Life on other planets

Q: What's the percentage of habitable planets in the universe? What's the current estimate on how may planets have life?

The astronomical framework for answering both of these questions is the Drake Equation, named after Frank Drake, Professor Emeritus at University of California, Santa Cruz:

N (Number of civilizations in our galaxy that we could communicate with) =
R_* (Rate of formation of stars that are suitable for life to develop) x
f_p (fraction of those stars that have planets) x
n_e (average number of planets per planet-hosting star that could host life) x
f_l (fraction of theoretically habitable planets on which life actually appears) x
f_i (fraction of life-bearing planets on which intelligent life develops) x
f_c (fraction of civilizations that develop a way to send signals into space) x
L (average length of time such civilizations send out signals)

From wide-field infrared surveys and more detailed studies of star-forming clouds of gas, we know R_* fairly well. In the Milky Way, about 1 Solar mass (2 x 10³³ g) of gas turns into stars per year, and that Solar mass is typically distributed among red dwarf stars, which we think are suitable for life. Looking back in time by observing more distant galaxies tells us that the Milky Way's star-formation rate probably peaked about two billion years after the Big Bang, and has been declining ever since. The fact that most of the Milky Way's stars are old could be positive because it takes so much time for intelligent life to evolve, but it could also be problematic if high-tech civilizations don't last very long.

Since 1995, astronomers have made substantial progress on f_p, the fraction of stars with planets. Analysis of data from NASA's Kepler mission shows that about 1/4 of Sun-like stars have small, rocky planets, 1-2 Earth radii, with orbital periods under 100 days (Marcy et al. 2014). Such short-period planets are closer to their stars than Venus is to the Sun, receive a much higher light flux than Earth, and might not be habitable. If you count larger planets or planets on more distant orbits, f_p goes up -- Cassan et al. (2012) find f_p = 0.62 for "super-Earths" of 5-10 Earth masses between 0.5 and 10 AU and Sumi et al. (2011) suggest two giant planets per star in the Galactic center (f_p = 1). We believe planets are an inevitable product of Sunlike star formation.

The number of habitable planets per planet-hosting star, n_e, is more difficult. About 1/5 of the candidate planets discovered by Kepler seem to be in multiple-planet systems. The true fraction of multiple planets is probably higher, as Kepler candidate planets may have undetected siblings in long-period orbits. (Some of these unseen siblings can be found by measuring changes in the planet candidate's eclipse timing.) Given the large range of planet masses and orbital periods that are currently undetectable, it is possible that every Sunlike star hosts multiple planets. However, planets are like porcupines. They can't get too close together unless they are protected by special orbital resonances, so it's hard to pack many of them into the "Goldilocks zone" where liquid water can exist. A gas giant that moves through the inner part of the planetary system can eject any rocky planets already there. Unfortunately, the Kepler mission didn't quite observe long enough to probe the Goldilocks zones of Sunlike stars. Extrapolating from their data on short-period planets, 1 < n_e < 2 is a likely range, making about 100 billion habitable planets!

This analysis centered on the Goldilocks zone of Sunlike stars leaves out many important points. First, low-mass red dwarfs, which are the most common stars in the Galaxy, are underrepresented in planet searches, mostly because they are so faint that we can only take detailed measurements if they're nearby. Several planet searches such as M2K and the HARPS survey are attempting to rectify this problem, and the HARPS team announced a "1.36 planets per red dwarf" result in March 2014 (Tuomi et al. 2014). Second and most importantly, the Goldilocks zone excludes possibilities like life in the subsurface ocean of an icy body (think Europa) or extremophiles that live in rocks or sea-floor vents. It may not be possible to empirically measure n_e since the chemical and orbital evolution history of each planet is different.

We may make some progress in measuring f_l in our lifetimes. One of the science goals of the James Webb Space Telescope is to probe the atmospheres of Earthlike planets. Astronomers are designing programs to search for biosignatures such as O₂ or methane, which we think are rarely geologically generated. We won't get a statistical measurement of f_l within the next 20 years, but we may detect a few Earthlike atmospheres. Again, this analysis ignores the possibility of un-Earth-like life.

f_i and f_c are matters of conjecture, though one could get an (Earth-centric) idea by dividing the length of time intelligent species have existed on Earth by Earth's age, 4.5 billion years. But how do you classify "intelligent" species? Could dolphins eventually evolve the capability to communicate with other civilizations? What about octopi and elephants, which seem to have excellent reasoning capabilities?

Finally, we are conducting an experiment with L right now. Time will tell if we can mitigate and adapt to climate change, avoid nuclear war and dodge asteroids long enough to communicate with another civilization.

-Sally Dodson-Robinson