Research Article
A Two-Tiered Approach to Assessing
the Habitability of Exoplanets
Dirk Schulze-Makuch,1,4 Abel Me´ndez,2 Alberto G. Faire´n,3 Philip von Paris,4 Carol Turse,1 Grayson Boyer,5
Alfonso F. Davila,3 Marina Resendes de Sousa Anto´nio,1 David Catling,6 and Louis N. Irwin7
Abstract
In the next few years, the number of catalogued exoplanets will be counted in the thousands. This will vastly
expand the number of potentially habitable worlds and lead to a systematic assessment of their astrobiological
potential. Here, we suggest a two-tiered classification scheme of exoplanet habitability. The first tier consists of
an Earth Similarity Index (ESI), which allows worlds to be screened with regard to their similarity to Earth, the
only known inhabited planet at this time. The ESI is based on data available or potentially available for most
exoplanets such as mass, radius, and temperature. For the second tier of the classification scheme we propose a
Planetary Habitability Index (PHI) based on the presence of a stable substrate, available energy, appropriate
chemistry, and the potential for holding a liquid solvent. The PHI has been designed to minimize the biased
search for life as we know it and to take into account life that might exist under more exotic conditions. As such,
the PHI requires more detailed knowledge than is available for any exoplanet at this time. However, future
missions such as the Terrestrial Planet Finder will collect this information and advance the PHI. Both indices are
formulated in a way that enables their values to be updated as technology and our knowledge about habitable
planets, moons, and life advances. Applying the proposed metrics to bodies within our Solar System for com-
parison reveals two planets in the Gliese 581 system, GJ 581 c and d, with an ESI comparable to that of Mars and
a PHI between that of Europa and Enceladus. Key Words: Habitability—Exoplanets—Index—Earth similarity—
Complexity—Life. Astrobiology 11, xxx–xxx.
1. Introduction
As of this writing, over 700 exoplanets have been de-tected, many in solar systems with multiple planets (for
an update, see http://exoplanet.eu/). All but one of the exo-
planets detected to date are larger than Earth and, for themost
part, considerably closer to their central star, which, given that
such planets are easier to detect, is to be expected. However,
more than 20 planets with minimum masses smaller than 10
Earth masses are known, and many of them could potentially
be terrestrial. Technologies already operational or under de-
velopment are hastening the discovery of smaller terrestrial
planets. On Hawaii, the Keck Interferometer will combine
light from the world’s largest optical telescopes to enable the
visualization of gas clouds, including large planets within
them, around distant stars. In Arizona, the Large Binocular
Telescope Interferometer is currently under construction.
These instruments will advance our understanding of the
proportion of planets that may be smaller andmore terrestrial
within the coming decades, at least for neighboring systems
in our own galaxy. The Kepler Mission and the Terrestrial
Planet Finder (TPF) are designed to detect Earth-sized planets
and directly measure gases consistent with life, such as ozone
andmethane, in terrestrial-like atmospheres on planets around
stars up to 50 light years away (Beichman et al., 2002; Basri
et al., 2005; Schulze-Makuch and Irwin, 2008). The possibility
of Earth-type habitable planets in other solar systems has been
suggested by plausiblemodels, as was the case for the 47UMa
planetary system (Cuntz et al., 2003) and Gliese 581 d (Selsis
et al., 2007; von Bloh et al., 2007; Schulze-Makuch and Guinan,
2010; von Paris et al., 2010; Wordsworth et al., 2010). With a
new generation of telescopes and missions on the way, the
1School of Earth and Environmental Sciences, Washington State University, Pullman, Washington, USA.
2Planetary Habitability Laboratory, University of Puerto Rico at Arecibo, Arecibo, Puerto Rico, USA.
3SETI Institute and NASA Ames Research Center, Moffett Field, California, USA.
4Institut fu¨r Planetenforschung, Deutsches Zentrum fu¨r Luft- und Raumfahrt (DLR), Berlin, Germany.
5Department of Chemistry and Biochemistry, Arizona State University, Tempe, Arizona, USA.
6Department of Earth and Space Sciences/Astrobiology Program, University of Washington, Seattle, Washington, USA.
7Department of Biological Sciences, University of Texas at El Paso, El Paso, Texas, USA.
ASTROBIOLOGY
Volume 11, Number 10, 2011
ª Mary Ann Liebert, Inc.
DOI: 10.1089/ast.2010.0592
1

discovery of many more exoplanets can be expected. That, in
turn, will drive the need for a classification scheme for as-
signing astrobiological potential for exoplanets based on es-
timates derived from quantitative data of their probability for
supporting life (e.g., Kaltenegger et al., 2010). A summary of
planetary parameters that can be directly observed for exo-
planets by current or proposed space mission is provided in
Table 1.
Estimates of astrobiological potential in general have been
dominated by the concept of the ‘‘circumstellar habitable
zone’’ (Kasting et al., 1993), which suggests that attention
should be focused on those worlds that could retain an at-
mosphere and liquid water—the implicit assumption being
that life is most likely to be found on planetary bodies with
those Earth-like conditions. Many have argued, however,
that focusing exclusively on a habitable zone biased by ter-
racentric assumptions is too restrictive for the full variety of
life that could exist (e.g., Darling, 2001; Grinspoon, 2003;
Schulze-Makuch and Irwin, 2004, 2008; Ward, 2005; Gaidos
et al., 2005). Therefore, the possibility that life could be found
under very different conditions from those of Earth needs to
be kept in mind (Schulze-Makuch and Irwin, 2002, 2006;
Bains, 2004; Benner et al., 2004).
As a practical matter, interest in exoplanets is going to
focus initially on the search for terrestrial, Earth-like planets,
as the deployment of the TPF illustrates. In a larger sense,
however, the task of astrobiology is to seek out life in the
Universe in all its forms. Therefore, the search for life on
other worlds is really divisible into a two-part question. The
first question is whether Earth-like conditions can be found
on other worlds, since we know empirically that those con-
ditions could harbor life. The second question is whether
conditions exist on exoplanets that suggest the possibility of
life, whether in a form known to us or not. We therefore
propose two different indices for addressing each of these
questions.
2. The Earth Similarity Index (ESI)
Similarity indices provide a powerful tool for categorizing
and extracting patterns from large and complex data sets.
They are relatively quick and easy to calculate and provide a
simple quantitative measure of departure from a reference
state, usually on a scale from zero to one. They are used in
many fields, including mathematics (e.g., set theory and
fuzzy logic), ecology (e.g., Sorensen similarity index), com-
puter imaging (e.g., structural similarity index), chemistry
(e.g., Jaccard-Tanimoto similarity index), and many others.
We propose the ESI as a measure of Earth-likeness. The
basic ESI expression is constructed from a weighted re-
formulation of the Bray-Curtis Similarity Index (Bloom,
1981) as
ESIx ¼ 1
x x0
xþ x0


 w
(1)
where x is a planetary property, xo is a terrestrial reference
value, w is a weight exponent, and ESIx is the similarity mea-
sure as a number between zero (no similarity) and one (iden-
tical). The weighting exponent is used to adjust the sensitivity
of the scale, to bring each calculated parameter equal to or
above 0.8 for Earth-like conditions (see below). The ESI for each
planetary property is then combined into a single ESI value by
using the geometric mean as the method of aggregation.
Application of the ESI to exoplanets requires only the use
of physical properties already available for many of them,
such as radius, mass, and temperature. We have found it
most instructive to distinguish among interior, surface, and
global similarities. The interior similarity is a measure of the
extent to which a planet has a rocky interior, while the sur-
face similarity is a measure of the ability to hold a temperate
surface like that of Earth. They are given by
ESII ¼ (ESIr  ESIq)1=2 (2)
Table 1. Basic Planetary Parameters That Can Be Directly Observed for Terrestrial
Extrasolar Planets by Current or Proposed Space Missions
Parameter Variable Method Missions Astrobiological relevance Reference
Semimajor axis A DI, AM, RV, TP, GM* Kepler, Gaia Surface temperature [1]
Eccentricity e DI, AM, RV Gaia Seasonal variations [1]
Orbital inclination I DI, TP (lower limit) Kepler, Gaia, TPF Seasonal variations [1]
Orbital period
(= semimajor axis)
Torb DI, AM, RV, TP Kepler, Gaia Surface pressure [1]
Mass M DI, AM, RV,{ GM JWST, Gaia, TPF Surface pressure
and temperature
[1]
Radius R DI, TP Kepler Composition [1]
Density q DI, TP Kepler, Gaia Seasonal variations [1]
Mean surface temperature Tsurf DI, TP* TPF, JWST, Kepler Stability of liquid water [1]
Ocean areas — DI TPF Stability of liquid water [2]
Atmospheric composition — DI, TP* TPF, JWST Bioelements [1]
Vegetation red edge — DI TPF Habitat distribution
and water cycle
[3]
*Under some conditions.
{Lower limit.
Planet detection methods are direct imaging (DI), astrometry (AM), radial velocity (RV), transit photometry (TP), and gravitational
microlensing (GM).
JWST, the James Webb Space Telescope.
[1] Jones (2008).
[2] Williams and Gaidos (2008).
[3] Arnold et al. (2009).
2 SCHULZE-MAKUCH ET AL.