Exoplanet Biosignatures: A Framework for Their Assessment

Brief

The paper formalizes P(life|D,C), Eq. 7, as the operative expression for exoplanet biosignature assessment, demanding four sequential observational inputs: host-star characterization, exoplanet internal-parameter characterization, biosignature detection, and false-positive exclusion. A worked numerical example (biotic likelihood 0.80, abiotic likelihood 0.25, prior 0.50) yields a posterior of 76%. Posterior probabilities map to five labeled confidence tiers from 'very likely' (90–100%) to 'very unlikely' (<10%). The decisive analytical insight is that as the abiotic false-positive probability P(D|C,no life) approaches zero, the posterior converges to 1 irrespective of the chosen prior, making false-positive exclusion the load-bearing step in any life-detection claim.

Metadata

Category

Search

Venue

Astrobiology

Type

Peer-reviewed

Year

2018

Authors

David C. Catling, Joshua Krissansen-Totton, Nancy Y. Kiang, David Crisp, Tyler D. Robinson, Shiladitya DasSarma, Andrew J. Rushby, Anthony Del Genio, William Bains, Shawn Domagal-Goldman

DOI

10.1089/ast.2017.1737

arXiv

1705.06381

Access

Open access

Length

931.6 K

Tags

biosignature, astrobiology, Bayesian-inference, exoplanet, false-positive, life-detection, SETI-adjacent

Key points

• Core Bayesian expression (Eq. 7) requires two likelihoods, P(D|C,life) and P(D|C,no life), weighted by a context-dependent prior P(life|C); the prior encodes stellar type, orbital parameters, and planetary habitability together.p.3
• Worked example: P(D|C,life)=0.80, P(D|C,no life)=0.25, prior=0.50 yields posterior P(life|D,C)=0.76 (76%).p.4
• Five proposed confidence levels span 'very likely' (90–100% posterior) to 'very unlikely' (<10%), giving a standardized vocabulary for reporting detection confidence.p.1
• When P(D|C,no life)→0, the denominator's second term vanishes and P(life|D,C)→1 regardless of prior choice, making false-positive exclusion mathematically decisive.p.6
• Technosignatures, SETI radio broadcasts, Dyson sphere IR excess, megastructures, are explicitly excluded from the framework's scope.p.2
• Proposed starting prior for an Earth-sized habitable-zone planet around a G-star: P(life|Earth-like)=0.5, grounded empirically in the two rocky HZ objects in the solar system (Earth inhabited, Mars not confirmed inhabited).p.6
• The framework's base-rate term <P(life)> maps directly onto the Drake equation's fl (fraction of habitable planets that develop life), expanded across stellar spectral classes weighted by their galactic occurrence frequencies.p.5
• Galactic occurrence rates used: M-stars 0.76, K-stars 0.12, G-stars 0.076, F-stars 0.03, A-stars 0.006, B-stars 0.0013, O-stars ~10^-5; P(life|O-stars) treated as negligible given O-star main-sequence lifetimes of <1–few Myr.p.5

Verbatim

• “But the extraordinary claim of life should be the hypothesis of last resort only after all conceivable abiotic alternatives are exhausted.”

  p.1

• “The opinions in this first approach range from P (life j Earth-like) being nearly unity to being vanishingly small.”

  p.5

Most interesting

• At low base rates, false-positive life identifications are mathematically anticipated even with a favorable likelihood ratio, a direct consequence of Bayesian base-rate suppression that the paper makes explicit with algebra.
• The only reason P(life|Earth-like) is known to be nonzero at all is the existence of life on Earth; for every other spectral-class prior, the paper acknowledges current scientific ignorance.
• The framework is deliberately analogous to Borucki et al.'s 1996 design of the Kepler mission, which proceeded despite near-total ignorance of Earth-sized-planet occurrence rates, the paper cites this as the correct historical precedent for acting under extreme prior uncertainty.
• The paper treats the habitable zone not as a binary boundary but as a probability density function, citing Zsom (2015), a framing that directly feeds uncertainty into P(life|C).
• O-type stars receive a prior of essentially zero for life because their main-sequence lifetimes (less than one to a few million years) are shorter than the ~3.5 billion years Earth required to produce a remotely detectable biosphere.
• The framework explicitly cannot detect subsurface biospheres (e.g., Europa's ocean) or cryptic surface biospheres too sparse to alter atmospheric chemistry, both are excluded not on scientific grounds but on practical detection grounds.