≡ Menu

Deep Time: Exoplanet Atmospheres in Perspective

As we improve our instrumentation, the search for worlds where life can flourish will generate more and more Earth-sized targets for extended investigation. Here time plays an interesting role, for our own planet seen two billion years ago would present a different aspect than the Earth of today. Atmospheres evolve, a fact that Lisa Kaltenegger has studied in a series of papers in recent years, working with colleagues at Cornell’s Carl Sagan Institute, where she is director. The result is a series of spectral templates applicable to Earth-like planets at various stages of evolution.

We have only one known example of a living planet to work with, so Kaltenegger’s atmospheric models are designed to match the Earth at different stages of development. The prebiotic Earth of 3.9 billion years ago is saturated with carbon dioxide, while what the paper refers to as Epoch 2, some 3.5 billion years ago, is a world without oxygen. Three more epochs can be defined covering the rise of atmospheric oxygen, from levels below 1 percent beginning about 2.4 billion years ago (the Grand Oxygenation Event) to 10 percent oxygen (the Neoproterozoic Oxygenation Event) and the modern Earth atmosphere with oxygen levels at 21 percent.

“Using our own Earth as the key, we modeled five distinct Earth epochs to provide a template for how we can characterize a potential exo-Earth – from a young, prebiotic Earth to our modern world,” Kaltenegger said. “The models also allow us to explore at what point in Earth’s evolution a distant observer could identify life on the universe’s ‘pale blue dots’ and other worlds like them.”

Image: Lisa Kaltenegger, director of the Carl Sagan Institute at Cornell and Associate Professor in Astronomy. Credit: Cornell University.

The resulting database draws on a solar evolution model that establishes the solar flux through the epochs described, all applied to a hypothesized planet with the same mass and radius as Earth orbiting at 1 AU from an evolving Sun. Note that in previous work, Kaltenegger and colleagues have developed reflection and emission spectra of Earth through its geological history, and have also applied these data to different classes of host stars. This model is different, in that it’s focused on transmission spectra, meaning we’re looking at what data a transiting planet would present to our spectrographs as the light of its host star passed through its atmosphere.

The high-resolution database that comes out of this work goes through visible wavelengths into the infrared (0.4–20 μm) through geological time. Transmission spectra of this kind are the current tool for probing exoplanet atmospheres and may well be the first derived from Earth-sized worlds in habitable zone orbits.

From the paper:

Throughout the atmospheric evolution of our Earth, different absorption features dominate Earth’s transmission spectrum…with CH4 and CO2 being dominant in early Earth models, where they are more abundant. O2 and O3 spectral features become stronger with increasing abundance during the rise of oxygen (Epoch 3–5). High-resolution (λ/Δλ = 100,000) spectral features that indicate life on Earth —the combination of O2 or O3 with a reducing gas like CH4 or N2O—can be detected for oxygen levels as low as 0.01 present atmospheric levels (0.21% O2), which correspond to a Neoproterozoic Earth model and a time about one to two billion years ago in Earth’s history.

The authors used a climate-photochemistry tool called EXO-Prime to compile the database, noting that its results have been validated for visible to infrared wavelengths when Earth is observed as an exoplanet. We’ve talked in the past about the EPOXI mission’s look back at Earth (see EPOXI: Clues to Terrestrial Worlds), but we also have data from Mars Global Surveyor and numerous Earthshine observations that have demonstrated the tool is robust. The results clearly show how different absorption features dominate the Earth’s spectrum through the course of atmospheric change.

Image: This artistic depiction shows exoplanet Kepler-62f, a rocky super-Earth size planet, located about 1,200 light-years from Earth in the constellation Lyra. Kepler-62f may be what a prebiotic Earth may have looked like. Other exoplanets may look similar. Credit: NASA Ames/JPL-Caltech.

For those wanting to explore these issues further, the high-resolution transmission spectra database can be found online at www.carlsaganinstitute.org/data. The authors see it as a tool for data interpretation as well as optimizing observation strategy and training data retrieval methods as new instruments become available. The work is applicable not only to the introduction of the ground-based Extremely Large Telescopes like the Giant Magellan Telescope and the Thirty Meter Telescope, but also space-based missions like the James Webb Space Telescope as well as future mission concepts like LUVOIR (Large UV Optical Infrared telescope) and HabEx (Habitable Exoplanet Observatory).

Kaltenegger notes that these telescopes will be identifying Earth-like planets out to about 100 light years:

“Once the exoplanet transits and blocks out part of its host star, we can decipher its atmospheric spectral signatures. Using Earth’s geologic history as a key, we can more easily spot the chemical signs of life on the distant exoplanets…Our transmission spectra show atmospheric features, which would show a remote observer that Earth had a biosphere as early as about 2 billion years ago.”

The paper is Kaltenegger et al., “High-resolution Transmission Spectra of Earth Through Geological Time,” Astrophysical Journal Letters Vol. 892, No. 1 (26 March 2020). Abstract.


Comments on this entry are closed.

  • Alexander Tolley March 31, 2020, 14:35

    Sample of one of teh transmission data files. Can someone explain why the y-value is “effective height” and how this is used?

    # Transmission spectra for Earth through geological time
    # wavelength (microns) effective height (km)
    0.4000176102 27.59539727
    0.4000208136 27.59522844
    0.4000240014 27.59502369
    0.400027205 27.59480083

    • wdk March 31, 2020, 21:35

      Alexander Tolley,
      A trial balloon or two: Scale height or optical depth.

      Either case, the unit that makes reductions of exp(-1) with altitude. One goes up and used with pressure or density reduction. The optical depth goes down and adjusted to wavelength.

      Scale height for a constant temperature region of the atmosphere? hsc = (R/m) T/g.

      In the lower atmosphere, the scale heights are around 8.5 km.

      Since an infra red wavelength is mentioned, it might be a measure of absorption in the atmosphere or optical depth. A better fit than
      scale height, but surprised that if this is “deep time”, that there is so little shift in the values, assumingthe atmospheric constituents changed considerably.

    • Ashley Baldwin April 1, 2020, 5:42

      I wonder if it represents the lowest practical atmospheric depth at which transmission spectra can be observed for the terrestrial planet type considered here ?

      One of the big draw backs of this observation method . The high incident angle light light transmitted through exoplanet atmospheres is subject to refraction outwards – limiting its penetration depth . Tyler Robinson modelled the transmission spectra for various atmosphere types including putative “exo Earth” / terrestrial style. Exactlythose considered here . For different spectral star classes . He showed that the lowest point at which these latter can be observed was around this sort of 27 km depth – for terrestrial atmospheres anyway . Significant as this is likely above the level of any troposphere, below which most of the atmosphere ( biosignatures and all) resides for Earth. This depth is to some extent related to the wavelength of the incident starlight . With longer IR peak emission wavelengths ( of later spectral class stars) reaching to lower depths . In general only light from later than M2 stars could penetrate below Earth dimension terrestrial tropopause levels before refracting outwards – and not by much.

      • Volucris April 2, 2020, 8:08

        Shouldn’t it refract towards towards surface, as lower depths mean higher densities, which mean lower speed of light (https://en.wikipedia.org/wiki/Atmospheric_refraction)? I suppose it would still set some sort of limit to maximum depth, as rays passing deeper in to the atmosphere would just strike the surface of the planet.

    • Mike Serfas April 3, 2020, 8:06

      When I see a link to an astronomy or physics paper the first thing I do is look it up in arxiv.org (generally I search an author name, since some journals have hit upon the idea of changing the title, but in this case it was the same) With https://arxiv.org/ftp/arxiv/papers/1912/1912.11149.pdf in hand I can see that an effective height is given as the ordinate in Figure 2, but the authors seem to expect readers to be familiar with the idea. A web search for “effective height” “spectroscopy” delivered https://www.sciencedirect.com/science/article/pii/S2405675817300295 which gives me some reason to infer that it is the apparent size of the transiting body based on the amount (area) of the star’s light its atmosphere blocks.

      • Alexander Michael Tolley April 3, 2020, 13:19

        Thank you. That reference was near perfect. Section 2.6 explaining effective height was most informative and answers my question.

  • Laintal March 31, 2020, 16:51

    Hi Paul

    Thanks for another interesting post.

    Here are three interesting papers related to the same general topic

    From Super-Earths to Mini-Neptunes

    Salty Oceans in Low Mass Habitable Planets and Global Climate Evolution

    Stellar Driven Evolution of Hydrogen-Dominated Atmospheres

  • Spaceman April 1, 2020, 9:48

    Deep time is a fascinating topic! I forget where I heard it said, but I am pretty sure it was Neil deGrasse Tyson who commented that Earth was like an “alien planet”. Indeed, there were times when earth was very a very different place than it is today. Back in the Hadean, for example, our clement world was an unrecognizable sterile, magma-covered orb battered by asteroids on a regular basis. Then, life arose (some say due to an exceedingly improbable chemical fluke), but life remained microscopic for ~2 billion years during the “boring billion” and, at least once– scientists believe– earth was completed covered in snow and ice aka “snowball earth.” Our perspective is both expansive and limited– limited in the sense that it is hard for us to imagine much changing as much does remain relatively static during the thin slice of time that is a human lifespan. Expansive in the sense that we have been able to use our unique cognition to peer back and discern earth’s past and to look forward to make educated guesses as to what the planet’s future will likely hold. Our planet will eventually be “alien” again. Searing hurricanes and boiling oceans in the lead up to 1 billions years from now, anyone? The question is: will the human species be here to witness these changes, or, will we have moved on to other worlds, or, will we have gone extinct much like the vast majority of other species. I personally take a neutral position on the view that intelligence is necessarily an asset when it comes to humanity’s longevity since there are those who make a good case that we will use our ingenuity to extend ourselves into the indefinite future yet there are also those who make an equally good case that our ingenuity has the potential to bring about our own extinction.

  • Geoffrey Hillend April 1, 2020, 16:10

    The incident angle is dependent on the phase angle. For example when just before the exoplanet goes behind the star it will have an emission and reflection spectrum. There is some reflection when the planet appears at it’s greatest elongation the star. Transmission spectroscopy happens when the exoplanet is directly in front of the star so the light passes through different layers of the atmosphere before it is blocked by the surface of the planet, so there is no reflection away from the planet. There is refraction, diffraction and Rayleigh scattering. The light is absorbed by different layers of thickness of the exoplanet’s atmosphere at different heights as it passes directly in front of it’s star and the flux re-emitted with a spectral signature. Consequently, with transmission spectroscopy we can get the wavelengths from the visible spectrum through the mid infra red of an Earth like exoplanet in front of a G class star.

    Also Earth’s atmosphere is opaque to the mid and far infraed since the ground state of carbon dioxide absorbs light in the mid infra red An exoplanet’s atmosphere looks larger in the infra red in front of it’s star since it is opaque to infra red. Consequently, there should be some absorption lines of infra red with an atmosphere which contains carbon dioxide like Earth, Venus and Mars. Earth’s atmosphere is transparent to the visible light and near infrared which reach the ground.

  • Geoffrey Hillend April 1, 2020, 20:50

    Excuse me. The refraction, diffraction and Rayleigh scattering happens with reflection and scattering off the top of the atmosphere. Seager, 2010, Exoplanet Atmospheres, p. 107, Scattered Light Spectrum.

    The transmission spectroscopy happens when some of the starlight passes through the atmosphere when the exoplanet is directly in front of the star. The atmospheric molecules like carbon dioxide, oxygen, nitrogen are mixed evenly throughout the entire height of the atmosphere, the troposphere and stratosphere, so all we need is the starlight to pass through the top of the troposphere, or stratosphere of the exoplanet and there still will be absorption or spectral lines. The exception is ozone which is concentrated in the upper stratosphere in our atmosphere.

    If we find in an exoplanet’s atmosphere a spectral signature absorption line at ten microns, there will be ozone there which means it is that there has to be oxygen in that atmosphere.

    • Andrei April 1, 2020, 22:24

      Yes, but it would not show how much oxygen, only the presence of minute amount as Venus got an Ozone layer too.

  • Michael Fidler April 2, 2020, 10:20

    I keep thinking about the green flash that are observed at sunset on earth and what about inversions? This could trap light below a layer and may release it as temperatures change from day to night. One of the most fascinating optical mirages I saw when stationed in Alaska was fata morgana: http://www.northernfocusphotography.com/img/s9/v96/p819364906-5.jpg
    The planets close to M dwarfs may have interference from planet wide aurorae.

    • Michael Fidler April 7, 2020, 9:56

      Saturn’s Great Atmospheric Mystery Explained by New Data From NASA’s Cassini.
      “New analysis of data from NASA’s Cassini spacecraft finds a viable explanation for what’s keeping the upper layers of Saturn, and possibly the other gas giants, so hot: auroras at the planet’s north and south poles. Electric currents, triggered by interactions between solar winds and charged particles from Saturn’s moons, spark the auroras and heat the upper atmosphere. ”


  • Geoffrey Hillend April 3, 2020, 1:26

    Andrei, I read online about the ozone layer on Venus, but it was detected with the European space agency’s Venus Express spacecraft in orbit around Venus from the night side of Venus by the occultation of distant stars or starlight through the upper atmosphere of Venus which is not possible from the distance of an exoplanet from Earth. Our spectroscopes from Earth can’t even detect that ozone layer which is very thin compared to Earth since Venus ozone is made from the UV photo dissociation of carbon dioxide. The same thing happens on Earth, but there is also a lot of molecular oxygen in our atmosphere which is also photo disocciated by UV photons or light. The two oxygen molecules in the O2 break apart into two O1 or single atoms. One of the single O1 atom combines with an O2 oxygen atom and we get O3 or ozone. There is very little free oxygen in Venus atmosphere. Consequently, I don’t think anyone or ET’s in another star system could not detect any ozone on Venus from that distance since Venus ozone layer The ozone layer on Earth is fifty kilometers wide, but very much thinner on Venus. Also I do think that if we can detect ozone on an exoplanet, there must be a lot of oxygen there to make the ozone.

  • anon86 April 9, 2020, 23:01

    Has anyone thought about whether intelligent life would produce a telltale signature in an atmosphere?

    Deep time is also a factor. Suppose an alien civilization existed at one point, but died. Is it possible that some permanent signature would be left that we could see? I mean, obviously if they were building Dyson spheres, that’s pretty clear. But how about a less advanced civilization?

  • anon86 April 11, 2020, 16:11

    Here’s another idea. Leave a message for an alien civilization that could be seen via their telescopes and would stay around for billions of years.

    I’m thinking paint a planet.