Quantifying the information impact of future searches for exoplanetary biosignatures

Abstract

One of the major goals for astronomy in the next decades is the remote search for biosignatures (i.e., the spectroscopic evidence of biological activity) in exoplanets. Here we adopt a Bayesian statistical framework to discuss the implications of such future searches, both in the case when life is detected and when no definite evidence is found. We show that even a single detection of biosignatures in the vicinity of our stellar system, in a survey of similar size to what will be obtainable in the next 2 decades, would affect significantly our prior belief on the frequency of life in the universe, even starting from a neutral or pessimistic stance. In particular, after such discovery, an initially agnostic observer would be led to conclude that there are more than [Formula: see text] inhabited planets in the galaxy with a probability exceeding 95%. However, this conclusion would be somewhat weakened by the viability of transfer of biological material over interstellar distances, as in panspermia scenarios. Conversely, the lack of significant evidence of biosignatures would have little effect, leaving the assessment of the abundance of life in the galaxy still largely undetermined.

Keywords: Bayesian analysis; biosignatures; exoplanets; origin of life.

Fig. 1.

Results for a survey searching for atmospheric biosignatures within a distance $R=100$ ly from Earth. Shown are the posterior PDF (A–C) and CCDF (D–F), updated in light of the evidence, starting from a noninformative, pessimistic, and optimistic prior (black dashed curves). The continuous curves refer to the posterior PDF and CCDF for the case $p_a=1$ (all habitable planets in the survey observed). The limit $p_{\mathrm{a}}=0$ is shown by red short-dashed curves in the case of detection, while the posteriors resulting from nondetection at $p_{\mathrm{a}}=0$ coincide with the priors. The shaded areas in the CCDF encompass the limiting cases $p_a=0$ and $p_a=1$, giving the range of probabilities that the mean number of life-bearing planets is larger than $\overline{k}$.

Fig. 2.

Bayes factor as a function of $p_a$, in a survey searching for atmospheric biosignatures within a distance $R=100$ ly. (A) Bayes factor from the comparison of the pessimistic vs. optimistic model (red) and pessimistic vs. noninformative model (blue), when no biosignature detection is made. (B) Bayes factor from the comparison of the optimistic vs. pessimistic model (red) and optimistic vs. noninformative model (blue), when exactly one biosignature detection is made.

Fig. 3.

Effect of assuming correlation between biosignatures due to a panspermia mechanism, after a detection is made in a survey with $R=100$ ly and $p_a=0.1$. (A) The fraction of life-harboring planets within a distance $R$ from Earth, $\overline{k}\left(R\right)$, over the total number $\overline{k}$ in the entire galaxy, as a function of the correlation length $\xi$ and correlation strength $\chi$. (B) The probability that the total number of life-bearing planets in the galaxy $\overline{k}$ is larger than a reference value $10^5$ in the entire galaxy, as a function of the correlation length $\xi$. (C) Bayes factor from the comparison of the optimistic vs. noninformative model.