≡ Menu

Colors of a Living World

Gliese 581d seems more and more to be considered a habitable zone planet, as Siddharth Hegde (Max Planck Institute for Astronomy) and Lisa Kaltenegger (Harvard-Smithsonian Center for Astrophysics) describe it in a new paper. They’re homing in on how to characterize a rocky exoplanet and point to HD 85512b and Gliese 667Cc as well as Gl581d as examples, but they also assume that we’ll be seeing more and more habitable zone worlds as the Kepler mission continues its work, so how we learn more about these planets becomes a big issue.

In the absence of missions like Terrestrial Planet Finder or ESA’s Darwin, which would allow us to analyze an exoplanetary atmosphere for biomarkers, what else can we do to find the places where life exists? Hegde and Kaltenegger look hard at a planet’s color to find the answer. Specifically, they’re interested in what’s known as a color-color diagram, which takes advantage of the fact that an object can be observed at a variety of wavelengths, with a different brightness becoming apparent in each band observed. ‘Color’ in this sense refers to the difference in brightness between different bands, easily plotted on a color-color diagram.

Image: Voyager 1’s famous image of the ‘pale blue dot’ that is our world. Can we use color information from direct images of exoplanets to learn which are most likely to house life? Credit: NASA.

Analyzing an exoplanet in visible light on a color-color diagram can reveal some of the basic physical properties of the planet, assuming cloud cover is not problematic. The new paper homes in on the kinds of environment on Earth that can support extreme forms of life and considers how we might identify equivalent environments on an exoplanet:

Small changes in temperature, pH or other physical and geochemical factors… can lead to such environments being dominant on a potentially habitable exoplanet, what could govern evolution of life. These various “extreme” surface environments on Earth have characteristic albedos in the visible waveband (0.4 µm – 0.9 µm) that could be distinguished remotely. We therefore explore the color signatures that are obtained from the surface environments inhabited by extremophiles as well as test our approach using measured reflection spectra of extremophiles.

Of course, detecting surface features in a reflection spectrum is not itself a detection of life, and the authors are quick to point out that their method is a diagnostic that has to be used in conjunction with a study of the exoplanetary atmosphere. But the paper is an interesting attempt to link the known characteristics of extremophile environments to observational astronomy, one that acknowledges that as we get to the point where we can study distant rocky worlds through actual imagery, we’ll be working at extremely low resolution at the limits of our instruments.

Nonetheless, there is much we can do to distinguish the percentage of the surface covered in water or vegetation or desert, a method that should allow us to prioritize the exoplanets best suited for follow-up spectroscopy. The method builds on prior studies of the vegetation red edge caused by the absorption of red light by photosynthesis, but expands that work to consider different life forms that may live on or below the surface. Piezophiles, for example, thrive under extreme oceanic pressure, while halophiles grow in high salt concentrations.

Although some extremophiles — lichens, bacterial mats and red algae — may be detected by direct albedo measurements, we would have no way of directly detecting many extremophiles in a reflection spectrum. Even so, we can do useful work: The idea here is to identify the kind of surface features that would be common in those environments that supported extremophiles living within them. And the range of characteristic surfaces that can be detected by these methods is large, ranging from water, snow and salt to sand, red-coated algae water and trees.

There are plenty of wild cards here, including the kind of star the planet orbits, which could have a profound effect on the signature of vegetation. As we detect rocky planets around different classes of star, we’ll have to adjust our methods accordingly. From the paper:

…the chlorophyll signature for planets around hot stars, may have a “blue-edge” to reflect some of the high energy radiation in order to prevent the leaves from overheating… Chlorophyll signature for planets orbiting cooler stars, may appear black due to the total absorption of energy in the entire visible waveband such that plants gain as much available light as possible for photosynthetic metabolism… Therefore, the positions of trees, microbial mats and lichens [on the diagram shown in the paper] are only valid for an Earth-analog planet orbiting around a Sun-like star and should be taken as guides. The albedo of vegetation and chlorophyll-bearing organisms for non-Sunlike stars requires further study.

Hegde and Kaltenegger’s paper points toward the first kind of work we’ll be able to perform on an exoplanet in the habitable zone once we’ve been able to acquire a direct image of it. By working with extremophiles, the researchers establish environmental limits for life on our own planet, a useful baseline for our first examinations of other terrestrial worlds. The basic filter photometry in visible light used here can provide a first step in probing these planets by identifying characteristic colors, linking them to environmental niches that support life. We would then await the space-based instruments needed to analyze the atmospheres of high-value targets.

The paper is Hegde and Kaltenegger, “Colors of Extreme ExoEarth Environments,” accepted for publication in Astrobiology (preprint).


Comments on this entry are closed.

  • jkittle October 5, 2012, 10:16

    I am not sure how to factor in the stellar color in the expectations. If earth were transported immediately tot he habitable some of a red dwarf, would it still appear “pale blue”? or if we sent it to circle a star higher temperature than our sun would the atmosphere initially appear blue and how long would it stay that way?
    I find it incredible we are now having discussions about the color of planets in habitable zones, 20 years ago there were still debates about the rarity or abundance of what we now call “exo-planets”
    SO- do not be discouraged about lack of speed in the manned space program. we are developing the technology for life support, advanced computation and robotics, sensors, and engineering in other disciplines. ( this is similar to the concept of “preadaptation” in evolution), Our knowledge of the solar system and the cosmos in general is growing faster than ever. We say that the only thing holding back exploration of the solar system is the willingness of the civilization to support this economically… well that is advancing too.
    Have some faith… we will get there, or at least our children ( genetic or intellectual) will roam the universe.

  • d.m.falk October 5, 2012, 10:47

    Something I wrote 2.5 years ago reflecting (as it were) on what it would be like on a red-dwarf exoplanet like Gliese 581d, to which inspired this.


    Something also to think about is that chlorophyll-using plant life on Earth actually reflect more near-infrared than they do green or yellow-green (I’m an infrared photographer, so this is in my ballpark ;) ), but on worlds like, say, a potential Gliese 581d, could range from “black” (not actually black, but somewhere in the near-infrared) to as high in the visible spectrum as yellow, with green and blue being far less common (but not unlikely) than the lower part of the visible spectrum.


  • FrankH October 5, 2012, 15:13

    This is a good ground laying paper, but one “minor” detail that they’ve overlooked is that *no* Earth-sized (less than 4 Earth mass) exoplanet in its star’s HZ has been directly imaged.

    There’s a considerable technological leap from where we are now to the level where we can image such a planet in the HZ of a nearby star; there’s another huge leap from there to the point where we can collect enough photons to have the luxury of sticking filters in the beam to get color magnitudes.

    They are also making the job much harder by looking in the range of visible light. The first telescopes capable of teasing the light of an exoplanet out of the glare of its primary are probably going to be IR only (or mainly) telescopes. They should expand their paper and look beyond the UBVRI filter range and into the near IR (JHKL), especially if these telescopes are space-based.

    Also, the light from most red dwarfs isn’t red. For most, the light will look about as “red” as that from an incandescent tungsten bulb. Terrestrial plants will grow under incandescents.

  • andy October 5, 2012, 15:54

    Gliese 581d seems more and more to be considered a habitable zone planet, as Siddharth Hegde (Max Planck Institute for Astronomy) and Lisa Kaltenegger (Harvard-Smithsonian Center for Astrophysics) describe it in a new paper.

    Ah, but does it actually exist, or is it an artifact of red noise in the measurements? Let’s go erasing some more planet candidates in the Gliese 581 system…

    Roman V Baluev (2012): “The impact of red noise in radial velocity planet searches: Only three planets orbiting GJ581?

  • Alex Tolley October 6, 2012, 12:28

    Some questions about the paper:
    1. It does not appear to show cases of planets that we can rule out: all the data is for terrestrial environments. Why didn’t it show cases of environments where life could be ruled out?
    2. Where would Mars be on figure 4 – probably in the extremophile zone.
    Venus would be ruled out because of cloud cover. Since we can have the data for Earth at a great distance, does our extensive cloud cover cause problems or not for the author’s approach.
    3. The authors make quite a lot of the vegetation red edge (VRE), although they speculate about the effect around stars with different spectra characteristics. Isn’t this rather naive? Why would we expect chlorophyll (and their accessory pigments), or something very like it, to evolve on these planets? This suggests to me that their highly desirable region II is unlikely to be found.

    Conversely, should a life form exist on Titan, would the method rule that sort of world out? We really need to determine the true range and forms of life in the solar system to ensure that we do not rule out habitats. When we search for life, we will need more universal bio-signatures, rather than focusing on Earth analogs. Of course if we found a world that was an Earth II, that would be really something, but I expect the universe to be far stranger, and interesting, than that.

  • william October 6, 2012, 18:23


    would you have any idea what the light colored band stretching
    across the photo is ? I’ve always wondered.

  • Paul Gilster October 6, 2012, 20:28

    Wikipedia says this about the ‘pale blue dot’ image:

    “The light band over Earth is an artifact of sunlight scattering in the camera’s optics, resulting from the small angle between the Earth and the Sun.”

  • Michael October 7, 2012, 9:11

    I believe it is caused by dust in a thin plane orbiting the Sun also known as Zodiacal light

  • wojciech October 7, 2012, 18:10

    A bit off-topic, but interesting nevertheless.
    Geoff Marcy will start a project to sort through Kepler data in hopes of finding Dyson Spheres(or more likely Dyson Swarms)


  • andy October 8, 2012, 13:55

    one “minor” detail that they’ve overlooked is that *no* Earth-sized (less than 4 Earth mass) exoplanet in its star’s HZ has been directly imaged.

    So what? Are you saying that we should not try to predict what we will see before we can see it?

    This is not something like string theory where no conceivable experiment can test the hypotheses.

    I’ll also note that it was a fairly short time from the first direct images of exoplanets to obtaining spectra of them.

  • FrankH October 9, 2012, 15:22

    Read the second paragraph. The problem is the size of the planet and the number of its photons that’ll be intercepted by a telescope.

    The few exoplanets that have been imaged directly are all very large gas giants or possibly even brown dwarfs; for spectra, we’re talking a couple of very hot gas giants (or brown dwarfs)…and it was all done in the IR (L band), nowhere near the visible bands mentioned in the paper.

    Right now we can barely image a hot gas giant; doing the same to a smaller, cooler Earth sized world is at least an order of magnitude harder.
    From there to collecting enough photons to get a good spectra is much harder still; probably beyond our current equipment.

    The Terrestrial Planet Finder would have been able to image nearby planets and maybe get a spectra, but it was canceled. I’m not sure that the James Webb telescope – if it’s ever launched – can detect a terrestrial planet without an occulting disk.

  • andy October 9, 2012, 19:19

    @FrankH: again, are you saying we should not try to predict these things? Bear in mind that in principle we know how to do the experiment to test this even though it is somewhat beyond current capabilities.

    Regarding gas giants, I seem to recall there have been several papers predicting the atmospheric conditions on such worlds and the visible spectra we would expect, since long before we had imaged even a single exoplanet.

    As for your assertion that all exoplanets imaged so far are imaged in the IR, there is the case of Fomalhaut b which is notable for being so far undetected in the IR but only detected in the visible. Admittedly this one is rather peculiar.

    Why are you so opposed to using theory to try to predict future observations? This is part of what theory is for…

  • FrankH October 10, 2012, 12:53

    Please read my posts again; I’m not against trying to predict or even looking for these planets. The problem I have is that the paper is proposing looking at wavelengths that are far, far beyond anything we are capable of detecting any time soon. It would be better to expand the data to the IR, where we at least have a glimmer of a chance of detecting a terrestrial world.

    Fomalhaut B is probably not a planet – at least not yet.