The uninhabitable planet: debunking the ESI

On July 8, 2014


It was the news for which everyone had been waiting.

At the end of last month, the popular press went wild. A planet had been discovered that was so much like Earth it was heralded as our best bet for supporting life. Positioned 16 light years away, Gliese (or GJ) 832c was a mere hop from home and there were rumours a popular coffee shop chain had already applied for planning permission.

And this was all utterly wrong.

The journal article, published by a team led by Robert Wittenmyer from the University of New South Wales in Australia, described a world boiling under its own stifling cloud cover. With an orbit that only skirted the region capable of maintaining water, and a mass sufficient to attract a thick atmosphere, the planet was designated a likely ‘super Venus’ and unsuitable for life. What was more, it was so close to its star that it risked being in a tidal lock, with one side doused in the heat of a perpetual day while its reverse remained shrouded in night.

It was a planet to inspire thoughts of Dante’s Inferno and the facts were clearly laid out in the freely available journal article. How then, did the press get it so completely wrong?

The answer lies in the use of a quantity denoted as the ‘Earth Similarity Index’ or ESI. As the product of the differences between the planet’s bulk properties and those of the Earth, the ESI is designed to indicate how ‘Earth-like’ a planet might be. The problem is that the resulting number is a weak comparison between the two objects and has absolutely no quantitative meaning for habitability.

Accessing the habitability of a world outside our Solar System is no easy task; a fact that lies at the heart of why the ESI is fallacious. Planetary scientists are typically working with only the mass (and sometimes radius) of the planet and the type of star it is orbiting.

The latter property can be used to define the ‘Habitable Zone’. Coined in 1959 by scientist Su-Shu Huang working at the Berkeley Astronomical Department, the Habitable Zone around a star is the location where water could exist on the surface of a planet, if that planet had a sufficient atmospheric pressure.

The caveats here are important: the Habitable Zone does not say there is water present, or that a planet exists that might be able to support life. It only marks out a region where the amount of stellar radiation would not boil nor freeze water. In short, very few planets found in the Habitable Zone of their star will be suitable for life, but if another Earth-like planet existed, it would be there. This makes the region a prime search area for future missions targeted towards habitability, including the forthcoming space telescope, JWST (predicted launch 2018), and future terrestrial planet-finders.

In the case of GJ 832c, the planet orbits just inside the inner edge of its Habitable Zone, providing the location of its edge is extended to its most generous estimate. The generosity requires scientists to include the possibility of water that is only supported during the early evolution of the planet, as is thought to have been true for Mars and Venus.


However, it is not its proximity to the edge of the Habitable Zone that kills the deal for GJ 832c. With a measured mass at least five times that of the Earth, GJ 832c is capable of attracting a thick atmosphere. Even if the gas constitution was similar to that on Earth, this would result in a high quantity of greenhouse gases that trap heat reflected from the planet’s surface. The surface temperature would therefore be kicked beyond that expected at the edge of the Habitable Zone to ensure no water (and therefore no known forms of life) exists. If that is not unpleasant enough, the higher planet mass could result in hydrogen and helium being retained in the planet’s atmosphere (they escape the Earth) to produce an utterly unusable mess.

Finally, GJ 832c is very close to its star, with an orbit of only 35.8 days. Since the star is small and relatively cool, the planet is saved from the cooked fate of Mercury, but the gravitational pull is likely to be sufficient to cause a tidal lock. In this situation, the planet will rotate an exact number of times per orbit. If that number happens to be one (as it is with the moon’s orbit about the Earth) then one side will be perpetually facing the star to create a strong temperature gradient across the planet. While this alone would not definitely rule out life, it is one more challenge on a task that is already basically impossible.

So why does the ESI come out so suggestive of a second homeland?

It is because it is not capable of taking into account any of the above factors.

The calculation of the ESI is based on four parameters: mean radius, bulk density, escape velocity and surface temperature. These are weighted by an exponent that reflects the range each property can take while maintaining a reasonable ‘Earth like’ condition, and then multiplied together to give a value between 0.0 and 1.0. Any value above 0.8 is considered a near-Earth match. GJ 832c has an ESI of 0.81.

The reason this holds no water (pun intended) boils down to three reasons:

The first is the small number of measurements that are being compared. When observing exoplanets, scientists can only measure the planet’s mass and radius, with the latter only being possible if the planet is transiting across its star. This means that only three of the four variables in the ESI can vary independently, since both density and escape velocity are calculated from the mass and radius. In the case of GJ 832c, which was detected from its star’s radial velocity, we only have a measurement for the minimum value of its mass, reducing our Earth comparison points to a poor two.

Within these three (or two) variables, a second problem arises in achieving an accurate value for the planet’s surface temperature. As with the location of the Habitable Zone, the temperature is estimated by calculating the radiation intensity from the star at the planet’s location. However —as we have just seen— this does not allow for the planet’s atmosphere. If we perform the calculation for our own Sun, the surface temperate estimates are low but reasonable for the Earth, but less than half the correct value for the closer Venus. This is due to the extra solar radiation boosting the greenhouse gases in Venus’s atmosphere, which spiral upwards to roast the planet’s surface.

The location of this ‘Runaway Greenhouse’ effect for an Earth-sized atmosphere is used to define the inner edge of the Habitable Zone, which places Venus outside it. However, the ESI calculation for Venus with its estimated temperature would give it an incredible 0.9 match with the Earth*.

“All conversations regarding habitability are worthless if we ignore the limited data present within our own Solar System,” points out Prof. Stephen Kane, a planetary scientist from San Francisco State University. “If we start defining Venusian planets as being habitable then we are no longer doing science.”

For GJ 832c, whose larger mass boosts the greenhouse effect, the discrepancy of its actual to estimated temperature is likely to be much worse.

The third problem with the ESI is that even if its parameters could be accurately and independently measured, they cannot be combined to determine habitability.

While the mass, radius and temperature certainly have a baring on the planet’s environment, they are overwhelmed by other factors that contribute to the support of life. For example, water is thought to have been delivered to the Earth by ice-rich meteorites scattered inwards by the outer planets during its formation. A different system of planets could bypass this process, allowing Earth’s twin to form but be uninhabitable to all known forms of life. Similarly, the Earth’s magnetic field protects it from harmful solar radiation, its distance from the Sun allows it to rotate hundreds of times per orbit to ensure an even distribution of heat and the Sun itself is a quiet star without
violent radiation outbursts that could overwhelm the Earth’s defences. These are a small fraction of the processes that will determine a planet’s suitability for life and none of them are included in the ESI.

The upshot of this is that a planet with an ESI of 0.1 is just as likely to support life as one with ESI 0.99. In the case of GJ 832c, the ESI estimates habitability based on two measurements, the first of which (the high mass) suggests the second (the temperature) is wildly inaccurate.

The ESI is neither used nor referred to in scientific publications, but its tantalising one-number answer has caused it to propagate rapidly through the majority of science news sites. The result is an article that is at best misleading, and at worst downright wrong.

Perhaps though, the most disappointing fact about the ESI is that it detracts from the real excitement of GJ 832c. With its high mass and thick atmosphere, the planet has been declared a ‘super Venus’, for which the only other example (Kepler 69c) was discovered last year.

Kane led the characterisation of Kepler 69c as a super Venus in a paper that identifies the planetary class.

“The Kepler telescope is discovering the Venus-analogues first,” Kane explains. “Understanding how common they are will help us to decode why the atmosphere of Venus so radically diverged from its sister planet, Earth.”

* The quoted value for the ESI of Venus is 0.44, but this is from using Venus’s true temperature, not the estimate that would be made in an exoplanet detection.


1. Right-side of image shows the region ‘Noctis Labyrinthus’ on Mars. Credit: NASA/JPL-Caltech/ASU

2. Orbit  of GJ 832c. Green marks the extent of the Habitable Zone, with dark green showing the optimistic estimates of its edge.  Credit: Stephen Kane

This post was amended on 21 July 2014 to remove a link to an Astrobiology Magazine article, which implied it incorrectly used the ESI to measure planet habitability.Their article on this subject correctly points out the issues with this index and references the lead author on the GJ 832c discovery.

About Elizabeth Tasker

Elizabeth Tasker is an astrophysicist at Hokkaido University in Japan. Her research looks at the formation of stars in simulations of galaxies like our own Milky Way. She writes the research blog for Hokkaido University's English website and keeps her own personal blog as testimony to exactly how confusing life can sometimes get in Japan. You can also find Elizabeth on Google+ and Twitter.

9 Responses to The uninhabitable planet: debunking the ESI

  1. will this ESI be based on the planet’s atmosphere? I’m guessing not. Most “habitability” ratings are assigned based on the effective temperature. By that definition, Earth is no habitable (it’s effective temperature is right at 0 C. It’s a back of the envelope calculation, but since it is still difficult to ascertain if a planet has an atmosphere, let alone what it is made of. I have always wondered if you swapped the positions of Venus and Mars, if one or both would be habitable.

  2. Since an exoplanet’s atmosphere cannot be yet detected, you are quite right that the ESI estimate of the temperature of a planet with an atmosphere is incorrect. This is one of the reasons it calculates a value so absurd for GJ 832c.

    However, even if it were possible to correctly measure the temperature of a planet, the ESI would still not be a quantitative measurement of habitability.

    There are many factors involved that include not just the simplest bulk properties of the planet, but also its history (the Earth was inhabitable for vast swaths of its own lifetime), the conditions in its Solar System (for tidal effects, the addition of water, ….) and a plethora of additions such as the magnetic field. We have no idea what half of these are, let alone what weighting to give them to produce a quantitative meaningful number.

    Therefore, even if the planet was next door to us, the ESI is not right.

  3. So, help me understand the science behind this piece. After repeating once and again throughout your article the cautionary words “capable”, “if was”, “would result”, “would therefore”, “expected”, “could result”, “if this happens to be”, and so on, you firmly conclude that the ESI “IS fallacious”, no doubt about it. In my opinion, you have to do your homework better, before trashing peer-reviewed literature in a blog.

    PS: About Prof. Kane comment, have we ever heard about extremophiles? In Venus, some cloud layers are well below 100C…

    • Hi Rose,

      Your point about my use of cautionary language is exactly why the ESI is fallacious. Habitability is an immensely complex problem with a large number of variables, whose extent and weighting we do not yet know. It is therefore impossible (no cautionary language necessary) to boil this down to a single quantitive number.

      (Stephen has answered your query about extremophiles below)

  4. I think the problem is that you did not read the original paper, where it is clearly stated that the ESI (Earth Similarity Index) has nothing to do with habitability, but only describes the similarity in regard to some planetary parameters. Thus, even our Moon has a relatively high ESI-value. The same paper refers to the PHI value (Planetary Habitability Index, in a two-tiered apporach) to address the habitability of a planet. Also, this kind of assessment is always limited due to a lack of data, as clearly stated in the paper. Thus, I would encourage you (and some of the media) to actually read the original paper: “Schulze-Makuch et al. (2011) A Two-Tiered Approach to Assessing the Habitability of Exoplanets. Astrobiology 10, , p. 1041-1052.

    • Hi Dirk,

      The problem is that the ESI is used widely (if not exclusively) in the press to measure habitability.

      I agree in the original paper, the ESI is presented as a comparison of the physical extent between planet and Earth. For exoplanets (as you note) that result is still limited by the lack of data. The issue, however, is that this not how the ESI is being presented.

      If we could disconnect “ESI” and “habitability”, it would be an excellent start.

  5. The original paper clearly states that the ESI (Earth Similarity Index) has nothing to do with habitability, but only describes the similarity in regard to some planetary parameters. Thus, even our Moon has a relatively high ESI-value and is clearly not habitable. The very same paper refers to the PHI value (Planetary Habitability Index, in a two-tiered approach) to address the habitability of a planet. This kind of assessment is of course always limited due to a lack of data, as clearly stated in the paper. Thus, I would encourage you (and some of the media) to read in detail the original paper: “Schulze-Makuch et al. (2011) A Two-Tiered Approach to Assessing the Habitability of Exoplanets. Astrobiology 10, , p. 1041-1052.

  6. Rose, I’m very familiar with extremeophiles (see Kane & Gelino, 2012, “The Habitable Zone and Extreme Planetary Orbits”, Astrobiology, 12, 940). It is unsurprising to find extremeophiles on Earth which has been contaminated with life for 4 billions years. The possability of life in the Venusian upper cloud layers is not a valid analogy.

  7. Dirk, no the problem isn’t with the paper (which I’ve not only read, but cited). The problem is that the PHL website describes the ESI as follows: “As a general rule, any planetary body with an ESI value over 0.8 can be considered an Earth-like planet. This means that the planet is rocky in composition (silicates) and has an atmosphere suitable for most terrestrial vegetation including complex life.” That is what the public and press are reading, not your paper. If you are serious about the ESI having nothing to do with habitability then I expect you must be even more dismayed than I am regarding how it is being blatantly misrepresented both on the PHL website and in popular media.

    By the way, if you find a way to actually get the media to read a science paper over a webpage then let me know. Also, please place your paper on the astro-ph preprint server so it is accessible.