≡ Menu

Enter the ‘Mirage Earth’

A common trope from Hollywood’s earlier decades was the team of explorers, or perhaps soldiers, lost in the desert and running out of water. On the horizon appears an oasis surrounded by verdant green, but it turns out to be a mirage. At the University of Washington, graduate student Rodrigo Luger and co-author Rory Barnes have deployed the word ‘mirage’ to describe planets that, from afar, look promisingly like a home for our kind of life. But the reality is that while oxygen may be present in their atmospheres, they’re actually dry worlds that have undergone a runaway greenhouse state.

This is a startling addition to our thinking about planets around red dwarf stars, where the concerns have largely revolved around tidal lock — one side of the planet always facing the star — and flare activity. Now we have to worry about another issue, for Luger and Barnes argue in a paper soon to be published in Astrobiology that planets that form in the habitable zone of such stars, close in to the star, may experience extremely high surface temperatures early on, causing their oceans to boil and their atmospheres to become steam.


Image: Illustration of a low-mass, M dwarf star, seen from an orbiting rocky planet. Credit: NASA/JPL.

The problem is that M dwarfs can take up to a billion years to settle firmly onto the main sequence — because of their lower mass and lower gravity, they take hundreds of millions of years longer than larger stars to complete their collapse. During this period, a time when planets may have formed within the first 10 to 100 million years, the parent star would be hot and bright enough to heat the upper planetary atmosphere to thousands of degrees. Ultraviolet radiation can split water into hydrogen and oxygen atoms, with the hydrogen escaping into space.

Left behind is a dense oxygen envelope as much as ten times denser than the atmosphere of Venus. The paper cites recent work by Keiko Hamano (University of Tokyo) and colleagues that makes the case for two entirely different classes of terrestrial planets. Type I worlds are those that undergo only a short-lived greenhouse effect during their formation period. Type II, on the other hand, comprises those worlds that form inside a critical distance from the star. These can stay in a runaway greenhouse state for as long as 100 million years, and unlike Type I, which retains most of its water, Type II produces a desiccated surface inimical to life.

Rodrigo and Barnes, extending Hamano’s work and drawing on Barnes’ previous studies of early greenhouse effects on planets orbiting white and brown dwarfs as well as exomoons, consider this loss of water and build-up of atmospheric oxygen a major factor in assessing possible habitability. From the paper:

During a runaway greenhouse, photolysis of water vapor in the stratosphere followed by the hydrodynamic escape of the upper atmosphere can lead to the rapid loss of a planet’s surface water. Because hydrogen escapes preferentially over oxygen, large quantities of O2 also build up. We have shown that planets currently in the HZs of M dwarfs may have lost up to several tens of terrestrial oceans (TO) of water during the early runaway phase, accumulating O2 at a constant rate that is set by diffusion: about 5 bars/Myr for Earth-mass planets and 25 bars/Myr for super-Earths. At the end of the runaway phase, this leads to the buildup of hundreds to thousands of bars of O2, which may or may not remain in the atmosphere.

Thus the danger in using oxygen as a biosignature: In cases like these, it will prove unreliable, and the planet uninhabitable despite elevated levels of oxygen. The key point is that the slow evolution of M dwarfs means that their habitable zones begin much further out than they will eventually become as the star continues to collapse into maturity. Planets that are currently in the habitable zone of these stars would have been well inside the HZ when they formed. An extended runaway greenhouse could play havoc with the planet’s chances of developing life.

Even on planets with highly efficient surface sinks for oxygen, the bone-dry conditions mean that the water needed to propel a carbonate-silicate cycle like the one that removes carbon dioxide from the Earth’s atmosphere would not be present, allowing the eventual build-up of a dense CO2 atmosphere, one that later bombardment by water-bearing comets or asteroids would be hard pressed to overcome. In both cases habitability is compromised.

So a planet now in the habitable zone of a red dwarf may have followed a completely different course of atmospheric evolution than the Earth. Such a world could retain enough oxygen to be detectable by future space missions, causing us to view oxygen as a biosignature with caution. Says Luger: “Because of the oxygen they build up, they could look a lot like Earth from afar — but if you look more closely you’ll find that they’re really a mirage; there’s just no water there.”

The paper is Luger and Barnes, “Extreme Water Loss and Abiotic O2 Buildup On Planets Throughout the Habitable Zones of M Dwarfs,” accepted at Astrobiology (preprint). The paper by Hamano et al. is “Emergence of two types of terrestrial planet on solidification of magma ocean,” Nature 497 (30 May 2013), pp. 607-610 (abstract). A University of Washington news release is also available.


Comments on this entry are closed.

  • Harry R Ray December 3, 2014, 13:41

    This is actually (kind of) GOOD news. Over a decade ago, Lissaur et al argued that Type Ones would have their ENTIRE atmosphere stripped away, and Type Twos would wind up with venus-like atmospheres (ie COLD venus’s). As planetary atmousphere science geta even more sophisticated (and hopfully more accurate) I hope this trend continues in favor of habitability of these kind of planets. Case-in-point: Kepler 186f, obviously a Type Two. If this was origionally a “WATER WORLD” a partial evaporation could lead to the emergiance of land masses, but, what would an overabundance of oxigen EARLY in its history mean for the EVOLUTION on that planet?

  • Ashley Baldwin December 3, 2014, 14:48

    This is an example of why we need large exoplanet “characterisation ” telescopes. We need the good SNRs and resolution over a wide bandwidth that they provide to avoid “false” bio signatures caused by abiotic oxygen production . So much simulation dependent astrophysics because there is no real data available. Until we have large aperture telescopes this is what will happen , starting with the TESS and JWST combination from 2018. Claim and counter claim . Too many exoplanets and not enough characterisation. Just wait till the Gaia and PLATO results are coming in. Oh for EChO !

  • Michael December 3, 2014, 16:27

    I am surprised it takes so long to form red dwarfs as they have very strong surface gravities. Jupiter (lower mass) is theorised to have formed in a few million years so why are RD so hard to form. Is the UV causing ionisation issues with the hydrogen/helium and slowing down the collapse?

    This is an interesting article on Red dwarfs.


  • Robert Feyerharm December 3, 2014, 17:23

    On the bright side, the Mirage Earth study establishes the possibility that bone dry, desert planets with no appreciable life may support a habitable environment with a breathable oxygen atmosphere. These kinds of world pop up in science fiction a lot (e.g., Tatooine and Arrakis) but I’ve always been skeptical that a planet with virtually no water, plate tectonics, or biosphere could evolve an oxygen atmosphere. Early settlers may need to guide a few comets to the planet to supply water, but otherwise these planets would be ideal for colonization.

  • Alex Tolley December 3, 2014, 18:40

    In the earlier, steam + O2 epoch, the H2O + O2 signatures would be even more misleading. I would hope that other information from the ‘scopes could be used to separate out true bio-signatures from false positives. Would the presence of oceanic water be useful, e.g. by making the overall color of the world match our “pale blue dot”?

  • Antonio December 3, 2014, 19:04

    Mi scuso, per la domanda, un po’ ingenua.
    Ma questi pianeti, così ricchi di ossigeno, in un futuro remoto, potrebbero essere “terraformati” e resi più abitabili, per noi esseri umani? Non so, con che tipo di tecnologia, ma solo il fatto di essere così ricchi di ossigeno, me li fanno apparire, potenzialmente interessanti.
    Noi vediamo, l’effetto serra causato dall’anidride carbonica e dal metano, anche sul nostro pianeta(oltre che su Venere)ma anche l’ossigeno, in grandi quantità, produce il medesimo effetto?
    Un saluto a tutti i lettori del “blog” da Antonio

    Google Earth translation:

    I apologize for the question, a bit ‘naive.
    But these planets, so rich in oxygen, in the distant future, could be “terraformed” and made more habitable for us humans? I do not know, with that kind of technology, but just the fact of being so rich in oxygen, I make them appear, potentially interesting.

    We see, the greenhouse effect caused by carbon dioxide and methane, also on our planet (and Venus), but also oxygen, in large quantities, produces the same effect?

    Greetings to all readers of the “blog” by Antonio

  • WILLIAM December 3, 2014, 21:07

    The problems that we can expect to see in finding different planets outside our system can be exemplified in this extremely interesting article:


  • Brett December 4, 2014, 3:33

    @Robert Feyerharm

    On the bright side, the Mirage Earth study establishes the possibility that bone dry, desert planets with no appreciable life may support a habitable environment with a breathable oxygen atmosphere.

    That’s only if the planet absorbed enough of the “ten times Venus’s atmosphere” envelope so that it was down below 1.6 bar. Above that, the oxygen atmosphere would be poisonous.

  • Ronald December 4, 2014, 5:26

    A few observations:
    – Wouldn’t such a lifeless but O2 rich planet be recognizable by the very fact that it’s O2 content is so extremely high?
    – Would Type I planets be habitable because of the much shorter runaway greenhouse period and resulting lower O2 levels plus presence of water? What conditions are needed for a Type I as opposed to Type II?
    – Also in reply to Antonio: I would expect that such a mirage planet, even with amiable atmospheric O2 level would not be habitable for ever, without introduction of CO2 assimilating (photosynthesizing) life (i.e. plants), to replenish the O2. Otherwise the O2 would be a finite resource.

  • Michael December 4, 2014, 14:06

    If UV greenhouse runaway processes create oxygen rich planets why has Venus little to no free oxygen. It’s gravity is strong enough to hold oxygen back surely, but photo-dissociation is mostly above the cold trap and the molecules get an extra kick to leave the planet. I believe the same effect will act on these other worlds, oxygen and hydrogen will leave the planet. There is unlikely to be an active dynamo to form a protective magnetic field either.

  • Robert Feyerharm December 4, 2014, 15:42

    @Brett –
    That’s only if the planet absorbed enough of the “ten times Venus’s atmosphere” envelope so that it was down below 1.6 bar. Above that, the oxygen atmosphere would be poisonous.

    Yes – you’d have to catch a planet with the right characteristics at the right time. For example, a geologically dead planet, either smaller or older than the Earth, which meets the conditions: 1) has already absorbed most CO2 into carbonates and is no longer venting CO2 into the atmosphere, and 2) has a weak magnetosphere that allows some of the excess oxygen to be drawn into space after photo-dissociation of water.

  • Eniac December 4, 2014, 21:40

    I have to say I am not convinced. According to here: http://en.wikipedia.org/wiki/Oxygen_cycle, the residence time of oxygen in the atmosphere is only 4500 years, and 99.5% is stored in the lithosphere. It seems to me that regardless of what happens with the other atmospheric components, any free oxygen will quickly be absorbed in the lithosphere. Unless, of course, there is photosynthesis, the only significant source of atmospheric oxygen on Earth.

    The authors are aware of this, and summarize it well in this passage:

    Because of its highly reactive nature, oxygen is not typically stable in terrestrial planet atmospheres and tends to quickly react at the surface by stripping electrons from reducing substances in a process known as oxidation. Oxygen on Earth is continuously produced by photosynthesis; in the absence of a steady source, continental weathering, volcanic outgassing of reducing gases, and oxidation of basalt via hydrothermal processes at oceanic ridges would quickly remove most of the atmospheric O2

    They then go on and speculate about many possible mechanisms that might slow or inhibit the removal of oxygen, but as far as I can tell that still leaves the abiotic oxygen atmosphere as a remote possibility, not the accomplished fact that has been represented here (and which, to be fair, their abstract suggests).

  • TVG December 5, 2014, 5:31

    How would this be affected by a young planet not yet being trapped in a tidal lock situation?

  • Robert Feyerharm December 5, 2014, 13:30

    @Eniac –

    That’s correct – you have to make a few more assumptions to arrive at a world with a dense oxygen atmosphere. On a geologically and biologically dead world with very little weathering and outgassing of reducing gases, once the surface has completely oxidized I imagine oxygen could accumulate in the atmosphere from photo disassociation of water vapor. But there are a lot of ifs. Peter Ward’s book “The Life and Death of Planet Earth” outlines scenarios where, after the oceans are cooked away by a more luminous sun in the distant future, Earth might be surrounded by a massive oxygen envelope.

  • Galacsi December 5, 2014, 19:25

    This slow accretion of planets around M stars looks almost like a strawman to me !
    Stars (and planets) accretion cannot be from gravity only , and electrostatic and magnetic forces surely innitiate it ( IMO).

  • andy December 6, 2014, 8:30

    See also another recent preprint, Ramirez & Kaltenegger “The Habitable Zones of Pre-Main-Sequence Stars” – they find similar results for late-type stars.

    As regards Venus, the results are consistent with it losing its water during accretion, and if the planet has a magma ocean (whose lifetime would be prolonged by the conditions of a runaway greenhouse) it should be quite good at consuming the atmospheric oxygen. Certainly I have seen results which suggest that the D/H ratio in the Venusian atmosphere can be explained by the occasional comet impact onto an initially dry planet rather than being the signature of evaporated oceans.

  • Dave Moore December 6, 2014, 15:44

    I had a thought on this:

    These papers look at water loss from the top of the atmospheric column.

    What about from the bottom?

    If, in particular a super Earth–but this may apply to any terrestrial type planet–a planet through a mixture of stellar insolation, internal heating (tidal and radioactive), and the adiabatic lapse rate of a gaseous column of thousands of bars develops a magma ocean, then, not only will the magma ocean absorb oxygen, but it will become saturated with any atmospheric volatile: carbon dioxide, water vapor, methane. I’m not sure how far down into the mantle this mixing will occur as rock saturated with volatiles as obviously less dense than unsaturated rock, but this process has the potential to suck large quantities of volatiles out of the atmosphere. After all, these are pressures greater than the bottom of Earth’s oceans, which is high enough to saturate crustal rock when it’s heated and subducted.

    I liken the process to making carbonated water in a soda bomb where you pressurize water with CO2, and at some point, the magma ocean will develop a crust, thereby acting as a bottle cap, retaining the volatiles for gradual release through volcanism. So, while the planet may initially have a desiccated surface, volcanism may produce oceans over time.

  • Rob Henry December 11, 2014, 1:41

    Dave, the magma ocean would you describe would be a true liquid not like Earth’s that can transmit S waves like a solid. As such it would be completely stratified into layers of different mineral density (thus with zero convection), and the most common oxidized ones such as ferric v ferrous ones would already be on top. One atmosphere of O2 is 10,000 kg per square metre at one g. O2 solubility in water is maximum at 0C where it is 15 g per cubic metre per bar. If that holds for fully oxidized rock (as we expect on top of our ocean) at much higher temperature, it would have to diffuse to a depth of 700 km before it could ever hold a comparable supply of O2 to the atmosphere, and several thousand before it could deplete it. All in all, I suspect our setup with a solid crust that can deliver partially reduced magma from below (and, ultimately from in contact with the iron core) is by far the better for absorbing free O2 given millions of years.

  • Eniac December 12, 2014, 0:19


    At its partial pressure, O2 is 2000 kg (2 tons) per square meter. Rock is about 50% oxygen, by mass, so a cubic meter of rock contains about 1-2 tons of oxygen. it seems to me that all the oxygen in the atmosphere, were it reacted with silicon and/or metals, would make a layer of rock no more than 2 m thick.

    Just another way to look at this. I am not sure at this point which one is more appropriate, but I think the solubility of O2 in water is hardly relevant, here.

  • Eniac December 12, 2014, 0:31

    We could consider another way to get an abiotic oxygen atmosphere fit for breathing: Imagine a very hot, tidally locked planet that was 2000 degrees or so on the day side. It is plausible that the intense heat on the day side could result in pyrolysis of magma, with silicon and metals raining back down and some of the free oxygen escaping to the night side. Temperatures on the night side might be low enough to be comfortable, and the oxygen supply might be sufficient to keep up a breathable atmosphere.

  • Eniac December 12, 2014, 0:37

    Yet another, semi-biotic way to make an oxygen atmosphere: Self-replicating probes have landed on a planet, covering it with metal structures. The refinement of metals from the rock using photovoltaic solar energy produces oxygen as a waste product, which forms an atmosphere that is continually maintained by the “mechanosphere” in the same manner as the biosphere does on Earth.

  • Rob Henry December 12, 2014, 18:04

    Sure Eniac, Si metal does have a low enough density to rise to the top, but you won’t get any of that in protoplanetary disks with C/O < circa 0.9. Also note 1atm is the standard reference pressure for gas, and the absorption is proportional to pressure. A stp O2 column on Earth is 10,000 kg, even if the current one is 2,000.

  • andy December 13, 2014, 14:35

    Dave, the magma ocean would you describe would be a true liquid not like Earth’s that can transmit S waves like a solid. As such it would be completely stratified into layers of different mineral density (thus with zero convection)

    Liquids can and do support convection, so I’m not entirely following your argument here. The magma ocean should be losing heat from the surface and being supplied with heat from its base, so what’s preventing convection?

  • Rob Henry December 14, 2014, 16:04

    Andy, as it becomes a true liquid near the very not surface, it becomes different liquids layered, not one nearly homologous liquid whose density variations are merely the result of pressure rise. Think of water atop mercury. Heating the system will never cause the mercury to rise a millimetre (let alone many kilometres) through the water because it just takes too much energy to work against gravity.

  • Rob Henry December 14, 2014, 16:23

    Further to the above, we know these materials are not completely soluble in each other near their melting points, or they would not tend to separate out into pure minerals so often. Looking at sea water, even tiny density changes associated with changes in salinity can shut down convection. They can also augment it so, under the right circumstances, this might even help convention WITHIN one of these layers, but even that is doubtful, the most important lesson being that a tiny permanent difference in density between any two layers is sufficient to bring it to a halt.

  • andy December 16, 2014, 5:50

    @Rob Henry: that argument would suggest that lava lakes do not convect. Observations indicate that they do. (There is a stagnant lid on top, but this is much thinner than your 700km requirement – we do not know of any lava lakes that deep)

  • Rob Henry December 16, 2014, 17:09

    @andy, do you have one particular system in mind, and why do you think that its visual surface layer is mixing well from its depths and above a point that S waves are supported? It is hard for me to comment, but you must note that liquids in planets are not like gas in stars – its takes far more heating energy for a liquid to expand its volume than the theoretical minimum.

    From minimum energy requirements alone, to raise a mineral through a 1000 km layer that is just 1% denser requires 100kJ/kg, and with the models you are used to this may look as if it is possible, but a liquid would need one, possibly two, orders of magnitude more energy than this. Heck some liquids even contract when heated! I have extreme difficulty in imagining how local heat retention building faster than conduction can dissipate it can build to an extra 1000 – 10,000 kJ/kg that would allow convection.

  • Rob Henry December 17, 2014, 18:50

    andy, I came across an interesting short video today that emphasises how very different in chemical nature mantle material is than erupted magma even that just a few hundred km from the surface. Note how all geologist seem to assume that to be the case, as though they realise it must have already separated out by the point it becomes a true liquid. To me, convection within the one fluid in the top portion is no problem at all.