Our recent focus on habitability addresses a significant problem. In order for astrobiologists to home in on the best targets for current and future telescopes, we need to be able to prioritize them in terms of the likelihood for life. I’ve often commented on how lazily the word ‘habitable’ is used in the popular press, but it’s likewise striking that its usage varies widely in the scientific literature. Alex Tolley today looks at a new paper offering a quantitative way to assess these matters, but the issues are thorny indeed. We lack, for instance, an accepted definition of life itself, and when discussing what can emerge on distant worlds, we sometimes choose different sets of variables. How closely do our assumptions track our own terrestrial model, and when may this not be applicable? Alex goes through the possibilities and offers some of his own as the hunt for an acceptable methodology continues.

by Alex Tolley

Artist illustrations of explanets in the habitable zone as of 2015. None appear to be illustrated as possible hanbitable worlds. This has changed in the last decade. Credit: PHL @ UPR Arecibo (phl.upr.edu) January 5, 2015 Source [1]

We now know that the galaxy is full of exoplanets, and many systems have rocky planets in their habitable zones (HZ). So how should we prioritize our searches to maximize our resources to confirm extraterrestrial life?

A new paper by Dániel Apai and colleagues of the Quantitative Habitability Science Working Group, a group within the Nexus for Exoplanet System Science (NExSS) initiative, looks at the problems hindering our quest to prioritize searches of the many possible life-bearing worlds discovered to date and continuing to be discovered with new telescope instruments.

The authors state that the problem we face in the search for life is:

“A critical step is the identification and characterization of potential habitats, both to guide the search and to interpret its results. However, a well-accepted, self-consistent, flexible, and quantitative terminology and method of assessment of habitability are lacking.”

The authors expend considerable space and effort itemizing the problems that have accrued: Astrobiologists and institutions have never defined “habitable” and “habitability” rigorously, even confounding “habitable” with “Habitable Zone” (HZ), and using “habitable” and “inhabited” almost interchangeably. They argue that this creates problems for astrobiologists when trying to plan how to develop strategies when determining which exoplanets are worth investigating and how. Therefore, defining terminology is important to avoid confusion. As the authors point out, researchers often assume that planets in the HZ should be habitable and those outside its boundaries uninhabitable, even though both assertions are untrue.

[I am not clear that the first assertion is claimed without caveats, for example, the planet must be rocky and not a gaseous world, such as a mini-Neptune.]

As our knowledge of exoplanets is data poor, it may not be possible to define whether a planet is habitable based on the available information, which leads to the imprecision of the term “habitable”. In addition, not only has “habitable” not been well-defined, but neither have the requirements for life been defined, which is more restrictive than the loose requirement for surface liquid water.

Ultimately, the root of the problem that hampers the community’s efforts to converge on a definition for habitability is that habitability depends on the requirements for life, and we do not have a widely agreed-upon definition for life.

The authors accept that a universal definition of life may not be possible, but that we can, however, determine the habitat requirements of particular forms of life.

The authors’ preferred solution is to model habitability with joint probability assessments of planetary conditions with already acquired data, and extended with new data. This retains some flexibility in the use of the term “habitable” in the light of new data.

Figure 1 below illustrates the various qualitative approaches to defining habitability. Adhering to any single definition is not possible for a universal definition. The paper suggests that a better approach is to use quantitative methods that are both rigorous, yet flexible in the light of new data and information.

Figure 1. Various approaches to defining “habitable”. Any single definition for “habitability” fails to meet the majority of the requirements.

The equation below illustrates the idea of quantifying the probability of a planet’s habitability as a joint probability of the known criteria:

The authors use the joint probability of the planet being habitable. For example, it is in the HZ, is rocky, and has water vapor in the atmosphere. Clearly, under this approach, if water vapour cannot be detected, the probability of habitability declines to zero.

Perhaps of even greater importance, the group also looks at habitability based on whether a planet supports the requirements for known examples of terrestrial life, whose requirements vary considerably. For example, is there sufficient energy to support life? If there is no useful light from the parent star in the habitat, energy must be supplied by geological processes, leading to the likelihood that only anaerobic chemotrophs could live under those conditions, for example, as hypothesized in dark, glacial-covered subsurface oceans..

The authors include more carefully defined terms, including: Earth-like life. Rocky Planet, Earth-Sized Planet, Earth-Like Planet, Habitable Zone, Metabolisms, Viability Model, Suitable Habitat for X, and lastly Habitat Suitability which they defined as:

The measure of the overlap between the necessary environmental conditions for a metabolism and environmental conditions in the habitat.

Because of the probability that life (at least some species of terrestrial life) will inhabit a planet, the authors suggest a framework where the Venn diagram of the probability of a planet being habitable intersects with the probability that the requirements are met for specific species of terrestrial life.

The Quantitative Habitability Framework (QHF) is shown in Figure 2 below.

Figure 2. Illustration of the basis of the Framework for Habitability: The comparison of the environmental conditions predicted by the habitat model and the environmental conditions required by the metabolism model.

The probability of the viability of an organism for each variable is a binary value of 0 for non-viability and 1 for viability. For archaea and temperature, this is:

The equation means that the probability of viability of the archaea at temperature T in degrees Kelvin is 1 if the temperature is between 257 K ( -16 °C) and 395 K (122 °C). Otherwise 0 if the temperature is outside the viable range. (Ironically, this is imprecise, as it is for a species of archaean methanogen extremophile, not all archaeans.)

This approach is applied to other variables. If any variable probability is zero, the joint probability of viability becomes 0.

The figure below shows various terrestrial organism types, mostly unicellular, with their known temperature ranges for survival. The model therefore allows for some terrestrial organisms to be extant on an exoplanet, whilst others would not survive.

Figure 3. Examples of temperature ranges for different types of organisms, taken for species at the extreme ranges of survival in the laboratory.

To demonstrate their framework, they work through models for archaean life on Trappist 1e and Trappist 1f, cyanobacteria on the same 2 planets, methanogens (that would include archaea) in the subsurface of Mars, and the subsurface ocean of Enceladus.

Figure 4 below shows the simplified model for archaea on Trappist 1e. The values and standard deviations used for the priors are not all explained in the text. For example, the mean surface pressure is set at 5 Bar for illustrative purposes, as no atmosphere has been detected for Trappist-1e. The network model for the various modules that determine the viability of an archaean is mapped in a). Only 2 variables, surface temperature and pressure, determine viability of the archaean prokaryotes in the modeled surface temperature. In more sophisticated models, this would be a multidimensional plot perhaps using principal component analysis (PCA) to show a 3D plot. The various assumed (prior) and calculated (result) values and their assumed distributions are shown as charts in b). The plot of viability (1) and non-viability (0) for surface temperature and pressure is shown in the 3D plot c. The distributions indicate the probability of suitability of the habitat.

Figure 4. QHF assessment of the viability of archaea/methanogens in a modeled TRAPPIST-1e-like planet’s surface habitat. a: Connections between the model modules. Red are priors, blue are calculated values, green is the viability model. b: Relative probability distributions of key parameters. c: The distribution of calculated viability as a function of surface temperature and pressure. The sharp temperature cutoff at 395K separates the habitat as viable or not.

The examples are then tabulated to show the probabilities of different unicellular life inhabiting the various example worlds.

As you can see, the archaea/methanogens on Trappist 1f have the highest probability of being present inhabiting that world if we assume terrestrial life represents good examples. Therefore, Trappist 1f would be prioritized given the information currently available. If spectral data suggested that there was no water in Trappist 1f’s atmosphere, this would reduce, and possibly eliminate, this world’s habitability probability, and with it the probability of it meeting the requirement of archaean life and hence reducing the overlap in the habitability and life requirement terms to nil.

My Critique of the Methodology

The value of this paper is that it goes beyond the usual “Is the exoplanet habitable?” with the usual caveats about habitability that apply under certain conditions, usually atmospheric pressure and composition. The habitable zone (HZ) around a star is calculated for the range of distances from the star where, with an ideal atmosphere composition and density, on a rocky surface, liquid water could be found. Thus, early Mars, with a denser atmosphere, could be habitable [2], and indeed, the evidence is that water was once present on the surface. Venus might also once have been habitable, positioned at the inner edge of our sun’s HZ, before a runaway greenhouse made the planet uninhabitable.

The concept of NASA’s “Follow the water” mantra is a first step, but this paper then points out this is only part of the equation when deciding the priority of expending resources on observing a prospective exoplanet for life. The Earth once had an anoxic atmosphere, making Lovelock’s early idea that gases in disequilibrium would indicate life, which quickly became interpreted as free oxygen (O2) and methane (CH4), largely irrelevant during this period before oxygenic photosynthesis changed the composition of the atmosphere.

Yet Earth was living within a few hundred million years after its birth, with organisms that predated the archaea and bacteria kingdoms [4]. Archaea are often methanogens, releasing CH4 into the early atmosphere at a rate exceeding that of geological serpentinization. Their habitat in the oceans must have been sufficiently temperate, albeit some are thermophilic, living in water up to 122 centigrade but under pressure to prevent boiling. If the habitability calculations include the important variables, then their methodology offers a rigorous way to determine the probability of particular terrestrial life on a prospective exoplanet.

The problem is whether the important variables are included. As we see with Venus, if the atmosphere was still Earthlike, then it might well be a prospective target. Therefore, an exoplanet on the inner edge of its star’s HZ might need to have its atmosphere modeled for stability, given the age of its star, to determine whether the atmosphere could still be earthlike and therefore support liquid water on the surface.

However, there may be other variables that we have repeatedly discussed on this website. Is the star stable or does it flare frequently? Does the star emit hard UV and X-rays that would destroy life on the surface by destroying organic molecules? Is the star’s spectrum suited for supporting photosynthesis, and if not, does it allow or prevent chemotrophs to survive? For complex life, is a large moon needed to keep the rotational axis relatively stable to prevent climate zones and circulation patterns from changing too drastically? Is the planet tidally locked, and if so, can life exist at the terminator, as we have no terrestrial examples to evaluate? The Ramirez paper [3] includes his modeling of Trappist 1e, using the expected synchronization of its rotation and orbital period, resulting in permanently hot and cold hemispheres.

While the authors suggest that the analysis can extend beyond species to ecosystems, and perhaps a biosphere, we really don’t know what the relevant variables are in most cases. Unicellular organisms are sometimes easily cultured in a laboratory, but most are not. We just don’t know what conditions they need, and whether these conditions exist on the exoplanet. It may be that the equation variables may be quite large, making the analyses too unwieldy to be worth doing to evaluate the probability of some terrestrial life form inhabiting the exoplanet.

A further critique is that organisms rarely can exist as pure cultures except in a laboratory setting with ideal culture media. Organisms in the wild exist in ecosystems, where different organisms contribute to the survival of others. For example, bacterial biofilms often comprise different species in layers allowing for different habitats to be supported, from anaerobes to aerobes.

The analysis gamely looks at life below the surface, such as lithophilic life 5 km below the surface of Mars, or ocean life in the subsurface Enceladan ocean. But even if the probability in either case was 100% that life was present, both environments are inaccessible compared to other determinants of life that we can observe with our telescopes. This would apply to icy moons of giant exoplanets, even if future landers established that life existed in both Europan and Enceladan subsurface oceans.

What about exoplanets that are not in near circular orbits, but more eccentric, like Brian Aldiss’ fictional “Heliconia”? How to evaluate their habitability? Or circumbinary planets where the 2 stars are creating differing instellation patterns as the planet orbits its close binary?. Lastly, can tidally locked exoplanets support life only at the terminator that supports the range of their known requirements, such as Ramirez’ modeling of Trappist 1e?

An average surface temperature does not cover either the extremes, for example the tropics and the poles, nor that water is at its densest at 277K (4 °C), ensuring that there is liquid water even when the surface is fully glaciated above an ocean. Interior heat can also ensure liquid water below an icy surface, and tidal heating can contribute to heating even on moons that have exhausted their radioactive elements. If the Gaia hypothesis is correct, then life can alter a planet to support life even under adverse conditions, stabilizing the biosphere environment. The range of surface temperatures is covered by the Gaussian distribution of temperatures as shown in Figure 4 b.

Lastly, while the joint probability model with Monte Carlo simulation to estimate the probability of an organism or ecosystem inhabiting the exoplanet is a relatively computationally lightweight model, it may not be the right approach with more variables added to the mix. The probabilities may be disjoint with a union of different subsets of variables with joint probabilities. In other words, rather than “and” intersections of planet and organism requirement probabilities, there may be an “or” union of probabilities. The modeled approach may prove brittle and fail, a known problem of such models, which can be alleviated to some extent by using only subsets of the variables. Another problem I foresee is that a planet with richer observational data may score more poorly than a planet with few data-supported variables, simply due to the joint probability model.

All of which makes me wonder if the approach really solves the terminology issue to prioritize exoplanet life searches, especially if a planet is both habitable and potentially inhabited. It is highly terrestrial-centric, as we would expect, as we have no other life to evaluate. If we find another life on Earth, as posited by Paul Davies’ “Shadow Biosphere,” [4] this methodology could be extended. But we cannot even determine the requirements of extinct animals and plants that have no living relatives, but flourished in earlier periods on Earth. Which species survived when Earth was a hothouse in the Carboniferous, or below the ice during the global glaciations? Where were those conditions outside the range of extant species? For example, post-glacial humans could not survive during the Eocene thermal maximum.

For me, this all boils down to whether this method can usefully help determine whether an exoplanet is worth observing for life. If an initial observation ruled out an atmosphere like any of those Earth has experienced in the last 4.5 billion years, should the search for life be immediately redirected to the next best target, or should further data be collected, perhaps to look for gases in disequilibrium? While I wouldn’t bet that Seager’s ‘MorningStar’ mission to look for life on Venus will find anything, if it did turn up microbes in the acidic atmosphere’s temperate zone, that would add a whole new set of possible organism requirements to evaluate, making Venus-like exoplanets viable targets for life searches. If we eventually find life on exoplanets with widely varying conditions, with ranges outside of terrestrial life, would the habitat analyses then have to test all known life from a catalog of planetary conditions?

But suppose this strategy fails, and we cannot detect life, for various reasons, including instrumentation limits? Then we should fall back on the method I last posted on, which reduces the probability of extant life on an exoplanet, but leaves open the possibility that life will eventually be detected.

The paper is Apai et al (2025)., “A Terminology and Quantitative Framework for Assessing the Habitability of Solar System and Extraterrestrial Worlds,” in press at Planetary Science Journal. Abstract.

References

1. Schulze-Makuch, D. (2015) Astronomers Just Doubled the Number of Potentially Habitable Planets. Smithsonian Magazine, 14 January 2015.
https://www.smithsonianmag.com/air-space-magazine/astronomers-just-doubled-number-potentially-habitable-planets-180953898/

2. Seager, S. (2013) “Exoplanet Habitability,” Science 340, 577.
doi: 10.1126/science.1232226

3. Ramirez R., (2024). “A New 2D Energy Balance Model for Simulating the Climates of Rapidly and Slowly Rotating Terrestrial Planets,” The Planetary Science Journal 5:2 (17pp), January 2024
https://doi.org/10.3847/PSJ/ad0729

4. Tolley, A. (2024) “Our Earliest Ancestor Appeared Soon After Earth Formed” https://www.centauri-dreams.org/2024/08/28/our-earliest-ancestor-appeared-soon-after-earth-formed

5. Davies P. C. W. (2011) “Searching for a shadow biosphere on Earth as a test of the ‘cosmic imperative,” Phil. Trans. R. Soc. A.369624–632 http://doi.org/10.1098/rsta.2010.0235