilat-Lohinger, M. Gudel, J. L.Grenfell, M. Godolt, and A. Nikolaou, Origin and evolution of the atmospheres of early Venus, Earthand Mars, Astron. Astrophys. Rev., 26, 1–72, 2018.

Leblanc, F. and R. E. Johnson, Role of molecular species in pick up ion sputtering of the Martian atmo-sphere, J. Geophys. Res., 107, 1–6, 2002.

Lu, W., X. Cang, and A. D. Howard, New Martian valley network volume estimate consistent with ancientocean and warm and wet climate, Nature Commun., 8, 15766, 2017.

Luhmann, J. G. and J. U. Kozyra, Dayside pickup oxygen ion precipitation at Venus and Mars: spatialdistributions, energy deposition and consequences, J. Geophys. Res., 96, 5457–5467, 1991.

Lundin, R., H. Lammer, and I. Ribas, Planetary magnetic fields and solar forcing: Implications for atmo-spheric evolution, Space Sci. Rev., 129, 245–278, 2007.

Lunine, J. I., J. Chambers, A. Morbidelli, and L. A. Leshin, The origin of water on Mars, Icarus, 165, 1–8,2003.

Lunine, J. I., D. P. O’Brien, S. N. Raymond, A. Morbidelli, T. Qinn, and A. L. Graps, Dynamical modelsof terrestrial planet formation, Adv. Sci. Lett. 4, 325–338, 2011.

Ma, Y., A. F. Nagy, I. V. Sokolov, and K. C. Hansen, Three-dimensional, multispecies, high spatial resolu-tion MHD studies of the solar wind interaction with Mars, J. Geophys. Res., 109, A07211, 2004.

Manning, C. V., C. P. McKay, and K. J. Zahnle, Thick and thin models of the evolution of carbon dioxideon Mars, Icarus, 180, 38–59, 2006.

Manning, C. V., Y. Ma, D. A. Brain, C. P. McKay, and K. J. Zahnle, Parametric analysis of modeled ionescape from Mars, Icarus, 212, 131–137, 2010.

Marty, B., The origins and concentrations of water, carbon, nitrogen and noble gases on Earth, EarthPlanet. Sci. Lett., 313, 56–66, 2012.

Matsui, T. and Y. Abe, Impact-induced atmospheres and oceans on Earth and Venus, Nature, 322, 526–528,1986.

McElroy, M. B. and T. M. Donahue, Stability of the Martian atmosphere, Science, 177, 986–988, 1972.

Modolo, R., G. M. Chanteur, E. Dubinin, and A. P. Matthews, Influence of the solar EUV flux on theMartian plasma environment, Ann. Geophysicae, 23, 1–12, 2005.

Morbidelli, A., J. Chambers, J. I. Lunine, J. M. Petit, F. Robert, G. B. Valsecchi, and K. E. Cyr, Sourceregions and timescales for the delivery of water on Earth, Meteor. Planet. Sci., 35, 1309–1320, 2000.

Nagy, A. F., J. Kim, and T. E. Cravens, Hot hydrogen and oxygen atoms in the upper atmospheres of Venusand Mars, Ann. Geophysicae, 8, 251–256, 1990.

30 H. Lammer et al.

Newkirk Jr, G., Solar variability on time scales of 105 years to 109.6 years, Geochim. Cosmochim. ActaSuppl., 13, 293–301, 1980.

Odert, P., H. Lammer, N. V. Erkaev, A. Nikolaou, H. I. M. Lichtenegger, C. P. Johnstone, K. G. Kislyakova,M. Leitzinger, and N. Tosi, Escape and fractionation of volatiles and noble gases from Mars-sizedplanetary embryos and growing protoplanets, Icarus, 307, 327–346, 2018.

Opik, E. J., Selective escape of gases, Geophys. J. Roy. Astron. Soc., 7, 490–509, 1963.

Pham San, Le Binh, O. Karatekin, and V. Dehant, Effects of meteorite impacts on the atmospheric evolu-tion of Mars, Astrobiology, 9, 45–54, 2009.

Phillips, R. J., B. J. Davis, S. Byrne, B. A. Campbell, L. M. Carter, R. M. Haberle, J. W. Holt, M. A.Kahre, D. C. Nunes, J. J. Plaut, N. E. Putzig, I. B. Smith, S. E. Smrekar, K. L. Tanaka, and T. N. Titus,SHARAD finds voluminous CO2 ice sequestered in the martian south polar lyered deposits, abstract�P34A-01, American Geophys. Union, Fall Meeting, 2010.

Quingzhu, Y., S. B. Jacobsen, K. Yamashita, J. Blichert-Toft, P. Telouk, and F. Albarede, A short timescalefor terrestrial planet formation from Hf-W chronometry of meteorites, Nature, 418, 949–952, 2002.

Ramirez, R. M., R. Kopparapu, M. E. Zugger, T. D. Robinson, R. Freedman, and J. F. Kasting, Warmingearly Mars with CO2 and H2, Nat. Geosci., 7, 59–63, 2014.

Raymond, S. N., T. Quinn, and J. I. Lunine, Making other Earth’s: dynamical simulations of terrestrialplanet formation and water delivery, Icarus, 168, 1–17, 2004.

Ribas, I., E. F. Guinan, M. Gudel, and M. Audard, Evolution of the solar activity over time and effects onplanetary atmospheres. I. High-energy irradiances (1–1700A), ApJ, 622, 680–694, 2005.

Rodriguez, J. A. P., A. G. Fairen, K. L. Tanaka, M. Zarroca, R. Linares, T. Platz, G. Komatsu, H. Miyamoto,J. S. Kargel, J. Yan, V. Gulick, K. Higuchi, V. R. Baker, and N. Glines, Tsunami waves extensivelyresurfaced the shorelines of an early Martian ocean, Nat. Sci. Rep., 6, 25106, 2015.

Rosenbauer, E. N. and J. W. Head III, Late Noachian fluvial erosionon Mars: cumulative water volumesrequired to carve the valley networks and grain size of bed-sediment, Planet. Space Sci., 117, 429–435,2015.

Segura, T. L., C. P. Mc Kay, and O. B. Toon, An impact-induced, stable climate on Mars, Icarus, 220,144–148, 2012.

Sekiya, M., K. Nakazawa, and C. Hayashi, Dissipation of the primordial terrestrial atmosphere due toirradiation of the solar EUV, Prog. Theoretical Phys., 64, 1968–1985, 1980.

Skumanich, A. and J. A. Eddy, Aspects of long-term variability in Sun and stars, in Solar Phenomena inStars and Stellar Systems, 349 pp., D. Reidel Publishing Co., Dordrecht, 1981.

Shematovich, V. I., D. E. Ionov, and H. Lammer, Heating efficiency in hydrogen-dominated upper atmo-spheres, Astron. Astrophys., 571, A94, 2014.

Shustov, B., M. Sachov, A. I. Gomez de Castro, I. Ana, and I. Pagano, WSO-UV ultraviolet mission forthe next decade, Astrophys. Space Sci., 320, 187–190, 2009.

Terada, N., Yu. N. Kulikov, H. Lammer, H. I. M. Lichtenegger, T. Tanaka, H. Shinagawa, and T. Zhang, At-mosphere and water loss form early Mars under extreme solar wind and extreme ultraviolet conditions,Astrobiology, 9, 55–70, 2009.

Tian, F., J. F. Kasting, and S. C. Solomon, Thermal escape of carbon from the early Martian atmosphere,Geophys. Res. Lett., 36, Issue 2, CiteID L02205, 2009.

Tu, L., C. P. Johnstone, M. Gudel, and H. Lammer, The extreme ultraviolet and x-ray Sun in time: High-energy evolutionary tracks of a solar-like star, Astron. Astrophys., 577, L3, 2015.

Valeille, A., M. R. Combi, V. Tenishev, S. W. Bougher, and A. F. Nagy, A study of suprathermal oxygenatoms in Mars upper thermosphere and exosphere over the range of limiting conditions, Icarus, 206,18–27, 2010.

Walsh, K. J., A. Morbidelli, S. N. Raymond, D. P. O’Brien, and A. M. Mandell, Alowmass for Mars fromJupiter’s early gas-driven migration, Nature, 475, 206–209, 2011.

Watson, A. J., T. M. Donahue, and J. C. G. Walker, The dynamics of a rapidly escaping atmosphere:applications to the evolution of Earth and Venus, Icarus, 48, 150–166, 1981.

Solar activity and Mars’ atmosphere evolution 31

Wetherill, G. W., Accumulation of terrestrial planets and implications concerning lunar origin, in Origin ofthe Moon, ed. by W. K. Hartmann, R. J. Phillips, and G. J. Taylor (Arizona Press, Tucson), pp. 519–550,1986.

Zahnle, K. J. and J. F. Kasting, Mass fractionation during transonic escape and implictions for loss of waterfrom Mars and Venus, Icarus, 66, 462–480, 1986.

Zahnle, K. J. and J. C. G. Walker, The evolution of solar ultraviolet luminosity, Rev. Geophys., 20, 280–292, 1982.

Zahnle, K. J., J. F. Kasting, and J. B. Pollock, Evolution of a steam atmosphere during Earth’s accretion,Icarus, 74, 62–97, 1988.

Zahnle, K. J., R. M. Haberle, D. C. Catling, and J. F. Kasting, Photochemical instability of the ancientMartian atmosphere, J. Geophys. Res., 113(E11), E11004, 2008.

Zent, A. P. and R. C. Quinn, Simutaneous adsorption of CO2 and H2O under Mars-like conditions andapplication to the evolution of the martian climate, J. Geophys. Res., 100, 5341–5349, 1995.