Information Theory Can Help Us Search for Life on Alien Worlds

Few questions are more intriguing than the possibility of life elsewhere in the universe. But since aliens are not visiting our planet, and we are not going to their faraway homes any time soon, indirect evidence for the existence of biology on distant worlds is our best bet for answers.

The problem is that planets and moons are not just much tinier but also much dimmer than their host stars, making their direct observation extremely challenging. Fortunately, creative astronomers have devised observational methods that detect planets orbiting distant stars on our cosmic shore and, remarkably, for obtaining the approximate chemical composition of their atmospheres. That’s where life comes in: if life exists at a global scale on a planet, it can leave signals in the atmosphere. Like fingerprints, different kinds of biological activity will leave specific atmospheric imprints. We see this, for instance, with the abundance of oxygen in our atmosphere produced by photosynthesis. The challenge for us is to decipher the message life leaves on alien atmospheres.

That’s going to take powerful telescopes but also new ways of thinking about how to decode the information hidden in the light they capture from alien worlds. We propose that the information theory we now rely on to sift signal from noise in modern communications offers tools astronomers can use to detect signs of biological activity in other worlds. The approach comes in two steps: after capturing the light from the exoplanet, we use information theory to search for chemicals associated with the presence of life. What in communications are letters of an alphabet making up a sentence, in astrobiology will be specific chemicals that exist in a distant world’s atmosphere.

Today, we can best infer an exoplanet’s chemical composition by transit spectroscopy: When a planet passes in front of its star, as seen from Earth, the planet’s atmosphere absorbs some of its starlight. The resulting absorption spectrum looks a bit like the profile of a jagged mountain range, with the valleys corresponding to different chemical elements that absorb the light coming from the star. From this, we can in principle infer whether there is any kind of biological activity there.

If we know how to look for it, that is. Things sound very promising, until we start asking harder questions. For starters, we are assuming that these biosignatures are for life as we know it here on Earth. Perhaps this is a good starting point, but how can we be sure that this is what we will find? Even so, what kinds of atmospheric biosignatures should we then look for? Also, if we are looking for some kind of Earth-like planet, at what stage in its evolutionary history are we catching it? Life has drastically changed Earth’s atmosphere since it first emerged 3.5 billion years ago. An alien astronomer looking at us at different geological times would notice radical changes. In that time span, life has massively increased Earth’s atmospheric oxygen and ozone, as well as triggering fluctuations in methane. Finally, host stars come in various sizes and temperatures, and will transform over their existence. Different host stars—from hot yellow ones to cooler red ones—at different life cycle stages affect their planets differently.

For all these reasons, we need to invoke multiple methods to search for signs of life in the atmosphere of other worlds. In a recent paper, we proposed information theory—a methodology to decode signal from noise in data transmissions of any kind—as one of those key tools. The analysis compared the spectral data from simulated exoplanets to our own Earth’s across a broad range of astrophysical and planetary contexts: at different stages of their evolution and orbiting around different host stars. The results suggest our tool can reliably analyze real data from current and future observations to search for alien biosignatures. The measure we adapted from information theory, known as the Jensen-Shannon divergence, directly compares any two absorption spectra and quantifies how alike (or not) they are. This “configurational entropy” measure is an excellent discriminator of subtle variations on a planet’s spectral patterns. A small value means the two compared worlds have very similar overall atmospheric compositions; a large one means they are quite different. We then used a finer diagnostic tool to compute the configurational entropy for a selection of specific wavelengths of light, the regions of the electromagnetic spectrum where the “spectral signatures”—the valleys on the mountain range of certain substances—are most easily seen. With this, we can zero in on specific compounds, like CO₂ or methane, or two compounds appearing together, like methane and ozone, and directly compare their abundances on the two worlds.

For us, an exoplanet is an “Earth analogue” not only when it has a radius and mass close to Earth’s, but also when its absorption spectrum is very close in information space to Earth’s across our planet’s multibillion-year history. This expands the concept of “Earth-like” widely used in astrobiology to look beyond our biosphere’s current state, into its distant past (and possible future) when life’s global fingerprints can be very different. We considered Earth during three stages of its evolution, from just after the Great Oxidation Event circa two billion years ago, with almost no oxygen in the atmosphere and small amounts in oceans and seabed rock, to 0.8 billion years ago when oxygen was at about 10 percent to, finally, modern amounts of 21 percent.

The information measure we propose can discriminate between worlds based on their spectroscopic signatures. We must still be careful, because a planet’s age and its star type can still confound our search for closely matching biosignatures. To confidently identify Earth-analogues we need to find exoplanets orbiting stars like our sun. But we expect life to be more creative than we currently assume based on what we see on Earth.

When we compare modern Earth’s spectral signatures of specific chemical compounds with those for Earth-like planets orbiting different host stars, we find that our method is well-suited to identify what could be attributed to biological activity and how it changes over the eons. The accuracy depends, of course, on how clean (or “low noise”) the spectra are, but even within JWST’s capabilities, the results are very promising. Also, our information measure can be applied as a comparison basis to any kind of world, not just Earth-analogues. This means we conceivably could identify truly alien varieties of biological activity that don’t conform to what we see on Earth, for example, by singling out its associated atmospheric chemistry as somehow unusual. Such openness and flexibility are vital in our search for life as we know it—and for life as we don’t.

We may not know whether we are alone in the cosmos yet, but if life is hiding elsewhere, we should be able to find it soon.

This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily those of Scientific American.

Few questions are more intriguing than the possibility of life elsewhere in the universe. But since aliens are not visiting our planet, and we are not going to their faraway homes any time soon, indirect evidence for the existence of biology on distant worlds is our best bet for answers.

The problem is that planets and moons are not just much tinier but also much dimmer than their host stars, making their direct observation extremely challenging. Fortunately, creative astronomers have devised observational methods that detect planets orbiting distant stars on our cosmic shore and, remarkably, for obtaining the approximate chemical composition of their atmospheres. That’s where life comes in: if life exists at a global scale on a planet, it can leave signals in the atmosphere. Like fingerprints, different kinds of biological activity will leave specific atmospheric imprints. We see this, for instance, with the abundance of oxygen in our atmosphere produced by photosynthesis. The challenge for us is to decipher the message life leaves on alien atmospheres.

That’s going to take powerful telescopes but also new ways of thinking about how to decode the information hidden in the light they capture from alien worlds. We propose that the information theory we now rely on to sift signal from noise in modern communications offers tools astronomers can use to detect signs of biological activity in other worlds. The approach comes in two steps: after capturing the light from the exoplanet, we use information theory to search for chemicals associated with the presence of life. What in communications are letters of an alphabet making up a sentence, in astrobiology will be specific chemicals that exist in a distant world’s atmosphere.

Today, we can best infer an exoplanet’s chemical composition by transit spectroscopy: When a planet passes in front of its star, as seen from Earth, the planet’s atmosphere absorbs some of its starlight. The resulting absorption spectrum looks a bit like the profile of a jagged mountain range, with the valleys corresponding to different chemical elements that absorb the light coming from the star. From this, we can in principle infer whether there is any kind of biological activity there.

If we know how to look for it, that is. Things sound very promising, until we start asking harder questions. For starters, we are assuming that these biosignatures are for life as we know it here on Earth. Perhaps this is a good starting point, but how can we be sure that this is what we will find? Even so, what kinds of atmospheric biosignatures should we then look for? Also, if we are looking for some kind of Earth-like planet, at what stage in its evolutionary history are we catching it? Life has drastically changed Earth’s atmosphere since it first emerged 3.5 billion years ago. An alien astronomer looking at us at different geological times would notice radical changes. In that time span, life has massively increased Earth’s atmospheric oxygen and ozone, as well as triggering fluctuations in methane. Finally, host stars come in various sizes and temperatures, and will transform over their existence. Different host stars—from hot yellow ones to cooler red ones—at different life cycle stages affect their planets differently.

For all these reasons, we need to invoke multiple methods to search for signs of life in the atmosphere of other worlds. In a recent paper, we proposed information theory—a methodology to decode signal from noise in data transmissions of any kind—as one of those key tools. The analysis compared the spectral data from simulated exoplanets to our own Earth’s across a broad range of astrophysical and planetary contexts: at different stages of their evolution and orbiting around different host stars. The results suggest our tool can reliably analyze real data from current and future observations to search for alien biosignatures. The measure we adapted from information theory, known as the Jensen-Shannon divergence, directly compares any two absorption spectra and quantifies how alike (or not) they are. This “configurational entropy” measure is an excellent discriminator of subtle variations on a planet’s spectral patterns. A small value means the two compared worlds have very similar overall atmospheric compositions; a large one means they are quite different. We then used a finer diagnostic tool to compute the configurational entropy for a selection of specific wavelengths of light, the regions of the electromagnetic spectrum where the “spectral signatures”—the valleys on the mountain range of certain substances—are most easily seen. With this, we can zero in on specific compounds, like CO₂ or methane, or two compounds appearing together, like methane and ozone, and directly compare their abundances on the two worlds.

For us, an exoplanet is an “Earth analogue” not only when it has a radius and mass close to Earth’s, but also when its absorption spectrum is very close in information space to Earth’s across our planet’s multibillion-year history. This expands the concept of “Earth-like” widely used in astrobiology to look beyond our biosphere’s current state, into its distant past (and possible future) when life’s global fingerprints can be very different. We considered Earth during three stages of its evolution, from just after the Great Oxidation Event circa two billion years ago, with almost no oxygen in the atmosphere and small amounts in oceans and seabed rock, to 0.8 billion years ago when oxygen was at about 10 percent to, finally, modern amounts of 21 percent.

The information measure we propose can discriminate between worlds based on their spectroscopic signatures. We must still be careful, because a planet’s age and its star type can still confound our search for closely matching biosignatures. To confidently identify Earth-analogues we need to find exoplanets orbiting stars like our sun. But we expect life to be more creative than we currently assume based on what we see on Earth.

When we compare modern Earth’s spectral signatures of specific chemical compounds with those for Earth-like planets orbiting different host stars, we find that our method is well-suited to identify what could be attributed to biological activity and how it changes over the eons. The accuracy depends, of course, on how clean (or “low noise”) the spectra are, but even within JWST’s capabilities, the results are very promising. Also, our information measure can be applied as a comparison basis to any kind of world, not just Earth-analogues. This means we conceivably could identify truly alien varieties of biological activity that don’t conform to what we see on Earth, for example, by singling out its associated atmospheric chemistry as somehow unusual. Such openness and flexibility are vital in our search for life as we know it—and for life as we don’t.

We may not know whether we are alone in the cosmos yet, but if life is hiding elsewhere, we should be able to find it soon.

This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily those of Scientific American.