Third Rocks from the Stars

The ill-fated protagonist of Shakespeare’s play Hamlet admonished his friend that: "There are more things in Heaven and Earth, Horatio, than are dreamt of in your philosophy." Today we might instead warn ourselves of the certainty that there are more kinds of Earths in the heavens than are dreamt of in our philosophy.

Artist concept of star system, HD70642.
Credit:John Rowe animation

Any mission to detect and spectroscopically characterize terrestrial planets around other stars must be designed so that it can characterize diverse types of terrestrial planets with a useful outcome. Such missions are now under study–the Terrestrial Planet Finder (TPF), by NASA, and Darwin by ESA, the European Space Agency. The principal goal of TPF/Darwin is to provide data to the biologists and atmospheric chemists. These investigators will evaluate the observations of a potentially broad diversity of objects in terms of evidence of life and the environmental conditions in which such life would be present.

The TPF/Darwin concept hinges on the assumption that one can screen extrasolar planets for habitability spectroscopically. For such an assumption to be valid, we must answer the following questions. What makes a planet habitable and how can that be studied remotely? What are the diverse effects that biota might exert on the spectra of planetary atmospheres? What false positives might we expect? What are the evolutionary histories of atmospheres likely to be? And, especially, what are robust indicators of life?

TPF/Darwin must survey nearby stars for planetary systems that include terrestrial sized planets in their habitable zones ("Earth-like" planets). Through spectroscopy, TPF/Darwin must determine whether these planets have atmospheres and establish whether they are habitable.

We define a habitable planet in the "classical" sense, meaning a planet having an atmosphere and with liquid water on its surface. The habitable zone therefore is that zone within which light from the planet’s parent star (its "Sun") is sufficiently intense to maintain liquid water at the surface, without initiating runaway greenhouse conditions that dissociate water and sustain the loss of hydrogen to space. The size of a planet can determine its capacity to sustain habitable conditions. Larger planets sustain higher levels of tectonic activity that also persists for a longer time. Tectonic activity sustains volcanism and also heats crustal rocks and recycles CO2 and other gases back into the atmosphere. These outgassing processes are required to ensure climate stability over geologic timescales.

There are interesting potential examples where liquid water might exist only deep below the surface, such as the Jovian moon Europa, or on present-day Mars. However, biospheres for which liquid water is present only in the subsurface might not be detectable by TPF/Darwin. Thus a planet having liquid water at its surface meets our operational definition of habitability, which is that habitable conditions must be detectable.

Although the "cross hairs" of the TPF/Darwin search strategy should be trained upon "Earthlike" planets, TPF/Darwin should also document the physical properties and composition of a broader diversity of planets. This capability is essential for the proper interpretation of potential biosignature compounds. For example, the presence of molecular oxygen in the atmospheres of Venus and Mars can indeed be attributed to non-biological processes, but only through a proper assessment of the conditions and processes involved. On the other hand, a planet might differ substantially from Earth yet still be habitable.


Our search for life elsewhere will inevitably deepen our understanding of life itself. Current definitions of life usually enumerate its key properties. For example, they cite the ability of cells and ecosystems to harvest energy, metabolize, replicate and evolve. Our definitions are based upon life on Earth yet they will affect our strategy to search for life elsewhere. Accordingly, we must distinguish between attributes of life that are truly universal versus those that solely reflect the particular history of our own biosphere.

Scene from a moon orbiting the extra-solar planet in orbit around the star HD70642.
Credit:David A. Hardy, (c)

Herein we assume that all life requires complex organic compounds that interact in a liquid water solvent. These assumptions do not seem overly restrictive, given that life is an information-rich entity that depends fundamentally upon the strong polarity of its associated solvent. Carbon compounds and structures appear to be unrivaled in their potential for attaining high information contents. Other plausible solvents cannot match the strong polar-nonpolar dichotomy that water maintains with certain organic substances; and this dichotomy is essential for maintaining stable biomolecular and cellular structures. However, our own biosphere utilizes only a small fraction of the number of potentially useful organic compounds. Alien life forms probably explored alternative possibilities, and so their discovery will increase the known diversity of life.

One major barrier to resolving these divergent views is that we know the history of only one biosphere. If we had other examples, we could directly compare them and begin to discern general principles of the origins and evolution of life. This circumstance creates a powerful scientific argument to look for life elsewhere.


The detection in situ of life is a strategy that might be viable within the Solar System, but not for extrasolar planets. Within our own Solar System, the search for extraterrestrial life and evidence about the origin of Earthly life will likely be confined to Mars, Europa, and Titan. Small bodies such as comets, asteroids, and meteorites offer insights concerning chemical "building blocks" for the origins of life. However, it seems feasible to detect biological signatures, or "biosignatures," by remote sensing. There are at least two types of biosignatures; spectral and/or polarization features created by biological products, and electromagnetic signals created by technology. The latter example of a biosignature requires SETI-like searches. Spectral biosignatures can arise from organic constituents (e.g., vegetation) and/or inorganic products (e.g., atmospheric O2).

One must bear in mind that the range of characteristics of rocky planets is likely to exceed our experiences with the four terrestrial planets and the Moon. While the nearly (but not quite) airless Moon and Mercury arguably represent the lifeless endmember case of terrestrial planets, there are always surprises. For example, Mercury appears, based on radar data, to support small polar caps of water ice, and the origin of the water appears to be exogenic impact of icy material followed by molecular migration to the poles. Were such a body to be in a planetary system in which the orbital plane happens to be face-on to the Earth, could that water ice signature be detectable in the near-infrared range, and, if so, what would one conclude about the habitability of such an object?

Habitability might be ruled out if the semimajor axis were too small (indeed, the planet might be missed altogether), but no laws of physics rule out a "Mercury" placed at the orbit of, say, Venus (0.7 AU). What would one conclude then?


Gas giant and terrestrial-sized planets can be easily distinguished by their apparent brightness (a function of area, albedo and phase) and orbital distance. Considering the surface area ratio of Jupiter to Earth and assuming the same albedo, Jupiter would be 120 times brighter than the Earth at the same orbital distance. Unless giant planet albedos are 10 to 100 times smaller than terrestrial planet albedos, confusion between giant and terrestrial-size planets is unlikely. A planet’s color [or "color-mass"] can also indicate whether it is a giant or terrestrial type, based on our experience with the particular spectral properties of the planets and atmospheres in our Solar System.

The simple observation that a planet exists at some distance from a star will determine whether the planet is in a predefined habitable zone of the star (which may or may not delineate where life is possible), but it in fact only provides a very rough estimate of the temperature. In general, if there is a greenhouse effect present (e.g., from CO2, H2O, CH4, or aerosols), then the surface temperature will be warmer than the effective temperature, which is determined by the stellar brightness, the star-planet distance, etc. Regarding the search for life, constraining a planet’s surface temperature holds much greater value than constraining the effective temperature. For example, both Venus and Earth have similar effective temperatures (220K and 255K, respectively), but vastly different surface temperatures (730K and about 290K, respectively), owing to the divergent greenhouse gas column abundances. At its surface, Venus is a cloud-covered, hot-house–with soil capable of melting lead. Visible and/or infrared spectra can help interpret these cases, but neither is able to penetrate clouds; therefore surface conditions may well be difficult to estimate.

Ultraviolet image of Venus obtained by Pioneer-1.
Image Credit: BNSC

The combination of infrared and visible observations is of course most valuable: neither region will yield all of the information, and either region will require modeling to interpret. There is, however, a concern that Earth is a peculiarly easy planet to interpret from external observations. Of the two larger terrestrial planets in our system one is roughly half cloud-covered (Earth) and the other completely cloud-covered (Venus). Planets that differ in their size, insolation or land-ocean fraction may well have more cloud cover and be much harder to interpret for that reason, as was the case for early studies of Venus. Both Venus and Titan are totally enshrouded by photochemical clouds, and spacecraft close flyby and in situ techniques have been required to determine the conditions in their lower atmospheres and at their surfaces. Planets with habitable surfaces that are hidden by deep, totally opaque atmospheres, or with only a small fraction of the surface exposed to view, probably cannot be recognized as habitable. We are limited to exploring habitability for only those planets that are habitable on a global scale, and which have mostly clear atmospheres in a significant part of the optical or infrared spectrum, or both.

Frost and Viking 2 Lander
White patches of frost on the ground are visible behind the Viking 2 Lander. Click to enlarge.Credit: NASA.

From albedo we can tell whether the cloud cover of the planet resembles that of the Moon, Mars, Earth or Venus. However, in the past, the Earth has been through a cold phase in which it had a high albedo (due to ice) and a low surface temperature. We could not distinguish such a snowball Earth from a Venus-like planet from albedo alone; spectra would be required to distinguish water ice from sulfuric acid droplets.


While our Solar System provides a good template for interpreting brightness and color mass, a much greater diversity of planets is likely and challenges their use for estimating planetary radii and masses. For example, Earth has had very different signatures throughout geologic time. The postulated "snowball Earth" during the Neoproterozoic Eon would have had a very high albedo. Earth’s atmosphere might have contained a larger methane component during the Archean Eon (prior to 2.5 b. y. ago), and therefore exhibited extremely strong methane spectral features. Another exception to the Solar System template is an icy Uranus just at the outer border of the habitable zone. Strong deviations from Solar System gas giant spectral signatures also are expected for young, hot Jupiters.

Oxygen: Poison or Predictor?

Among biosignatures, oxygen is one of the most intriguing: in excess, oxygen will poison plant life, but this vegetative waste product is life-giving to animal life–including during the transformation of Earth’s biosphere with rise of more complex life forms. There may be only a very small range of planetary conditions that might produce a false positive answer for oxygen–a large ice-bound planet or a small planet with a thin atmosphere but receiving a Venus-like insolation. Because non-biological oxygen is quickly absorbed by rocks, the need to regenerate oxygen continuously points to biology at work unless ice covers the rock-surfaces or acidic volcanoes react with oxygen to remove it. Detection of O2 or its photolytic product O3 merits highest priority.

In summary, the first and best-known aspect of a planet from infrared observations is its size. From the size, insolation and integrated emission we can determine both the albedo and a temperature associated with the emitting layer. Estimates of planet size and albedo can definitely be determined from mid-infrared observations. [The preferred wavelength ranges are 7 to 25 microns in the mid-IR, and 0.5 to about 1.1 microns in the visible to near-IR.] Surface temperature determination is only possible if there is a planet with a substantial fraction that is cloud free.

Wildcards: Water?

Liquid H2O is not a bioindicator but it is considered essential to life. Absorption by water poses an interesting challenge for interpretation. Water absorption is observed in the near-infrared spectra of cool giant stars and brown dwarfs. The bands are the same as those seen in Earth’s spectrum, but they are somewhat broader and therefore they modify the apparent shape of the continuum between bands. A fairly precise measure of temperature is needed to take advantage of information about the strength of water bands.

Substantial carbon dioxide CO2 indicates an atmosphere and oxidation state typical of a terrestrial planet. Abundant methane CH4 might require a biological source, yet abundant CH4 also can arise from a crust and upper mantle more reduced than that of Earth.

It would be advisable to use this process to test against observations of Earth, Mars and Venus as well as simulations of a "smaller" Venus and "larger" Mars to explore the quality of the results. Indeed, it is a necessary step in this work that both visible and infrared processes be validated. The resulting library of spectral features should provide a useful and quantitative starting point for modeling of equivalent features identified in the light from extra-solar terrestrial planets.

Related Web Pages

The University of California Planet Search Project
Astrobiology Magazine New Planets
Transit Search
Extrasolar Planets Encyclopedia
Planet Quest (JPL)
Kepler Mission
Eddington Mission
Darwin Mission
Herschel Mission
Space Interferometry Mission